WCRI Research Brief · 2017-08-03 · Page 1 of 2 WCRI Research Brief: Longer-Term Dispensing of...

of 2

WCRI Research Brief: Longer-Term Dispensing of Opioids, 4th Edition This report examines the prevalence and trends of longer-term dispensing of opioids in 26 state workers’ compensation systems. It also documents how often the services recommended by medical treatment guidelines are used for monitoring and managing the care delivered to the injured workers who received opioids on a longer-term basis.

Long-term opioid therapy for chronic non-cancer pain is a controversial topic, and there is little evidence about the effectiveness of such treatment on functional recovery or return to work. Unnecessary opioid prescriptions may lead to opioid addiction, overdose, and diversion, which is a top priority public health topic in the United States. The public concern is also shared by the workers’ compensation community. In recent years, many states have made legislative or regulatory changes, within and outside workers’ compensation, to address issues related to overuse and misuse of opioids. Some policy changes were also made at the federal level, including the Centers for Disease Control and Prevention (CDC) guidelines for prescribing opioids for chronic pain.

This study tracks the prevalence of longer-term dispensing of opioids in 26 study states. Claims receiving longer-term dispensing of opioids are those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury. Trends reported are based an average of 24 months of experience for claims with injuries occurring in 2010 and 2013, with prescriptions filled through March 31, 2015.

Major Findings:

• The frequency of claims that received opioids on a longer-term basis decreased more than 4 percentage points in Kentucky and New York, between 2010 and 2013 claims

Research Questions: • What was the recent trend in the

longer-term dispensing of opioids in the 26 study states?

• How did the prevalence of longer-term dispensing of opioids in my state compare with others?

• What policy tools are available that might help reduce unnecessary opioid use?

• How often were drug testing and other guideline-recommended services provided to injured workers receiving longer-term opioids?

of 2

with 24-month experience. The same measure decreased 2–3 percentage points in several other states (Kansas, Massachusetts, Michigan, Minnesota, and Tennessee). Noticeable decreases in the longer-term dispensing of opioids were also seen in several other states, including California, Florida, and Texas, with reductions of 1–2 percentage points.

• Among claims with injuries in 2013 that were observed over a two-year time period ending March 2015, longer-term dispensing of opioids was most prevalent in Louisiana—1 in 6 injured workers with opioid prescriptions were identified as receiving longer-term opioids. Compared with most study states, the number was also higher in California, Georgia, North Carolina, Pennsylvania, South Carolina, and Texas—1 in 10 to 1 in 12 injured workers received longer-term opioids. By contrast, about 1 in 25 injured workers with opioid prescriptions received them on a longer-term basis in Indiana, Kansas, Missouri, Nevada, New Jersey, and Wisconsin.

• The study continued to find that fewer than expected injured workers who received opioids on a longer-term basis had certain services (i.e., drug testing, psychological evaluation and treatment, etc.) that are recommended by treatment guidelines for chronic opioid management. For example, in 19 out of 26 states, less than 10 percent of injured workers with longer-term opioids received psychological evaluations.

Data & Methods: This study uses data comprising over 400,000 nonsurgical workers’ compensation claims with more than seven days of lost time, and over 2 million prescriptions are associated with these claims from 26 states. These claims had injuries in 2010 and 2013 and received on average up to 24 months of medical treatment. The sample of claims in the study represents 36–69 percent of workers’ compensation claims in each state.

The 26 states in the study are Arkansas, California, Connecticut, Florida, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maryland, Massachusetts, Michigan, Minnesota, Missouri, Nevada, New Jersey, New York, North Carolina, Pennsylvania, South Carolina, Tennessee, Texas, Virginia, and Wisconsin.

LONGER-TERM DISPENSING OF OPIOIDS,

4TH EDITION

Dongchun Wang

WC-17-29

August 2017

WORKERS COMPENSATION RESEARCH INSTITUTE CAMBRIDGE, MASSACHUSETTS

copyright © 2017 workers compensation research institute

1

COPYRIGHT © 2017 BY THE WORKERS COMPENSATION RESEARCH INSTITUTE ALL RIGHTS RESERVED. NO PART OF THIS BOOK MAY BE COPIED OR

REPRODUCED IN ANY FORM OR BY ANY MEANS WITHOUT WRITTEN PERMISSION OF THE WORKERS COMPENSATION RESEARCH INSTITUTE.

ISBN 978-1-61471-701-0

PUBLICATIONS OF THE WORKERS COMPENSATION RESEARCH INSTITUTE DO NOT NECESSARILY REFLECT THE OPINIONS OR POLICIES

OF THE INSTITUTE’S RESEARCH SPONSORS.


L O N G E R - T E R M D I S P E N S I N G O F O P I O I D S , 4 T H E D I T I O N_____________________________________________________________________________________________

2

ACKNOWLEDGMENTS

The author would like to thank the technical reviewers, Dr. Jeffery Harris, Eric Taylor, and David Feldman.

Their thoughtful comments and suggestions not only helped us to improve the accuracy and clarity of the

final report but also are valuable for our future research. Thanks also go to several practitioners who provided

valuable comments on this study, including Dr. Dan Hunt, Dr. Gary Franklin, John Pedrick, Dr. Kirsten

Koos, Marilyn Hoffmeister, and Mary Colvin.

This is an update of a previously published study, which reflects the important contributions made by

many people, especially the coauthors of the previous study, Dr. Kathryn Mueller and Dr. Dean Hashimoto,

and Dr. Rick Victor for his guidance during the early stages of the project. Critical to the study was the

indispensable assistance provided by Dr. Philip Borba and his colleagues at Milliman, Inc., Eric Harrison, and

other colleagues at WCRI. Their contributions, including construction and quality assurance of the drugs

database, programming support, and update of drugs policies, made the study possible.

Thanks go to Andrew Kenneally, the communications director at WCRI, for his efforts in disseminating

the research findings. My gratitude goes to Ramona Tanabe, the executive vice president and counsel of the

Institute, and Dr. John Ruser, the president and CEO of the Institute, for their valuable input and guidance

that shaped the final report.

Thanks also go to Sarah Solorzano and Elizabeth Hopkins for their superior administrative assistance

that helped to improve the readability and accuracy of the report, and Sarah Solorzano, who managed the

review and publication process.

Of course, any errors or omissions that remain in the report are the responsibility of the author.

Dongchun Wang

Cambridge, Massachusetts August 2017


3

_____________________________________________________________________________________________L O N G E R - T E R M D I S P E N S I N G O F O P I O I D S , 4 T H E D I T I O N

TABLE OF CONTENTS

List of Tables 5

List of Figures 6

Executive Summary 7

1. Introduction 17

Scope of the Report 17 Organization of the Report 18

2. Data and Methods 19

Data and Representativeness 19 Identifying Opioid Prescriptions 22 Identifying Claims with Longer-Term Opioids 23 Services Recommended for Chronic Opioid Management by Medical Guidelines 24 Measuring Frequency of Claims Receiving Opioids and Longer-Term Opioids 26 Sensitivity Analysis for Claim Selection and Case Mix 30 Limitations and Caveats 31

3. Longer-Term Dispensing of Opioids 33

4. Implications and Conclusions 55

Technical Appendix A: A Brief Summary of Factors That May Influence the Prescribing of Opioids 58

Technical Appendix B: Medical Treatment Guideline Principles for Chronic Opioid Management 71

Technical Appendix C: Sensitivity Analysis of Claim Selection and Case Mix 76

Glossary 86

References 87



4

LIST OF TABLES

2.1 Claims and Prescriptions Included in the Study / 20

2.2 Federal Classification of Controlled Substances as of 2015 / 22

2.3 CPT4 and HCPCS Codes Used to Identify Specific Services Recommended for Managing Long-

Term Use of Opioids / 27

3.1 Frequency of Claims Receiving Opioids among Nonsurgical Claims with More Than 7 Days of Lost

Time, 2013/2015 / 36

3.2 Trends in the Percentage of Nonsurgical Claims with Opioids That Received Opioids on a Longer-

Term Basis, 2010/2012–2013/2015 / 38

3.3 Trends in the Frequency of Drug Testing for Claims Receiving Longer-Term Opioids / 47

3.4 Utilization of Drug Testing Services, among Nonsurgical Claims with Longer-Term Opioids That

Received Drug Testing, Pooled 2010/2012–2013/2015 Claims / 49

3.5 Utilization and Costs of Drug Testing Services, among Nonsurgical Claims Receiving Longer-Term

Opioids That Had Drug Testing / 50

3.6 Frequency of Claims Receiving Psychological Services, 2011/2013–2013/2015 Nonsurgical Claims

Identified as Receiving Longer-Term Opioids / 53

3.7 Frequency of Claims Receiving Active Physical Therapy Services, 2011/2013–2013/2015 Nonsurgical

Claims Identified as Receiving Longer-Term Opioids / 54

TA.A1 State Prescription Drug Monitoring Programs / 60

TA.A2 Relationship between Statewide Prescription Drug Monitoring Programs Enrollment, Utilization,

and the Overall Use of Opioids and Schedule II Opioids among the 26 Study States / 63

TA.B1 Summary of Medical Treatment Guideline Recommendations for Chronic Opioid Management /

73

TA.C1 Results of Logistic Regression of Longer-Term Opioid Dispensing on Case-Mix Variables, 2013/2015

Claims / 80

TA.C2 Trends in the Prevalence of Longer-Term Dispensing of Opioids, Adjusted and Unadjusted,

2013/2015 / 83

TA.C3 Consistency in the Results between Longer-Term Opioids and Chronic Opioids, 2013/2015 / 84



5

LIST OF FIGURES

A Changes in the Prevalence of Longer-Term Opioid Dispensing between 2010/2012 and 2013/2015 /

9

B Percentage of Claims with Opioids That Received Opioids on a Longer-Term Basis, 2013/2015 / 13

3.1 Percentage of Claims with Pain Medications That Received Opioids, 2013/2015 / 34

3.2 Percentage of Claims with Opioids That Received Opioids on a Longer-Term Basis, 2013/2015 / 35

3.3 Percentage of Claims with Opioids That Did Not Receive Opioids in Three Months Postinjury but

Had Opioids on a Longer-Term Basis, 2013/2015 / 35

3.4 Changes in the Prevalence of Longer-Term Opioid Dispensing between 2010/2012 and 2013/2015 /

37

TA.C1 Assessing Potential Bias of Selecting Claims with More Than 7 Days of Lost Time / 77

TA.C2 Assessing Potential Bias of Selecting Nonsurgical Claims with More Than 7 Days of Lost Time / 78

TA.C3 Assessing Potential Selection Effect on Longer-Term Dispensing of Opioids / 79

TA.C4 Percentage of Claims with Opioids That Were Identified as Receiving Longer-Term Opioids,

Unadjusted and Adjusted Results for 2013/2015 / 82

TA.C5 Percentage of Claims with Longer-Term Opioids and Average MEA per Claim with Opioids,

2013/2015 / 85



6

EXECUTIVE SUMMARY

This report examines the prevalence and trends of longer-term dispensing of opioids in 26 state workers’

compensation systems.1 It also documents how often the services recommended by medical treatment

guidelines are used for monitoring and managing the care delivered to the injured workers who received

opioids on a longer-term basis. The information is useful for (1) state officials who are interested in learning

if the prevalence of long-term opioids is unusually high in their state and what policy tools are available to

reduce unnecessary opioid use, (2) payors and managed care organizations looking to set priorities for opioid

management programs, (3) injured workers and worker advocates looking to understand the use of opioids

and related issues in their state, and (4) providers who wonder what the prescribing norms in their state may

be and what could be done to better manage long-term opioid therapy. This updated study also serves as a

tool to monitor the results of recent policy changes regarding the use and long-term use of opioids.

Long-term opioid therapy for chronic non-cancer pain is a controversial topic, and there is little evidence

about the effectiveness of such treatment on functional recovery or return to work.2 Unnecessary opioid

prescriptions may lead to opioid addiction, overdose, and diversion, which is a top priority public health

topic in the United States. The public concern is also shared by the workers’ compensation community. In

recent years, many states have made legislative or regulatory changes, within and outside workers’

compensation, to address issues related to overuse and misuse of opioids. Some policy changes were also

made at the federal level, including the Centers for Disease Control and Prevention (CDC) guidelines for

prescribing opioids for chronic pain.3

SUMMARY OF MAJOR FINDINGS

Over the study period, we observed considerable decreases in the prevalence of longer-term dispensing of

opioids in many of the states studied.4 For several states, noticeable decreases were first observed in the 2016

1 The prevalence of longer-term dispensing of opioids is measured as the percentage of claims receiving longer-term opioids, i.e., those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury. The 26 states are Arkansas, California, Connecticut, Florida, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maryland, Massachusetts, Michigan, Minnesota, Missouri, Nevada, New Jersey, New York, North Carolina, Pennsylvania, South Carolina, Tennessee, Texas, Virginia, and Wisconsin. These states represent more than 70 percent of the workers’ compensation benefits paid in the United States. Note that the definition of longer-term opioids used in this report is largely consistent with the definition of chronic opioids used in WCRI’s Interstate Variations in Use of Opioids, 4th Edition, which presented the results on chronic opioids for 17 of the 26 states (Thumula, Wang, and Liu, 2017). This report covers all 26 states in terms of longer-term use of opioids. 2 Although several studies have documented some benefits of long-term opioid therapy for limited pain relief (see a more detailed discussion in Wang, Mueller, and Hashimoto, 2011), no studies have been published that support chronic opioid use for improved function or rapid return to work. For patients with occupational injuries, several studies found that a higher use of opioids may lead to addiction, increased disability or work loss, and even death (Kidner, Mayer, and Gatchel, 2009; Franklin et al., 2005; and Volinn, Fargo, and Fine, 2009). 3 See Technical Appendix A for a description of recent policy changes related to opioids. Chapter 3 also provides some background information on policy changes in the states that experienced a considerable decrease in the prevalence of longer-term opioid use. 4 The results are based on nonsurgical claims with more than seven days of lost time that had opioids over the study period from 2010/2012 to 2013/2015. 2013/2015 represents claims with injuries arising from October 1, 2012, to September 30, 2013, with prescriptions filled through March 31, 2015. See Chapter 2 for more details.



7

edition of this study.5 These decreases might have been associated with numerous opioid policies and

initiatives that were in place over the study period that were aimed at addressing issues related to opioid

prescribing and chronic opioid management.6 However, longer-term dispensing of opioids was more

prevalent in Louisiana and several other states in 2013/2015.7 We continued to find that fewer than expected

injured workers who received opioids on a longer-term basis had certain services recommended by treatment

guidelines for chronic opioid management.

TRENDS IN THE FREQUENCY OF CLAIMS WITH LONGER-TERM OPIOIDS

Over the study period from 2010/2012 to 2013/2015, the prevalence of longer-term opioid dispensing

decreased in most study states (Figure A).8 In this report, we highlight the states where the decrease was

significant9 and group the states in terms of the magnitude of the reduction.

Substantial decreases were seen in Kentucky and New York, where the percentage of claims that received

opioids on a longer-term basis decreased more than 4 percentage points.

The decreases in the prevalence of longer-term opioid dispensing were considerable in Kansas,

Massachusetts, Michigan, Minnesota, and Tennessee, with 2–3 percentage point reductions.

Several other states, including California, Florida, and Texas, also saw noticeable decreases in the longer-

term dispensing of opioids, with reductions of 1–2 percentage points.

Note that in the 2016 edition of this study, we reported an increase in the frequency of longer-term

opioid dispensing in Indiana and Wisconsin. With additional data covering one more year, we found that the

increase observed in 2012/2014 did not continue.

5 In the 2016 edition of this study, we saw a noticeable decrease, between 2010/2012 and 2012/2014, in the prevalence of longer-term opioid dispensing in a number of study states, including Maryland, Michigan, New Jersey, New York, North Carolina, and Texas (Wang, 2016). 6 Since there have been many changes in the policies and practices, it is difficult to attribute the reduction in the frequency of longer-term opioid dispensing to any policy initiatives. While examining the impact of changes in opioid prescribing policies and chronic pain management on the prevalence of longer-term opioid use is beyond the scope of this study, we provide the reader with some background information about recent policy changes to facilitate the interpretation of the results, without drawing a causal relationship between specific policy changes and the results. See Chapter 3 and Technical Appendix A for more details. 7 There have been studies examining factors in the early stage of treatment that may be associated with receiving chronic opioid therapy. These factors include early opioid use, the presence of psychological and psychiatric conditions, and substance use disorders (SUDs). These early factors that may likely influence the likelihood of receiving long-term opioids should be addressed in practice to reduce the prevalence of claims receiving long-term opioids. In addition, most treatment guidelines for chronic pain recommend careful screening of patients before prescribing chronic opioid therapy, which is also an important tool to address the prevalence of longer-term opioid use. 8 In the 2016 edition of this report, we observed a noticeable decrease in the prevalence of longer-term opioid dispensing in several states, including Maryland, Michigan, New Jersey, New York, North Carolina, and Texas. Note that in this study, we highlight the states in which we observed a more than 1 percentage point change and the change was statistically significant at the 95 percent confidence level. In the Executive Summary, we only highlight the states where the decrease in the prevalence of longer-term opioid dispensing was substantial (more than 2 percentage points) and statistically significant at the 95 percent level. 9 For the states highlighted, the decrease was statistically significant at the 95 percent confidence level.



8

Figure A Changes in the Prevalence of Longer-Term Opioid Dispensinga between 2010/2012 and 2013/2015

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2010/2012 refers to claims with injuries occurring from October 1, 2009, through September 30, 2010, and prescriptions filled through March 31, 2012. Similar notation is used for other years.

a We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury. See Chapter 2 for more details.

* States with an asterisk (*) are those for which the observed change was statistically significant at the 95 percent confidence level.

The considerable reductions in the prevalence of longer-term opioid dispensing might have been

associated with changes in opioid policies and initiatives aimed at reducing opioid prescriptions. In the past

few years, numerous legislative and regulatory changes in opioid policies have been made at the state level

within and outside workers’ compensation, including

a mandatory check of prescription history in the state prescription drug monitoring program (PDMP)

database at the point of prescribing and dispensing opioids;

treatment guidelines for prescribing opioids and managing chronic opioid therapy;

drug formularies;

mandatory provider education focusing on appropriate opioid prescribing and pain management; and

state laws regulating pain clinics and dispensing of opioids by physicians.

More recently, a number of states passed comprehensive legislation to address opioid overuse prevention

and intervention. Efforts were also made in several states to encourage inter-agency collaborations and

collaborations among state agencies and stakeholders to address opioid issues in a coordinated way.10

Research on opioid policies and their impact has also been conducted to support evidence-based

policymaking.

Here we highlight the states for which we observed a large and significant decrease in the prevalence of

10 More detailed information about opioid policies and initiatives can be found in Chapter 3 and Technical Appendices A and B.

-5%

-4%

-3%

-2%

-1%

0%

1%

2%

KY* NY* TN* MI* KS* MA* MN* AR* CT TX* FL* CA* NV NC MD NJ PA IL MO GA WI IN IA VA SC LA

% P

oin

t C

han

ge



9

opioid use over the study period.11 For these states, we also briefly describe key policy changes over the study

period that might be helpful for interpreting the results.

A substantial reduction in the prevalence of longer-term opioids was seen in Kentucky and New York. In

Kentucky, 6.4 percent of nonsurgical claims with opioids received longer-term opioids in 2013/2015, a

decrease from 11.0 percent in 2010/2012, which translates to a 42 percent reduction in the prevalence of

longer-term opioid dispensing. The figure decreased 4.2 percentage points in New York, from 11.3

percent in 2010/2012 to 7.1 percent in 2013/2015, which translates to a reduction of approximately 37

percent (see Figure A). As a result, the frequency of longer-term dispensing of opioids in these two states

in 2013/2015 was closer to the median of the 26 study states (Figure B), moving from the higher end in

2010/2012. It is worth noting that over the same period, the percentage of nonsurgical claims that

received opioids for pain relief also experienced the largest decrease among the study states—a 16

percentage point decrease in Kentucky and a 9 percentage point decrease in New York.12 The large

reductions in the use and longer-term use of opioids may be associated with the comprehensive reforms

in these two states over the study period. In Kentucky, House Bill 1, which went into effect in July 2012,

includes provisions for pain clinic regulation, rules for opioid prescribing and dispensing, and

mandatory checks of the state PDMP database (Kentucky All Schedule Prescription Electronic Reporting

[KASPER]). New York passed legislative mandates that require prescribers to check the PDMP database

at the point of prescribing opioids. More recent policy changes include the publication of non-acute

medical treatment guidelines. According to the 2013 report by the Trust for America’s Health that

provides ratings on state opioid policies, both Kentucky and New York received a rating of 9 out of 10.13

The frequency of longer-term opioid use also decreased considerably in Kansas, Massachusetts,

Michigan, Minnesota, and Tennessee. The percentage of nonsurgical claims with opioids that were

identified as receiving longer-term opioids decreased more than 2 percentage points between 2010/2012

and 2013/2015 in these states.14 The percentage reductions were approximately 25 to 35 percent. In

recent years, there were many changes and initiatives regarding opioid policies in these states.

Kansas: In 2013/2015, 3.4 percent of the claims with opioids were identified as having longer-term

opioid use, down from 5.6 percent in 2010/2012. The Kansas PDMP, also known as K-TRACS, has

been operational since 2011, and the number of prescribers registered with the PDMP program has

increased steadily, with nearly one-third of all prescribers enrolled as of December 2014.15 Kansas

adopted the Model Policy for the Use of Controlled Substances for the Treatment of Pain,

established by the Federation of State Medical Boards (FSMB).16 The Kansas workers’

11 The states are highlighted and discussed individually if the reduction was more than 2 percentage points and statistically significant at the 95 percent confidence level. 12 These results can be found in Chapter 3. 13 The states with the highest scores were New Mexico and Vermont (neither are included in this study). See Trust for America’s Health (2013). 14 The decreases were statistically significant at the 95 percent confidence level. 15 See the 2016 Pew Charitable Trusts report entitled Prescription Drug Monitoring Programs: Evidence-Based Practices to Optimize Prescriber Use, which is available at http://www.pewtrusts.org/~/media/assets/2016/12/prescription_drug_monitoring_programs.pdf?la=en. 16 The guidelines consist of evaluating the patient, developing a treatment plan, obtaining informed consent and agreement for treatment, periodically reviewing the treatment plan, consulting with the patient, maintaining good medical records, and complying with controlled substances laws and regulations. See FSMB (2013). The Model Policy was being updated as of December of 2016.



10

compensation regulatory agency, the Department of Labor, Division of Workers’ Compensation,

has adopted the Work Loss Data Institute’s Official Disability Guidelines—Treatment in Workers’

Comp (ODG) as the standard of reference for evidence-based medicine used in caring for injured

workers.

Massachusetts: A steady decrease was seen—the percentage of nonsurgical claims with longer-term

opioids decreased from 8.8 percent in 2010/2012 to 6.7 percent in 2013/2015. The state PDMP

program became operational in 2010, and prescribers are required to register with the PDMP

program. In subsequent years, efforts have been made to enhance the utility of the PDMP system

to facilitate online checking of the PDMP database. Effective March 2013, the Massachusetts

Department of Industrial Accidents Office of Health Policies made major revisions to its chronic

pain treatment guidelines and mandated the use of the guidelines in utilization review. Continuing

medical education on opioid prescribing was also required for license renewal, effective February

2012. These policy changes may have contributed to the considerable reduction in the prevalence

of longer-term opioid dispensing in the state. In March 2016, Governor Charlie Baker signed into

law a landmark opioid legislation (House Bill 4056), which is expected to have an impact on opioid

utilization, in response to a continued increase in opioid overdoses.

Michigan: The frequency of longer-term opioid dispensing decreased from 7.5 percent in

2010/2012 to 5.0 percent in 2013/2015. This drop was after an increase of more than 2 percentage

points from 2008/2010 to 2010/2012.17 This considerable decrease coincided with several policy

initiatives over the study period, including third-party payors’ ability to access to the state PDMP

database, the Michigan Workers’ Compensation Agency’s rules aimed at managing opioid

prescriptions beyond 90 days for non-cancer related chronic pain, and changes in the fee schedule

that encourage compliance with the rules for better management of chronic opioid therapy.

Initiatives were also started by the Advisory Committee on Pain and Symptom Management

(ACPSM) in Michigan to improve pain management.18

Minnesota: In 2013/2015, 5.0 percent of the claims with opioids were identified as receiving

longer-term opioids, a decrease from 7.0 percent in 2010/2012 (an approximately 29 percent

reduction). The considerable decrease may be associated with several opioid policies and initiatives

in the state. The state PDMP became operational in 2010, and Minnesota’s legislation only

mandates that prescribers in opioid treatment programs check the PDMP database. Effective July

13, 2015, Minnesota’s administrative rules require compliance with a set of guidelines governing

long-term treatment with opioid analgesic medication for workers’ compensation injuries.

Minnesota’s Opioid Prescribing Work Group (OPWG) published A Protocol for Addressing Acute

Pain, which provides guidelines for prescribing opioids. In May 2016, the legislature passed a law

that requires more medical professionals in Minnesota to sign up for the PDMP to encourage the

use of the PDMP database.19

17 Wang (2014). 18 The ACPSM, charged with studying pain issues in the state and making recommendations to improve the care of patients with pain, developed model core curricula, continuing education (CE) recommendations to provide guidance on required CE hours and content for competent prescribing for the state professional boards, and pain management tool kits. 19 See https://www.revisor.mn.gov/bills/text.php?number=SF1440&version=4&session=ls89&session_year=2016 &session_number=0.



11

https://www.revisor.mn.gov/bills/text.php?number=SF1440&version=4&session=ls89&session_year=2016%20&session_number=0

Tennessee: The frequency of receiving longer-term opioids decreased from 7.2 percent in

2010/2012 to 4.5 percent in 2013/2015, which translates to a decrease of approximately 37 percent.

In Tennessee, prescribers must check the PDMP database when first prescribing opioids and

benzodiazepines for more than seven days and at a three-month interval thereafter if prescribing

continues.20 In November 2012, Tennessee’s legislature passed Senate Bill 3315, which amended

the definition of utilization review to explicitly include Schedule II, III, and IV drugs being used for

pain management. These policy changes and other initiatives in and outside the workers’

compensation system may explain the decrease in longer-term dispensing of opioids. In 2016, the

Bureau of Workers’ Compensation adopted the ODG drug formulary as part of a comprehensive

set of treatment guidelines adopted at the same time.21

Several other states also saw a noticeable decrease in the prevalence of longer-term opioid dispensing

over the study period, including California, Florida, and Texas, as well as Arkansas, Georgia, and Nevada.

In California, Florida, and Texas, the percentage of nonsurgical claims with opioids that received opioids

on a longer-term basis decreased by 1–2 percentage points over the three study years.22 There have been a

number of reforms in each of these states, many of which were implemented after the study period. For

these states, a further reduction in the frequency of longer-term opioid use may be expected. In several

states, the frequency of longer-term opioid use appeared to have peaked in 2011/2013. Among these

states, noticeable reductions in the same measure were seen in Arkansas, Georgia, and Nevada in the

latest two study years. We will continue to monitor the opioid trends for these and other states.

Indiana and Wisconsin: In 2010/2012, 3.8 percent of the nonsurgical claims with opioids received

opioids on a longer-term basis in these two states. The measure increased to 5.7 percent in Indiana and

5.1 percent in Wisconsin in 2012/2014.23 With additional data covering one more year, we found that for

2013/2015 claims, these two states returned to the 2010/2012 level. Indiana and Wisconsin continued to

be among the lower group of states on the prevalence of longer-term opioid dispensing, with 4 percent or

fewer nonsurgical claims receiving opioids on a longer-term basis. We are not aware of any major policy

changes over the study period that might help explain the temporary fluctuation.

PREVALENCE OF LONGER-TERM OPIOID DISPENSING

Among claims with more than seven days of lost time that did not have surgery and received opioid

prescriptions, there was a large difference across the 26 study states in the prevalence of longer-term opioid

dispensing.24

In Louisiana, the prevalence of longer-term dispensing of opioids peaked in 2011/2013, at 18.4 percent.

20 See PDMP Center of Excellence at Brandeis University (2014b) and Clark et al. (2012). 21 The Bureau also subscribes to the recommendations of the State of Tennessee, Department of Health, Chronic Pain Guidelines, the Bureau of Workers’ Compensation Pain Management/Opioid Guidelines Appendix, according to the 2016 International Association of Industrial Accident Boards and Commissions (IAIABC) report on drug formularies. 22 The decreases were statistically significant at the 90 percent confidence level. 23 The results are consistent with what we reported in the previous edition of this study (Wang, 2016). 24 We selected nonsurgical claims with more than seven days of lost time to better monitor trends within a state in the use and longer-term use of opioids and describe interstate variations in the prevalence of longer-term opioid use in a meaningful way. In Chapter 2, we provide the rationale for selecting this subset of claims for the study and discuss, in Technical Appendix C, some potential concerns about this selection and our sensitivity test, based on which we conclude that the claim selection is unlikely to distort the findings in this report.



12

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

KS NJ MO WI IN NV IA AR TN FL MN MI IL MD CT KY VA MA NY GA CA NC TX PA SC LA

% o

f Cla

ims

wit

h O

pio

ids

The figure decreased slightly in the two subsequent years. However, the frequency of longer-term

dispensing of opioids was still high relative to the other study states—approximately one in six injured

workers with opioids was identified as receiving longer-term opioids (Figure B).25

The frequency of longer-term opioid use was also higher in California, Georgia, North Carolina,

Pennsylvania, South Carolina, and Texas,26 when compared with the other states in the study. In these

states, 1 in 10–12 injured workers who received opioids and did not have surgery was identified as

receiving longer-term opioids (Figure B).

Kentucky and New York had substantial reductions in the prevalence of longer-term opioid dispensing,

moving these states closer to the median study state. In 2010/2012, these two states were the second

highest, with 11.0 and 11.3 percent of claims with longer-term dispensing of opioids, respectively. In

2013/2015, the figure decreased to 6.4 percent in Kentucky and 7.1 percent in New York (Figure B,

Figure 3.4, and Table 3.2).

By contrast, about 1 in 25 injured workers with opioids received them on a longer-term basis in Indiana,

Kansas, Missouri, Nevada, New Jersey, and Wisconsin. Figure B Percentage of Claims with Opioids That Received Opioids on a Longer-Term Basis,a 2013/2015

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015. a We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury. See Chapter 2 for more details.

25 It should be noted that our definition of longer-term opioid use is based on the number of visits during which an opioid prescription was filled between the 7th and 12th months postinjury. The number for Louisiana might have been closer to that in New York and Pennsylvania, the next highest states on the same measure, if the definition was based on morphine equivalent daily dose and duration of opioids. However, the difference would not be large enough to change the ranking for Louisiana. See Chapter 3 for more discussion. 26 In Texas, efforts have been made within the workers’ compensation community to address issues related to utilization of opioids and other prescription drugs. A study by the Texas Department of Insurance (TDI) found a substantial decrease in the use of opioids and other not-recommended drugs after the adoption of the guideline-based closed pharmacy formulary (TDI, Texas Workers’ Compensation Research and Evaluation Group, 2013).



13

HOW SERVICES RECOMMENDED BY TREATMENT GUIDELINES WERE USED AMONG CLAIMS WITH LONGER-TERM OPIOIDS

Medical treatment guidelines for chronic opioid management recommend the use of monitoring and

management services, such as periodic drug screening and testing and psychological evaluation and

treatment.27

The frequency of drug testing increased in many study states over the entire study period, based on the

results of 17 states with at least 100 claims with longer-term opioids in the study sample. 28,29 Claim

frequency of drug testing in the latest study year was stable for most states except for a few states where

the use of drug testing changed noticeably.30 For the 2013/2015 nonsurgical claims with longer-term use

of opioids across the states, the claim frequency of drug testing ranged from 22 to 59 percent.

The results on drug testing raise two questions: (1) Are drug testing services provided to all injured

workers who are on chronic opioid therapy and should have drug tests as recommended by most

treatment guidelines? (2) Is there evidence of excessive utilization of drug tests for some injured workers

who received drug testing?

The rate of use for drug testing services appeared to be low, at least for the states where the

percentage of claims with longer-term opioids that had drug testing was lower. Among the

nonsurgical claims receiving longer-term opioids, only 22–26 percent had drug testing services in

Illinois, Massachusetts, and Michigan.31 In these states, the frequency of drug testing was likely

lower than what is recommended by treatment guidelines.

At the same time, we saw unusually frequent drug tests among a small percentage of injured

workers receiving longer-term opioids. For example, the top 5 percent of 2013/2015 claims that

received longer-term opioids had at least 7–10 visits for drug tests in most study states, about 11–

12 visits in the states with the highest rates (Louisiana, Maryland, and Massachusetts) and 3–4

visits in the states with the lowest rates (Michigan and Minnesota).32 The top 5 percent of drug

testing visits (i.e., unique dates), based on the number of different drug tests, had at least 10–13

different drug tests in most study states.33 All chronic pain treatment guidelines recommend a

baseline drug test before the patient is prescribed chronic opioid therapy, and most guidelines

recommend up to four visits for drug tests per year, depending on whether the patient is at low risk

or high risk of opioid misuse. The number of visits for drug testing appeared to be especially high

in Louisiana, Maryland, and Massachusetts.

27 Technical Appendix B provides a summary of guideline recommendations for chronic opioid management. 28 The 17 states selected for reporting the frequency and utilization of drug testing services are California, Connecticut, Florida, Georgia, Illinois, Louisiana, Massachusetts, Maryland, Michigan, New Jersey, New York, North Carolina, Pennsylvania, South Carolina, Tennessee, Texas, and Virginia. Also note that in this study, drug screening and testing services were identified as paid services that were provided in a nonhospital or hospital setting. Chapter 2 provides more details and a list of the drug screening and testing services we captured. 29 We only included these 17 states to ensure reliability of the results. 30 Several states saw considerable increases in the claim frequency of use of drug testing (Connecticut, Florida, Illinois, North Carolina, and Tennessee). The figure decreased in the latest study year in Michigan and Virginia. 31 The results for the other states reported can be found in Chapter 3. 32 The estimated number of visits per claim with longer-term opioids and drug testing was based on a pooled sample of claims with injury years from 2010 to 2012, with an average of 24 months of experience. Five states are not included in this analysis: Arkansas, Iowa, Kansas, Missouri, and Nevada. 33 All 26 states were included in the per-visit analysis of drug testing.



14

We found low use of psychological services among injured workers with longer-term opioids. In 19 out

of 26 states, less than 10 percent of injured workers with longer-term opioids received psychological

evaluations and 5 percent or fewer received psychological treatment in all states except for Iowa,

Minnesota, Texas, and Wisconsin.34 Even in Texas, the state with the highest use of these services, only

about one in three injured workers with longer-term opioids had a psychological evaluation and one in

eight received psychological treatment. Medical treatment guidelines recommend psychological

evaluations for most patients with chronic pain. The low rate of use for psychological evaluations35 may

be associated with under-diagnosis and, consequently, under-treatment for injured workers who had

longer-term opioids and certain psychological comorbidities.

DATA AND APPROACH

This study uses data comprising 432,380 nonsurgical workers’ compensation claims with more than seven

days of lost time36 that received at least one prescription for pain medications paid under workers’

compensation in 26 states. More than 2.3 million prescriptions for pain medications (including opioid and

non-opioid pain medications) were associated with these claims. The claims represent injuries occurring in

three accident years from 2010 to 2013 (e.g., accident year 2010 covering claims with injuries from October 1,

2009, to September 30, 2010). Corresponding to each accident year, we include prescriptions filled through

March 31 of each year from 2012 to 2015. This provides us with an average of 24 months of experience for

each accident year reported. The data sources represent 36–69 percent of workers’ compensation claims in

each study state.

To examine longer-term use of opioids, we identified a subset of claims that had opioids within the first

three months after the injury and had three or more visits to fill prescriptions for opioid pain medications

between the 7th and 12th months after the injury. Using Current Procedural Terminology (CPT®) codes and

the Healthcare Common Procedure Coding System (HCPCS) codes, we also identified several medical

evaluation and treatment services recommended by medical treatment guidelines for chronic opioid

management. We measured the frequency in use of these services to examine compliance with treatment

guidelines.37

LIMITATIONS AND CAVEATS

Several limitations should be noted. First, the claims used for this study may not be representative of all

claims in some states. For a few states, we did not obtain data from some payors with relatively large market

34 The results are based on a pooled sample of claims with longer-term use of opioids across the three years from 2010/2012 to 2012/2014. See Chapter 3 for more details. It should be noted that the measure of the use of psychological evaluations might be understated to the extent that is does not capture, if any, psychological service components of an office visit for evaluation and management (see Chapter 2 for a more detailed discussion). 35 This measure might be understated to the extent that certain psychological evaluations during regular office visits, which are likely to be performed by a medical provider for routine management and evaluation instead of specialty care, are not captured by the procedure codes in the billing. See Chapter 2 for more details. 36 See Chapter 2 for more details about the choice of nonsurgical claims with more than seven days of lost time for the study. 37 See Chapter 2 for more details. Also see Technical Appendix B for a summary of guideline recommendations for chronic opioid management.



15

shares.38 Second, the interstate comparisons in this study were not adjusted for differences across states in the

mix of cases and injury severity.39 We examined the differences in demographics and injury/industry mix, and

the differences in these factors were unlikely to be large enough to affect the comparative results.40 Third, the

reader should be reminded that the prevalence of longer-term opioid dispensing is reported based on

nonsurgical claims with more than seven days of lost time that received opioid prescriptions paid under

workers’ compensation at the time of evaluation.41 While excluding surgical claims from the analysis is likely

to undercount the incidence of longer-term opioid use overall, the exclusion of claims with seven or fewer

days of lost time may overestimate the prevalence of longer-term opioids. Because of this, we caution the

reader that the results cannot be simply extrapolated to all claims in a state. Note that the data used for this

analysis are based on an average of 24 months of experience. Although this may affect the ultimate rankings

for some states on the utilization metrics,42 it is unlikely to affect the metrics on the frequency of use and

longer-term use of opioids.

38 We do not list the states because of confidentiality concerns. 39 As part of sensitivity analysis, we estimated the likelihood of receiving longer-term opioids holding the mix of cases constant. The case mix is a set of variables in proportion that characterize the injured workers, claims, and types of injuries; these variables include age, gender, marital status, weekly wage, tenure with preinjury employer, and high-level injury groups. We did not attempt to control for medical severity using administrative data. See a more detailed discussion in Technical Appendix C. 40 See a more detailed discussion in Chapter 2 and Technical Appendix C. 41 We chose to report the prevalence of longer-term opioid dispensing based on this subset of data for various reasons, which are discussed in Chapter 2 and Technical Appendix C. 42 Certain types of opioids, especially long-acting opioids, are typically used more often at a later stage of medical treatment. As a result, we expect that in some states the utilization of opioids per claim would increase as claims age, especially in those states with higher proportions of longer-term opioid users compared with the typical state.



16

1

INTRODUCTION

Unnecessary opioid prescriptions may lead to overdose, addiction, and diversion, which is a top priority

public health topic in the United States. The public concern regarding overuse and abuse, which often result

in emergency room visits and even overdose deaths, is shared by the workers’ compensation health care

community. Opioids have been widely prescribed for and filled by injured workers—about 55–85 percent of

injured workers with more than seven days of lost time received opioids for pain relief.1 In the past few years,

many states have made legislative or regulatory changes, within and outside workers’ compensation, to

address prevention and intervention of opioid overuse and misuse, including mandatory use of the state

prescription drug monitoring program (PDMP) databases at the point of prescribing and/or dispensing,

provider education, treatment guidelines for prescribing opioids and pain management, drug formularies,

and laws that regulate pain clinics and limit physician dispensing of stronger opioid medications.2 Some

policy changes were also made at the federal level, including the Centers for Disease Control and Prevention

(CDC) guidelines for prescribing opioids for chronic pain.

This report tracks the prevalence and trends of longer-term opioid dispensing for 26 study states. It also

shows how medical treatment guidelines for managing chronic pain and chronic opioid therapy were

followed in practice. The report may serve as a tool to monitor the results of many policy changes for the

states included in the study.

SCOPE OF THE REPORT

This fourth edition report is based on the same framework developed in earlier Workers Compensation

Research Institute (WCRI) studies. It covers 26 states with prescriptions filled through March 2015.3 With

additional data covering one more year, the study period was expanded to cover the time period from

1 The percentage is for claims with or without surgery as a product of how often injured workers received pain medications (85–94 percent) and how often injured workers received opioids when they were prescribed pain medications (59–87 percent) (Wang and Liu, 2011). 2 See Technical Appendix A for a detailed discussion of these changes. 3 The 26 states are Arkansas, California, Connecticut, Florida, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maryland, Massachusetts, Michigan, Minnesota, Missouri, Nevada, New Jersey, New York, North Carolina, Pennsylvania, South Carolina, Tennessee, Texas, Virginia, and Wisconsin. We do not cover Oklahoma because of a representativeness concern. Added to this edition are two new states, Kentucky and Nevada. These states are geographically diverse and represent a significant share of the U.S. population, a wide range of industries, and a variety of benefit structures and policies for workers’ compensation pharmaceuticals. The 26 states include a wide range of states where medical costs per claim were lower, medium, or higher compared with the national average, and they represent various changes in opioid policies.



17

2010/2012 to 2013/2015.4 In addition, this study also examines compliance with medical treatment guidelines

for long-term opioid therapy over the study period. A separate report presents the results of interstate

comparisons and state trends in the overall use of opioids among the same 26 states.5

Several issues are not addressed in either of these reports. These include (1) what factors explain the

interstate variations and trends in the use and longer-term use of opioids and to what extent; (2) whether the

interstate differences are attributable to a relatively small number of heavy prescribers or arise because the

typical providers in the state differ from the typical providers in other states; (3) who are more likely to

receive long-term opioid therapy for what types of injuries; and (4) how the longer-term use of opioids affects

the use of other medical services and return to work.

This report is based on claims with more than seven days of lost time that did not have surgery.6 Future

studies may expand to examine longer-term use of opioids among surgical claims and claims with seven or

fewer days of lost time. An analysis of policies and medical practices regarding psychological and other co-

morbidities among injured workers with long-term opioid use is also important to consider for future

studies.

ORGANIZATION OF THE REPORT

This report is organized into four chapters. Chapter 2 describes the data and methods relevant to this report.

Chapter 3 reports key findings on the longer-term dispensing of opioids and compliance with medical

treatment guidelines for chronic opioid management in each state. Chapter 4 discusses the implications of the

results and the need for future studies.

Technical Appendix A briefly discusses some key factors in the public policy environment that might

influence the utilization and prescribing patterns of opioids. Technical Appendix B provides a summary of

medical treatment guideline recommendations for chronic opioid management. Technical Appendix C

covers several possible technical issues and findings from our sensitivity analysis.

4 The underlying data include prescriptions for nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers’ compensation payor. 2010/2012 refers to claims with injuries occurring from October 1, 2009, through September 30, 2010, and prescriptions filled through March 31, 2012. Similar notation is used for other years. 5 See Thumula, Wang, and Liu (2017). 6 We chose this limited scope for two reasons: (1) the definition and identification of longer-term opioid use are very different between surgical and nonsurgical cases; and (2) claims with more than seven days of lost time are those that account for a small proportion of all claims but represent a large share of workers’ compensation medical costs with significant implications for indemnity, total costs, and worker outcomes.



18

2

DATA AND METHODS

This chapter describes the data and methods relevant to this report. For the reader who is interested in more

information about some specific aspects associated with our study, the technical appendices provide more

details.

DATA AND REPRESENTATIVENESS

This study includes 432,380 nonsurgical claims with more than seven days of lost time that received at least

one prescription for pain medication paid under workers’ compensation, and more than 2.3 pain medication

prescriptions associated with those claims1,2 (Table 2.1). Those claims are from 26 states: Arkansas, California,

Connecticut, Florida, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maryland, Massachusetts,

Michigan, Minnesota, Missouri, Nevada, New Jersey, New York, North Carolina, Pennsylvania, South

Carolina, Tennessee, Texas, Virginia, and Wisconsin. The claims represent injuries occurring in three

accident years from 2010 to 2012 (e.g., accident year 2010 covers claims with injuries from October 1, 2009,

to September 30, 2010). Corresponding to each accident year, we included prescriptions filled through March

31 of each year from 2012 to 2015. This provides us with an average of 24 months of experience for each

accident year reported.

The analysis data were extracted from the WCRI Detailed Benchmark/Evaluation (DBE) database, which

consists of detailed prescription transaction data collected from workers’ compensation payors and their

medical bill review and pharmacy benefit management vendors, representing 36–69 percent of workers’

compensation claims in each state.

1 We chose to use claims with more than seven days of lost time for the analysis for several reasons. First, because these claims provided a similar set of cases across states in terms of disability for work-related injuries, they helped to make the interstate comparisons of the utilization and prescribing patterns more meaningful. Second, these claims received more prescriptions and experienced a wider range of opioid therapy compared with those that had only seven or fewer days of lost time. Focusing on these claims helps identify more meaningful interstate variations in the utilization and prescribing patterns of opioids. Third, the claims with more than seven days of lost time also accounted for the majority of the workers’ compensation medical costs—a key consideration in public policy decisions. It is possible that selecting claims with more than seven days of lost time may filter in a subset of claims that may be more serious for some states and less serious for others. If that occurs, the results of interstate comparisons of the utilization of opioids may be biased. However, we did not see strong evidence suggesting that this was likely to occur in our data (see Technical Appendix C for more details). 2 In this study, we also focus on nonsurgical claims because opioids may be prescribed to patients with surgery for different reasons, especially for post-surgical care. Evaluation of opioid prescriptions may need to take into account the timing in relation to surgery as well as injury severity and case experience. By focusing on nonsurgical claims, we make sure that the results that describe the use and prescribing of opioids are meaningful. Future studies may examine the use of opioids among surgical cases to provide a more complete picture.



19

AR CA CT FL GA IA IL IN KS KY LA MA MD MI MN MO NC NJ NY NV PA SC TN TX VA WI

% of all claims in each state represented by data sources included in the study 39% 43% 66% 37% 40% 38% 50% 44% 41% 44% 39% 52% 43% 44% 46% 44% 40% 60% 36% 49% 41% 45% 45% 69% 57% 36%

Pain medications 2,631 93,724 8,697 38,933 14,722 5,183 24,738 8,809 4,623 6,356 5,510 9,156 11,401 13,324 7,874 8,962 12,328 17,353 4,011 21,878 24,309 6,753 12,228 52,872 8,416 7,588

Opioids 2,253 62,956 5,396 26,827 10,898 3,857 14,257 6,616 3,496 4,546 4,595 6,774 7,236 8,606 6,123 6,228 9,709 9,372 3,134 15,309 16,815 5,341 9,068 40,123 6,587 5,921

Pain medications 13,346 627,980 38,687 201,356 84,407 21,876 105,098 40,289 20,685 36,855 50,208 47,480 43,519 56,840 34,947 32,410 67,781 56,988 18,567 123,320 142,688 39,204 58,227 324,729 42,956 30,386

Opioids 9,240 304,352 21,522 98,019 48,277 13,355 53,302 25,101 13,222 23,304 34,889 31,002 23,590 32,343 23,104 18,941 43,068 26,685 11,436 74,015 85,899 24,577 34,818 192,538 26,566 19,910

Table 2.1 Claims and Prescriptions Included in the Study

Number of nonsurgical claims with more than 7 days of lost time that received prescriptions for …

Among nonsurgical claims with more than 7 days of lost time, number of prescriptions for …

Note: Underlying data are claims that had injuries arising from October 1, 2009, to September 30, 2013, and prescriptions paid for by a workers' compensation payor filled through March 31, 2015.

Definitions:

Pain medications: Prescription medications for pain relief, including opioid and non-opioid medications, including over-the-counter strength pain medications.

Opioids: Opioid analgesics that are often prescribed by physicians for pain relief. Unlike other non-opioid pain medications, opioids are classified at both the federal and state level as controlled substances because they have the potential for producing psychological or physical dependence.



20

The data available for each prescription identify the specific medication prescribed, the date on which the

prescription was filled, amounts charged and paid, the number of pills (for orally-administered opioids), and

the strength of the medication in milligrams. The specific medication prescribed was identified by National

Drug Code (NDC). Note that we have seen improvement in the data for the days of supply, a key element

based on which one can measure the duration of opioid prescriptions and derive the morphine equivalent

daily dose (in conjunction with the quantity and strength of opioid medications). The improvement,

especially among pharmacy-dispensed opioid prescriptions in the latest study year, made it possible to report

these two measures for selected states in another WCRI study (see Thumula, Wang, and Liu, 2017). However,

because the data were not consistently and adequately populated across all data sources for all study states, we

continued to use the empirical definition to report longer-term use of opioids for all 26 states in this study.

It is worth noting that there was substantial variation across states in the percentage of nonsurgical

workers’ compensation claims with more than seven days of lost time that received at least one prescription

paid under workers’ compensation (from 28 percent in Massachusetts to 70 percent in California in

2012/2014). Several reasons may explain this variation. For example, some initial prescriptions may be paid

for by non-workers’ compensation payors,3 and some patients may be using over-the-counter medications

that they already have to treat their work injury. We tested the magnitude of the potential bias introduced by

these missing prescriptions on interstate variations of our key measures and found that this was unlikely to be

material in how we characterized a state in this analysis as low, medium, or high.4

Unlike other WCRI benchmark reports, the claims included in this study may not necessarily be

representative of the total population of claims in some states. This occurs for two reasons. First, the

reporting of detailed prescription data was less complete than other benchmarking data for a few data sources

in some states, which resulted in additional exclusions of data sources from this study.5 This occurred when a

data source in a state did not have complete and adequate data on NDCs and quantities for prescriptions—

the two data elements critical for constructing benchmark metrics for this study. Although we included

several additional data sources in this edition, the exclusions of some data sources may affect the

representativeness of the data if the claims from those excluded data sources were very different in some way.

Second, we did not obtain data from one or more important data sources for a few states, which may affect

the representativeness of our data for these states.6 Because of potentially less representative data for this

study compared with the CompScope™ benchmarks reports, we did not apply the same market segment

weightings. For this study, we combined the claims in the voluntary, residual, and self-insured markets so that

the weighting was applied only for the states with a state fund for the market segment representation.

3 This is consistent with a study of workers’ compensation cases of New York state employees, all of whom were covered by group health insurance that included prescription coverage (Stapleton et al., 2001). By combing the record of the state fund and group health insurer’s interviews with workers, the study found that 21 percent of all drug expenditures for those injured workers were paid by the state fund, 69 percent by group health insurance, and 9 percent by the worker without reimbursement. See a more detailed discussion in Wang and Liu (2011). 4 To test the bias we assumed that the rate of filling opioid prescriptions among claims without prescriptions was similar to the rate among claims with one prescription. Similar findings were observed when we set the rate of opioid fills to the rate among claims with one or two prescriptions. 5 To ensure the accuracy of the utilization metrics used for interstate comparisons and the trend analysis, we excluded a few data sources in some states whose data showed unreasonably high or low values for certain measures. 6 We do not provide more detailed information regarding the states and data sources due to confidentiality concerns.



21

IDENTIFYING OPIOID PRESCRIPTIONS

We identified opioid prescriptions based on the therapeutic classification developed by Medi-Span®,7 which

we linked to prescription transactions through NDCs. There are five schedules of controlled substances,

classified by the Drug Enforcement Administration (DEA) under federal law, which are based on a drug’s

medical usefulness and abuse potential. We also identified the schedule associated with each opioid

prescription based on an indicator in the Medi-Span® database. Table 2.2 provides the definitions of the five

schedules and examples of specific drugs classified in each schedule as of 2015. For example, oxycodone

(OxyContin®) and oxycodone-acetaminophen (Percocet®) are classified as Schedule II opioids. Note that

hydrocodone-acetaminophen (brand names including Vicodin® and Lortab®) and other hydrocodone-

combined products were reclassified by the DEA as Schedule II opioids, effective October 6, 2014,8 because of

the widespread use of the drug and the potential for abuse, addiction, and overdose. These hydrocodone-

combined products were classified as Schedule III opioids over the earlier years of the study period. Tramadol

(Ultram® and Ultracet®) was classified as a Schedule IV controlled substance, effective August 18, 2014.9 The

change in scheduling tramadol also occurred during the study period.

Table 2.2 Federal Classification of Controlled Substances as of 2015

Schedule Criteria for Classification Examples of Specific Drugs

Schedule I The drug or other substance has high potential for abuse and has no currently accepted medical use in treatment in the United States.

Heroin, marijuana,a lysergic acid diethylamide (LSD), and methaqualone

Schedule II The drug or other substance has high potential for abuse, which may lead to severe psychological or physical dependence, and has a currently accepted medical use in treatment in the United States.

Morphine (Avinza®), fentanyl (Duragesic®), oxycodone HCL (OxyContin®), oxycodone-acetaminophen (Percocet®), methadone,b and hydrocodone-acetaminophen (Vicodin®, Lortab®)c

Schedule III The drug or other substance has less potential for abuse than the drugs or substances in Schedules I and II and has a currently accepted medical use in treatment in the United States. Abuse of the drug or substance may lead to moderate or low physical dependence or high psychological dependence.

Codeine-acetaminophen (Tylenol® with codeine), buprenorphine (Suboxone®)

Schedule IV The drug or other substance has a low potential for abuse relative to drugs in Schedule III and has a currently accepted medical use in treatment in the United States. Abuse of the drug or substance may lead to limited physical or psychological dependence relative to the drugs or other substances in Schedule III.

Tramadol (Ultram®),d propoxyphene-N w/APAP (Darvon®)e

Schedule V The drug or other substance has a low potential for abuse relative to the drugs or other substances in Schedule IV and has a currently accepted medical use in treatment in the United States. Abuse of the drug or substance may lead to limited physical dependence relative to the drugs or substances in Schedule IV.

Cough medicine with codeine (Robitussin®AC)

continued

7 According to Medi-Span®’s Therapeutic Classification System, a hierarchical classification scheme, the first two digits of the 10-digit generic product indicator classifies general drug products. We identified opioid prescriptions based on drug group 65 for opioid analgesics. See Medi-Span® (2005). 8 The final rule reclassifying hydrocodone-combination products can be found at http://www.gpo.gov/fdsys/pkg/FR-2014-08-22/pdf/2014-19922.pdf. 9 The final rule placing tramadol into Schedule IV is available at http://www.gpo.gov/fdsys/pkg/FR-2014-07-02/pdf/2014-15548.pdf.



22

Table 2.2 Federal Classification of Controlled Substances as of 2015 (continued)

a Marijuana is listed as a Schedule I controlled substance at the federal level, along with other opioids that have high potential for abuse and are not for medical use. However, this drug has been recognized for medical use, by laws, in 23 states plus the District of Columbia in the United States (information available at http://medicalmarijuana.procon.org/view.resource.php?resourceID=000881).

b Methadone may be prescribed as a Schedule II analgesic for chronic pain because it is inexpensive. However, its use has been discouraged because of a high risk of overdose death. The drug can also be used for weaning the patient from high dose opioids, but is less likely to be present in our data because, under the Controlled Substances Act, it is not lawful to prescribe opioid drugs for the purpose of detoxification of opioid addiction without being registered as a Narcotic Treatment Program (NTP). NTPs may only use drugs approved for this purpose, such as methadone, and must comply with federal and state methadone program regulations. c The hydrocodone-combined drug products were classified as Schedule III opioids during most of the study period. In October 2014, the Drug Enforcement Administration moved hydrocodone-combined products, including Vicodin® and Lortab®, to Schedule II, the category of medically accepted drugs with the highest potential for abuse, mainly because of the rise in hydrocodone abuse and trafficking in the last several years. d In August 2014, the Drug Enforcement Administration scheduled tramadol products (including Ultram® and Ultracet®) as Schedule IV controlled substances. e Darvon® was voluntarily recalled from the market due to safety reasons in November 2010.

Source: Drug Enforcement Administration, U.S. Department of Justice. 2015. Available at http://www.justice.gov/dea/pr/multimedia-library/publications/drug_of_abuse.pdf.

IDENTIFYING CLAIMS RECEIVING LONGER-TERM OPIOIDS

To examine chronic opioid therapy and management, we identified claims receiving longer-term opioids as

those that had opioids within the first three months after the injury and had three or more visits10 to fill

opioid prescriptions between the 7th and 12th months after the injury. This subset of claims was identified

based on the detailed transaction data for opioid prescriptions filled over the specified period of time. With

the assumption that one opioid prescription likely represents at least a 30-day supply of opioids, this

empirical definition is largely consistent with a widely-agreed on definition for long-term opioid use that is

based on morphine equivalent daily dose and duration of opioid prescriptions.11 Although we have seen

improvement in the completeness of the key data fields, the data were still less adequate among physician-

dispensed prescriptions. For this reason, we continue to report the prevalence and trend of longer-term

opioid dispensing based on our empirical definition (described above) for all 26 states.

Since the claims receiving longer-term opioid prescriptions were identified based on the number of

opioid prescriptions, not morphine equivalent daily dose and duration, we might have identified

proportionally more cases as having longer-term opioids for the states where stronger opioids were used

rarely but weaker strength opioids were more often prescribed.12 However, even for weaker strength opioids,

three or more prescriptions over a six-month period (following the initial six months of treatment) would

normally be considered, clinically, as having longer-term opioids.

We also identified a smaller percentage of nonsurgical claims with opioids that did not receive opioids

10 The number of visits here is the number of unique dates on which at least an opioid prescription was filled. 11 The assumption we made about a 30-day supply seems reasonable, on average, based on what we saw in the data on the quantities of the prescriptions filled after six months postinjury. Several studies outside workers’ compensation have examined long-term or high-dose use of opioids. These studies identified cases with long-term use similarly to this study in terms of timing and duration of opioid use (Sullivan et al., 2008; Morasco et al., 2010; Braden et al., 2010). Because our definition is based on the number of fill dates of opioid prescriptions rather than days of supply, we labeled the category we identified as longer-term rather than long-term. Future studies will revisit this definition once we have consistently available data on days of supply. 12 An alternative definition for claims receiving longer-term opioids would be those claims whose daily dosage, during the 7th through 12th months after the injury, exceeded 30 milligrams of morphine equivalent opioids, which is the minimum daily dosage for chronic opioid therapy as suggested by the Canadian guidelines (National Opioid Use Guideline Group, 2010). Although this definition takes into account the strength of opioids, thus making the results more comparable, data limitations prevented us from using it.



23

within three months postinjury, but otherwise exhibited the same pattern of longer-term opioids as described

above. Several reasons may explain this subset of claims. For example, some injured workers might have late

onset pain and start pain treatment much later in time.13 These cases should not be considered as having

longer-term opioids. If a significant number of such cases are in this subset of claims, including these claims

would overstate the prevalence of longer-term opioid dispensing. In addition, some physicians who follow the

current practice guidelines may not prescribe opioids early in the course of treatment, but use it later when

other treatments (e.g., non-opioid pain medications, over-the-counter painkillers, or other services that may

be helpful for pain relief) have failed. It is also possible that some injured workers received opioids as initial

treatment but the opioid prescriptions were not paid for by a workers’ compensation payor. Since we do not

see a clear pattern for this group of claims, and there is a potential concern of overstating the prevalence of

longer-term opioid prescriptions, we separated the two types of longer-term opioids and focused on the first

type, which received opioids early.14

For the analysis of longer-term dispensing of opioids, we included a small percentage of claims with an

unusually high amount of opioids.15 We randomly reviewed some of the cases in this category and did not

find evidence suggesting data issues in the drug transactions for these claims. Our review of the detailed

transaction data suggested that these were the prescriptions (often with large quantities) for the same or

different opioid medications that were filled by the injured workers at a high frequency over a longer period

of time. Since heavy and prolonged use of opioids is an important part of the issues related to longer-term

dispensing of opioids, we included these claims in the analysis.16

SERVICES RECOMMENDED FOR CHRONIC OPIOID MANAGEMENT BY MEDICAL GUIDELINES

In this report, we also examine the use of some key services recommended by medical treatment guidelines

for chronic opioid management, including drug screening testing, psychiatric evaluation and treatment, and

active physical therapy.17 We identified these recommended services using CPT and Healthcare Common

Procedure Coding System (HCPCS) codes, which are listed in Table 2.3.

We identified the guideline recommended services as those paid for by a workers’ compensation payor

regardless of whether the service was provided in a hospital or nonhospital setting. The drug testing codes

13 Clinically, it is rare to see patients who do not have pain symptoms at the initial stage of the treatment who later develop pain for the same injury. We do not have an estimate of the percentage of cases with opioids that might have late onset pain. However, the physician co-authors of the 2012 study (Wang, Hashimoto, and Mueller) believe that it is rare for clinicians to see patients with late onset pain. 14 The claims with longer-term opioids with early use of opioids are referred to as claims receiving longer-term opioids or Type I claims. The group of claims that had the pattern of longer-term opioids but did not have opioids within three months after the injury is referred to as Type II claims in our discussion. In the analysis presented in Chapter 3, we focus on Type I claims. 15 The morphine equivalent amount of opioids was considered unusually high for a claim if the estimated daily dosage (i.e., the total amount of morphine equivalent opioids received by the claim divided by the duration of receiving opioid prescriptions) for the claim exceeded 120 milligrams of morphine equivalent opioids per day. The 120-milligram threshold is the maximal daily dosage typically recommended by most guidelines (e.g., Oregon guidelines for prescribing narcotics [Oregon Health and Science University, 2006]). 16 Note that in WCRI’s Interstate Variations in Use of Opioids, 4th Edition, these claims with an unusually high amount of opioids were excluded from the interstate comparison and trend analysis of the overall utilization of opioids. See more detailed descriptions in Thumula, Wang, and Liu (2017). 17 Technical Appendix B provides a summary of the general recommendations for chronic opioid management from several widely-accepted treatment guidelines at the national and state levels.



24

identified are those that have been adopted by most payors since 2010.18 Note that CPT 2015 made major

changes to the coding of drug testing services, with the intention to adequately differentiate specific drug tests

and describe individual drug testing procedures. These new CPT codes are included in the identification

program for drug testing services.

In this edition of the study, we continue to discuss unusually frequent drug tests among a small

percentage of claims with longer-term opioid prescriptions.19 Due to sample size concerns for the cohort of

claims with longer-term opioids that received drug testing, we pooled the claims across four injury years from

2010 to 2013 (with an average of 24 months of experience). By doing so, we highlight the patterns of

unusually frequent use of drug tests among a small number of claims, for all except a few states.20 We

illustrate the increases in the claim frequency and per-claim utilization and costs of the drug testing services,

for selected states, to highlight this emerging issue. More rigorous analysis with additional data is needed to

examine the patterns in use of drug testing services, for claims receiving longer-term opioids.

For psychiatric and psychological services recommended by treatment guidelines, there have also been

major revisions and additions to the CPT coding since 2013, which is reflected in the revised mappings for

this edition (see Table 2.3). In this edition, we added several new CPT and HCPCS codes for psychological

testing. Because these codes are either rarely seen in the data or most likely used with other codes, this

addition did not result in a material difference from what was reported in the 2016 report in the frequency of

claims with psychological evaluation or psychological treatment/report. Note that most CPT and HCPCS

codes identified for psychological and psychiatric evaluations and treatment are likely to be billed by

providers with a specialty in psychological and psychiatric services (e.g., psychologists, psychiatrists, social

workers, and other health care professionals). Since a treating physician may perform an initial psychological

evaluation or provide a consultation during a regular office visit, the frequency of psychological services may

be understated to the extent that these office visits with an element of psychological services were not

captured. Because the low frequency in use of psychiatric and psychological services was observed based on a

small sample of claims with longer-term opioid prescriptions that received the services for some states, we

pooled the claims across three injury years from 2010 to 2013 (with an average of 24 months of experience) to

report the use of these services. By doing so, we highlight the likely low use of these services recommended by

treatment guidelines for chronic opioid management across the states studied.

Active physical therapy includes the therapeutic exercises and activities that encourage patient

participation and promote functional recovery (Table 2.3). These exercises and activities are recommended

by guidelines for chronic opioid management as part of comprehensive treatment for patients with chronic

pain, which may be supervised or attended by a medical provider or self-directed. Since self-directed exercises

are not considered treatment, we do not observe these activities. However, the education and training

regarding the exercises are likely to be part of therapeutic exercises supervised or attended by the provider, so

we are likely to capture the claim frequency of having these exercises, based on the set of active physician

therapy codes we identified.

18 The drug testing codes added since 2011 include G0431 and G0434. More details can be found at https://www.cms.gov/Outreach-and-Education/Medicare-Learning-Network-MLN/MLNMattersArticles/downloads/SE1105.pdf. 19 This was first added in the 2016 report (Wang, 2016). 20 For Arkansas, Iowa, Kansas, Missouri, and Nevada, the sample size does not support a reliable analysis even after we pool the sample claims across the three injury years from 2010 to 2012.



25

MEASURING FREQUENCY OF CLAIMS RECEIVING OPIOIDS AND LONGER-TERM OPIOIDS

Several metrics were used in this report to measure the frequency of claims receiving opioids and longer-term

opioids, among claims with more than seven days of lost time that did not have surgery but received pain

medication. The measures include the percentage of claims with pain medications that had opioids21 and the

percentage of claims with opioids that were identified as receiving longer-term opioid prescriptions. We

chose to report the prevalence of opioid dispensing and longer-term dispensing based on this subset of claims

for several reasons. First, the patterns of receiving longer-term opioids are very different between surgical and

nonsurgical claims. Measuring longer-term use separately for surgical and nonsurgical claims helps to

produce clinically meaningful results. Second, since nonsurgical claims are mostly claims with

musculoskeletal injuries that tend to be less serious, compared with surgical cases, analyzing nonsurgical

claims can help identify potential issues with opioid utilization and also help to monitor the results of opioid

policy changes.22 Future studies may examine surgical claims to provide a full picture of opioid utilization.

Third, claims with more than seven days of lost time are those that account for a small proportion of all

claims but represent a large share of workers’ compensation medical costs with significant implications for

indemnity, total costs, and worker outcomes. Focusing on this set of claims also helps to capture interstate

variations and identify states that are likely to have more prevalent dispensing and longer-term dispensing of

opioids. Fourth, we focus on claims with pain medications because receiving pain medications may be used to

indicate that the injured worker suffered from pain, for which opioids and non-opioid pain medications were

prescribed. One may be concerned about potential bias of the comparative results due to the claim selection,

which is addressed later in this chapter and in Technical Appendix C. The metrics of opioid dispensing and

longer-term dispensing, and measures of claims receiving guideline-recommended services, were constructed

based on a weighting method, so the results reflect the claim experience in each state for all market segments

included.23

The measures to categorize use of opioids are based on claims with pain medications, including both

opioids and prescription non-opioid pain medications. It should be noted that over-the-counter non-

prescription pain medications are not included in our data. Because of this, one may be concerned that if

practice patterns varied widely across the states in the use of prescription versus non-prescription pain

medications, it might imply a variable level of severity in the claims included in the study. However, the use of

21 By doing so, we focus on injured workers who received opioids and/or non-opioid pain medications for pain relief. Since the percentage of injured workers receiving at least one prescription varied by state and over time due to certain external factors (e, g., variable coverage of health insurance outside workers’ compensation systems), claims with a prescription paid under workers’ compensation provide a robust base to measure the frequency of claims receiving opioids. 22 In this study, we also focus on nonsurgical claims because opioids may be prescribed to patients with surgery for different reasons, especially for post-surgical care. We define nonsurgical claims as claims that did not have a major surgery during the study period. Major surgery is a WCRI-defined service group that is a subset of the surgery section of the CPT manual. This service group includes invasive surgical procedures, as opposed to surgical treatments and pain management injections (which are also included in the surgery section of the CPT manual). The most frequent surgeries in this service group include (but are not limited to) arthroscopic surgeries of the shoulder or knee, laminectomies, laminotomies, discectomies, lumbar fusion, carpal tunnel surgeries, neuroplasty, and hernia repair. Evaluation of opioid use may need to take into account the timing in relation to surgery as well as injury severity and case experience. By focusing on nonsurgical cases, we make sure that the results that describe the use and prescribing of opioids are meaningful. However, in doing this, one may be concerned that states with higher surgery rates would have fewer serious cases among nonsurgical claims, thus distorting the comparisons. We assessed the extent of this potential concern and concluded that selecting nonsurgical cases is unlikely to bias the results of interstate comparisons. See Technical Appendix C for a more detailed discussion. Future studies may examine the use of opioids among surgical cases to provide a more complete picture. 23 We included the voluntary market, the residual market, the self-insured market, and state funds where they exist. For this study, we combined the voluntary market, residual market, and self-insured claims for weighting purposes.



26

prescription and non-prescription pain medications should not be considered a marker for severity for

several reasons. First, non-prescription pain medications can be taken at varying dosages, depending upon

the degree of pain. Second, a physician, without regard to the injury severity, may prescribe a pain medication

to a patient who is less able to pay the out-of-pocket cost of a non-prescription pain medication. Third,

physicians may prescribe and dispense pain medications at their offices for economic reasons, also regardless

of injury severity. In addition, we do not believe that variations in the use of non-prescription pain

medications should be a serious concern because most of the claims with more than seven days of lost time

received prescriptions for pain medications (85–94 percent).24 Table 2.3 CPT4 and HCPCS Codes Used to Identify Specific Services Recommended for Managing Long-Term Use of Opioids

CPT or HCPCS Code Definition

Drug screening

New codes for presumptive (immunoassay) drug test codes:

80301 Drug screen, presumptive, single drug class from drug class A

80302 Drug screen, presumptive, single drug class from drug class B

Active codes for drug screening during the study period:

80100 Drug screen, qualitative; multiple drug classes chromatographic method, each procedure

80101 Drug screen, qualitative; single drug class method (e.g., immunoassay, enzyme assay), each drug class

80102 Drug confirmation, each procedure, used to confirm each drug class

80103 Drug testing in deceased

80104 Drug screen, multiple

Drug testing

New codes for definitive drug test codes:

80320 Alcohol(s)

80321, 80322 Alcohol biomarkers

80323 Alkaloids, NOS

80324–80326 Amphetamines

80329–80331 Analgesics, non-opioid

80332–80334 Antidepressants, serotonergic

80335–80337 Antidepressants, tricyclic

80338 Antidepressants, NOS

80339–80341 Antiepileptics, NOS

80342–80344 Antipsychotics, NOS

80345 Barbiturates

80346–80347 Benzodiazepines

80348 Buprenorpine

80349 Cannabinoids, natural

80350–80352 Cannabinoids, synthetic

80353 Cocaine

80354 Fentanyls

80355 Gabapentin

80356 Heroin metabolites

80357 Ketamine and norketamine

80361 Opiates

80362–80364 Opioids and opiate analogs

continued

24 The percentage of claims with prescriptions that had prescriptions for pain medications was reported in Wang and Liu (2011).



27

Table 2.3 CPT4 and HCPCS Codes Used to Identify Specific Services Recommended for Managing Long-Term Use of Opioids (continued)


80358 Methadone

80359 Methylenedioxyamphetamines

80360 Methylphenidate

80365 Oxycodone

80366 Pregabalin

80368 Hypnotics, sedative (non-benzodiazepines)

80369–80370 Skeletal muscle relaxants

80371 Stimulants, synthetic

80372 Tapentadol

80373 Tramadol

80374 Stereoisomer (enantiomer) analysis

80375–80377 Drug(s) not otherwise specified

Active codes for drug testing during the study period:

80154 Benzodiazepines

80184 Phenobarbital

80299 Quantitation of drug, not elsewhere specified

82055 Alcohol (ethanol); any specimen except breath

82065 Urine alcohol chemical

82075 Alcohol (ethanol); breath

82101 Alkaloid, urine, quantitative

82145 Amphetamine or methamphetamine

82205 Barbiturates, not elsewhere specified

82486 Chromatography, qualitative, column (e.g., gas liquid or HPLC), analyte not elsewhere specified

82487 Chromatography, quantitative, column (e.g., gas liquid or HPLC); paper, 1-dimensional, analyte not elsewhere specified

82488 Chromatography, quantitative, column (e.g., gas liquid or HPLC); paper, 2-dimensional, analyte not elsewhere specified

82489 Chromatography, quantitative, column (e.g., gas liquid or HPLC); thin layer, analyte not elsewhere specified

82491 Chromatography, quantitative, column (e.g., gas liquid or HPLC); single analyte not elsewhere specified, single stationary and mobile phase

82492 Chromatography, quantitative, column (e.g., gas liquid or HPLC); multiple analytes, single stationary and mobile phase

82520 Chromatography, quantitative, column (e.g., gas liquid or HPLC); cocaine or metabolite

82541 Column chromatography/mass spectrometry, analyte not elsewhere specified; qualitative, single stationary and mobile phase

82542 Column chromatography/mass spectrometry, analyte not elsewhere specified; quantitative, single stationary and mobile phase

82646 Dihydrocodeinone

82649 Dihydromorphinone

83520 Immunoassay, not otherwise specified (DIA)

83805 Assay of meprobamate (anti-psychotic medication)

83840 Methadone

83912 Molecular diagnostics; interpretation and report

83921 Organic acid, single, quantitative

83925 Opiates (e.g., morphine, meperidine), each procedure

83992 Phencyclidine (PCP)

84022 Phenothiazine (anti-psychotic medication)

G0430a Drug screen, qualitative; single drug classes by high complexity test method (e.g., immunoassay, enzyme assay), per patient encounter

G0431a Drug screen, qualitative; multiple drug classes by high complexity test method (e.g., immunoassay, enzyme assay), per patient encounter

G0434 Drug screen, qualitative; multiple drug classes other than chromatographic method, each patient encounter

continued



28



Psychological and psychiatric evaluations

90801 Psychiatric diagnostic interview examination

90802 Interactive psychiatric diagnostic interview examination using play equipment, physical devices, language interpreter, or other mechanisms of communication

90791, 90792 Psychiatric diagnostic evaluation with or with medical

90785

Interactive psychiatric diagnostic interview exam using play equipment, physical devices, language interpreter, or other mechanisms of communication (add-on code for 90791 or 90792 or other psychotherapy codes)

90805, 90817 Individual psychotherapy with evaluation and management, 20–30 minutes



90811, 90824 Interactive individual psychotherapy with evaluation and management, 20–30 minutes

90813, 90827 Interactive individual psychotherapy with evaluation and management, 45–50 minutes

90815, 90829 Interactive individual psychotherapy with evaluation and management, 75–80 minutes 90833, 90835, 90836, 90838 Psychotherapy with evaluation and management (new codes since 2013)

90862 Pharmacologic management

96101–96105 Psychological testing performed by psychologist, physician, qualified health care professional/technician, or computer

96110, 96111 Developmental testing; limited/extended, with interpretation and report

96116 Neurobehavioral status exam (clinical assessment of thinking, reasoning, and judgement), per hour of the psychologist's or physician's time, both face-to-face time with the patient and time interpreting test results

96118 Neuropsychological testing, per hour of the psychologist's or physician's time, both face-to-face time with the patient and time interpreting test results

96119 Neuropsychological testing, with qualified health care professional interpretation and report, administered by technician, per hour

96120 Neuropsychological testing, administered by a computer, with qualified health care professional interpretation and report

96150

Health and behavior assessment (e.g., health-focused clinical interview, behavioral observations, psychophysiological monitoring, health-oriented questionnaires), each 15 minutes face-to-face with the patient, initial assessment

96151

Health and behavior assessment (e.g., health-focused clinical interview, behavioral observations, psychophysiological monitoring, health-oriented questionnaires), each 15 minutes face-to-face with the patient, re-assessment

H0002, H0004 Behavioral health screening/treatment (rare in data)

H0031, H0032 Mental health assessment/service plan by a non-physician

Psychological and psychiatric treatment and report

90804, 90816 Individual psychotherapy, 20–30 minutes



90810, 90823 Interactive individual psychotherapy, 20–30 minutes



90832 Psychotherapy, 30 minutes (new code since 2013)



90839, 90840 Psychotherapy for crisis

90846, 90847, 90849 Family psychotherapy

90853 Group psychotherapy

90857 Interactive group psychotherapy (old code prior to 2013; use 90853 after 2013)

90857–90876 Individual psychophysiological therapy incorporating biofeedback training by any modality (face-to-face with the patient), with psychotherapy (depending on length of service)

continued



29



90882 Environmental intervention for medical management purposes on a psychiatric patient's behalf with agencies, employers, or institutions

90887 Interpretation or explanation of results of psychiatric, other medical exam and procedures, or other accumulated data to family or other responsible person or advising them how to assist the patient

90899 Unlisted psychiatric service or procedure

96152 Health and behavior intervention, each 15 minutes, face-to-face; individual

96153 Health and behavior intervention, each 15 minutes, face-to-face; group (2 or more patients)

96154 Health and behavior intervention, each 15 minutes, face-to-face; family (with the patient present)

96155 Health and behavior intervention, each 15 minutes, face-to-face; family (without the patient present)

H codes Behavioral or mental health services, psychiatric supportive treatment, psychiatric rehabilitation services in various settings (extremely rare in data)

Active physical therapy

97110 Therapeutic procedure, one or more areas, each 15 minutes; therapeutic exercises to develop strength and endurance, range of motion and flexibility

97112

Therapeutic procedure, one or more areas, each 15 minutes; neuromuscular reeducation of movement, balance, coordination, kinesthetic sense, posture, and/or proprioception for sitting and/or standing activities

97113 Therapeutic procedure, one or more areas, each 15 minutes; aquatic therapy with therapeutic exercises

97116 Therapeutic procedure, one or more areas, each 15 minutes; gait training (includes stair climbing)

97150 Therapeutic procedure(s), group (2 or more individuals)

97530 Therapeutic activities, direct (one-on-one) patient contact by the provider (use of dynamic activities to improve functional performance), each 15 minutes

97545 Work hardening/conditioning; initial 2 hours

97546 Work hardening/conditioning; each additional hour (list separately in addition to code for primary procedure)

a New HCPCS codes that have been used by the Centers for Medicare & Medicaid Services to replace the CPT codes since 2010.

Key: CPT: Current Procedural Terminology; FDA: Food and Drug Administration; HCPCS: Healthcare Common Procedure Coding System; HPLC: high-performance liquid chromatography; NOS: not otherwise specified.

Source: American Medical Association, 2006.

SENSITIVITY ANALYSIS FOR CLAIM SELECTION AND CASE MIX

Nonsurgical claims with more than seven days of lost time that received opioids were used for the analysis.

Since the selection is based on three variables reflecting the differences across states in claim type and how

medical services were delivered to injured workers, one may be concerned that such a selection may bias the

results of interstate comparisons if more severe cases were selected for some states and less severe cases were

selected for others.25

One way to assess the existence and extent of this potential selection issue is to examine how a selection

variable is correlated with key utilization measures among the subset of cases selected. For example, if a lower

proportion of claims is selected in a state, the subset of claims selected in the state may have proportionally

more serious claims compared with a state with a higher proportion of selected claims and, therefore, have a

higher utilization of opioids. In this case, the selection of claims from a rather similar claim set would

25 This is a valid concern. At the state level, variations in the mix of cases and injury severity are relatively small, based on the findings from the CompScope™ reports on the result of case-mix adjustment and the worker outcomes reports on the interstate variation of injury severity (Yang et al., 2009; Belton and Liu, 2009). Because the proportion of the claims selected by each of these variables varied widely across the states, the selection of a subset of claims at each step might change the mix of cases and severity. See Technical Appendix C for more details.



30

introduce differences in the mix of cases and severity across states. Without clear evidence of such a

correlation, the selection is unlikely to be a serious concern. If this occurs, however, one has to assess how

sensitive the results are to potential selection. We looked at the correlation at three different points of

selection: (1) claims with more than seven days of lost time, (2) nonsurgical claims, and (3) claims that

received opioids. The results of our analysis suggest that the potential bias due to the selection of the subset of

claims is unlikely to be a serious concern.

For nonsurgical claims with more than seven days of lost time that received opioids, we report the

prevalence or claim frequency of longer-term opioid dispensing, which is an unadjusted measure not

controlling for selection and case mix. We did a correlation analysis, similar to the one above, on the

percentage of claims with opioids that received opioid prescriptions on a longer-term basis to see whether it

could be used to characterize the prevalence of longer-term opioid dispensing across the study states (see

Figure TA.C3). The results suggest that states with a higher percentage of claims with opioids, which resulted

in a less serious mix of claims, had proportionally more claims identified as receiving longer-term opioid

prescriptions.26 This implies that holding severity and case mix constant would make the percentage of claims

with longer-term opioid use even higher, which strengthens the comparative results in the unadjusted

measures. We also did a sensitivity analysis based on a logistic regression on the frequency of longer-term

opioid use across the states and over time within a state. The regression-based results (Figure TA.C4; Tables

TA.C1–C2), although they adjust away some of the differences, are largely consistent with the results for the

unadjusted measures.

LIMITATIONS AND CAVEATS

Two limitations should be noted. First, unlike other WCRI benchmark reports (the CompScope™ series, for

example), the claims used for this study may not be necessarily representative of all claims in some states for

the reasons we discussed earlier in this chapter.

Second, unlike other WCRI benchmark reports, the benchmark metrics reported in this study are not

adjusted for differences in case mix and injury severity across states. However, we examined the differences in

demographics and injury/industry mix and did a sensitivity test based on a logistic regression on the

percentage of claims with opioids that were identified as receiving longer-term opioids. The results confirmed

the characterization of the states in terms of high, medium, and low on the prevalence of longer-term opioid

dispensing. The differences in case mix across the states and over time within a state were not large enough to

make material differences in the comparative results.27 Nonetheless, the reader should keep this in mind when

interpreting the results.

The reader is also reminded that the prevalence of longer-term opioid dispensing is based on nonsurgical

claims with more than seven days of lost time that received at least one opioid prescription paid under

workers’ compensation at the time of evaluation. Based on the findings of our sensitivity analysis, these claim

selections are unlikely to distort the results in this report. However, we caution the reader not to simply

extrapolate the results based on nonsurgical claims with more than seven days of lost time to all claims in a

state. This is because excluding surgical claims from the analysis is likely to undercount the incidence of

longer-term opioid use overall while the exclusion of claims with seven or fewer days of lost time may

overestimate the prevalence of longer-term opioids. The combined effect may be different depending on the

26 See Figure TA.C3 in Technical Appendix C and the related discussion. 27 See Technical Appendix C for more details.



31

state.

Note that the data used for this analysis are based on an average of 24 months of experience, which

means that we include claims that had from 18 to 30 months of experience postinjury. This gives us adequate

data to identify claims receiving longer-term opioids. Although the data used may not necessarily be sufficient

to capture the full utilization of opioids,28 this selection is unlikely to affect the results on the frequency of use

and longer-term dispensing of opioids and other frequency metrics used in the study.29

28 In a National Council on Compensation Insurance, Inc. (NCCI) study, the authors found that the narcotics share of all prescriptions increased steadily when claims became more mature until about the eighth year postinjury (Lipton, Laws, and Li, 2009). The same study also looked at the narcotics share by costs per narcotic prescription, where the high-cost group would presumably include more prescriptions for stronger and long-acting narcotics. The study found that the high-cost narcotic prescriptions grew from 9 percent of all narcotic prescriptions in the first year to 45 percent in the 12th year postinjury. 29 This is because the identification of longer-term opioids is based on the detailed drug transactions in the first year of treatment, and the data for claims with an average of 24 months of experience are for claims with detailed data covering all drug transactions from 18 months (for claims with the shortest time window) to 30 months (for claims with the longest time window).



32

3

LONGER-TERM DISPENSING OF OPIOIDS

This chapter presents the updated results on the prevalence and trends of longer-term dispensing of opioids.1

It also provides evidence on how closely medical treatment guidelines were followed among the states studied.

In the 2016 edition of this study, we reported a considerable decrease in the frequency of longer-term

opioid dispensing in a number of states, including Maryland, Michigan, New Jersey, New York, North

Carolina, and Texas.2 With additional data covering one more year, we continued to observe significant

decreases in the prevalence of longer-term opioids in an increasing number of states, especially in Kentucky

and New York as well as several other study states (Kansas, Massachusetts, Michigan, Minnesota, and

Tennessee). However, longer-term use of opioids continued to be prevalent in Louisiana3 and several other

study states, among 2013/20154 claims with more than seven days of lost time that did not have surgery but

received opioid prescriptions.

Figure 3.1 shows the percentage of nonsurgical claims with more than seven days of lost time that

received prescriptions for opioids for pain relief. In most study states, the frequency of claims receiving opioid

prescriptions was between 60 and 80 percent. The figure was slightly higher in Arkansas, Louisiana, and South

Carolina (80–83 percent) and lower in several states, including Illinois and New Jersey (52–56 percent).5 It

should be noted that even in the states with the lowest percentage of claims receiving opioids, more than half

of the nonsurgical claims received opioids for pain relief.

Figure 3.2 shows that among 2013/2015 nonsurgical claims with opioids, longer-term dispensing of

opioids was most prevalent in Louisiana, where 1 in 6 injured workers with opioids had them on a longer-

1 The prevalence of longer-term dispensing of opioids is measured based on claims receiving longer-term opioids, i.e., those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury. Throughout the report, the term prevalence of longer-term dispensing of opioids is used interchangeably with frequency of claims with longer-term opioids. Note that the definition of longer-term opioids is largely consistent with the definition of chronic opioids presented in WCRI’s Interstate Variations in Use of Opioids, 4th Edition (Thumula, Wang, and Liu, 2017). Since the identification of claims with chronic opioids was based on duration and daily dose of opioids, and the data for these elements were less complete for physician-dispensed prescriptions, the interstate variations report presented the results on chronic opioids for 17 of the 26 states. This report covers all 26 states in terms of longer-term opioid dispensing. 2 The decreases were also echoed by the 2016 study on interstate variations in opioids that examined the overall use of opioids and prescribing patterns of opioids across states, based on the same underlying data. See Thumula, Wang, and Liu (2016). 3 In Louisiana, the revised Workers’ Compensation Act in 2009 resulted in the creation of medical treatment guidelines, including guidelines for chronic pain disorders, effective January 2011. The treatment guidelines were updated in June 2011. 4 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015. 5 We measure the frequency in use of opioids based on claims with pain medications, which indicates that the opioids were prescribed for pain relief. Among the nonsurgical claims with prescriptions, over 90 percent had prescriptions for pain medications in most of the 25 study states (Table 3.1).



33

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

NJ IL MD KY CT MI NY CA PA FL TN MO IN GA MA IA VA KS TX NV WI NC MN SC LA AR

% o

f Cla

ims

term basis.6,7 The figure was 1 in 10–12 in California, Georgia, North Carolina, Pennsylvania, South Carolina,

and Texas. In several states, by contrast, longer-term dispensing of opioids was seen less frequently—

approximately 1 in 25 injured workers who did not have surgery but had opioid prescriptions received

opioids on a longer-term basis in Indiana, Kansas, Missouri, Nevada, New Jersey, and Wisconsin (Figure

3.2).8

Figure 3.1 Percentage of Claims with Pain Medications That Received Opioids, 2013/2015

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015.

Some injured workers did not receive opioids within the first quarter postinjury but exhibited the same

pattern of longer-term opioids as those identified as having longer-term opioid prescriptions.9 Figure 3.3

shows that in most states studied, less than 3 percent of claims with opioids fell into this category. The figure

was higher (between 3 and 6 percent) for California, Connecticut, Louisiana, Massachusetts, New York, and

6 It should be noted that compared with Pennsylvania and South Carolina, the next highest states on this measure, Louisiana’s number might be slightly higher because of our definition of longer-term opioids. Louisiana had relatively less frequent use of stronger Schedule II opioids than Pennsylvania and South Carolina, but a higher number of prescriptions and visits to fill prescriptions for opioids. However, the difference in the pattern of opioid use would not be enough to change Louisiana’s ranking on this measure. 7 There have been studies examining factors in the early stage of treatment that may be associated with receiving chronic opioid therapy. These factors include early opioid use, presence of psychological and psychiatric conditions, and substance use disorders (SUDs). These early factors that may likely influence the likelihood of receiving long-term opioids should be addressed in practice to reduce the prevalence of claims receiving long-term opioids. In addition, most treatment guidelines for chronic pain recommend careful screening of patients for chronic opioid therapy, which is also an important tool to address the prevalence of longer-term opioid use. 8 One may be concerned that a percentage of claims with longer-term opioids that is based on claims with opioids might be misleading on whether a state has an overall higher or lower prevalence of longer-term opioid use. This would be a valid concern if the percentage of claims receiving opioids varied substantially across the states and if a low percentage of claims receiving opioids appeared to be correlated with a higher percentage of claims with opioids that were identified as having longer-term opioids. However, this is not a serious concern based on what we found in Figure TA.C3, which shows that states with higher percentages of claims receiving opioids tended to have higher percentages of claims identified as having longer-term opioids. See Chapter 2 and Technical Appendix C for more detailed discussions. 9 Several reasons might explain this subset of claims (referred to as Type II claims). For example, some injured workers might have had late onset pain and therefore received opioids later. Physicians who follow treatment guidelines might not have prescribed opioids for low back pain at the beginning of the treatment and prescribed opioids later, after all other methods had failed. It is also possible that some of these workers filled opioid prescriptions within three months postinjury, but the prescriptions were paid for by a non-workers’ compensation payor, and thus did not appear in our data. These are possibilities that may be investigated further in future studies.



34

0%

3%

6%

9%

12%

15%

18%

21%


% o

f Cla

ims

wit

h O

pio

ids

0%

3%

6%

9%

12%

15%

18%

NV WI AR NJ MO FL MI VA KS TN MN TX GA IN IA KY NC MD SC IL CT CA NY PA MA LA

% o

f Cla

ims

wit

h O

pio

ids

Pennsylvania. This group of claims could contribute to a higher rate of longer-term opioid dispensing in the

states. However, we do not focus on this group of claims in this report because of a potential concern of

overstating the prevalence of longer-term opioid dispensing, as discussed in Chapter 2.10 Table 3.1 provides

the underlying data for Figures 3.1 through 3.3.

Figure 3.2 Percentage of Claims with Opioids That Received Opioids on a Longer-Term Basis,a 2013/2015


Figure 3.3 Percentage of Claims with Opioids That Did Not Receive Opioids in Three Months Postinjury but Had Opioids on a Longer-Term Basis,a 2013/2015


10 We chose not to focus on this group of claims because it is unclear as to the nature of the treatment provided for this group of claims. Further investigations are needed to understand the medical conditions and treatment received by these injured workers to help address the management of longer-term opioid therapy when it occurs later in a claim.



35

AR CA CT FL GA IA IL IN KS KY LA MA MD MI MN MO NC NJ NV NY PA SC TN TX VA WI26-State Median

Total number of nonsurgical claims that had Rx for pain medications 529 24,845 2,286 9,716 3,678 1,281 5,712 2,080 1,114 1,173 1,302 2,206 2,518 2,474 2,007 2,070 2,940 4,255 1,044 4,738 5,636 1,504 2,276 13,424 1,998 1,784

% of nonsurgical claims with Rx that had pain medications 94% 92% 86% 93% 93% 88% 89% 90% 91% 91% 92% 88% 92% 91% 83% 90% 91% 88% 92% 88% 89% 91% 93% 94% 89% 88% 94%

% of nonsurgical claims with pain medications that had opioids 85% 66% 59% 68% 73% 73% 56% 72% 75% 59% 80% 73% 59% 64% 78% 70% 77% 52% 76% 64% 67% 80% 70% 76% 74% 76% 73%

% of nonsurgical claims with opioids that were identified as receiving longer-term

opioids (Type I)a 4.4% 8.6% 6.2% 4.9% 8.4% 4.1% 5.5% 3.9% 3.4% 6.4% 17.9% 6.7% 5.9% 5.0% 5.0% 3.7% 8.7% 3.6% 3.9% 7.1% 9.3% 10.1% 4.5% 9.1% 6.5% 3.8% 5.7%

% of nonsurgical claims with opioids that exhibited the pattern of receiving longer-

term opioids (Type II)b 1.1% 3.4% 3.1% 1.4% 1.9% 1.9% 2.5% 1.9% 1.6% 2.1% 5.8% 4.2% 2.1% 1.5% 1.7% 1.4% 2.1% 1.3% 1.0% 3.7% 3.9% 2.3% 1.7% 1.8% 1.6% 1.1% 1.9%

Table 3.1 Frequency of Claims Receiving Opioids among Nonsurgical Claims with More Than 7 Days of Lost Time, 2013/2015

Key: Rx: prescriptions.

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2012/2014 refers to claims with injuries occurring from October 1, 2011, through September 30, 2012, and prescriptions filled through March 31, 2014.

a We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury (i.e., Type I claims). See Chapter 2 for more details.

b The Type II claims are those that did not have opioids within the first three months after the injury but had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury. See Chapter 2 for more details.



36

-5%

-4%

-3%

-2%

-1%

0%

1%

2%

KY* NY* TN* MI* KS* MA* MN* AR* CT TX* FL* CA* NV NC MD NJ PA IL MO GA WI IN IA VA SC LA

% P

oin

t C

han

ge

Figure 3.4 shows changes in the prevalence of longer-term dispensing of opioids between 2010/2012 and

2013/2015 across the 26 study states. Table 3.2 provides data underlying Figure 3.4 and the year-by-year data

for the percentage of claims with opioids that received opioids on a longer-term basis. It also includes data on

the frequency of claims receiving opioid prescriptions over time. Note that for most states, the claim

frequency of longer-term opioid dispensing peaked in 2010/2012, but in several states, the figure continued to

grow and peaked in 2011/2013. Figure 3.4 shows the magnitude of the decrease in terms of percentage point

difference, and Table 3.2 provides more details, from which the reader may examine the year-to-year changes

and compute the magnitude of the reduction in the prevalence of longer-term opioid dispensing in terms of

percentage change.11

Figure 3.4 Changes in the Prevalence of Longer-Term Opioid Dispensinga between 2010/2012 and 2013/2015


a We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury. See Chapter 2 for more details.

* States with an asterisk (*) are those for which the observed change was statistically significant at the 95 percent confidence level.

11 Both percentage point change and percentage change are valid measures to indicate the magnitude of change. While the percentage change measures the magnitude of change in the total volume, it may be very sensitive to small samples or denominators in a way that distorts the significance of the change.



37

ARa CAb CT FLb GAa IA IL INc KSb KYb LA MAb MD MIb MNb MO NCd NJ NVa NYb PA SC TNb TXb VA WIc 26-State Median

2010/2012 536 15,154 1,125 6,701 2,412 943 3,386 1,593 897 983 1,192 1,719 1,482 1,985 1,383 1,452 2,401 2,139 775 3,627 4,297 1,258 1,977 9,360 1,610 1,373

2011/2013 513 15,288 1,268 6,799 2,473 1,011 3,621 1,654 894 992 1,173 1,681 1,634 1,856 1,612 1,564 2,409 2,410 763 3,652 4,334 1,359 1,959 10,190 1,695 1,597

2012/2014 529 16,012 1,422 6,747 2,674 962 3,477 1,588 875 967 1,115 1,597 1,545 1,760 1,563 1,522 2,515 2,420 798 3,382 4,123 1,315 1,844 10,401 1,594 1,503

2013/2015 447 16,285 1,355 6,581 2,679 941 3,202 1,505 831 692 1,043 1,620 1,481 1,578 1,566 1,458 2,273 2,214 797 3,034 3,785 1,203 1,593 10,174 1,489 1,363

2010/2012 88% 68% 62% 71% 73% 73% 57% 76% 76% 75% 85% 75% 65% 64% 77% 69% 80% 57% 82% 73% 70% 79% 75% 76% 79% 78% 75%

2011/2013 85% 68% 64% 69% 75% 75% 59% 77% 78% 76% 84% 74% 66% 66% 79% 71% 80% 54% 79% 72% 71% 79% 76% 76% 81% 80% 75%

2012/2014 86% 67% 63% 67% 74% 77% 57% 74% 74% 71% 83% 73% 62% 65% 77% 69% 77% 54% 75% 69% 68% 80% 74% 76% 79% 78% 74%

2013/2015 85% 66% 59% 68% 73% 73% 56% 72% 75% 59% 80% 73% 59% 64% 78% 70% 77% 52% 76% 64% 67% 80% 70% 76% 74% 76% 73%

% point change (2010/2012–2013/2015) -3 -2 -3 -3 -1 1 -1 -4 -1 -16 -5 -1 -6 0 1 2 -3 -5 -6 -9 -3 1 -5 0 -4 -1 -2.9

2010/2012 6.2% 10.2% 7.9% 6.5% 8.5% 4.1% 6.0% 3.8% 5.6% 11.0% 16.5% 8.8% 7.2% 7.5% 7.0% 3.9% 10.0% 4.6% 5.5% 11.3% 10.3% 9.5% 7.2% 10.8% 6.3% 3.8% 7.2%

2011/2013 8.2% 9.4% 7.2% 6.2% 10.1% 4.2% 6.5% 5.0% 5.4% 11.3% 18.4% 8.0% 7.3% 6.5% 6.3% 3.5% 8.6% 4.2% 6.7% 10.4% 10.0% 10.2% 5.6% 9.2% 7.5% 3.7% 7.3%

2012/2014 5.4% 9.7% 7.1% 5.6% 8.2% 4.7% 5.3% 5.7% 4.8% 8.7% 17.6% 6.8% 5.7% 5.5% 6.6% 3.5% 8.0% 3.5% 5.0% 9.7% 9.0% 8.3% 5.6% 8.9% 6.4% 5.1% 6.0%

2013/2015 4.4% 8.6% 6.2% 4.9% 8.4% 4.1% 5.5% 3.9% 3.4% 6.4% 17.9% 6.7% 5.9% 5.0% 5.0% 3.7% 8.7% 3.6% 3.9% 7.1% 9.3% 10.1% 4.5% 9.1% 6.5% 3.8% 5.7%

% point change (2010/2012–2013/2015) -1.7 -1.5 -1.7 -1.6 -0.1 0.1 -0.5 0.1 -2.2 -4.7 1.4 -2.1 -1.3 -2.5 -2.0 -0.2 -1.3 -1.1 -1.5 -4.2 -1.0 0.6 -2.6 -1.6 0.2 0.0 -1.4

% point change (2011/2013–2013/2015) -3.8 -0.8 -1.1 -1.3 -1.7 0.0 -1.0 -1.1 -2.0 -5.0 -0.5 -1.2 -1.5 -1.5 -1.4 0.1 0.1 -0.7 -2.8 -3.3 -0.7 -0.1 -1.1 0.0 -1.0 0.1 -1.1

c For Indiana and Wisconsin, we reported an increase in the prevalence of longer-term opioid dispensing in the 2016 edition of this report. With additional data covering one more year, we found that the increases in these two states previously reported appeared to be temporary. Changes in the longer-term use of opioids will continue to be monitored in future editions of this study. d For North Carolina, we reported a significant decrease in the prevalence of longer-term opioid dispensing in the 2016 edition of this report. With additional data covering one more year, we found that the 2013/2015 number increased slightly but not statistically significantly. We will continue to monitor the changes with more recent data. e We identified claims with longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury (i.e., Type I claims). See Chapter 2 for more details.

Table 3.2 Trends in the Percentage of Nonsurgical Claims with Opioids That Received Opioids on a Longer-Term Basis, 2010/2012–2013/2015

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2010/2012 refers to claims with injuries occurring from October 1, 2009, through September 30, 2010, and prescriptions filled through March 31, 2012. Similar notation is used for other years. For the analysis of longer-term use of opioids, we included a small number of claims with unusually high amounts of opioids. See Chapter 2 for more details. a In Arkansas, Georgia, and Nevada, the frequency of longer-term opioid dispensing peaked in 2011/2013, and the reduction in the frequency between 2011/2013 and 2013/2015 was significant at the 95 percent level. Note that in several other states, the frequency of longer-term opioid dispensing also peaked in 2011/2013, but the decrease in the frequency over the three-year period was not statistically significant. b The decrease in the frequency of longer-term opioid dispensing was statistically significant at the 95 percent level over the four-year period in California, Florida, Kansas, Kentucky, Massachusetts, Michigan, Minnesota, New York, Tennessee, and Texas.

Total number of nonsurgical claims that received opioids

% of nonsurgical cases with opioids that received longer-term opioids (Type I)e

% of nonsurgical cases with pain medications that received opioids



38

Overall, the prevalence of longer-term opioid dispensing decreased considerably between 2010/2012 and

2013/2015 in most study states. A substantially large decrease (more than 4 percentage points) was seen in

Kentucky and New York. The same measure decreased considerably (between 2 and 4 percentage points) in

several other states, including Kansas, Massachusetts, Michigan, Minnesota, and Tennessee. In California,

Florida, and Texas, the decrease was to a lesser degree (1–2 percentage points) but still statistically significant

at the 95 percent confidence level.12 We also saw that the prevalence of longer-term opioid dispensing in

several states (Connecticut, Illinois, Maryland, New Jersey, North Carolina, and Pennsylvania) decreased over

the study period, but the percentage point change was not statistically significant at the 95 percent confidence

level (see Figure 3.4 and Table 3.2). Note that in several states, the frequency of longer-term opioids peaked in

2011/2013. Among those states, we saw noticeable reductions in the measure between 2011/2013 and

2013/2015 in Arkansas, Georgia, and Nevada (Table 3.2). By contrast, the claim frequency of longer-term

opioids changed little in Iowa, Missouri, and Virginia, and fluctuated in Indiana, South Carolina, and

Wisconsin. The figure increased slightly in Louisiana.13 We discuss these findings in detail in this chapter; in

the Executive Summary, we only highlighted states where we saw large and significant changes in the

frequency of claims receiving longer-term opioid prescriptions.

Considerable decreases in the prevalence of longer-term opioid dispensing may have been associated

with numerous opioid policies and initiatives in the past few years that were aimed at addressing the epidemic

of opioid misuse and abuse. At the state level, most prominent policy changes include (1) a mandatory check

of patient prescription history from the state PDMP database at the point of prescribing and dispensing

opioid prescriptions, (2) clinical and practice guidelines for prescribing opioids and managing chronic opioid

therapy adopted by states in and outside workers’ compensation, (3) regulations that require the completion

of continuing medical education (CME) on opioid prescription and pain management, (4) drug formularies

in several workers’ compensation jurisdictions, and (5) state laws regulating pain clinics and preventing

doctor shopping. More recently, several states took a more comprehensive approach to addressing issues

related to the growth of opioid overdoses and deaths, through legislation, leadership of the state executive

branch, and coordination among different agencies and stakeholders. Research on the impact of opioid

policies has also been conducted to support evidence-based policymaking. When discussing the results, we

provide some background information about recent opioid policies and initiatives to facilitate the

interpretation of the results without suggesting a causal relationship between policy changes and the

reduction in prevalence. More detailed descriptions about key policy factors that may influence the use and

longer-term use of opioids can be found in Technical Appendix A.

Over the study period, we saw a substantial decrease in the frequency of claims receiving longer-term

opioids in Kentucky and New York. In Kentucky, 6.4 percent of nonsurgical claims with opioids received

longer-term opioids in 2013/2015, a decrease from 11.0 percent in 2010/2012, which translates to a 42 percent

reduction. The figure decreased 4.2 percentage points in New York, from 11.3 percent in 2010/2012 to 7.1

percent in 2013/2015, which translates to a reduction of approximately 37 percent. As a result, the frequency

of longer-term use of opioid in these two states was closer to the median of the 26 study states in the latest

study period (Figure 3.2). It is worth noting that the decrease in the prevalence of longer-term opioids in

these two states was coupled with a large decrease in the percentage of nonsurgical claims that received

12 The frequency of longer-term opioid dispensing in Arkansas and Nevada also decreased at a similar magnitude, but the change between 2010/2012 and 2013/2015 was not statistically significant at the 95 percent confidence level. In these two states and Georgia, the frequency of longer-term opioids peaked in 2011/2013, and the reduction in the frequency in the latest two years was significant at the 95 percent level. 13 The increase was not statistically significant at the 95 percent confidence level.



39

opioids (Figure 3.4 and Table 3.2).14

The substantial reductions in the frequency of claims receiving longer-term opioids may be associated

with the comprehensive reforms in these two states over the study period that were aimed at reducing opioid

prescriptions and overdoses. In Kentucky, House Bill 1, which went into effect in July 2012, includes

provisions for pain clinics regulation, rules for opioid prescribing and dispensing, and a mandatory check of

the state PDMP database (Kentucky All Schedule Prescription Electronic Reporting [KASPER]). Kentucky

was one of the seven states that participated in the National Governors Association (NGA) Prescription Drug

Abuse Reductions Policy Academy launched in 2012.15 New York passed legislative mandates that require

prescribers to check the PDMP database at the point of prescribing opioids. Effective August 2013, health care

practitioners must consult the PDMP when prescribing or dispensing controlled substances in Schedules II–

IV, with limited exceptions.16 More recent policy changes also include the publication of non-acute medical

treatment guidelines.17 According to the 2013 report by the Trust for America’s Health that provides ratings

on state opioid policies, both Kentucky and New York received a rating of 9 out of 10.18

Considerable reductions were also seen in Kansas, Massachusetts, Michigan, Minnesota, and

Tennessee. The percentage of nonsurgical claims with opioids that were identified as having longer-term

opioids decreased more than 2 percentage points between 2010/2012 and 2013/2015, reductions of

approximately 24 to 39 percent. In recent years, there have been many changes and initiatives in opioid

policies in these states. For each state, we highlight the key policies and changes in the prevalence of longer-

term opioids.

Kansas: In 2013/2015, 3.4 percent of the claims with opioids were identified as having longer-term

opioids, down from 5.6 percent in 2010/2012. The Kansas PDMP, also known as K-TRACS, has been

operational since 2011 and the number of prescribers registered with the PDMP program has increased

steadily, with nearly one-third of all prescribers enrolled as of December 2014.19 Kansas adopted the

Model Policy for the Use of Controlled Substances for the Treatment of Pain, established by the

Federation of State Medical Boards (FSMB).20 The Kansas workers’ compensation regulatory agency, the

Department of Labor, Division of Workers’ Compensation, has adopted the Work Loss Data Institute’s

Official Disability Guidelines—Treatment in Workers’ Comp (ODG) as the standard of reference for

14 The percentage of nonsurgical claims with pain medications that received opioid prescriptions decreased by 16 percentage points in Kentucky (from 75 percent in 2010/2015 to 59 percent in 2013/2015) and 9 percentage points in New York (from 73 to 64 percent). 15 Seven states participated when the academy was first launched in 2012, including Alabama, Arkansas, Colorado, Kentucky, New Mexico, Oregon, and Virginia. Michigan, Minnesota, Nevada, North Carolina, and Vermont were among the states participating in the academy in 2014. The academy has since published several papers to discuss lessons learned from the participating states and provide strategic action plan recommendations. See NGA’s paper Reducing Prescription Drug Abuse: Lessons Learned from an NGA Policy Academy. 16 Information about New York’s I-STOP legislation can be found at http://www.ag.ny.gov/sites/default/files/press-releases/2012/ISTOP%20REPORT%20FINAL%201.10.12.pdf. 17 Effective December 15, 2014, the New York Workers’ Compensation Board published the state non-acute pain medical treatment guidelines in an effort to address issues related to long-term opioid use in the state workers’ compensation system. The Board has also been working to coordinate its medical treatment guideline implementation with the I-STOP legislation. 18 The states with the highest scores were New Mexico and Vermont (neither are included in this study). See Trust for America’s Health (2013). 19 See the 2016 Pew Charitable Trusts report entitled Prescription Drug Monitoring Programs: Evidence-Based Practices to Optimize Prescriber Use, which is available at http://www.pewtrusts.org/~/media/assets/2016/12/prescription_drug_monitoring_programs.pdf?la=en. 20 The guidelines consist of evaluating the patient, developing a treatment plan, obtaining informed consent and agreement for treatment, periodically reviewing the treatment plan, consulting with the patient, maintaining good medical records, and complying with controlled substances laws and regulations. See FSMB (2013). The Model Policy was being updated as of December of 2016.



40

evidence-based medicine used in caring for injured workers.

Massachusetts: The percentage of nonsurgical claims with opioids that received longer-term opioid

prescriptions decreased from 8.8 percent in 2010/2012 to 6.7 percent in 2013/2015. The state PDMP

system became operational in 2010, and prescribers are required to register with the PDMP program.

Prescribers are also encouraged to check the online PDMP database when prescribing opioids. In

subsequent years, additional efforts were made to enhance the utility of the PDMP system. Effective

March 2013, a major update was made to the Chronic Pain Treatment Guidelines, which are required to

be used in drug utilization review. Massachusetts’ Medical Board also requires three credits of opioid

education and pain management training as a condition for license renewal, effective February 2012. In

addition, the largest group health insurer in Massachusetts, Blue Cross Blue Shield, recently implemented

an industry-leading policy to restrain abuse of opioid prescriptions, which may likely influence the

prescribing practices of physicians treating workers’ compensation patients. These opioid policies and

initiatives likely help explain the considerable reduction in the longer-term opioid use in the state. In

March 2016, the Governor signed into law a landmark opioid legislation, which contains several

provisions to address the prevention and intervention of opioid overdoses, in response to a continued

increase in opioid overdoses. The 2016 legislation is expected to have an impact on the utilization of

opioid prescriptions.

Michigan: Five percent of nonsurgical claims with opioids received longer-term opioids in 2013/2015, a

decrease from 7.5 percent in 2010/2012. This translates to an approximately 33 percent decrease in the

number of nonsurgical claims with opioids that received opioids on a longer-term basis. This drop was

after an increase of more than 2 percentage points from 2008/2010 to 2010/2012, as reported in a

previous edition of this study.21 Over the study period, there were several policy initiatives, which might

have contributed to the decrease. Among all states with state PDMP programs, Michigan is the only state

that allows third-party payors to access PDMP data.22 Effective December 26, 2014, the amended rules by

Michigan’s Workers’ Compensation Agency require that opioid treatment beyond 90 days for non-

cancer related chronic pain should not be reimbursed unless detailed physician reporting requirements

and other processes are met. The new rules also provide incentives for compliance with the requirement

by allowing a provider to bill for the additional services required for reporting beyond 90 days and for

accessing the state PDMP (Michigan Automated Prescription System [MAPS]) or other PDMPs in the

treating jurisdiction.23 On September 4, 2013, Governor Rick Snyder issued a proclamation recognizing

September as Pain Awareness Month, which raised awareness of opioid prescribing and pain

management.24 On June 23, 2016, Governor Snyder created the Prescription Drug and Opioid Abuse

Commission and abolished the Advisory Committee on Pain and Symptom Management (ACPSM),

which was created to study pain issues in the state and make recommendations to improve care of

patients with pain.25

21 See Wang (2014). 22 Michigan’s PDMP rules can be found at http://www.legislature.mi.gov/(S(mj5rcwuqhzd25a3uea5l0l5d))/mileg.aspx?page=getobject&objectname=mcl-333-7333a&query=on&highlight=7333a. 23 For more details, please refer to the amended rules, which are available at http://www.michigan.gov/documents/lara/2014-029_LR_Final_Health_Care_Services_476952_7.pdf. 24 More details are available at http://www.michigan.gov/lara/0,4601,7-154-61805-311785--,00.html. 25 The ACPSM was charged with developing model core curricula, continuing education (CE) recommendations to provide guidance on required CE hours and content for competent prescribing for the state professional boards, and pain management tool kits.



41

http://www.legislature.mi.gov/(S(mj5rcwuqhzd25a3uea5l0l5d))/mileg.aspx?page=getobject&objectname=mcl-333-7333a&query=on&highlight=7333a

http://www.michigan.gov/lara/0,4601,7-154-61805-311785--,00.html

Minnesota: In 2013/2015, 5 percent of the claims with opioids were identified as having longer-term

opioid use, a decrease of 2 percentage points (or approximately 29 percent) from 7 percent in 2010/2012.

The state PDMP became operational in 2010. Minnesota’s legislation only mandates that prescribers in

opioid treatment programs check the PDMP database. Other related policies include a doctor shopping

law, pharmacy lock-in program, etc. Effective July 13, 2015, Minnesota’s administrative rules require

compliance with a set of guidelines governing long-term treatment with opioid analgesic medications for

workers’ compensation injuries.26 Minnesota’s Opioid Prescribing Work Group (OPWG) played a

crucial role in the state’s response to the crisis of opioid overuse. In April 2014, the OPWG published A

Protocol for Addressing Acute Pain, which provides guidelines for prescribing opioids.27 The OPWG will

help shape the opioid-prescribing improvement and monitoring program, created in 2015 at the

direction of the governor and the legislature to reduce the spread of opioids.28 In May 2016, the

legislature passed a law that requires more medical professionals in Minnesota to sign up for the PDMP

to encourage the use of the PDMP database.29

Tennessee: The frequency of claims receiving longer-term opioids decreased from 7.2 percent in

2010/2012 to 4.5 percent in 2013/2015, which translates to a decrease of approximately 37 percent

(Figure 3.2 and Table 3.2). In Tennessee, prescribers must check the PDMP database when first

prescribing opioids and benzodiazepines for more than seven days and at three-month intervals

thereafter if prescribing continues.30 In November 2012, Tennessee’s legislature passed Senate Bill 3315,

which amended the definition of utilization review to explicitly include Schedule II, III, and IV drugs

being used for pain management. The provision requires the parties involved to participate in utilization

review if opioids are prescribed for pain management to an injured or disabled employee for a period of

time exceeding 90 days from the initial prescription. Effective October 1, 2013, a new Tennessee

legislation requires that prescriptions for opioids or benzodiazepines not be dispensed in quantities

greater than a 30-day supply. The law also encourages mandatory urine drug testing of patients on long-

term drug therapy. In July 2014, the Tennessee Board of Medical Examiners voted to adopt as policy the

Department of Health guidelines for chronic pain (Clinical Practice Guidelines for Outpatient

Management of Chronic Non-Malignant Pain). These changes in opioid policies likely contributed to the

considerable reduction in longer-term dispensing of opioids in the state. In 2016, the Bureau of Workers’

Compensation adopted the ODG drug formulary as part of a comprehensive set of treatment guidelines

adopted at the same time. The Bureau also subscribes to the recommendations of the State of Tennessee,

Department of Health, Chronic Pain Guidelines, the Bureau of Workers’ Compensation Pain

Management/Opioid Guidelines Appendix, according to a 2016 International Association of Industrial

Accident Boards and Commissions (IAIABC) report.31

26 Minnesota workers’ compensation guidelines for chronic opioid management can be found at https://www.revisor.mn.gov/rules/?id=5221.6110. 27 The protocols include avoiding prescribing more than a three-day supply (or 20 pills) of low-dose, short-acting opioids in general; never prescribing long-acting/extended-release preparations for acute pain; and maximizing appropriate non-opioid therapies. See the OPWG protocol for addressing acute pain and prescribing opioids at https://www.icsi.org/_asset/xbhfdk/opioidprotocolmnmedicine.pdf. 28 More information about the OPWG and its initiatives can be found at https://mn.gov/dhs/partners-and-providers/news-initiatives-reports-workgroups/minnesota-health-care-programs/opioid-work-group/. 29 See https://www.revisor.mn.gov/bills/text.php?number=SF1440&version=4&session=ls89&session_year= 2016&session_number=0. 30 See PDMP Center of Excellence at Brandeis University (2014b) and Clark et al. (2012). 31 See IAIABC (2016).



42

https://www.revisor.mn.gov/bills/text.php?number=SF1440&version=4&session=ls89&session_year=2016&session_number=0

In California, Florida, and Texas, we also saw noticeable decreases in the frequency of claims receiving

longer-term opioids between 2010/2012 and 2013/2015. In these states, the percentage of nonsurgical claims

with opioids that received opioids on a longer-term basis decreased between 1 and 2 percentage points, which

translates to a reduction of approximately 15–25 percent.32 The following discussions are focused on these

states, which were not individually discussed in the Executive Summary. We also provide some background

information about state policies and initiatives regarding prescribing opioids to facilitate the interpretation of

the results.33

California: The frequency of longer-term opioid dispensing was 8.6 percent in 2013/2015, a decrease

from 10.2 percent in 2010/2012, which translates to a reduction of about 15 percent. In 2013/2015, the

figure for California was nearly 3 percentage points above the median of the 26 study states. There were

several opioid-related policy changes in California. While policy changes made during the study period

are expected to have a direct impact on the opioid prescribing and dispensing results in this study, policy

changes made shortly after the study period may also have an impact because of the awareness of the

issues in the process of making changes. In California, medical services delivered to injured workers were

required to go through mandatory utilization review (UR) and independent medical review (IMR).

Effective January 1, 2013, Senate Bill 863 requires the use of evidence-based medicine to guide treatment

decisions and treatment dispute settlements by independent medical reviewers and to improve workers’

access to network physicians.34 In California, the chronic pain treatment guidelines were first adopted by

the Division of Workers’ Compensation in 2009. The regulatory efforts led to the publication of the

updated chronic pain medical guidelines and the first-time adoption of opioid treatment guidelines in

July 2016, by the Division of Workers’ Compensation, California’s Department of Industrial Relations.35

The guidelines apply to any treatment requests made on or after July 29, 2016. In March 2017, the first

draft of the medical treatment utilization schedule (MTUS) drug formulary was released, and it was

subsequently revised. The revised drug formulary is anticipated to take effect for services on or after

January 1, 2018.36,37 The Controlled Substance Utilization Review and Evaluation System (CURES), the

state PDMP program, was operational in 2005, but it was not until recently that California passed

funding bills to establish a state fund to pay for CURES. In September 2016, Senate Bill 482 was also

32 The changes in the frequency of longer-term opioids in these states were statistically significant at the 95 percent confidence level. 33 The information about opioid policy changes provided here and in Technical Appendix A is intended to be used by the reader who would like to interpret the results in the context of policy changes. The investigation of the relationship between opioid policy changes and the frequency of longer-term dispensing of opioids is beyond the scope of this report. 34 Among Senate Bill 863’s goals was the implementation of evidence-based medicine guidelines for treatment decisions. Evidence of opioid abuse prompted legislation mandating the adoption of an evidence-based workers’ compensation drug formulary by July 1, 2017. See http://www.dir.ca.gov/DIRNews/2016/2016-73.pdf. 35 The regulation can be found at http://www.dir.ca.gov/DIRNews/2016/2016-78.pdf, and the guidelines are available at https://www.dir.ca.gov/dwc/DWCPropRegs/MTUS-Opioids-ChronicPain/MTUS-Opioids-ChronicPain.htm. 36 The effective date of the drug formulary was originally set for July 1, 2017. In May 2017, there was some stakeholder input toward recommending a delay in the implementation to address additional comments. On July 19, 2017, the revised drug formulary was released for public comment over a 15-day review period, and the new rules will be effective January 1, 2018. 37 Assembly Bill 1124 requires the California’s Division of Workers’ Compensation to implement a drug formulary no later than July 1, 2017, in light of findings from several studies. A carefully structured formulary can reinforce the MTUS guidelines to encourage prescribing of medically appropriate drugs and reduce the administrative burdens associated with the UR/IMR process (Wynn et al., 2016). A 2014 report by the California Workers’ Compensation Institute (CWCI) noted prescription drug costs as one of the fastest growing areas of workers’ compensation medical benefits in California, partly because of the wide array of drugs available and a dramatic increase in pain management therapies (Swedlow, Hayes, and David, 2014). The study looked at formularies used in both Texas and Washington and found that drug costs could be reduced between 12 and 42 percent, approximately $124 to $420 million in savings annually.



43

passed and signed by Governor Jerry Brown, requiring prescribers to check CURES prior to prescribing

opioids.

Florida: In 2013/2015, the prevalence of longer-term opioid dispensing was 4.9 percent, nearly 1

percentage point below the median of the 26 study states (5.7 percent). The figure in 2010/2012 was 6.5

percent (7.2 percent for the median state). Most opioid prescriptions were highly concentrated among a

small percentage of Florida prescribers. According to the researchers at Johns Hopkins Bloomberg School

of Public Health, 40 percent of total opioid prescriptions dispensed in the state between 2010 and June

2011 were prescribed by 4 percent of prescribers in Florida, accounting for 67 percent of the opioid

volume.38 In 2011, Florida passed two laws aimed at curbing opioid prescriptions. Effective July 2011,

Florida’s legislation prohibited dispensing Schedule II and III controlled substances from offices and

clinics. In September 2011, the state PDMP system became operational. According to a study by Johns

Hopkins researchers, in the first year after the two Florida laws went into effect, the state’s top opioid

prescribers wrote significantly fewer prescriptions for this type of pain medication.39 Effective December

24, 2015, the Florida Board of Pharmacy adopted a rule on standards of practice for the filling of

controlled substance prescriptions and electronic prescribing, and requires two hours of mandatory

continuing education.40

Texas: In 2013/2015, 9.1 percent of workers who did not have surgery but received opioid prescriptions

were identified as having longer-term opioids, a decrease from 10.8 percent in 2010/2012. The Texas

Division of Workers’ Compensation adopted a closed pharmacy formulary, which is based on ODG. The

formulary went into effect on September 1, 2011, for new claims with dates of injury on or after that date,

and became effective on September 1, 2013, for legacy claims with dates of injury before September 1,

2011. According to a study by the Texas Department of Insurance (TDI), fewer opioids and other not-

recommended drugs are being prescribed after the reform (TDI, Texas Workers’ Compensation Research

and Evaluation Group, 2013). In 2010, Texas also passed legislation to regulate pain clinics, requiring

pain management clinics that supply more than 50 percent of their patients with controlled substances to

register with the state medical board.41

The prevalence of longer-term opioid dispensing peaked in 2011/2013 in several study states. Among

these states, Arkansas, Georgia, and Nevada saw a significant decrease in the latest two years of the study

period.

Arkansas: In 2013/2015, 4.4 percent of nonsurgical claims with opioids received opioid prescriptions on a

longer-term basis. The figure was 6.2 in 2010/2012, and it increased to 8.2 percent in 2011/2013.

Arkansas established the state PDMP in 2011, and the program was fully functioning in 2013.

Recognizing the need for more comprehensive policies related to opioid misuse and abuse in the state,

the 90th General Assembly introduced and passed several bills during the 2015 legislative session that

38 See Chang, et al. (2016). 39 In a study that examined how the state’s laws impact prescribing and use of opioids in 2010 through September 2012, the researchers reported reductions among prescribers and patients who had the highest prescribing and usage of opioids. They estimated that opioid prescriptions by Florida’s top opioid prescribers fell 6.2 percent, and the total volume prescribed by this group dropped 13.5 percent; the number of patients also dropped by 5.1 percent. The top 4 percent included 1,526 providers out of 38,465 in the state. Among the remaining 96 percent, prescriptions slipped a mere 0.7 percent (Rutkow et al., 2015). Also see Chang et al. (2016). Other states where similar laws are in place include Alabama, Florida, Georgia, Indiana, Kentucky, Louisiana, Mississippi, Ohio, Tennessee, Texas, and West Virginia. 40 See the Guidelines for Opioid Management of Pain Using Dangerous Drugs and Controlled Substances, which is available at http://druglibrary.org/schaffer/asap/florida_guide.htm. 41 See Lyapustina et al. (2016).



44

complement existing law.42 The 2015 legislation provides a comprehensive approach to the prevention

and intervention of opioid overdoses, including access to the state PDMP database for various

stakeholders (prescribers and their employees, professional licensing boards, and certified law

enforcement investigators); mandatory checks of prescription history, using the PDMP, for chronic pain

patients every six months; random urine drug testing and pill counts; and the requirement for all patients

to be evaluated by a physician every six months. Arkansas was one of the few states that participated in

the Prescription Drug Abuse Reduction Policy Academy, a year-long program in strategic planning

aimed at reducing prescription drug abuse.43

Georgia: At 10.1 percent, the frequency of claims receiving longer-term opioid prescriptions peaked in

2011/2013 (it was 8.5 percent in 2010/2012). A large decrease was seen in 2012/2014, with the percentage

of claims with longer-term opioids at 8.2 percent in 2012/2014, a nearly 2 percentage point reduction

from the peak year. In Georgia, the state PDMP system became operational in 2013. The state medical

board adopted Guidelines for the Use of Controlled Substances for Treatment of Pain in 2008. Laws

setting dosage limits and prohibiting doctor shopping were also passed.

Nevada: The frequency of claims receiving longer-term opioids decreased from 6.7 percent in 2011/2013

to 3.9 percent in 2013/2015. The figure in 2013/2015 was also lower than that in 2010/2012 (5.5 percent),

but the decrease was not statistically significant. Nevada is one of the states that had a PDMP in place in

earlier years, but prescriber checks of the state PDMP database were not mandatory until October 2015.

The state workers’ compensation agency adopted the American College of Occupational and

Environmental Medicine (ACOEM) guidelines for chronic opioid management and evidence-based

formulary. In the 2014 regulation, the state adopted the Federation of State Medical Boards’ Model

Policy by referencing it in the state regulation. Nevada also participated in the National Governors

Association’s Prescription Drug Abuse Reduction Policy Academy, which was first launched in 2012 to

assist states with strategic action plans that meet the unique needs of each state to reduce prescription

drug abuse.44 The Academy-proposed plan recommendations helped Nevada pass a comprehensive

reform that is expected to have a significant impact on opioid overdose prevention and intervention.45

In several states, the prevalence of longer-term opioid dispensing decreased over the study period by 1

percentage point or more, including Connecticut, Illinois,46 Maryland, North Carolina, New Jersey, and

Pennsylvania. However, the noticeable decreases in these states were not statistically significant at the 95

percent confidence level. All these states had a number of opioid-related policies implemented during and

after the study period. We will continue to monitor trends in these states in the use and longer-term use of

42 Prior to 2015, Arkansas passed a law in 2011 to establish the state PDMP and enable access to PDMP information to authorized individuals and agencies. In 2013, laws were also passed to prohibit retrieving drugs through forgery, fraud, or deceit to prevent doctor shopping. See the report published by the Arkansas Center for Health Improvement entitled Prevention of Opioid Pain Reliever Misuse in Arkansas, which is available at http://www.achi.net/Docs/333/. 43 The academy was first launched in 2012 with the goal to help the participating states in developing and implementing comprehensive and coordinated strategies that take advantage of all available tools and resources to address the growing opioid problem. The academy identified a set of key strategies that states can adopt, including better use of the state PDMP database, provider and patient education, law enforcement, and partnership among key stakeholders. See NGA (2012) available at http://www.nga.org/files/live/sites/NGA/files/pdf/1209ReducingRxDrugsBrief.pdf. 44 Seven states were participants when the academy was first launched in 2012, including Alabama, Arkansas, Colorado, Kentucky, New Mexico, Oregon, and Virginia. The academy has published several papers since then to discuss lessons learned from the participating states and provide strategic action plan recommendations. See NGA (2015). 45 The state legislation mandates that prescribers check the prescription history in the state PDMP database when prescribing opioids for new patients, effective October 2015. The state also requires at least one hour of CME on opioid prescribing and abuse prevention for license renewal. 46 In Illinois, the frequency of longer-term opioid dispensing decreased by only 0.5 percentage points between 2010/2012 and 2013/2015, but a 1 point decrease was seen between 2011/2013 and 2013/2015.



45

opioids in future editions of this study.

By contrast, the prevalence of longer-term dispensing of opioids changed little in Indiana, Iowa,

Missouri, South Carolina, Virginia, and Wisconsin. The year-to-year figures fluctuated over the study period

in Indiana, South Carolina, and Wisconsin. Note that in the 2016 edition of this report, we discussed a

noticeable increase in the percentage of nonsurgical claims with opioids that were identified as having longer-

term opioids in Indiana and Wisconsin. We suspected that the results might be due to a temporary

fluctuation instead of a long-term trend. With additional data covering one more year, we found that the

same measure for 2013/2015 claims was back to the 2010/2012 level for the two states. In 2013/2015, Indiana

and Wisconsin continued to be among the states with the lowest prevalence of longer-term opioid use,

among the 26 study states (Figure 3.2). We are not aware of any major changes over the study period that

might help explain the temporary fluctuation. In Louisiana, the frequency of longer-term opioid use

increased slightly between 2010/2012 and 2013/2015, but the trend was not statistically significant (Figure

3.4). The prevalence of longer-term opioid use peaked in 2011/2013, and we saw a slight decrease in the figure

in the latest two study years, but it was still higher than in 2010/2012 (Table 3.2).

UTILIZATION OF TREATMENT GUIDELINE-RECOMMENDED SERVICES

Because of the potential risks of heavy and prolonged use of opioids, most guidelines recommend careful

screening of patients prior to the use of chronic opioid therapy and close monitoring and management

through drug testing and psychological evaluation and treatment.47 Guidelines also recommend that opioids

be used as part of comprehensive care, including active physical therapy and exercises to promote timely

recovery.

Table 3.3 shows the percentage of claims receiving longer-term opioids that ever received drug testing,

for selected states with large enough sample sizes.48 Among these states, the claim frequency of receiving drug

testing ranged from 22 percent in Michigan to 59 percent in Tennessee for 2013/2015 claims with longer-term

opioids. This was an increase from what we saw in 2010/2012 (15 percent in Michigan to 43 percent in

Georgia), and the increase in claim frequency was seen in all states reported except for Georgia and South

Carolina (0–3 percentage point change).

In the latest two study years, the claim frequency of drug testing appeared to level off in most of the 17

states except for a few. In Connecticut, Florida, Illinois, North Carolina, and Tennessee, the frequency of

receiving drug tests continued to increase in the latest year, while the rate of use appeared to fluctuate in

Michigan and Virginia. In 2013/2015, the frequency of drug testing was still low, at least in several states with

a lower rate of claims receiving drug testing. However, we saw an unusually high utilization of drug testing

among the top 5 percent of claims with longer-term opioids receiving drug testing.

47 Guidelines recommend that all patients be screened for potential alcohol and drug abuse problems and psychological issues, since these patients are less likely to succeed with chronic opioid treatment and need close monitoring. Chronic opioid management requires a comprehensive treatment approach with clear functional goals agreed upon between the physician and patient. Guidelines also recommend careful monitoring, and management includes random urine drug screening, periodic assessment and evaluation of function and side effects, and tapering of opioid medication when the goals and patient behavior expectations are not met. See Technical Appendix B for a summary of recommendations by guidelines for chronic opioid management. 48 Due to a small sample size concerns, we provide the results on drug testing only for states with at least 100 claims with longer-term use of opioids. See Chapter 2 for a discussion regarding sample size and state selection for reporting.



46

Table 3.3 Trends in the Frequency of Drug Testinga for Claims Receiving Longer-Term Opioidsb

CA CT FL GA IL KY LA MA MD MI NC NY PA SC TN TX VA 17-State Median

Number of cases that had urine drug testing

2010/2012 1,793 113 449 247 208 137 198 155 148 163 243 450 446 130 163 1,010 104

2011/2013 1,725 123 436 278 244 149 216 140 157 140 207 432 437 146 124 935 132

2012/2014 1,781 126 405 253 187 101 199 114 116 114 204 370 372 114 128 929 104

2013/2015 1,592 106 338 250 185 59 190 111 111 101 198 233 357 130 89 931 99

% of claims receiving longer-term opioids that had urine drug testing

2010/2012 38% 27% 40% 43% 19% 19% 24% 15% 30% 15% 31% 32% 20% 32% 39% 35% 24% 30%

2011/2013 46% 21% 43% 41% 22% 31% 42% 30% 38% 20% 35% 44% 28% 43% 40% 39% 40% 39%

2012/2014 49% 34% 44% 44% 22% 34% 44% 23% 48% 26% 33% 51% 35% 31% 42% 41% 40% 40%

2013/2015 49% 40% 48% 42% 26% 32% 43% 25% 46% 22% 40% 48% 37% 35% 59% 44% 33% 40%

% point change 2010/2012–2013/2015c 11 13 7 0 7 13 19 10 16 7 9 16 17 3 20 9 9 1

Notes: The underlying data include nonsurgical claims that are identified as receiving longer-term opioids. These claims had more than seven days of lost time and prescriptions filled over the defined period and paid for by a workers' compensation payor. 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015. Similar notation is used for other years. There are 17 states included in this table. States excluded are those with fewer than 100 claims receiving longer-term opioids in the study sample.

a Drug testing services were identified using CPT and HCPCS codes; see Chapter 2 and Table 2.3 for the definition of drug testing services. In this analysis, all recommended services were identified as those paid services provided and billed by hospital and nonhospital providers.

b We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury (i.e., Type I claims). See Chapter 2 for more details. c Percentage point changes shown may not agree with the difference in the percentage of claims with longer-term use of opioids that had drug testing between 2010/2012 and 2013/2015 due to rounding.

Key: CPT: Current Procedural Terminology; HCPCS: Healthcare Common Procedure Coding System.

There are two important questions regarding the use of drug testing services: (1) Was the frequency in

use of these recommended drug testing services low? (2) How were the services for drug testing utilized

among the injured workers with longer-term opioids who received drug testing?

On one hand, although the optimal rate of use for drug testing among claims receiving longer-term

opioids is unknown, the rate of use for drug testing services appeared to be low, at least for the states with a

lower percentage of claims with longer-term opioid prescriptions that had drug testing. In Illinois,

Massachusetts, and Michigan, the figure was 22–26 percent. In these states, the use of drug testing for

managing chronic opioid therapy was likely low.49 In the states with the highest rates on this measure

(California, Florida, Maryland, New York, and Tennessee), 46–59 percent of injured workers with longer-

term use of opioids received drug testing. The increase in the frequency of drug testing over the study period

might reflect an increased compliance with treatment guideline recommendations as well as an increased

supply of drug testing kits.50

49 Treatment guidelines recommend drug screening for all patients who are prescribed chronic opioid therapy and random urine drug testing after the patient has started chronic opioid therapy (see Technical Appendix B for more discussion). Because of this, we expect to see a high rate of drug testing among this group of claims. Recognizing that some drug testing services may be provided but not paid for by workers’ compensation payors and some claims we identified as having longer-term opioids may not be patients with chronic pain who are prescribed chronic opioid therapy, we take a conservative approach to discussing the compliance with treatment guidelines regarding drug testing. 50 It is also possible that the increase in the frequency of drug testing was in part due to more serious claims with longer-term opioids as a result of the decreases in the percentage of claims with longer-term opioids between 2010/2012 and 2012/2014. If this was the case, we would expect to see a decrease in the frequency of longer-term opioid dispensing being correlated with an increase in the use of drug testing. Our analysis showed a negative correlation, but the size was small.



47

On the other hand, we saw unusually frequent drug tests performed for a small percentage of injured

workers with longer-term opioids who received drug testing.51 Table 3.4 provides the average number of visits

per claim for drug testing and the average number of tests done per visit for the nonsurgical claims with

longer-term opioids that received drug testing. It also shows the utilization of drug tests at the median and

several percentiles on the higher end (90th, 95th, and 98th percentiles). For example, the top 5 percent of the

2013/2015 claims in most study states that had longer-term opioids received at least 7 to 10 visits for drug

testing, depending on the state.52 The figure was 11–12 visits in states with the highest rates (Louisiana,

Maryland, and Massachusetts) and 3–4 visits in the states with lowest rates (Michigan and Minnesota).

Among the top 5 percent of claims with drug tests in most states, 10–13 different drug tests were performed

during a single visit; the number ranged from 6 in Kansas to 15 in Texas. All chronic pain treatment

guidelines recommend a baseline drug test before the patient is prescribed chronic opioid therapy, and most

guidelines recommend up to four visits for drug tests per year, depending on whether the patient is at low risk

or high risk of opioid misuse. The number of visits and services for drug testing was high, especially in

Louisiana, Maryland, and Massachusetts, for the top 5 percent of claims with longer-term opioid

prescriptions. The results may reflect changes in the behavior of some medical providers and suppliers in

response to the financial incentives embedded in the practice of in-office drug testing, which should be

investigated.

51 According to chronic pain treatment guidelines, urine drug testing should be done as a screening test and performed once the patient is on chronic opioid therapy—once a year for low-risk patients and three to four times a year for high-risk patients. See Technical Appendix B for treatment guideline recommendations. 52 The estimated number of visits (i.e., unique dates) for drug tests was based on 17 states for which we had a large enough sample. We excluded Arkansas, Indiana, Iowa, Kansas, Massachusetts, Minnesota, Missouri, and Nevada from this estimation due to small cell sizes.



48

AR CA CT FL GA IA IL IN KS KY LA MA MD MI MN MO NC NJ NV NY PA SC TN TX VA WIMedian

State

Mean – 3.9 2.9 3.2 3.1 – 2.8 2.1 – 2.8 3.7 3.0 4.2 1.9 1.5 – 2.7 3.0 – 3.3 2.6 2.5 3.3 2.7 2.6 2.0 2.8

Median – 2.8 1.8 2.0 1.9 – 1.8 1.3 – 1.9 2.0 1.8 3.5 1.5 1.3 – 1.8 1.5 – 2.3 1.5 1.8 2.1 2.0 2.0 1.4 1.8

90th percentile – 8.5 5.8 7.0 7.0 – 5.8 5.8 – 5.5 8.1 7.5 8.3 3.5 2.8 – 5.3 8.0 – 7.3 5.3 4.8 7.6 5.5 5.0 4.3 5.8

95th percentile – 10.5 8.8 10.5 8.8 – 8.5 8.0 – 9.3 11.0 11.3 12.0 4.3 3.5 – 8.3 9.8 – 10.8 7.8 7.0 10.5 7.8 6.5 5.5 8.8

98th percentile – 14.5 13.5 14.8 12.8 – 14.3 9.3 – 14.0 16.8 14.0 13.5 5.3 3.5 – 12.8 16.3 – 14.3 13.5 10.5 13.4 10.8 9.3 5.5 13.5

Mean 6 4 6 6 6 6 3 4 3 4 4 5 6 6 4 3 5 6 5 6 5 6 5 6 5 4 5

Median 5 2 6 6 5 4 2 2 2 2 2 3 6 5 2 3 4 5 4 5 4 7 3 4 5 4 4

90th percentile 11 9 11 12 12 12 9 11 5 8 11 11 10 12 8 8 11 11 10 11 10 11 11 13 10 8 11

95th percentile 11 11 13 12 14 12 10 11 6 10 13 12 12 13 10 9 13 12 11 12 12 12 12 15 11 10 12

98th percentile 11 12 14 13 15 12 12 12 6 12 16 13 14 15 10 9 15 13 11 13 14 14 14 18 12 12 13

Key: –: not reported due to small sample size.

Table 3.4 Utilization of Drug Testing Services, among Nonsurgical Claims with Longer-Term Opioids That Received Drug Testing, Pooled 2010/2012–2013/2015 Claimsb


b Claims included are those that were identified as receiving longer-term opioids and received drug testing services. See Chapter 2 for more details.

a In this analysis, all recommended services for drug testing were identified as those paid services provided and billed by hospital and nonhospital providers.

Number of visits during which drug tests were performed

Number of drug tests per visit

a



49

Table 3.5 illustrates the utilization and costs of drug testing services among claims with longer-term

opioids in the four study years for states with a large enough sample. As the table shows, the average number

of visits for drug testing increased substantially (by more than 25 percent) between 2010/2012 and 2013/2015

for several states—Florida, Illinois, Louisiana, New York, Pennsylvania, South Carolina, Tennessee, and

Texas. Increases were also seen in the number of drug tests per visit; they were significant in Louisiana, North

Carolina, Pennsylvania, South Carolina, Tennessee, and Texas. Substantial increases in costs per claim for

drug testing were also seen over the same period in several states. By 2013/2015, the average payment per

claim for drug testing exceeded $1,000 for all states included in the analysis; it was the highest in Louisiana at

$2,520 per claim. One may wonder if the substantial increase in the utilization of drug testing services reflects

a shift in severity between 2010/2012 and 2013/2015 claims. Since we observed considerable decreases in the

frequency of longer-term opioid use in a number of states (Figure 3.4), it is possible that the 2013/2015 claims

were, on average, more severe than the 2010/2012 claims. If so, we would expect to see that the states with

decreased frequency of longer-term opioid dispensing had larger increases in the utilization and costs of drug

testing services. However, this is not supported by the data (Table 3.3 and Figure 3.4).

Table 3.5 Utilization and Costs of Drug Testinga Services, among Nonsurgical Claims Receiving Longer-Term Opioids That Had Drug Testingb

CA FL GA IL LA MD NC NY PA SC TN TX 12-State Median

% of cases that had urine drug testing

2010/2012 38% 40% 43% 19% 24% 30% 31% 32% 20% 32% 39% 35% 32%

2011/2013 46% 43% 41% 22% 42% 38% 35% 44% 28% 43% 40% 39% 41%

2012/2014 49% 44% 44% 22% 44% 48% 33% 51% 35% 31% 42% 41% 43%

2013/2015 49% 48% 42% 26% 43% 46% 40% 48% 37% 35% 59% 44% 44%

% point change 2010/2012–2013/2015c 11 7 0 7 19 16 9 16 17 3 20 9 12

Average number of visits during which drug tests were performed per claim

2010/2012 3.4 2.7 2.6 1.9 2.7 3.8 2.6 2.8 1.8 1.8 2.3 2.5 2.6

2011/2013 3.9 3.0 2.6 2.9 3.2 3.8 2.0 2.6 2.1 2.6 3.3 2.7 2.8

2012/2014 4.0 3.0 3.4 3.2 4.4 4.1 3.0 3.7 3.0 2.2 4.0 2.8 3.3

2013/2015 3.8 3.4 2.7 2.7 4.4 4.3 2.8 4.2 2.9 2.5 3.7 3.2 3.3

% change 2010/2012–2013/2015 12% 27% 7% 39% 64% 12% 10% 50% 59% 42% 60% 28% 28%

Average number of drug tests per visit

2010/2012 4.0 5.7 5.5 4.0 4.7 5.5 3.8 5.8 4.3 4.2 3.7 4.9 4.5

2011/2013 3.7 4.7 5.2 3.1 4.0 5.9 4.4 5.5 5.1 5.5 5.3 6.3 5.2

2012/2014 3.6 6.8 6.4 3.8 4.6 5.1 5.4 6.8 5.7 7.2 5.0 7.1 5.6

2013/2015 4.0 6.1 5.5 4.0 5.7 6.5 5.2 5.5 5.3 5.7 5.0 7.0 5.5

% change 2010/2012–2013/2015 1% 7% 1% 1% 21% 18% 35% -5% 25% 38% 34% 42% 25%

Average payment per claim for drug testing services

2010/2012 $1,486 $767 $951 $923 $960 $1,929 $746 $884 $644 $540 $822 $922 $903

2011/2013 $1,872 $805 $1,117 $1,104 $1,188 $2,150 $696 $829 $909 $1,024 $1,451 $1,087 $1,095

2012/2014 $2,043 $1,070 $1,646 $1,596 $2,262 $1,593 $1,373 $1,181 $1,513 $1,132 $1,732 $1,477 $1,553

2013/2014 $1,655 $1,362 $1,314 $1,003 $2,520 $1,970 $1,150 $1,245 $1,407 $1,017 $1,861 $1,613 $1,384

% change 2010/2012–2013/2015 11% 78% 38% 9% 163% 2% 54% 41% 119% 88% 126% 75% 53%

continued



50

Table 3.5 Utilization and Costs of Drug Testinga Services, among Nonsurgical Claims Receiving Longer-Term Opioids That Had Drug Testingb (continued)


a In this analysis, all recommended services were identified as those paid services provided and billed by hospital and nonhospital providers.

b Claims included are those that were identified as receiving longer-term opioids and received drug testing services. See Chapter 2 for more details.

c Percentage point changes shown may not agree with the difference in the percentage of claims with longer-term use of opioids that had drug testing between 2010/2012 and 2013/2015, due to rounding.

The results on drug testing suggest two issues: (1) fewer than expected injured workers with longer-term

opioids received drug testing, and (2) for some injured workers who received drug testing, the frequency and

per-claim utilization of drug tests was unusually high. Further analysis of these issues is needed with

additional data to monitor how the guideline recommendations were followed and to detect any behavioral

changes in the use of drug testing services.

We continued to find that few claims with longer-term opioids received psychological evaluations and

treatment, services recommended by treatment guidelines for chronic opioid management. Table 3.6 shows

the percentage of claims with longer-term opioids that ever received these recommended services, based on a

pooled sample of claims over the three-year period from 2010/2012 to 2013/2015.53 For psychological

evaluations, Texas had the highest frequency of use of these services, but only about one in three claims with

longer-term opioids had psychological evaluations in Texas. In all the other states, the figure ranged from 4 to

18 percent, with the median state at 7 percent. The use of psychological treatment services was infrequent in

all study states. We should caution the reader that the psychological evaluation measure might be understated

to the extent that certain psychological evaluations during regular office visits were not captured by the

procedure codes, which we used to identify psychological evaluation and treatment.54 If the rate of use for

psychological evaluations was low, certain psychological comorbidities could be under-diagnosed among the

injured workers with longer-term use of opioids, and consequently, these injured workers might not be

treated for their psychological comorbidities.

An important purpose of opioid therapy is to facilitate active physical therapy, including therapeutic

exercises and activities that encourage patient participation and promote functional recovery. These exercises

and activities are recommended by guidelines for chronic opioid management as part of comprehensive

treatment for patients with chronic pain. For example, the ACOEM guidelines state that a therapeutic

exercise prescribed should address specific treatment goals and be time limited, with a transition to an

independent exercise program as part of a healthy lifestyle. The purpose of supervised exercise therapy is

symptom reduction, functional improvement, and education of the patient so that he or she can

independently manage the program, according to ACOEM.55 We found that while a vast majority of injured

53 For this edition, we decided to combine the data for 2010/2012, 2011/2013, and 2012/2014 claims with longer-term use of opioids to support the analysis on claim frequency of psychological services. Unlike what was presented in the previous editions, we took a conservative approach recognizing several potential issues associated with this analysis, including small sample sizes for several states, relatively low probability of receiving psychological services, and possible reasons for a potential understatement. 54 If a treating physician performs a psychological evaluation in an office setting and bills the service as an office visit for evaluation and management, we cannot use the office visit code to identify psychological services because of the lack of differentiation from other routine office visits. See Chapter 2 for a more detailed discussion of this limitation. 55 Since independent exercises are not considered treatment, we did not observe these activities in the data. However, because the educational and training components of the self-directed exercises are likely to be part of therapeutic exercises



51

workers with longer-term opioids received active physical therapy across the states studied, more than 10

percent of claims with longer-term opioids did not receive such services in most study states (Table 3.7). In

the states with the lowest rate of use—it appeared that one-fifth of the injured workers who had longer-term

use of opioids did not receive active physical therapy.

It is worth noting that some claims receiving longer-term opioids based on our definition may not have

had chronic pain but instead had episodic pain or short-term pain from a source other than the original

injury. If true, these cases are not considered cases with chronic pain; therefore, chronic opioid therapy

should not be provided. Even if we assume that there were a considerable number of such cases, the frequency

of using recommended services for chronic opioid management was low. On the other hand, those injured

workers who had episodic or short-term pain and received opioids on a longer-term basis might be

unnecessarily exposed to the risk of side effects, addiction, and even overdose death. For those injured

workers, more accurate clinical diagnoses of underlying medical conditions, initial screening for chronic

opioid therapy, and judicial prescribing of opioids may be helpful to avoid potentially unnecessary harm.

Our findings raise several questions about longer-term dispensing of opioids that need to be further

investigated: How many injured workers who are identified as having longer-term opioids do not benefit

from chronic opioid therapy? Can unnecessary long-term opioid prescriptions be prevented and if so, how?

For injured workers with chronic pain who receive longer-term opioids, do they have better recovery and

return to work sooner compared with those who do not receive longer-term opioid therapy? What are the

alternative treatments for injured workers with chronic pain? More rigorous methodology and additional

data are needed to answer these questions.

supervised or attended by the provider, we are likely to capture the claim frequency of having these exercises, based on the set of active physician therapy codes we identified (see Table 2.3).



52


Number of claims 139 6,892 468 1,629 1,029 165 824 301 168 445 802 520 532 517 380 218 852 409 165 1,484 1,612 520 503 3,806 439 244

% of claims receiving psychological evaluation services 7% 13% 8% 6% 7% 13% 7% 9% 5% 4% 11% 7% 5% 11% 18% 3% 5% 7% 5% 7% 5% 7% 7% 30% 7% 13% 7%

% of claims receiving psychological treatment 1% 3% 2% 3% 3% 10% 3% 5% 4% 1% 5% 3% 3% 5% 9% 2% 2% 4% 3% 5% 2% 3% 2% 12% 3% 11% 3%

Table 3.6 Frequency of Claims Receiving Psychological Services,a 2011/2013–2013/2015 Nonsurgical Claims Identified as Receiving Longer-Term Opioidsb

Notes: The data are based on a pooled sample of nonsurgical claims that were identified as receiving longer-term opioids. These claims had more than seven days of lost time and prescriptions filled over the defined period and paid for by a workers' compensation payor. 2011/2013 refers to claims with injuries occurring from October 1, 2010, through September 30, 2011, and prescriptions filled through March 31, 2013. Similar notation is used for other years. The pooled sample includes claims across three years, from 2011/2013 to 2013/2015, that were identified as receiving longer-term opioids.

a The psychological services were identified using CPT codes; see Chapter 2 and Table 2.3 for the definition of psychological services. In this analysis, all recommended services were identified as those paid services provided and billed by hospital and nonhospital providers.

b We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury (i.e., Type I claims). See Chapter 2 for more details.

Key: CPT: Current Procedural Terminology.



53


Number of claims 139 6,892 468 1,629 1,029 165 824 301 168 445 802 520 532 517 380 218 852 409 165 1,484 1,612 520 503 3,806 439 244

% of claims receiving active physical therapy services 79% 91% 86% 89% 87% 82% 88% 87% 73% 78% 82% 85% 89% 86% 91% 81% 88% 88% 75% 90% 85% 87% 87% 87% 86% 81% 87%

Table 3.7 Frequency of Claims Receiving Active Physical Therapy Services,a 2011/2013–2013/2015 Nonsurgical Claims Identified as Receiving Longer-Term Opioidsb

Key: CPT: Current Procedural Terminology.

Notes: The data are based on a pooled sample of nonsurgical claims that were identified as receiving longer-term opioids. These claims had more than seven days of lost time and prescriptions filled over the defined period and paid for by a workers' compensation payor. 2011/2013 refers to claims with injuries occurring from October 1, 2010, through September 30, 2011, and prescriptions filled through March 31, 2013. Similar notation is used for other years. The pooled sample includes claims across three years, from 2011/2013 to 2013/2015, that were identified as receiving longer-term opioids.

a The active physical therapy services were identified using CPT codes; see Chapter 2 and Table 2.3 for the definition of active physical therapy services. In this analysis, all recommended services were identified as those paid services provided and billed by hospital and nonhospital providers.

b We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury (i.e., Type I claims). See Chapter 2 for more details.



54

4

IMPLICATIONS AND CONCLUSIONS

Long-term use of opioids has been controversial, and there is little quality evidence about the effectiveness of

long-term opioid therapy for chronic non-cancer pain in terms of function recovery or return to work.1

Studies found that for patients with occupational injuries, a higher use of opioids may lead to addiction,

increased disability or work loss, and even death.2 Since the CDC declared a prescription opioid epidemic in

2011,3 an increasing number of states have made legislative and regulatory changes, within and outside

workers’ compensation, to prevent misuse and overuse of prescription opioids.4 The most common changes

that have been enacted include mandatory requirements for physicians and pharmacies to check state PDMP

databases prior to prescribing and dispensing opioids, mandatory requirements for CME on appropriate

opioid prescribing and pain management, and state laws that regulate pain clinics and limit physicians’ ability

to dispense controlled substances.5 In addition, many states have adopted or updated clinical guidelines for

prescribing opioids and managing patients with chronic opioid therapy, in and outside workers’

compensation systems. All guidelines require treating physicians to screen their patients, through physical

examination, medical and psychosocial assessment, and history of drug and alcohol use, and identify those

who are most likely to benefit from chronic opioid therapy while mitigating the potential risk of opioid

misuse and abuse. Guidelines also require conducting drug testing at baseline and during chronic opioid

therapy to prevent misuse, documenting progress of the treatment, evaluating for psychological/psychiatric

issues, and referring the patients for whom chronic opioid treatment may not be effective to specialists for

alternative pain management.6 More recently, several state workers’ compensation systems also adopted a

closed formulary aimed at addressing the utilization of certain opioid medications.

We found that the prevalence of longer-term opioid dispensing decreased substantially in Kentucky and

New York between 2010/2012 and 2013/2015. The substantial decrease in the claim frequency of longer-term

opioids was also coupled with a large decrease in the percentage of claims that received opioids. Kentucky and

1 Although several studies have documented some benefits of long-term opioid therapy for limited pain relief (see a more detailed discussion in Wang, Mueller, and Hashimoto, 2011), no quality studies have been published that support chronic opioid use for improved function or rapid return to work. For patients with occupational injuries, several studies found that a higher use of opioids may lead to addiction, increased disability or work loss, and even death (Kidner, Mayer, and Gatchel, 2009; Franklin et al., 2005; and Volinn, Fargo, and Fine, 2009). 2 See Kidner, Mayer, and Gatchel (2009) and Franklin et al. (2005). 3 See the CDC’s press release on November 1, 2011, which is available at http://www.cdc.gov/media/releases/2011/p1101_flu_pain_killer_overdose.html. Also see CDC Grand Rounds: Prescription Drug Overdoses—a U.S. Epidemic published on January 13, 2012, which is available at http://www.cdc.gov/mmwr/preview/mmwrhtml/mm6101a3.htm. 4 The 2011 report by the CDC showed that the number of prescription drug overdose deaths has exceeded the number of deaths due to traffic accidents. See CDC (2011), available at http://www.cdc.gov/nchs/data/databriefs/db81.pdf. 5 See Technical Appendix A for more details. 6 See Technical Appendix B for more details.



55

New York had comprehensive opioid reforms over the study period. In Kentucky, House Bill 1, which went

into effect in July 2012, includes a comprehensive set of provisions for pain clinic regulation, rules for opioid

prescribing and dispensing, and mandatory checks of patients’ prescription history using the state PDMP

(KASPER). New York passed a legislative mandate that requires prescribers to check the state PDMP database

at the point of prescribing opioids. Opioid policy changes in New York also include the publication of non-

acute medical treatment guidelines in 2014 that were aimed at addressing issues with long-term opioid use in

the state workers’ compensation system. The substantial reduction in the dispensing and longer-term

dispensing of opioids in these two states may be associated with these comprehensive opioid policies that

were in place during the study period. The prevalence of longer-term opioid dispensing also saw considerable

reduction over the study period in several other states, including Kansas, Massachusetts, Michigan,

Minnesota, and Tennessee. A noticeable decrease on the same measure was also observed in California,

Florida, Maryland, New Jersey, and Texas. In these states, there have been many changes in opioid policies

and initiatives that are aimed at reducing opioid prescriptions. Many of these states adopted opioid policies,

after the study period, which have been reported to be effective for reducing unnecessary opioid

prescriptions. Many of these policies are expected to have a positive impact on the prevalence of use and

longer-term use of opioids. We will continue to monitor the trends in this area, with additional data covering

more recent years.

However, longer-term opioid dispensing continued to be more frequent in Louisiana, where one in six

injured workers who did not have surgery but had opioids received them on a longer-term basis. The

frequency of claims receiving longer-term opioids in Louisiana peaked in 2011/2013 and decreased slightly in

the latest two study years, but it was still higher than in 2010/2012 and 2012/2014. In California, Georgia,

North Carolina, Pennsylvania, South Carolina, and Texas, the frequency of longer-term opioids was

noticeably higher than the median of the 26 states. There have been studies examining factors in the early

stage of treatment that may be associated with receiving chronic opioid therapy. These factors include early

opioid use, presence of psychological and psychiatric conditions, and substance use disorders (SUDs). These

early factors that may likely influence the likelihood of receiving long-term opioids should be addressed in

practice to reduce the prevalence of claims receiving long-term opioids. In addition, most treatment

guidelines for chronic pain recommend careful screening of patients for chronic opioid therapy, which is also

an important tool to address the prevalence of longer-term opioid use. Policymakers in these states should

look at the opioid policies that were in place to see if the adopted policies are effective and if any additional

measures need to be considered to reduce unnecessary opioid prescriptions.

We saw considerable increases in the frequency of drug testing services over the study period among

injured workers receiving opioids on a longer-term basis, but the increases slowed down in the latest study

year, and in a few states the rate of drug testing decreased. While the frequency of claims using these drug

testing services appeared to be low (considering that almost all chronic opioid guidelines recommend drug

screening/testing for all patients with long-term opioid therapy), at least for some states, we saw an unusually

high utilization of drug tests among the top 5 percent of injured workers who had longer-term use of opioids

and received drug testing. The results may suggest that some physicians conducted more drug testing than

needed in response to the economic incentives imbedded in office testing. We continued to observe a low rate

of use of psychological evaluation and treatment among claims with longer-term use of opioids. Medical

treatment guidelines recommend psychological evaluation for most patients with chronic pain. The

infrequent use of psychological evaluations may result in under-diagnosis of psychological comorbidities

among injured workers with longer-term use of opioids. If this is the case, those injured workers receiving

longer-term opioids who have psychological comorbidities may be under-treated.



56

CONCLUDING REMARKS

By tracking the changes in the prevalence of longer-term opioid dispensing, this report helps to monitor the

possible impact of public policy changes on long-term opioid prescribing and dispensing. While the report

highlights a considerable decrease in the prevalence of opioid use in a number of states, it also discusses

prevalent prescribing and dispensing of opioids on a longer-term basis observed in several states. The findings

provide a benchmark that policymakers and stakeholders can use to better target their efforts to address

possible overuse and diversion of opioids in their states. The information provided can also help payors target

efforts to better manage the use of opioids while providing appropriate care to injured workers and reducing

unnecessary risks to patients and unnecessary costs to employers. This study may also be used as an

important educational tool for the community of workers’ compensation medical providers in each state to

compare their practice patterns with the norms seen across the 26 study states. Some providers may modify

their practice patterns after seeing the practice norms. Limited in the scope of what policy questions this

report can answer, this update leaves many important policy questions unanswered. Future studies should

focus on providing insights as to which policies or initiatives are effective at reducing unnecessary opioid

prescriptions that may put some injured workers at risk for potential harm.



57

TECHNICAL APPENDIX A:

A BRIEF SUMMARY OF FACTORS THAT MAY

INFLUENCE THE PRESCRIBING OF OPIOIDS

The reader may want to know what might explain the results we observed for each state. In this report, we do

not identify and examine the impact of key factors underlying the utilization and prescribing patterns of

opioids. Instead, we provide in this appendix some background information about several possible factors

that may influence prescribing and utilization patterns of opioids, including information about the legal and

regulatory environment for prescribing opioids for some state workers’ compensation health care delivery

systems.1 Our intention is to provide the reader with a policy context that may help facilitate the

interpretation of the patterns observed.

The development of policies regarding controlled substances reflects a legislative and regulatory effort to

strike a balance between providing necessary pain relief and minimizing the risk of abuse and diversion of

opioids. In the past few years, there were numerous legislative and regulatory changes in opioid policies that

were aimed at reducing unnecessary use of opioids. Some of the policy changes were at the federal level,

including up-scheduling certain controlled substances2 and Risk Evaluation and Mitigation Strategies

(REMS) programs. Many changes were at the state level, in and outside workers’ compensation systems. Most

of the policy changes have been focused on changing provider behavior for safe opioid prescribing and

dispensing. Numerous guidelines have also been adopted by different agencies to address not only the

management of chronic opioid use but also opioid prescribing for acute and sub-acute pain. More recently,

an increasing number of states took a comprehensive approach to addressing opioid issues in a coordinated

way, focusing on the prevention and intervention of opioid overdoses. The policy discussion we provide in

this appendix is focused on the prevention of opioid misuse and abuse.

FEDERAL LAWS REGARDING PRESCRIPTIONS OF OPIOIDS

The federal Controlled Substance Act (CSA),3 as Title II of the Comprehensive Drug Abuse Prevention and

Control Act of 1970, established a classification structure by categorizing controlled substances into five

schedules based on their medicinal value and potential for abuse, addiction, and dependency. Controlled

substances in Schedules II through V can be used for medical purposes. Table 2.2 in Chapter 2 provides the

definition of each schedule and examples of specific opioid medications that are classified in each schedule.

1 Other non-policy factors are addressed in Technical Appendix C, where we discuss the case-mix adjustment for measuring the prevalence of longer-term opioid dispensing as well as other technical issues. 2 Examples include the change of hydrocodone-combined drugs from Schedule III to Schedule II and the change of tramadol from unscheduled to Schedule IV. 3 The CSA requires any pharmacy, hospital, physician, manufacturer, or distributor that works with any of the substances listed under the CSA to register with the DEA. The DEA has the authority to regulate transactions and monitor the movement of controlled substances from manufacturer and wholesale distributors to the retail level. The transaction data are available for use in investigations of illegal diversions from manufacturers and wholesalers to retail distributors, such as physicians and pharmacists, who receive unusual quantities of certain drugs. See United States General Accounting Office (2002) and Kraman (2004).



58

Prescribing and dispensing of controlled substances in each of these schedules is regulated under federal law.

For example, refills are prohibited for substances in Schedule II, whereas prescription drugs in Schedules III–

V can be refilled. In 2014, the DEA reclassified two opioids that are among the drugs most frequently

prescribed to injured workers, hydrocodone-acetaminophen and tramadol. Effective October 2014,

hydrocodone-combined products were rescheduled from Schedule III to the more restrictive Schedule II.

Tramadol, which was not previously scheduled, was reclassified to Schedule IV effective August 2014. Jones,

Lurie, and Throckmorton (2016) found that dispensing of prescriptions of hydrocodone-combination

products decreased by 22 percent in the 12 months after the rescheduling compared with 12 months prior to

rescheduling. Considering the implementation of these federal rule changes toward the end of the study

period, we may observe the full impact of the up-scheduling of these two drugs.

REMS programs are another key policy tool at the federal level aimed at mitigating the risk of misuse and

abuse of opioid medications while ensuring the accessibility of those drugs to patients who may need them.

REMS programs are developed by drug sponsors, and the Food and Drug Administration (FDA) reviews and

approves these programs. For example, in April 2011, the FDA announced REMS programs for all long-

acting opioids and oral fentanyl products. The REMS program for transmucosal immediate-release fentanyl

(TIRF), consisting of oral fentanyl products, was approved in December 2011. Under the TIRF REMS Access

program, prescribers are required to have a special certification to continue prescribing oral fentanyl

products, which is likely to result in a reduction of prescriptions for oral fentanyl products, such as Actiq®

and Fentora®.4 The REMS program for long-acting and extended-release opioids (ER/LA REMS) was

implemented in July 2012 and includes the following key components: (1) prescriber training, (2) medication

guide and new drug labeling, (3) patient education, and (4) evaluation of collected program data.5 While

prescriber training is on a voluntary basis and physicians can continue to prescribe long-acting opioids

without the additional training, the implementation of the program may help raise awareness of the potential

danger of long-acting opioid medications and, therefore, reduce the inappropriate use of such drugs as

OxyContin® and Kadian®, to some extent. Some recent studies provided evidence suggesting that some

REMS programs were effective in improving prescribers’ knowledge, attitudes, and self-reported clinical

practice in safe opioid prescribing (Alford et al., 2015; Donovan et al., 2016).

STATE-LEVEL POLICIES AND INITIATIVES AIMED AT REDUCING UNNECESSARY OPIOID

PRESCRIPTIONS

At the state level, the legal and regulatory environment aimed at preventing the abuse and diversion of

opioids includes, but is not limited to, statewide PDMPs, mandatory provider education and physician

licensing, opioid prescribing and pain policies, and state workers’ compensation laws and regulations for

pharmaceuticals.

PDMPs maintain statewide electronic databases of prescriptions dispensed for controlled substances.

Information collected by PDMPs may be used to support access to legitimate medical use of controlled

substances; identify or prevent drug abuse and diversion; facilitate identification of prescription drug-

addicted individuals and enable intervention and treatment; outline drug use and abuse trends to inform

4 Please see FDA (2014), available at http://www.fda.gov/downloads/Drugs/DrugSafety/ PostmarketDrugSafetyInformationforPatientsandProviders/UCM289730.pdf. 5 Please see FDA (2015), available at http://www.accessdata.fda.gov/drugsatfda_docs/rems/ERLA_opioids_2015-10-23_FDA_Blueprint.pdf.



59

https://www.fda.gov/downloads/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/UCM289730.pdf.

public health initiatives; or educate individuals about prescription drug use, abuse, and diversion.6 As of June

2017, all states but Missouri have enacted PDMP legislation.7

The state PDMPs vary widely with respect to what information is contained in the database, who should

report to the system in what time frame, who can and should access the database for what purposes, and

whether the information can be shared with other state PDMPs. Table TA.A1 provides a summary of some

key aspects of state PDMPs for the states included in this study.

Table TA.A1 State Prescription Drug Monitoring Programsa

State Year Current Program Operational

Year Initial Program Operational

Year Enacted

Frequency of Reporting

Prescribers Required to Check Prior to Prescribing in Certain Conditions

Schedule of Drugs Covered

Data Collected from Dispensing Practitioners

Arkansas 2013 2013 2011 Weekly Yes, for the treatment of chronic pain II, III, IV, V No

California 2005 (CURES) 1939 1939 Weekly Yes, effective September 2016 II, III, IV Yes

Connecticut 2008 2008 2006 Daily Yes, effective October 2015 II, III, IV, V Yes

Florida 2011 2011 2009 Weekly No II, III, IV Yes

Georgia 2013 2013 2011 Weekly Yes, for doctors in pain management clinics II, III, IV, V No

Illinois 2011 1968 1961 Daily No II, III, IV, V Yes

Indiana 2008 (INSPECT) 1998 1997 Daily

Yes, required for opioid treatment programs II, III, IV, V Yes

Iowa 2009 2009 2006 Weekly No II, III, IV No

Kansas 2011 2011 2008 Daily No II, III, IV Yes

Kentucky 2005 (KASPER) 1999 1998 Daily Yes, effective July 2012 II, III, IV, V Yes

Louisiana 2009 2008 2006 Daily Yes, effective August 2014 for Schedule II drugs II, III, IV, V Yes

Maryland 2013 2013 2011 Daily Yes, effective July 2018 II, III, IV, V Yes

Massachusetts 2010 1994 1992 Daily Yes, effective January 2016 II, III, IV, V No

Michigan 2003 1989 1988 Daily No II, III, IV, V Yes

Minnesota 2010 2010 2007 Daily

Yes, required for methadone outpatient clinics II, III, IV Yes

Missouri No prescription drug monitoring program

Nevada 1997 1996 1995 Daily Yes, effective November 2015 II, III, IV Yes

New Jersey 2011 2011 2008 Daily Yes, effective October 2015 II, III, IV, V No

New York 2013 1973 1972 Daily Yes, effective June 2013 II, III, IV, V Yes

North Carolina 2007 (CSRS) 2007 2005 3 business days

Yes, required for opioid treatment programs II, III, IV, V Yes

continued

6 See Finklea, Sacco, and Bagalman (2014). 7 Information is available at http://www.pdmpassist.org/pdf/PDMPProgramStatus2015_v5.pdf.



60

Table TA.A1 State Prescription Drug Monitoring Programsa (continued)

State Year Current Program Operational

Year Initial Program Operational

Year Enacted

Frequency of Reporting

Prescribers Required to Check Prior to Prescribing in Certain Conditions

Schedule of Drugs Covered

Data Collected from Dispensing Practitioners

Pennsylvania 2014 1973 1972 3 business days Yes, effective June 2015 II, III, IV, V Yes

South Carolina 2008 2008 2006 Daily Yes, for Medicaid II, III, IV Yes

Tennessee 2013 2006 2003 Daily Yes, effective April 2013 II, III, IV, V Yes

Texas 2008 1982 1981 Weekly No II, III, IV, V No

Virginia 2006 2003 2002 Weekly Yes, effective March 2015 II, III, IV Yes

Wisconsin 2013 2013 2010 Daily Yes, effective April 2017 II, III, IV, V Yes

a Information included is based on state profiles, which are available at http://www.pdmpassist.org/content/state-profiles.

Definition: Operational: program currently collecting prescription data and can respond to requests for reporting by those authorized to make these requests.

Key: CSRS: Controlled Substance Reporting System; CURES: Controlled Substance Utilization Review and Evaluation System; INSPECT: Indiana Scheduled Prescription Electronic Collecting & Tracking; KASPER: Kentucky All Schedule Prescription Electronic Reporting.

In recent years, an increasing number of states made legislative mandates requiring prescribers to register

and use the PDMP database. These state-level policy changes have been tracked by the PDMP Center of

Excellence at Brandeis University (PDMP COE).8 According to the PDMP COE 2016 briefing, 23 states had

mandatory registration with the state PDMP, as of December 2015.9 Prior to 2012, only Arizona and Utah

required prescribers to register with the PDMP, without mandating use of the PDMP database. In 2009,

Nevada’s legislation required use of the PDMP with a more subjective trigger of “reasonable belief that the

patient may be seeking the controlled substances.” Oklahoma in 2010 required a prescriber to check the

PDMP only when prescribing or dispensing methadone. Since 2012, mandates for the use of state PDMPs

have accelerated with more obligatory requirements for prescribers to check the prescription history in the

PDMP database at the initial and continued prescribing of opioids. Kentucky was the first state to adopt a

comprehensive mandate that requires all prescribers to check a patient’s prescription history. Nevada and

Oklahoma expanded their mandates of checking the state PDMP database in 2015. Michigan did not have a

prescriber PDMP use mandate over the study period, but it was the only state that allowed third-party payors

to access PDMP data. The reports noted that a rapid increase in mandatory use of PDMP databases coincided

with decreases in multiple provider episodes and opioid prescriptions.10 For example, Kentucky observed an

8.5 percent decline in opioid doses dispensed in the first year after requiring prescriber enrollment and use of

KASPER. After Tennessee’s PDMP went into effect in April 2013, opioid prescriptions in the state decreased

by over 7 percent between August 2012 and July 2013. New York also observed a 9.5 percent decrease in

8 The PDMP COE has published three editions of the briefing that describes the development of state PDMP programs and documents evidence on the effectiveness of PDMPs across states. The original edition of the briefing was published in November 2013, focusing on data and experience in Kentucky. The second edition of the briefing, Mandating PDMP Participation by Medical Providers: Current Status and Experience in Selected States, was published in 2014. In its third edition, the PDMP COE’s briefing describes the recent history and current status of prescriber mandates, as well as outcomes in selected states. It also discusses policy and implementation issues for states considering mandates. See the third edition, PDMP Prescriber Use Mandates: Characteristics, Current Status, and Outcomes in Selected States, published in May 2016, at http://www.pdmpassist.org/pdf/COE_documents/Add_to_TTAC/COE briefing on mandates 3rd revision.pdf. 9 The number increased to 30 states as of May 2016, as noted in the same report. 10 The PDMP COE’s 2016 report also provides information on the implementation and impact of the PDMP use mandates on health care practices, prescriptions of specific drugs, and outcomes.



61

http://www.pdmpassist.org/pdf/COE_documents/Add_to_TTAC/COE%20briefing%20on%20mandates%203rd%20revision.pdf

opioid prescriptions between the fourth quarters of 2012 and 2013 after implementing the I-STOP legislation

in July 2013. A recently published study by Dowell et al. (2016) also shows positive results of reforms

mandating prescribers to review the state PDMP. The authors observed substantial reductions in the amount

of opioids and opioid overdose deaths in states that simultaneously implemented PDMP prescriber mandates

and regulated pain clinics, compared with states without these reforms.11

In a report jointly prepared by PDMP COE and the Pew Charitable Trusts, the researchers describe

evidence-based practices that were shown to increase prescriber utilization of PDMPs.12 The report noted that

prescriber use mandates were one of the important features of PDMPs, associated with a rapid increase in

PDMP utilization. Other practices include allowing delegates to access the PDMP, unsolicited reporting of at-

risk patients, enhanced user interfaces, integration of PDMPs with health information exchanges, among

others—all features that make it easier for prescribers to utilize the information. The authors reported wide

variations in the status of the adoption of these evidence-based practices and noted wide variations in PDMP

utilization by prescribers of controlled substances across the 49 states with PDMPs.

In this study, we assessed how the PDMP enrollment and utilization in 2014 correlated with the changes

in the frequency of opioid use over the study period between 2010/2012 and 2013/2015. We found some

correlations overall based on the 26-state data (Table TA.A2). We found a strong positive correlation between

PDMP utilization and the decrease in frequency of opioid use. PDMP enrollment was also correlated with the

change in frequency of opioid use over the study period to a lesser extent. We also observed positive

correlations between PDMP enrollment and utilization and the decrease in the share of pain medication

prescriptions that were for Schedule II opioids. However, the results we saw only suggest that there might be

an association between the policy mandate and the decreased use of opioids in workers’ compensation

systems. More rigorous analysis is needed to evaluate the impact of PDMP policies on the use of opioids

received by injured workers.

11 Contrary to recent evidence of increases in heroin overdose deaths following reforms addressing opioid prescribing, the authors found reductions in heroin overdose deaths following the implementation of these reforms, although the findings were statistically insignificant. 12 The information on state regulations mandating the enrollment and use of PDMPs is from the 2016 report published by the Pew Charitable Trusts, Prescription Drug Monitoring Programs: Evidence-Based Practices to Optimize Prescriber Use. The report showed that, as of December 2014, there was substantial variation across states in prescribers’ enrollment and utilization of PDMPs, with several states experiencing increases in PDMP utilization in recent years. See Appendix A of that study.



62

AR CA CT FL GA IA IL IN KS KY LA MA MD MI MN MO NC NJ NV NY PA SC TN TX VA WI

Correlation with change in % of

claims with pain medications that

had opioids

Correlation with change in % of Rx for

pain medications that were for

Schedule II opioids

Change in % of claims with pain medications that had opioids, 2010/2012 to 2013/2015 -4% -2% -3% -3% -1% 1% -1% -4% -1% -16% -5% -1% -6% 0% 1% 2% -3% -5% -6% -9% -3% 1% -5% 0% -4% -1%

Change in % of Rx for pain medications that were for Schedule II opioids, 2010/2012 to 2013/2015 -6% -6% -5% -8% -3% 0% 1% -2% -5% -13% -2% -7% -8% -3% -3% 2% -3% -7% -5% -10% -5% 0% -2% -5% -5% 1%

% of DEA-registered prescribers enrolled in the PDMP as of

December 2014a 30 5 14 21 n/a 36 32 47 32 86 31 55 15 64 32 n/a 54 68 n/a 98 n/a 23 85 17 42 18 -0.59 -0.35

Queries per DEA-registered

prescriber in 2014a 45 10 9 28 n/a 12 10 62 15 224 54 10 9 72 16 n/a 59 31 n/a 152 n/a 18 104 9 28 8 -0.84 -0.52

Key: DEA: Drug Enforcement Administration; n/a: not available; PDMP: prescription drug monitoring program; Rx: prescriptions.

Note: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015. Similar notation is used for other years.

a Source: Appendix E of Prescription Drug Monitoring Programs: Evidence-Based Practices to Optimize Prescriber Use (The Pew Charitable Trusts, 2016).

Table TA.A2 Relationship between Statewide Prescription Drug Monitoring Programs Enrollment, Utilization, and the Overall Use of Opioids and Schedule II Opioids among the 26 Study States



63

Because of the dynamic nature of opioid policy changes, research that evaluates the impact of PDMPs has

been evolving. Some researchers documented earlier evidence of the positive impact of PDMPs on reducing

“doctor shopping” (see a review article by Worley, 2012) and slowing down the rapid growth of misuse and

abuse (as reported in Reifler et al., 2012, for example). Other researchers found limited effectiveness of

PDMPs partly due to the diversity of key program features of the state PDMPs as well as certain confounding

factors, including differences in prescribing practices and economic conditions across states (see a review

article by Finklea, Sacco, and Bagalman, 2014).13 However, studies using more recent data have been

reporting evidence of the positive impact of PDMPs in reducing opioid prescriptions as well as opioid

overdose deaths. For example, Bao et al. (2016) used National Ambulatory Medical Care survey data from

2001 to 2010 across 24 states that implemented a PDMP during the study period. They found an immediate

decrease in the rate of prescribing of opioids, especially Schedule II opioids, by 30 percent. Patrick et al.

(2016) found that PDMPs resulted in 1.12 fewer opioid overdose deaths per 100,000 people on average in the

year following the implementation of the PDMP, using data from 1999 to 2013 across 35 states that adopted a

PDMP during the study period. Greater reductions were seen in states monitoring more schedules of drugs

and updating the PDMP at least weekly, compared with states that did not have these characteristics.

Researchers are also investigating the potential unintended consequences of PDMPs, and some recent studies

found a noticeable increase in the use of non-prescription opioids (e.g., heroin).14

Mandatory provider education for safe opioid prescribing is among the policy initiatives aimed at

preventing misuse and overuse of opioids. As of July 2014, five states (Kentucky, Massachusetts, New Jersey,

Texas, and Utah) adopted legislation or regulations mandating continuing education about pain management

for licensees.15 More states required continuing medical education on opioid prescribing and chronic pain

management as part of the criteria for license renewal. According to the Federation of State Medical Boards’

board-to-board review, a number of states adopted CME requirements for opioid prescribing and/or chronic

opioid management, including California, Florida, Kentucky, Maryland, Massachusetts, New Hampshire,

New Mexico, Oklahoma, South Carolina, and Vermont. Several other states also encourage prescriber

education for pain management by providing online courses and resources for responsible opioid prescribing

and by recognizing the credits earned through training.

The authority for regulating medical practices also belongs to the states. State laws govern the prescribing

and dispensing of prescription drugs by licensed health care professionals through delegating the

responsibility of regulating physicians to state medical boards and delegating the responsibility of regulating

pharmacy practices to state boards of pharmacy. For example, Ohio passed legislation requiring its medical

and pharmacy licensing boards to adopt rules mandating the use of its PDMP (Clark et al., 2012).

State laws also address the prevention of prescription drug overdose, including setting prescription drug

limits,16 regulating pain clinics,17 limiting physician dispensing of controlled substances, and prohibiting

13 For example, a study by Brady et al. (2014) reported that the implementation of state PDMPs did not show a significant impact on annual morphine milligram equivalents dispensed per capita. However, the data used for that study only captured the state PDMPs up to 2008. Since 2008, several of the 31 states included in that analysis have made significant improvements to their PDMPs. Based on a review of 47 PDMP websites for overdose contents from December 2012 to October 2013, Green et al. (2015) found that most PDMPs did not provide relevant tools and materials to address overdose and related issues. Green et al. recommended a more comprehensive public health orientation for PDMPs that explicitly and publicly articulates their application and role in overdose prevention, which may increase PDMP effectiveness and use. 14 See Finklea, Sacco, and Bagalman (2014). 15 See the Pain & Policy Study Group (PPSG) report entitled Achieving Balance in State Pain Policy: A Progress Report Card (CY 2013). The report is available at http://www.acscan.org/content/wp-content/uploads/2014/07/PRC-2013.pdf. 16 Laws setting prescribing or dispensing limits for controlled substances including quantity and day supply limits. See the CDC guidelines.



64

doctor shopping.18 For example, Florida’s 2011 legislation bans physician from dispensing Schedule II and III

opioids. Kentucky’s 2013 rule change limits physician dispensing of Schedule II and III opioids to 48 hours of

supply. In workers’ compensation jurisdictions, issues with opioid prescribing and dispensing are typically

addressed in reimbursement rules, treatment guidelines, and drug formularies, which are described in the

following section.

There has been an on-going effort to evaluate state policies related to opioid prescribing and pain

management. The Trust for America’s Health published a report in 2013 that provided ratings on state opioid

policies.19 The ratings were based on the status of opioid polices in individual states capturing opioid policies

in 10 main areas, including existence of state PDMPs, mandatory use of PDMPs, doctor shopping laws,

policies expanding the coverage of substance abuse services (a.k.a. Medicaid expansion), prescriber education,

good Samaritan laws to provide immunity from criminal charges for individuals seeking help for themselves

or others who experience an overdose, support of rescue drug use, physician exam requirement, laws

requiring an ID prior to dispensing a controlled substance, and pharmacy lock-in programs. The National

Governors Association also made an effort to evaluate the effectiveness of different opioid policies and

provide recommendations to participating states to address opioid issues in a comprehensive and coordinated

way.20 Several state reports also showed positive results of other state opioid policies. For example, Johnson et

al. (2014) reported a 27 percent decrease in opioid overdose death rates in Florida between 2010 and 2012,

after a series of changes regulating the use of opioids in the state during 2010 and 2011. Lyapustina et al.

(2016) reported that Texas’ pill mill law was associated with declines in the average morphine equivalent daily

dose and monthly opioid prescriptions. The reductions were more pronounced among prescribers and

patients with higher rates of prescribing and utilization prior to the law.

TREATMENT GUIDELINES FOR PRESCRIBING OPIOIDS AND PAIN MANAGEMENT

A number of treatment guidelines for prescribing opioids and pain management have been developed at the

national and state levels. The widely-accepted national guidelines include the general treatment guidelines by

the American Pain Society and the American Academy of Pain Management, the occupational medical

treatment guidelines by the American College of Occupational and Environmental Medicine (ACOEM), and

the Official Disability Guidelines (ODG). The Federation of State Medical Boards also published the Model

Policy on the Use of Opioid Analgesics in the Treatment of Chronic Pain to provide a resource for use by

state medical boards in educating their licensees about appropriate opioid prescribing and avoiding over- and

under-treatment of patients with pain.21 More recently, the CDC published its guidelines for prescribing

17 Laws regarding pain clinic regulation include state oversight of pain management clinics and specific requirements for licensure or ownership of a pain management clinic. See the CDC guidelines. 18 All states have a “general” fraud statute that adopts verbatim or with slight alteration the provision in the Uniform Narcotic Drug Act of 1932 or the Uniform Controlled Substances Act of 1970. These statutes prohibit obtaining drugs, including through “doctor shopping,” by any or all of the following means: fraud, deceit, misrepresentation, subterfuge, or concealment of material fact. This resource distinguishes between general statutes and laws categorized as “specific” doctor shopping laws. See the CDC guidelines. 19 See Prescription Drug Abuse: Strategies to Stop the Epidemic (Trust for America’s Health, 2013). 20 See several reports published by NGA (2012, Six Strategies for Reducing Prescription Drug Abuse; 2014, Reducing Prescription Drug Abuse: Lessons Learned from an NGA Policy Academy; 2016, Finding Solutions to the Prescription Opioid and Heroin Crisis: A Road Map for States). 21 The Model Policy was published in 2013, and an updated and expanded edition was published in 2014. See Responsible Opioid Prescribing: A Clinician’s Guide (FSMB, 2014). The 2014 edition includes information from the 2013 model guidelines, which can be used as educational material for the category I CME required for license renewal. The 2013 Model Policy is available at http://www.fsmb.org/Media/Default/PDF/FSMB/Advocacy/pain_policy_july2013.pdf. Note that the model guidelines were being updated in 2016, entitled FSMB Guidelines for the Chronic Use of Opioid Analgesics.



65

opioids for chronic pain.

At the state level, many states, within and outside the workers’ compensation system, either developed

state proprietary guidelines or adopted existing guidelines or a modified version to address the state’s specific

needs. In a published report, IAIABC tracked the states that adopted treatment guidelines for pain

management and highlighted key issues that should be addressed in chronic opioid treatment guidelines.22 A

number of states we studied, including Connecticut, Louisiana, Massachusetts, Minnesota, and New York,

have adopted separate guidelines for chronic pain and opioid management during the study period. Many of

these states have separate chronic opioid guidelines while addressing opioid prescribing for acute pain in the

treatment guidelines for different parts of the body and/or conditions. Massachusetts had a major update to

the chronic pain treatment guidelines in 2012, and the guidelines were further updated in May 2016. In

Massachusetts, the regulation requires that the chronic pain treatment guidelines be used in the drug

utilization review. The New York State Workers’ Compensation Board adopted non-acute pain medical

treatment guidelines effective December 2014 in an effort to address issues related to long-term opioid use in

the state workers’ compensation system. Connecticut recently updated its medical treatment protocols to

include guidelines for prescribing opioids for acute and sub-acute pain. In 2016, California had a major

update to the chronic pain medical treatment guidelines first published in 2009 and adopted its first opioid

treatment guidelines to address opioid prescribing. Washington State and Colorado, not included in this

study, were the first states that had guidelines for chronic opioid management.23 Some of the recently updated

guidelines in California, Connecticut, and Washington include recommendations about prescribing opioids

for acute and sub-acute pain. In general, they recommend a limited use of opioids for severe acute pain or use

of opioids when non-opioid therapy is ineffective. For example, Washington guidelines recommend the

lowest necessary dose of immediate release opioids for the shortest duration (less than two weeks) for acute

pain.24 The 2016 CDC guidelines for prescribing opioids for chronic pain also noted that for acute pain, three

days or less is sufficient and more than seven days is rarely needed.

During and after the study period, several workers’ compensation jurisdictions adopted the ODG or

ACOEM guidelines. These guidelines generally discourage the use of opioids initially,25 except for traumatic

cases or those with severe pain, and if opioids are prescribed, prescriptions are usually for two weeks,

according to ACOEM. The 2014 update of the ACOEM guidelines also sets the maximum dose at 50

milligrams of morphine equivalents (MEQ) and recommends tapering if the dose exceeds 50 MEQ. The

ACOEM guidelines also recommend checking the PDMP database when prescribing opioids, and similar

recommendations can be found in several other guidelines.26 Outside workers’ compensation, opioid

guidelines have been adopted by the medical boards and the boards of other health care professionals in a

number of states. In Nevada, for example, the state medical board adopted opioid guidelines by referencing

the Model Policy for the Use of Controlled Substances for the Treatment of Pain, while the state workers’

22 According to IAIABC’s 2013 report, 20 states developed or adopted opioid treatment guidelines. Among those states, 7 states adopted either ACOEM or ODG and 13 states had state-specific guidelines. The report identified several key issues that should be addressed in state opioid guidelines, including comprehensive coverage of opioid management for acute and chronic pain, monitoring and evaluation of patients with chronic opioid therapy, documentation of improvement for continuing opioid treatment, and discontinuation of chronic opioid therapy in some circumstances. See IAIABC (2013). 23 The Washington State guidelines for prescribing opioids have also been used by the CDC to advise prescribing physicians on the use of opioids for treating pain. We have used these guidelines as a reference in our analysis. 24 See Part II. Prescribing Opioids in the Acute and Subacute Phase of the Washington Agency Medical Directors’ Group Interagency Guideline on Prescribing Opioids for Pain. 25 Opioids are recommended for post-operative pain and for fractures and other conditions likely to result in significant pain. 26 See Appendix B for more details.



66

compensation jurisdiction adopted ACOEM guidelines.

Long-term use of opioids has long been controversial. The Cochrane reviews27 pointed out that the

evidence in support of opioid use for chronic non-cancer pain is weak or questionable.28 Numerous

guidelines have been developed, within and outside workers’ compensation, to ensure appropriate use and

management of chronic opioid therapy. They all advise similar approaches. Patients should be carefully

screened for signs of aberrant drug behavior and other risk factors such as comorbid psychiatric conditions.

Chronic opioid management should only be offered after other therapies have failed and the patient has

moderately severe pain from a defined physical condition. Technical Appendix B provides a summary of

guideline recommendations for chronic opioid management.

The impact of medical treatment guidelines on opioid use has been examined with some mixed results

(see Haegerich et al., 2014, for a systematic review). Several studies examined the impact of Washington’s

opioid dosing guideline and found a decrease in opioid use after the implementation of the opioid dosing

guideline. For example, Garg et al. (2013) reported a 53 percent decrease in chronic opioid users between the

first quarter of 2004 and the fourth quarter of 2010, after the implementation of the state opioid dosing

guidelines on safe prescribing for chronic non-cancer pain. The authors also reported that these chronic

opioid users were 35 percent less likely to receive a dosage greater than a 120-milligram morphine equivalent

daily dose in the post-guideline period compared with the pre-guideline period. Similar findings were

reported in Franklin et al. (2012).

STATE WORKERS’ COMPENSATION DRUG FORMULARIES AND OTHER RELEVANT LAWS AND

REGULATIONS

Workers’ compensation laws and regulations that are likely to influence prescribing behavior and impact the

use of opioids include, but are not limited to, pharmacy fee schedules, physician dispensing restrictions,

pharmacy formularies, preauthorization requirements, and drug utilization review. State laws and regulations

regarding provider choice are also likely to affect the utilization of opioid medications as well as other medical

care delivered to injured workers. We provide a few examples to illustrate approaches that workers’

compensation communities use to deal with issues related to opioid utilization.

The Texas Division of Workers’ Compensation adopted a closed pharmacy formulary, which is based on

ODG. The formulary went into effect on September 1, 2011, for new claims with dates of injury on or after

that date, and became effective on September 1, 2013, for legacy claims with dates of injury before September

1, 2011. According to a study by TDI, fewer opioids and other not-recommended drugs are being prescribed

after the reform (TDI, Texas Workers’ Compensation Research and Evaluation Group, 2013). In the workers’

compensation system, two single-payor states, Washington and Ohio, also mandated drug formularies.29,30

The experiences from these states with earlier implementation of drug formularies suggest that evidence-

27 Cochrane Reviews provide systematic reviews of primary research in evidence-based health care and health policy. Published online in The Cochrane Library, Cochrane Reviews investigate the effect of interventions for prevention, treatment, and rehabilitation. They also assess the accuracy of a diagnostic test for a given condition in a specific patient group and setting. See more details at http://www.cochrane.org/evidence. 28 According to a Cochrane study, there is only weak evidence suggesting that patients on long-term opioid therapy experience clinically significant pain relief. However, multiple side effects are common, causing many patients to discontinue use. It is unclear whether this type of therapy functionally benefits most patients. See Noble et al. (2010). Most studies show that only around 50 percent of patients tolerate the side effects of opioids and related medications well and benefit from opioid therapy for pain relief. Depending on the diagnoses and other agents available for treatment, the incremental benefit of chronic opioid use can be small (Cepeda et al., 2007; Laudau et al., 2007; and Noble et al., 2010). 29 See the Washington drug formulary (http://www.lni.wa.gov/ClaimsIns/Files/Providers/DrugFormulary.pdf). 30 See the Ohio drug formulary (https://www.bwc.ohio.gov/downloads/blankpdf/OAC4123-6-21.3Appendix.pdf).



67

based drug formularies are expected to help address issues related to the use of opioids.

In January 2016, Tennessee adopted comprehensive workers’ compensation medical treatment

guidelines, and a drug formulary was adopted at the same time. In March 2017, the first draft of the medical

treatment utilization schedule (MTUS) drug formulary was released by the Division of Workers’

Compensation in California after the legislative mandate on implementing an evidence-based drug

formulary.31 In May 2017, there was some stakeholder input toward recommending a delay in the

implementation to address additional comments and provide education sessions to providers and injured

workers about the formulary. On July 19, 2017, the revised drug formulary was released for public comment

over a 15-day review period, and the revised drug formulary will be effective January 1, 2018.

Earlier in November 2012, Tennessee’s legislature passed Senate Bill 3315, which amended the definition

of utilization review to explicitly include Schedule II, III, and IV drugs being used for pain management. The

provision requires the parties involved to participate in utilization review if opioids are prescribed for pain

management to an injured or disabled employee for a period of time exceeding 90 days from the initial

prescription. Effective October 1, 2013, a new Tennessee legislation requires that a prescription for opioids or

benzodiazepines may not be dispensed in quantities greater than a 30-day supply. The new law also

encourages mandatory urine drug testing of patients on long-term drug therapy.

Effective December 26, 2014, the amended rules by Michigan’s Workers’ Compensation Agency require

that opioid treatment beyond 90 days for non-cancer related chronic pain should not be reimbursed unless

detailed physician reporting requirements and other processes are met. The new rules also provide incentive

for compliance with the requirement in ways that a provider may bill the additional services required for

reporting beyond 90 days and may bill for accessing the state PDMP (MAPS) or other PDMPs in the treating

jurisdiction.32

Several states, including California, Connecticut, Louisiana, Massachusetts, Minnesota, New York, and

Oklahoma, have adopted or updated treatment guidelines for chronic opioid pain management, specifically

requiring treating physicians to screen their patients prior to conducting chronic opioid therapy, conduct

random drug screening and testing, and refer the patient to specialists for pain management and

psychological/psychiatric consultation. A brief summary of individual state treatment guidelines is provided

in Technical Appendix B.

Several states changed reimbursement rules to limit physician dispensing of opioid prescriptions. In

Pennsylvania, for example, reimbursements for physician-dispensed Schedule II and III opioids are limited to

15 days of supply for nonsurgical cases and 30 days for surgical cases. In North Carolina, Schedule II and III

opioids are limited to a 5-day supply commencing with the initial treatment.

Across the states studied, we saw a spectrum of policies regarding provider choice, which is expected to

have an impact on overall medical care and opioid use.33 Many states allow injured workers to choose their

providers with or without limitation, which include, among the states studied, Illinois, Louisiana, and

31 Assembly Bill 1124 requires the California’s Division of Workers’ Compensation to implement a drug formulary no later than July 1, 2017, in light of findings from several studies. A carefully structured formulary can reinforce the MTUS guidelines to encourage prescribing of medically appropriate drugs and reduce the administrative burdens associated with the UR/IMR process (Wynn et al., 2016). A 2014 report by the California Workers’ Compensation Institute (CWCI) noted prescription drug costs as one of the fastest growing areas of workers’ compensation medical benefits in California, partly because of the wide array of drugs available and a dramatic increase in pain management therapies (Swedlow, Hayes, and David, 2014). The study looked at formularies used in both Texas and Washington and found that drug costs could be reduced between 12 and 42 percent, approximately $124 to $420 million in savings annually. 32 For more details, please refer to the amended rules, which are available at http://www.michigan.gov/documents/lara/2014-029_LR_Final_Health_Care_Services_476952_7.pdf. 33 See Tanabe (2015).



68

Maryland (with strictly employee choice) and California, Georgia, New York, and Pennsylvania (with limited

employee choice of providers). It is worth noting that two of the three states with strictly employee choice

were among the lowest on the frequency of claims receiving longer-term opioid prescriptions.

MEDICAL PRACTICES AND HEALTH CARE DELIVERY SYSTEM

Geographic differences in medical practices and health care delivery systems may also play an important role

in shaping interstate variations in opioid use. Multiple studies reported that higher concentrations of active

physicians and surgeons in a region are strongly correlated with the amount of opioids prescribed (Curtis et

al., 2006a; Han et al., 2012; McDonald et al., 2012). Some states or regions may also have a higher

concentration of pain clinics and doctors who specialize in pain treatment than others. In states where

patients have easier access to clinics specializing in the treatment of pain, the prescribing patterns may differ

from states where there are few pain clinics. For example, some occupational medicine clinics are affiliated

with academic medical centers that also have pain clinics. This arrangement facilitates referrals of patients to

providers who specialize in pain treatment. If pain specialists prescribed opioids more frequently on a longer-

term basis compared with non-specialists, this could increase the use of opioids in these states.

In states where more workers’ compensation medical care is provided by hospital-affiliated clinics, the

prescribing patterns may be influenced indirectly by certain requirements of the Joint Commission on

Accreditation of Healthcare Organizations,34 which regulates hospital accreditation. Doctors who practice in

hospital-based or hospital-affiliated programs may be more likely to be influenced by these requirements,

compared with doctors who are in private practice or those who work for commercial occupational medicine

networks. It is worth noting that a higher level of involvement with chiropractic care may also contribute to a

lower rate of opioid use in some states at the aggregate level because chiropractors cannot prescribe

medications.

Opioid prescribing patterns may also be influenced by programs implemented by Medicaid and group

health insurers, especially when the same doctors are treating patients covered under different insurance

policies. For example, Blue Cross Blue Shield, the largest group health insurer in Massachusetts, implemented

a program aimed at reducing the risk of opioid addiction in 2012.35 While this policy change influenced the

prescribing behaviors of the doctors who treated the patients covered by the group health policy, the same

doctors may also have changed their prescribing practices when treating workers’ compensation patients. In

recent years, state Medicaid programs also implemented various opioid management strategies, including

quantity limits, prior authorization requirements, etc. Twelve states required prescribers to check PDMPs

before prescribing opioids.36 These Medicaid policy changes may also have had a spill-over effect on the

patterns of opioid prescriptions in workers’ compensation health care. Since these state-level and local

34 The Joint Commission is an independent not-for-profit organization that accredits and certifies more than 20,000 health care organizations and programs in the United States. More information can be found at www.jointcommission.org. 35 Under the Blue Cross Blue Shield (BCBS) program, first prescriptions of short-acting opioids are limited to a 15-day supply and prescriptions written for longer than a 30-day supply must be accompanied by a medical authorization before coverage is approved. Preauthorization is also required before prescribing long-acting opioids for the first time. See http://www.bluecrossma.com/bluelinks-for-employers/whats-new/special-announcements/opioid-management.html. Since the inception of the program in 2012, Blue Cross Blue Shield found an 18 percent decline in opioid doses prescribed and a 50 percent reduction in the prescriptions for long-acting opioids such as OxyContin®. Note that similar programs were implemented by BCBS of Michigan and BCBS of Tennessee. 36 See Table 19 in Kaiser Family Foundation/National Association of Medicaid Directors report Implementing Coverage and Payment Initiatives Results from a 50-State Medicaid Budget Survey for State Fiscal Years 2016 and 2017 (Smith et al., 2016).



69

policies tend to be different in different states, the extent of the impact on opioid prescriptions may vary

across states.

Pharmacy benefit managers (PBMs) also offer a greater number of services to help payors manage the

utilization of pharmaceuticals, especially opioids, which may also impact the use of opioids. It is worth noting

that workers’ compensation fee schedules for other medical services may also have an indirect impact on the

use of opioids. For example, a low fee schedule rate for surgery might incentivize physicians to treat marginal

cases with opioid therapy instead of surgery.

PHARMACEUTICAL MARKET TRENDS

Several changes affecting the availability of opioids commonly used by workers’ compensation claimants have

occurred in recent years—such as the introduction of safer tamper-resistant formulations of opioids, generic

availability of commonly-used opioids, and product recalls.

During the last five years, several tamper-resistant formulations of opioids, like controlled-release

oxycodone (OxyContin®), immediate-release oxycodone hydrochloride (Oxecta®), and extended-release

oxymorphone (Opana® ER), were released. Physicians may feel more comfortable prescribing these safer

formulations, which may increase their use relative to opioids for which tamper-resistant formulations are

not currently available. Several other tamper-resistant formulations of commonly-used opioids are currently

in the pipeline. Propoxyphene, a commonly-used opioid, was withdrawn from the market in November 2010.



70

TECHNICAL APPENDIX B:

MEDICAL TREATMENT GUIDELINE PRINCIPLES FOR

CHRONIC OPIOID MANAGEMENT

A number of extensive medical treatment guidelines have been developed to ensure appropriate use of

chronic opioid therapy. Table TA.B1 provides a summary of the general guideline recommendations in

several selected key areas based on a set of medical treatment guidelines we identified and reviewed for long-

term opioid management of chronic non-cancer pain. These selected guidelines were developed by

governmental or other national entities representing medical practice standards, or they were adopted and

implemented by states within and outside workers’ compensation systems. Most of these guidelines were

developed based on available scientific evidence by multi-disciplinary groups of practitioners. The

congruence among the recommendations speaks to a general consensus regarding chronic opioid

management. In recent years, several states have developed medical treatment guidelines or protocols for

injured workers with chronic pain. Recommendations from these guidelines or protocols for the selected

areas are also summarized in Table TA.B1.

The medical treatment guidelines all advise a similar approach. Prior to receiving chronic opioid therapy,

patients should be carefully screened through a comprehensive baseline assessment, including physical

examination, pain and medical history, and psychosocial assessment. Many guidelines also recommend initial

assessment of past prescription drug use by checking the state PDMP database for signs of aberrant drug-

related behavior and other risk factors. Patients with a history of drug or alcohol abuse, or other psychiatric

conditions, are less likely to benefit from chronic opioid treatment and require close management by

professionals who have expertise in addiction and pain control. Chronic opioid therapy should only be

offered after other therapies have failed and the patient has moderately severe pain from a defined physical

condition.

After extensive counseling, a detailed consent form is required before initiating a therapeutic trial.

Treatment guidelines also provide clear descriptions of various side effects of chronic opioid therapy, which

the patient should be educated about, including constipation, nausea, hyperalgesia (an increased pain

response to low level painful stimuli), endocrine changes, sexual dysfunction, increases in sleep apnea if

present, and cognitive dysfunction, especially initially. Patients should also know the indications for tapering

off of opioids, which include aberrant drug behavior, lack of progress toward functional goals, inadequate

response to the opioid chosen, and suspected hyperalgesia. Patients should also be educated about the risk of

diversion and having their medication kept in a locked location, according to treatment guidelines.

A short-term therapeutic trial is recommended initially, and short-acting opioids are generally prescribed

for the trial. Although, several years ago, the trend was to shift to long-acting opioids if a trial was successful,

there appears to be no strong evidence that long-acting opioids are preferable. All patients should be regularly

monitored with an assessment of function as well as pain. A successful trial normally results in only a 2–3

point decrease on a 10 point pain scale.

In terms of dosage control, most medical treatment guidelines define high-dose use in morphine



71

equivalents and suggest caution before using higher doses. The 2014 ACOEM guidelines set the maximum

dose to be exceeded with caution at a morphine equivalent daily dose (MED) of 50 milligrams and

recommend tapering if above 50 milligrams MED. Note that the 50-milligram maximum MED was based on

two large population studies at Kaiser and the Veterans Administration, and higher limits recommended by

other guidelines may understate the potential harm of high-dose opioid therapy. The CDC guidelines for

prescribing opioids for chronic pain published in 2016 and the 2017 revision of the Canadian Guideline for

Opioid Therapy and Chronic Non-Cancer Pain also recommend a maximum of 50 milligrams MED at the

initial stage of chronic opioid therapy and a threshold of 90 milligrams MED for chronic opioid therapy.

These guidelines recommend tapering and referral to multidisciplinary pain management programs if the

patient exceeds the maximum daily dose of 90 milligrams of morphine equivalents.

Most guidelines address the use of drug testing and checking the state PDMP database as tools to manage

potential risk of misuse and abuse of prescription opioids. Almost all guidelines recommend drug screening

for all patients before initiating chronic opioid trials. During chronic opioid therapy, random urine drug

testing is recommended by all treatment guidelines. However, guidelines vary in terms of frequency and when

to have urine drug testing, depending on whether the patient is at high- or low-risk of potential misuse and

abuse of opioids.

Opioid rotation is sometimes tried if a patient is no longer benefiting from a specific opioid, but the

benefit of this rotation is unclear. Opioid therapy is tapered for aberrant drug behavior, inability to meet

therapeutic goals, or significant side effects. Patients with suspected opioid-induced hyperalgesia are also

likely to benefit from tapering.1

1 A number of studies have discussed opioid-induced hyperalgesia (Chu, Clark, and Angst, 2006; Hay et al., 2009; and Silverman, 2009). While findings of the clinical prevalence of opioid-induced hyperalgesia are not available, treatment of this condition involves reducing the opioid dosage, tapering, or switching to a new type of treatment for pain (Lee et al., 2011). An example of new drug treatment for controlling opioid-induced hyperalgesia was discussed by Kim (2013) at the American Academy of Pain Medicine (AAPM) 29th Annual Meeting in Ft. Lauderdale, FL.



72

American College of Occupational and Environmental Medicine (2014)

Substance abuse screening and psychiatric evaluation for most cases.

Begin with weaker acetaminophen-combination products. Only progress when necessary.

Recommended 50 MEQ Taper if above 50–100 MEQ.

For all patients, baseline and randomly.

Yes Recommend functional restoration and behavioral interventions under specific circumstances.

Canadian Guideline for Opioids for

Chronic Non-Cancer Pain (2017)aEmphasizes treating physician’s psychiatric and substance abuse evaluation; select patients with no current and past substance abuse and psychiatric disorders who have persistent pain after optimized nonopioid therapy.

Use of benzodiazepines discouraged. Step-wise progression starting with short acting and then moving to long acting, if desired (2010).

Review of prescription drug monitoring data is suggestedinitially and periodically.

50 MED for patients with chronic non-cancer pain who are beginning chronic opioid therapy;90 MED for chronic opioid therapy—consider tapering if exceeded.

Discusses pre-therapy and follow up drug screening. No specific numbers.

Yes Seek consultation for psychiatric patients. Refer to a multi-disciplinary pain program if the patient experiences serious changes during tapering.

Centers for Disease Control and Prevention (CDC) Guideline for Prescribing Opioids for Chronic Pain—United States, 2016

Clinicians should consider opioid therapy only ifexpected benefits for both pain and function are anticipated to outweigh risks to the patient.

When starting opioid therapy for chronic pain, cliniciansshould prescribe immediate-release opioids instead ofextended-release/long-acting (ER/LA) opioids.

Review PDMP data when starting opioid therapy for chronic pain and periodicallyduring opioid therapy, ranging from every prescription to every 3 months.

Lowest effective dosage when starting; Evaluate benefits and risks when increasing dosage to ≥ 50 MED; should avoidincreasing dosage to ≥ 90 MED.

Urine drug testing before starting opioidtherapy; consider urine drug testing at least annually to assess for prescribed medications as well as other controlled prescription drugs and illicit drugs.

Continue opioid therapy only if there is clinicallymeaningful improvement in pain and function thatoutweighs risks to patient safety.

Discuss with patients known risks andrealistic benefits of opioid therapy and patient andclinician responsibilities for managing therapy.

Colorado Treatment Guidelines (2010) and Fee Schedule Rule 18 (2011, effective 2012)

Psychological evaluation for all chronic pain patients. Screen for substance abuse.

After successful trial, one long acting and one short acting “rescue.” Not more than 2 opioids to be prescribed.

Yes 120 MEQ For all patients before beginning therapy and annually, randomly. More frequently per clinician discretion.

Yes Psychological treatment, active therapy, and interdisciplinary therapy.

Connecticut Medical Protocols for Nonacute Pain (2012)

The opioid management protocol was created 7/1/2012. The psychological pain assessment and treatment protocol was created 2/5/2016.

Screening for potential comorbidities such as opioid addiction, drug/alcohol problems, and depression.

Long acting not to be used for acute pain. Begin trials with short acting.

Physicians prescribing opioids required to register with the state PDMP (known as CPMRS). If beyond 4 weeks, checking is recommended.

90 MEQ For all patients prior to beginning therapy as a baseline; drug testing up to 2 times per year for stable low-risk patients and more frequently for high-risk patients.

Proper documents required for improvements in pain, functioning, and work capacity.

Evaluation/consultation and treatment by a physician with appropriate specialty training in pain management should be considered.

Louisiana Pain Medical Treatment Guidelines (2011)

Clinical evaluation for history of alcohol or substance abuse, signs of preinjury psychological dysfunction, pre-existing psychiatric conditions.

Long acting not recommended for acute pain.

n/a n/a Clinician may order random drug testing if deemed appropriate to monitor drug compliance.

All patients on chronic opioids should have an informed, witnessed consent, including withdrawal symptoms, requirement for drug testing, and necessity of tapering.

Active therapeutic exercise program to improve patient strength, endurance, and flexibility.

continued

Table TA.B1 Summary of Medical Treatment Guideline Recommendations for Chronic Opioid Management

GuidelineRecommended Screening

Initial Trial and Drug Choice (short acting versus long acting)

Use of PDMP Database

Maximum Dose to Be Exceeded with Caution

Urine Drug Screening/Testing

Requires Recording of Functional Status with Each Visit

Recommended Co-Therapies



73

Massachusetts Chronic Pain Treatment Guideline (2016)

Clinical evaluation for medical and psychosocial history, substance use/abuse, psychological evaluation, and other factors that may affect treatment outcome.

n/a Prescribing physicians need to register to utilize the state PDMP.

90 MED Baseline drug screen required, use of random drug screening 2–4 times per year.

All patients on long-term opioids must have a written, informed agreement including side effects, withdrawal symptoms, requirement for drug testing, necessity of tapering, and reasons for termination of prescriptions.

Requires strategic plan to mitigate the risk for patients with total opioid daily dose greater than 50 MED or history of overdose/substance disorder or concurrent use of benzodiazepines.All patients diagnosed with chronic pain should be referred for a psychosocial evaluation and concomitant interdisciplinary rehabilitation treatment.

New York Workers' Compensation Board Non-Acute Pain Medical Treatment Guidelines, First Edition (2014)

Requires trial and transitioning period before long-term opioid therapy.

Use limited to maximum of two opioids. A long-acting opioid for maintenance of painrelief and a short-acting opioid for limited rescueuse when pain exceeds the routine level.

Prescribing physicians must comply with I-STOP and other related statutes and regulations.

Lowest possible effective dose.

Random urine drug testing at least once a year, more frequent as deemed appropriate by the prescribing physician.

Update the patient informed consent and patient understanding for opioid treatment when opioid dose, type, or patient conditionchanges.

Continuing review of overall therapy plan, assessment of pain relief, functional status, and risk factors; all efforts to encourage active exercises.

Official Disability Guidelines (2011)

Screening for opioid risk. Psychosocial assessment.

Long acting recommended for continuous pain.

Recommended 100 MED Random drug screening, especially for high-risk patients.

Yes Multidisciplinary pain clinic.

Tennessee Department of Health Clinical Practice Guidelines for Outpatient Management of Chronic Non-Malignant Pain (2014)

Screening to assess the patient's risk for misuse, abuse, and addiction using a validated risk assessment tool.

n/a n/a Patient with > 120 MED should be referred to a pain specialist.

Recommended before initiating opioid therapy. Unannounced urine drug testing or comparable oral fluids test should be done twice a year at a minimum.

n/a On-going evaluation and documentation of risk factors.

Utah Clinical Guidelines on Prescribing Opioids (2009)

Screening for substance abuse and consultation if psychological issues.

Long acting not to be used for acute pain. Begin trials with short acting, then progress to long acting.

Check PDMP 120–200 MED For all patients prior to beginning therapy. Randomly with frequency based on clinical judgement.

Yes Previous active therapy, psychological therapy if diagnosis identified.

Washington State Interagency Guidelines on Prescribing Opioids

for Pain (2015)a

Added guidelines for acute and subacute pain; chronic part is similar to the 2010 version.

Screening for substance abuse and psychological conditions and referrals for treatment as needed.

Generally, do not combine with sedative-hypnotics. Appendix implies long acting with “rescue” short acting.

Check the state's PDMP to ensure controlled substance history is consistent with prescribing record, at the acute, subacute, and chronic phases.

120 MED All patients at baseline. Low risk once per year. Medium risk 2 times per year. High risk 3–4 times per year.

Yes Discusses referrals including psychological, as needed.

continued

Table TA.B1 Summary of Medical Treatment Guideline Recommendations for Chronic Opioid Management (continued)

GuidelineRecommended Screening

Initial Trial and Drug Choice (short acting versus long acting)

Use of PDMP Database

Maximum Dose to Be Exceeded with Caution

Urine Drug Screening/Testing

Requires Recording of Functional Status with Each Visit

Recommended Co-Therapies



74

a These guidelines, or updated versions of guidelines, were effective after the study period, which was up through March 2015.

Key: CPMRS: Connecticut Prescription Monitoring and Reporting System; I-STOP: Internet System for Over-Prescribing Act; MED: morphine equivalent daily dose; MEQ: morphine equianalgesic conversion; n/a: not available; PDMP: prescription drug monitoring program.

Sources: Hegmann et al., 2014; Busse, 2017; Centers for Disease Control and Prevention, 2016; Colorado Department of Labor and Employment, 2011; Work Loss Data Institute, 2011; Tennessee Department of Health, 2014; Utah Department of Health, 2009; and Washington State Agency Medical Directors Group, 2015. Also included are the workers' compensation guidelines for chronic pain and opioid management for Connecticut, Louisiana, and Massachusetts.

Table TA.B1 Summary of Medical Treatment Guideline Recommendations for Chronic Opioid Management (continued)



75

TECHNICAL APPENDIX C:

SENSITIVITY ANALYSIS OF CLAIM SELECTION AND CASE MIX

In this report, the analysis is based on nonsurgical claims with more than seven days of lost time that received

opioids. While the analysis based on this subset of claims is clinically more meaningful and policy relevant,

the data were selected based on three variables that might introduce a potential selection bias. Here, we

discuss potential issues related to such a selection and the results of our sensitivity analysis.

First, we chose to focus on claims with more than seven days of lost time for this study. Since the

percentage of claims that had more than seven days of lost time varied across the states, one may be

concerned that the injury severity and case mix could be very different across the states. Conceivably, a state

with a lower percentage of claims with more than seven days of lost time may have proportionally more

severe claims than another state where the percentage was higher.1 If this was the case, the selection would

make interstate comparisons less meaningful. However, previous WCRI studies concluded that differences

across states in injury severity and case mix among claims with more than seven days of lost time were not

large enough to materially affect the comparative results.2 We also looked at the dispensing of opioids to see

how it was correlated with the percentage of claims with more than seven days of lost time and did not find

evidence suggesting that it should be a concern (Figure TA.C1).

1 This is based on an assumption that the mix of cases and injury severity for all claims tend to be similar across states. 2 For example, a WCRI study, based on survey data of worker outcomes, reported that the injury severity for injured workers with more than seven days of lost time was similar among the 11 states surveyed (Belton and Liu, 2009). The WCRI CompScope™ multistate benchmarks adjusted for differences in the mix of cases and other factors across the states and assessed the impact of the case-mix adjustment (Yang et al., 2009). That study found that the difference in the mix of cases across states had only a small impact on the results—not large enough to change how the states were characterized as higher, medium, or lower. The impact was 1–2 percent for most states, with the exception of California and Texas at 3–4 percent.



76

AR

CA

CTFL

GA

IA

IL

INKS

KY

LA

MAMDMI

MN

MO

NCNJ

NVNY

PA

SC

TN

TX

VA

WI

0

2

4

6

8

10

12

14

16

0% 5% 10% 15% 20% 25% 30% 35% 40%

Ave

rag

e N

um

ber

of O

ffic

e V

isit

s p

er C

laim

wit

h

Mo

re T

han

7 D

ays

of

Lost

Tim

e

% of Claims with More Than 7 Days of Lost Time

AR

CA

CTFL

GAIA

IL

INKS

KY

LA

MA

MD

MI

MN

MO

NC

NJ

NVNY

PA

SC

TNTXVA

WI

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 5% 10% 15% 20% 25% 30% 35% 40%

% o

f C

laim

s w

ith

Mo

re T

han

7 D

ays

of L

ost

Tim

e Th

at R

ecei

ved

a P

ain

Med

icat

ion

Th

at H

ad O

pio

ids

% of Claims with More Than 7 Days of Lost Time

Figure TA.C1 Assessing Potential Bias of Selecting Claims with More Than 7 Days of Lost Time

Note: The underlying data include claims with more than seven days of lost time that had injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015.

Second, because post-operative opioid use is very different from the opioid use among nonsurgical cases,

we chose to use nonsurgical cases as the base to make the interstate comparisons more meaningful. However,

a potential concern may be that since surgery rates varied widely across the states (Yee, Pizer, and Fomenko,

2015), the nonsurgical criterion might result in a higher proportion of more severe cases for the states with

lower surgery rates, and vice versa. We looked at the percentage of cases that did not have surgery and how it

was correlated with the percentage of cases with pain medications that received opioids. Figure TA.C2 shows

a somewhat negative relationship between the percentage of nonsurgical claims with more than seven days of



77

AR

CA

CT

FLGAIA

IL

INKS

KY

LA

MA

MD

MI

MN

MONC

NJ

NV

NYPA

SC

TN

TX

VA

WI

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

% o

f No

nsu

rgic

al C

laim

s w

ith

Pai

n M

edic

atio

ns

That

Had

Op

ioid

s

% of Claims with More Than 7 Days of Lost Time That Did Not Have Surgery

lost time and the percentage of these claims with pain medications that had opioids. This negative

relationship, if existent, is likely to reinforce the results of the interstate comparisons. For example, the

surgery rates were much lower in California and higher in Wisconsin, which may suggest that the nonsurgical

claims in California included in the analysis were somewhat more severe compared with those in Wisconsin.

In the presence of a strong selection effect, the percentage of claims with opioids would be higher than that in

Wisconsin, which is the opposite of what we see. Had we included more comparable nonsurgical claims for

these two states, we would have seen fewer claims in California receiving opioids than we observed.

Figure TA.C2 Assessing Potential Bias of Selecting Nonsurgical Claims with More Than 7 Days of Lost Time

Note: The underlying data include nonsurgical claims with more than seven days of lost time that had injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015.

Third, we examined the interstate variations in the dispensing and longer-term dispensing of opioids

among nonsurgical cases that received opioids. Since injured workers with pain can be treated in various

ways, depending on the treating physician’s diagnosis and choice of treatment (including using prescription

non-opioid pain medications), a potential concern could be that different physicians may have different

thresholds for prescribing opioids. For example, for the same injured worker with a pain score of 7 on a scale

of 1 to 10, a physician in State A may not think that the pain would be severe enough to warrant an opioid

prescription, while a physician in State B may well prescribe opioids for pain relief. If this reflects the practice

norms in the two states, on average, cases with opioids in State A would be more severe than those in State B,

due to physician selection. Considerable variation across the states in the percentage of cases with pain

medications that received opioids may raise a concern about the comparability of the states’ results. However,

we did not see a strong correlation between the percentage of claims with opioids that had longer-term

dispensing of opioids and the percentage of nonsurgical cases with pain medications that received opioids

(Figure TA.C3).



78

AR

WI

CA

CT

GA

MO

IL

MI

LA

MA

NC

MDMN

NY

NJ

VA

PA

TN

KY

TX

KS

SC

IN

IAFL

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

% o

f No

nsu

rgic

al C

laim

s w

ith

Op

ioid

s Th

at W

ere

Iden

tifi

ed a

s H

avin

g L

on

ger

-Ter

m D

isp

ensi

ng

of

Op

ioid

s

% of Nonsurgical Claims with Pain Medications That Had Opioids

Based on the sensitivity analysis, we concluded that the selection of nonsurgical claims with more than

seven days of lost time with opioid prescriptions will not distort the results of interstate variations in the

prevalence of longer-term opioids. However, we caution the reader not to simply extrapolate the results based

on nonsurgical claims with more than seven days of lost time to all claims in a state. This is because excluding

surgical claims from the analysis is likely to undercount the incidence of longer-term opioid dispensing

overall, while the exclusion of claims with seven or fewer days of lost time may overestimate the prevalence of

longer-term opioids. The combined effect may be different depending on the state.

Figure TA.C3 Assessing Potential Selection Effect on Longer-Term Dispensing of Opioids

Note: The underlying data include nonsurgical claims with more than seven days of lost time that had injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015.

One other possible concern is the potential bias of an analysis based on a subset of claims with

prescriptions paid under workers’ compensation since large variation was observed across states in the

percentage of claims with prescriptions paid under workers’ compensation. Based on a sensitivity test we

conducted, we concluded that an analysis based on claims with prescriptions paid under workers’

compensation is unlikely to distort the results per claim with prescriptions.3

For nonsurgical claims with more than seven days of lost time that received opioids, we report the

prevalence or claim frequency of longer-term opioid use, which is an unadjusted measure, not adjusted for

case mix and severity. To ensure that the comparative results are credible and meaningful, we did a sensitivity

analysis based on a logistic regression on the frequency of claims receiving longer-term opioid prescriptions

across the states and over time within a state. For this analysis, we included nonsurgical claims with seven

3 See Wang and Liu (2011) for a detailed discussion of the issue, analysis, and conclusions from the analysis.



79

days of lost time that received opioids. The binary dependent variable is whether a claim is identified as

having longer-term dispensing of opioids. The independent variables used for case-mix adjustment include

the injured worker’s age, gender, and marital status; tenure with the preinjury employer; average weekly wage;

injury type; and industry group.4 A logistic regression was run for 2013/2015 claims with state dummies for

the fixed state effect and for trends; a similar regression was run separately for each state included in the

study. Table TA.C1 presents the regression results for the 2013/2015 claims.

Table TA.C1 Results of Logistic Regression of Longer-Term Opioid Dispensing on Case-Mix Variables, 2013/2015 Claims

DF Estimate

Standard Error Pr > ChiSq

Intercept 1 -5.878 ** 0.234 <.0001

State fixed effect (reference=Minnesota)

Arkansas 1 -0.266 0.258 0.304

California 1 0.481 ** 0.070 <.0001

Connecticut 1 0.059 0.150 0.692

Florida 1 -0.092 0.088 0.296

Georgia 1 0.444 ** 0.099 <.0001

Illinois 1 -0.003 0.106 0.979

Indiana 1 -0.384 ** 0.164 0.020

Iowa 1 -0.178 0.193 0.356

Kansas 1 -0.412 * 0.223 0.065

Kentucky 1 0.168 0.178 0.345

Louisiana 1 1.268 ** 0.108 <.0001

Maryland 1 -0.044 0.147 0.766

Massachusetts 1 0.078 0.126 0.536

Michigan 1 -0.091 0.154 0.554

Missouri 1 -0.326 * 0.173 0.059

Nevada 1 -0.226 0.206 0.273

New Jersey 1 -0.605 ** 0.153 <.0001

New York 1 0.054 0.100 0.589

North Carolina 1 0.542 ** 0.104 <.0001

Pennsylvania 1 0.648 ** 0.087 <.0001

South Carolina 1 0.655 ** 0.124 <.0001

Tennessee 1 -0.097 0.145 0.505

Texas 1 0.653 ** 0.073 <.0001

Virginia 1 0.289 ** 0.131 0.028

Wisconsin 1 -0.465 ** 0.188 0.013

continued

4 We did not attempt to adjust for severity using the administrative data because even if we group the International Classification of Diseases, 9th Revision (ICD-9) codes at a more detailed level, we may not be able to adequately characterize the medical severity due to coding and issues inherent in the ICD-9 codes.



80

Table TA.C1 Results of Logistic Regression of Longer-Term Opioid Dispensing on Case-Mix Variables, 2013/2015 Claims

DF Estimate

Standard

Error Pr > ChiSq

Injured worker's characteristics

Age 1 0.108 ** 0.010 <.0001

Age squared 1 -0.001 0.000 <.0001

Male 1 0.039 0.035 0.268

Single 1 0.074 ** 0.033 0.024

Tenure with preinjury employer 1 -0.019 ** 0.002 <.0001

Average weekly wage 1 0.000 0.000 0.880

Injury group (reference=fractures)

Inflammations, lacerations, and contusions 1 -0.148 * 0.079 0.059

Knee derangement 1 0.353 * 0.195 0.070

Neurological pain of the spine 1 1.651 ** 0.064 <.0001

Neurological pain of the upper extremity 1 0.226 0.232 0.331

Sprains, strains, and non-specific pain of the spine 1 0.919 ** 0.064 <.0001

Other sprains, strains, and non-specific pain 1 0.115 0.072 0.110

Skin 1 -0.863 ** 0.223 0.000

Other type of injuries 1 0.504 ** 0.077 <.0001

Industry group (reference=manufacturing)

Construction 1 0.266 ** 0.059 <.0001

Clerical and professional 1 -0.015 0.075 0.846

Trade 1 -0.024 0.056 0.666

High-risk services 1 -0.100 * 0.052 0.054

Low-risk services 1 -0.060 0.059 0.307

Other industries 1 -0.147 ** 0.070 0.036

Missing values 1 -0.336 ** 0.102 0.001

Key: * statistically significant at the 10 percent level; ** statistically significant at the 5 percent level; ChiSq: chi squared; DF: degree of freedom; Pr: probability.

Figure TA.C4 shows the percentage of claims receiving longer-term opioids, both unadjusted and

adjusted results, among the 2013/2015 nonsurgical claims with opioids. The regression-based results,

although they adjusted away some of the differences, support the comparative findings in how the study states

are characterized as high, medium, and low on the prevalence of longer-term opioid use. Case mix appears to

have had a relatively bigger effect for Louisiana, Massachusetts, and New York among the states with higher

prevalence of longer-term opioid use. Among these three states, Louisiana was still the highest on the adjusted

measure while Massachusetts and New York were fairly typical of the 26 states (Table TA.C1). Nevada was

among the lowest on the unadjusted numbers, while the adjusted number appeared to be higher but close to

the median of the 26 states.



81

Figure TA.C4 Percentage of Claims with Opioids That Were Identified as Receiving Longer-Term Opioids,a Unadjusted and Adjusted Results for 2013/2015

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions through March 31, 2015. a We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury (i.e., Type I claims). See Chapter 2 for more details.

The same model was also run for each of the study states to check if the trends in the prevalence reported

were sensitive to differences in the case mix between 2010/2012 and 2013/2015 claims. Table TA.C2 provides

the adjusted and unadjusted results on this measure for the 26 study states, using the 2010/2012 case mix as a

reference. In general, the case-mix adjustment tended to reduce the magnitude of change in the prevalence

rate, but the difference in the case mix was not large enough to explain all the changes. The unadjusted results

for the states with a noticeable decrease were largely consistent with the adjusted results.

0%

4%

8%

12%

16%

20%

0%

4%

8%

12%

16%

20%


Unadjusted Case-Mix Adjusted



82


2013/2015 447 16,285 1,355 6,581 2,679 941 3,202 1,505 831 692 1,043 1,620 1,481 1,578 1,566 1,458 2,273 2,214 797 3,034 3,785 1,203 1,593 10,174 1,489 1,363

Unadjusted -1.7 -1.5 -1.7 -1.6 -0.1 0.1 -0.5 0.1 -2.2 -4.7 1.4 -2.1 -1.3 -2.5 -2.0 -0.2 -1.3 -1.1 -1.5 -4.2 -1.0 0.6 -2.6 -1.6 0.2 0.0 -1.5

Adjusted -0.5 -1.3 -0.7 -0.8 -0.3 -0.1 -0.4 -0.2 -0.8 -3.5 -0.3 -1.4 -1.0 -1.2 -1.2 -0.2 -1.4 -0.6 -0.2 -3.2 -0.8 0.0 -0.9 -1.4 0.2 0.2 -0.9

Table TA.C2 Trends in the Prevalence of Longer-Term Dispensing of Opioids, Adjusted and Unadjusted, 2013/2015

Total number of nonsurgical cases that had opioids

Percentage point difference in the prevalence of longer-term opioid dispensing (Type I)a

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015. Similar notation is used for other years. For the analysis of longer-term use of opioids, we included a small number of claims with unusually high amounts of opioids. See Chapter 2 for more details.

a We identified claims receiving longer-term opioids as those that had opioids within the first three months after the injury and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the injury (i.e., Type I claims). See Chapter 2 for more details.



83

In the main report, we also noted that the definition of longer-term opioids is largely consistent with the

definition of chronic opioids presented in WCRI’s Interstate Variations in Use of Opioids, 4th Edition

(Thumula, Wang, and Liu, 2017). The difference between the two measures is that the identification of claims

with longer-term opioids was based on the number of visits during which opioid prescriptions were filled and

the identification of claims with chronic opioids was based on widely-accepted data elements of duration and

daily dose of opioids received by injured workers. Since the data elements were less complete for physician-

dispensed opioid prescriptions, the interstate variation report only presented the results on chronic opioids

for 17 of the 26 states. This report covers all 26 states. We should note that the comparative results are largely

consistent between the two definitions, as shown in Table TA.C3. Among the 17 states for which results on

chronic opioids were published, the correlation coefficients between the two measures were over 90 percent

(0.914 for the percentage of claims with opioid prescriptions that had more than 50 MED of opioids supply

for at least 60 days and 0.974 for the percentage of claims with opioid prescriptions that had at least 60 days of

opioids supply in any 90-day period). Michigan and Texas were exceptions. Texas was higher than the 17-

state median in the percentage of claims with longer-term opioids but fairly typical based on the definition of

chronic opioids. Michigan was slightly higher in terms of chronic opioid dispensing but was at the 17-state

median based on the definition of longer-term opioids. Note that in measuring the longer-term dispensing of

opioids, we included a small percentage of “extreme” claims that had unusually high numbers of opioid

prescriptions.5 These claims were not included in the data underlying the benchmark metrics used for the

interstate variations report, which might explain some of the differences for these two states.

Table TA.C3 Consistency in the Results between Longer-Term Opioids and Chronic Opioids, 2013/2015

Among claims with opioids that had days of supply populated for all opioid Rx

MO NJ KS MN IA WI AR NV VA TX NC MA KY SC MI NY LA 17-State Median

Correlation Coefficient

% claims with opioid Rx that had at least 60 days of opioids supply in any 90-day period Mean value 7% 7% 8% 9% 9% 9% 12% 12% 13% 15% 15% 15% 15% 18% 18% 19% 29% 13% 0.974

% point above/below median -6 -5 -4 -4 -4 -3 -1 -1 0 2 3 3 3 5 5 6 16

MO IA MN AR NJ KS TX WI KY MI NV VA NC MA SC NY LA 17-State Median

% claims with opioid Rx that had more than 50 MED of opioids supply for at least 60 days Mean value 1% 2% 2% 2% 2% 2% 3% 3% 3% 3% 4% 4% 4% 4% 4% 5% 6% 2.8% 0.914

% point above/below median -1.6 -1.3 -1.0 -0.7 -0.6 -0.4 -0.2 0.0 0.0 0.2 0.8 0.9 1.0 1.1 1.2 2.2 2.9

KS NJ MO WI NV IA AR MN MI KY VA MA NY NC TX SC LA 17-State Median

% claims with opioid Rx that were identified as receiving longer-term opioids

Mean value 3.4% 3.6% 3.7% 3.8% 3.9% 4.1% 4.4% 5.0% 5.0% 6.4% 6.5% 6.7% 7.1% 8.7% 9.1% 10.1% 17.9% 5.0%

% point above/below median -1.6 -1.5 -1.4 -1.2 -1.1 -0.9 -0.6 0.0 0.0 1.3 1.5 1.7 2.1 3.7 4.1 5.1 12.8

Notes: The underlying data include nonsurgical claims with more than seven days of lost time that had prescriptions filled by injured workers over the defined period and paid for by a workers' compensation payor. 2013/2015 refers to claims with injuries occurring from October 1, 2012, through September 30, 2013, and prescriptions filled through March 31, 2015. The 17 states included in this table are those where the days of supply is complete for all opioids prescriptions for a majority of claims in the state and claims with complete days of supply are substantially representative of all claims with opioids.

Key: MEA: morphine equivalent amount; MED: morphine equivalent daily dose in milligrams; Rx: prescriptions.

5 See the “Identifying Longer-Term Opioids” section of Chapter 2 for more discussion.



84

Wh

morph

consist

average

states,

because

more S

contrib

Figure

Notes: Tinjuredoccurri

b We idhad thrChapte

Key: ME

0%

4%

8%

12%

16%

20%

hen comparing

ine equivalent

ent between th

e MEA of opioi

which is diffe

e when injured

Schedule II opio

buting to the hi

e TA.C5 Percen 2013/2

The underlying da workers over theng from October

entified claims recree or more visits ter 2 for more detai

EA: morphine equ

KS NJ MO W

% of

g the results be

amount (MEA

he two measur

ids per claim in

rent from wh

d workers in th

oids (which ar

igher MEA per

ntage of Claims2015

ata include nonsue defined period a1, 2012, through S

ceiving longer-terto fill opioid prescils.

ivalent amount.

WI IN NV IA AR

f Claims with Long

etween the per

A) of opioids p

res, with a cor

n New York an

at the percent

hese two states

re more potent

r claim. Massac

s with Longer-T

rgical claims with nd paid for by a wSeptember 30, 20

rm opioids as thoscriptions between

R TN FL MN M

ger-Term Opioids

rcentage of clai

er claim with o

relation coeffic

nd Pennsylvani

tage of claims

received opioi

and have a hig

chusetts had a s

Term Opioids a

more than sevenworkers' compens

13, and prescripti

se that had opioidn the 7th and 12th

MI IL MD CT KY

ims with longe

opioids, we fou

cient of 0.893

ia was consider

with longer-t

id prescription

gher morphine

similar pattern

and Average M

days of lost time ation payor. 2013ions through Marc

ds within the first h months after the

KY VA MA NY G

Average MEA pe

er-term opioids

und that the re

(Figure TA.C5

rably higher th

term opioids i

ns, they receive

e equivalent co

n but to a lesser

MEA per Claim w

that had prescrip3/2015 refers to clach 31, 2015.

three months aftee injury (i.e., Type

GA CA NC TX P

r Claim

s and the avera

esults were larg

5). Note that t

han in most stu

indicates. This

ed proportiona

onversion facto

r extent.

with Opioids,

ptions filled by aims with injuries

er the injury and I claims). See

0

1,00

2,00

3,00

4,00

PA SC LA

age

gely

the

udy

s is

ally

or),

s

00

00

00

00



85

GLOSSARY

Chronic opioid management refers to the clinical management of chronic opioid therapy, the prolonged

opioid treatment used for patients with chronic pain after the sub-acute stage of injury.

Controlled substances are prescription drugs and illegal drugs that have a potential for producing

psychological or physical dependence. They are classified into five schedules by the federal

government.

Claims with longer-term opioids refers to claims that had opioids within the first three months after the injury

and had three or more visits to fill opioid prescriptions between the 7th and 12th months after the

injury. Because our definition is based on the number of fill dates of opioid prescriptions rather than

days of supply, which indicates duration of opioid consumption, we labeled the category we

identified as longer-term rather than long-term dispensing of opioids. We also refer to them as Type I

claims. We refer to claims that did not receive opioids within the first quarter postinjury but

exhibited the same pattern of longer-term dispensing of opioids as those identified as Type I claims

as Type II claims.

Narcotic is a legal term that was used to classify substances such as opioids, under the Single Convention on

Narcotic Drugs, 1961, and the U.S. Controlled Substance Act, according to the Pain & Policy Studies

Group Resource Guide (PPSG, 1998). Since marijuana and cocaine are also legally classified as

narcotics, the term opioids has been increasingly used in health care policy discussions when referring

to narcotics prescribed for pain relief. In this report, we use the two terms interchangeably.

Opioid denotes both natural (codeine, morphine [Avinza®]) and synthetic (methadone [Methadose®],

fentanyl [Duragesic®]) drugs whose pharmacological effects are mediated by specific receptors in the

nervous system.

Prescription drug diversion is the resale or redistribution of prescription drugs obtained from medical sources

in the illegal market.

Schedule II opioids are opioids that are classified as Schedule II controlled substances, which are of the highest

abuse potential among the controlled substances for medical use. There are five schedules of

controlled substances classified by the Drug Enforcement Administration under federal law. The

classifications are based on a drug’s medical usefulness and abuse potential.



86

REFERENCES

Alford, D., L. Zisblatt, P. Ng, S. Hayes, S. Peloquin, I. Hardesty, and J. White. 2015. SCOPE of pain: An

evaluation of an opioid risk evaluation and mitigation strategy continuing education program. Pain

Medicine. Doi:10.1111/pme.12878.

American Medical Association. 2006. Current procedural terminology CPT 2006. Chicago, IL: American

Medical Association.

Arkansas Center for Health Improvement. 2015. Prevention of opioid pain reliever misuse in Arkansas. Little

Rock, AR.

Bao, Y., Y. Pan, A. Taylor, S. Radakrishnan, F. Luo, H. Pincus, and B. Schackman. 2016. Prescription drug

monitoring programs are associated with sustained reductions in opioid prescribing by physicians.

Health Affairs 35 (6): 1,045–1,051.

Belton, S., and T. Liu. 2009. Comparing outcomes for injured workers in Michigan. Cambridge, MA: Workers

Compensation Research Institute.

Braden, J., J. Russo, M. Fan, M. Edlund, B. Martin, A. DeVries, and M. Sullivan. 2010. Emergency department

visits among recipients of chronic opioid therapy. Archives of Internal Medicine 170 (16): 1,425–

1,432.

Brady, J., H. Wunsch, C. DiMaggio, B. Lang, J. Giglio, and G. Li. 2014. Prescription drug monitoring and

dispensing of prescription opioids. Public Health Reports 129 (2): 139–147.

Busse, J. (ed.). 2017. The 2017 Canadian guideline for opioids for chronic non-cancer pain. Hamilton, Ontario:

Michael G. DeGroote National Pain Centre at McMaster University and funded by Health Canada

and the Canadian Institutes of Health Research.

Centers for Disease Control and Prevention (CDC), United States Department of Health and Human

Services. 2010. Unintentional drug poisoning in the United States.

———. 2011. Drug poisoning deaths in the United States, 1980–2008. NCHS Data Brief. No. 81. Retrieved

from http://www.cdc.gov/nchs/data/databriefs/db81.pdf (accessed February 23, 2016).

———. 2012. CDC grand rounds: Prescription drug overdoses — a U.S. epidemic. Morbidity and Mortality

Weekly Report 61 (01): 10–13.

———. 2016. CDC guideline for prescribing opioids for chronic pain. Retrieved from

https://www.cdc.gov/drugoverdose/prescribing/guideline.html (accessed January 11, 2017).

Centers for Medicare & Medicaid Services (CMS), United States Department of Health and Human Services.

2012. MLN matters®: Information for Medicare fee-for-service health care professionals. Number:

SE1105 Revised. Retrieved from http://www.cms.gov/Outreach-and-Education/Medicare-Learning-

Network-MLN/MLNMattersArticles/downloads/SE1105.pdf (accessed February 23, 2016).

Cepeda, M., F. Camargo, C. Zea, and L. Valencia. 2007. Tramadol for osteoarthritis: A systematic review and

metaanalysis. The Journal of Rheumatology 34: 543–555.

Chang, H., T. Lyapustina, L. Rutkow, G. Alexander. 2016. Impact of prescription drug monitoring programs

and pill mill laws on high-risk opioid prescribers: A comparative interrupted time series analysis.

Drug and Alcohol Dependence 165 (June 2016). DOI: 10.1016/j.drugalcdep.2016.04.033.

Chou, R., G. Fanciullo, P. Fine, J. Adler, J. Ballantyne, P. Davies, M. Donovan, D. Fishbain, K. Foley, J. Fudin,

A. Gilson, A. Kelter, A. Mauskop, P. O’Connor, S. Passik, G. Pasternak, R. Portenoy, B. Rich, R.

Roberts, K. Todd, and C. Miaskowski (for the American Pain Society–American Academy of Pain



87

Medicine Opioids Guidelines Panel). 2009. Opioid treatment guidelines: Clinical guidelines for the

use of chronic opioid therapy in chronic noncancer pain. The Journal of Pain 10 (2): 113–130.

Chu, L., D. Clark, and M. Angst. 2006. Opioid tolerance and hyperalgesia in chronic pain patients after one

month of oral morphine therapy: A preliminary prospective study. The Journal of Pain 7 (1): 43–48.

Clark, T., J. Eadie, P. Kreiner, and G. Strickler. 2012. Prescription drug monitoring programs: An assessment of

the evidence for best practices. The Prescription Drug Monitoring Program Center for Excellence.

Heller School for Social Policy and Management, Brandeis University. Prepared for The Pew

Charitable Trusts.

Colorado Department of Labor and Employment, Division of Workers’ Compensation. 2007a. Rule 17,

exhibit 1: Low back pain medical treatment guidelines.

———. 2007b. Rule 17, exhibit 8: Cervical spine injury medical treatment guidelines. Retrieved from

https://www.colorado.gov/pacific/sites/default/files/MTG_Ex8_CSI.pdf (accessed April 20, 2017).

———. 2011. Rule 17, exhibit 9: Chronic pain disorder medical treatment guidelines.

Coomer, N., R. Yang, S. Landes, E. Radeva, C. Telles, and R. Tanabe. 2010. CompScope™ medical benchmarks:

Technical appendix, 10th edition. Cambridge, MA: Workers Compensation Research Institute.

Curtis, L., J. Stoddard, J. Radeva, S. Hutchison, P. Dans, A. Wright, R. Woosley, and K. Schulman. 2006a.

Geographic variations in the prescription of Schedule II opioid analgesics among outpatients in the

United States. Health Services Research 41: 837–855.

———. 2006b. Correction to “Geographic variations in the prescription of Schedule II opioid analgesics

among outpatients in the United States.” Health Services Research 41: 856–858.

Donovan, A., G. Wood, D. Rubio, H. Day, and C. Spagnoletti. 2016. Faculty communication knowledge,

attitudes, and skills around chronic non-malignant pain improve with online training. Pain

Medicine. Advance online publication. Doi: 10.1093/pm/pnw029.

Dowell, D., K. Zhang, R. Noonan, and J. Hockenberry. 2016. Mandatory provider review and pain clinic laws

reduce the amounts of opioids prescribed and overdose death rates. Health Affairs 35 (10): 1,876–

1,883.

Drug Enforcement Administration (DEA), United States Department of Justice. 2013. Schedules of controlled

substances: Placement of tramadol into Schedule IV. Retrieved from

http://www.deadiversion.usdoj.gov/fed_regs/rules/2013/fr1104.htm (accessed February 23, 2016).

———. 2015. Drugs of abuse. Retrieved from http://www.justice.gov/dea/pr/multimedia-

library/publications/drug_of_abuse.pdf (accessed February 23, 2016).

Dunn, K., K. Saunders, C. Rutter, C. Banta-Green, J. Merill, M. Sullivan, C. Weisner, M. Silverberg, C.

Campbell, B. Psaty, and M. Von Korff. 2010. Opioid prescriptions for chronic pain and overdose: A

cohort study. Annals of Internal Medicine 152: 85–92.

Federation of State Medical Boards (FSMB). 2004. Model policy for the use of controlled substances for the

treatment of pain. Retrieved from

http://www.painpolicy.wisc.edu/sites/www.painpolicy.wisc.edu/files/model04.pdf (accessed

February 23, 2016).

———. 2013. Model policy on the use of opioid analgesics in the treatment of chronic pain. Retrieved from

http://www.fsmb.org/Media/Default/PDF/FSMB/Advocacy/pain_policy_july2013.pdf (accessed

March 27, 2017).

Finklea, K., L. Sacco, and E. Bagalman. 2014. Prescription drug monitoring programs. Congressional Research

Service. Retrieved from https://www.fas.org/sgp/crs/misc/R42593.pdf (accessed February 23, 2016).

Food and Drug Administration (FDA). U.S. Department of Health and Human Services. 2014. Transmucosal



88

immediate release fentanyl (TIRF), risk evaluation and mitigation strategy (REMS). Retrieved from

http://www.fda.gov/downloads/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsand

Providers/UCM289730.pdf (accessed February 23, 2016).

———. 2015. Introduction for the FDA blueprint for prescriber education for extended-release and long-acting

opioid analgesics. Retrieved from

http://www.accessdata.fda.gov/drugsatfda_docs/rems/ERLA_opioids_2015-10-

23_FDA_Blueprint.pdf (accessed February 23, 2016).

Franklin, G. 2014. Opioids for chronic noncancer pain: A position paper of the American Academy of

Neurology. Neurology 83: 1,277–1,284.

Franklin, G., J. Mai, J. Turner, M. Sullivan, T. Wickizer, and D. Fulton-Kehoe. 2012. Bending the prescription

opioid dosing and mortality curves: Impact of the Washington State opioid dosing guideline.

American Journal of Industrial Medicine 55 (4): 325–331.

Franklin, G., J. Mai, T. Wickizer, J. Turner, D. Fulton-Kehoe, and L. Grant. 2005. Opioid dosing trends and

mortality in Washington State workers’ compensation, 1996–2002. American Journal of Industrial

Medicine 48: 91–99.

Garg, R., D. Fulton-Kehoe, J. Turner, A. Bauer, T. Wickizer, M. Sullivan, and G. Franklin. 2013. Changes in

opioid prescribing for Washington workers’ compensation claimants after implementation of an

opioid dosing guideline for chronic noncancer pain: 2004 to 2010. Journal of Pain 14 (12): 1,620–

1,628.

Garrettson, M, and C. Ringwalt. 2013a. An evaluation of the North Carolina Controlled Substances Reporting

System: Part II impact evaluation. Injury Prevention Research Center, University of North Carolina

at Chapel Hill.

———. 2013b. An evaluation of the North Carolina Controlled Substances Reporting System: Part I user surveys.

Injury Prevention Research Center, University of North Carolina at Chapel Hill.

Green, T., S. Bowman, C. Davis, C. Los, K. McHugh, and P. Friedman. 2015. Discrepancies in addressing

overdose prevention through prescription monitoring programs. Drug and Alcohol Dependence 153:

355–358.

Haegerich T., L. Paulozzi, B. Manns, and C. Jones. 2014. What we know, and don’t know, about the impact

of state policy and systems-level interventions on prescription drug overdose. Drug and Alcohol

Dependence 145: 34–47.

Haffajee, R., A. Jena, S. Weiner. 2015. Mandatory use of prescription drug monitoring programs. Journal of

the American Medical Association 313 (9): 891–892.

Han, H., P. Kass, B. Wilsey, and C. Li. 2012. Individual and county level factors associated with use of

multiple prescribers and multiple pharmacies to obtain prescriptions in California. PLoS ONE 7(9):

e46246.

Hay, J., J. White, F. Bochner, A. Somogyi, T. Semple, and B. Rounsefell. 2009. Hyperalgesia in opioid-

managed chronic pain and opioid-dependent patients. The Journal of Pain 10 (3): 316–322.

Hegmann, K., M. Weiss, K. Bowden, F. Branco, K. DuBrueler, C. Els, S. Mandel, D. McKinney, R. Miguel, K.

Mueller, R. Nadig, M. Schaffer, L. Studt, J. Talmage, R. Travis, T. Winters, M. Thiese, and J. Harris.

2014. ACOEM practice guidelines: Opioids for treatment of acute, subacute, chronic, and

postoperative pain. Journal of Occupational and Environmental Medicine 56 (12): e143–59.

International Association of Industrial Accident Boards and Commissions (IAIABC). 2013. Reducing

inappropriate opioid use in treatment of injured workers: A policy guide. Madison, WI: IAIABC.

———. 2016. A discussion on the use of a formulary in workers’ compensation. Madison, WI: IAIABC.



89

https://www.fda.gov/downloads/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/UCM289730.pdf

Joint Commission on Accreditation of Healthcare Organizations. 2002. Examples of compliance: Pain

assessment and management. Oakbrook Terrace, IL: Joint Commission Resources.

Johnson, H., L. Paulozzi, C. Porucznik, K. Mack, and B. Herter. 2014. Decline in drug overdose deaths after

state policy changes—Florida, 2010–2012. Morbidity and Mortality Weekly Report 63 (26): 569–574.

Jones, J., P. Lurie, and D. Throckmorton. 2016. Effect of US Drug Enforcement Administration’s

rescheduling of hydrocodone combination analgesic products on opioid analgesic prescribing.

JAMA Internal Medicine 176 (3): 399–402.

Kidner, C., T. Mayer, and R. Gatchel. 2009. Higher opioid doses predict poorer functional outcome in

patients with chronic disabling occupational musculoskeletal disorders. The Journal of Bone and Joint

Surgery 91: 919–927.

Kim, Y. 2013. Possible new treatment for opioid-induced hyperalgesia. American Academy of Pain Medicine

(AAPM) 29th Annual Meeting in Ft. Lauderdale, FL.

Kraman, P. 2004. Drug abuse in America – Prescription drug diversion. The Council of State Governments:

Trends Alert. Retrieved from http://www.csg.org/knowledgecenter/docs/TA0404DrugDiversion.pdf

(accessed February 23, 2016).

Laudau C., W. Carr, A. Razzetti, N. Sessier, C. Munera, and S. Ripa. 2007. Buprenorphine transdermal

delivery system in adults with persistent noncancer-related pain syndromes who require opioid

therapy: A multicenter, 5-week run-in and randomized, double-blind maintenance-of-analgesia

study. Clinical Therapeurics 29 (10): 2,179–2,193.

Laws, C. 2012. Narcotics in workers compensation. NCCI Research Brief. Boca Raton, FL: National Council on

Compensation Insurance, Inc.

Lee, M., S. Silverman, H. Hansen, V. Patel, and L. Manchikanti. 2011. A comprehensive review of opioid-

induced hyperalgesia. Pain Physicians 14 (2): 145–161.

Lipton, B., C. Laws, and L. Li. 2009. Narcotics in workers compensation. NCCI Research Brief. Boca Raton, FL:

National Council on Compensation Insurance, Inc.

Luo, X., R. Pietrobon, and L. Hey. 2004. Patterns and trends in opioid use among individuals with back pain

in the United States. Spine 29: 884–890. Discussion 891.

Lyapustina, T., L. Rutkow, H. Chang, M. Daubresse, A. Ramji, M. Faul, E. Stuart, and G. Alexander. 2016.

Effect of a “pill mill” law on opioid prescribing and utilization: The case of Texas. Drug and Alcohol

Dependence 159: 190–197.

Mahmud, M., B. Webster, T. Courtney, S. Matz, J. Tacci, and D. Christiani. 2000. Clinical management and

the duration of disability for work-related low back pain. Journal of Occupational and Environmental

Medicine 42: 1,178–1,187.

Massachusetts Department of Public Health, Bureau of Health Care Safety and Quality, Drug Control

Program. 2010. Prescription monitoring program. Slides presented at the Public Health Council

meeting on August 11, 2010.

Massachusetts Department of Public Health, Bureau of Substance Abuse Services. 2010. Opioid overdose

prevention strategies in Massachusetts. Retrieved from

http://media.masslive.com/breakingnews/other/opioid-strategies-feb2010.pdf (accessed February 23,

2016).

Massachusetts Executive Office of Labor and Workforce Development. 2012. Treatment guidelines. Retrieved

from http://www.mass.gov/lwd/workers-compensation/hcsb/tg/ (accessed February 23, 2016).

McDonald, D., K. Carlson, and D. Izrael. 2012. Geographic variation in opioid prescribing in the U.S. The

Journal of Pain 13 (10): 988–996.



90

Medi-Span®. 2005. Master drug data base (MDDB®) v2.5: Documentation manual. Indianapolis, IN: Wolters

Kluwer Health, Inc.

Morasco, B., J. Duckart, T. Carr, R. Deyo, and S. Dobscha. 2010. Clinical characteristics of veterans prescribed

high doses of opioid medications for chronic non-cancer pain. Pain 151 (3): 625–632.

National Governors Association (NGA). 2012. Six strategies for reducing prescription drug abuse. Washington,

DC.

———. 2014. Reducing prescription drug abuse: Lessons learned from an NGA policy academy. Washington,

DC.

———. 2015. National Governors Association Policy Academy on Prescription Drug Abuse Prevention: State of

Nevada draft plan recommendation. Washington, DC.

———. 2016. Finding solutions to the prescription opioid and heroin crisis: A road map for states. Washington,

DC.

National Opioid Use Guideline Group. 2010. Canadian guideline for safe and effective use of opioids for chronic

non-cancer pain. Retrieved from http://nationalpaincentre.mcmaster.ca/opioid/documents.html


New York State Office of the Attorney General. 2012. Internet system for tracking over-prescribing (I-STOP), a

proposal addressing New York’s prescription drug abuse and drug diversion epidemic. Retrieved from

http://www.ag.ny.gov/sites/default/files/press-

releases/2012/ISTOP%20REPORT%20FINAL%201.10.12.pdf (accessed February 23, 2016).

Noble, M., J. Treadwell, S. Tregear, V. Coates, P. Wiffen, C. Akafomo, and K. Schoelles. 2010. Long-term

opioid management for chronic noncancer pain. Cochrane Database of Systematic Reviews. Issue 1.

Art. No.: CD006605. DOI: 10.1002/14651858.CD006605.pub2.

North Carolina Program Evaluation Division. 2014. North Carolina needs to strengthen its system for

monitoring and preventing the abuse of prescribed controlled substances: Final report to the Joint

Legislative Program Evaluation Oversight Committee. Retrieved from

http://www.ncleg.net/PED/Reports/documents/Rx/Rx_Report.pdf (accessed February 23, 2016).

Oklahoma Department of Health. 2014. Oklahoma opioid prescribing guidelines for health care providers in the

office-based setting. Oklahoma City, OK.

Oregon Health and Science University. 2006. Opioids and chronic non-malignant pain: A clinicians’ handbook.

Part II, Section 3: Converting opioids after a successful trial: Equianalgesic dosing. Portland, OR:

Oregon Health and Science University.

Pain & Policy Studies Group (PPSG). 1998. Resource guide: Information about regulatory issues in pain

management.

———. 2014. Achieving balance in state pain policy: A progress report card (CY 2013). University of Wisconsin.

Retrieved from http://www.acscan.org/content/wp-content/uploads/2014/07/PRC-2013.pdf


Passik, S. 2009. Issues in long-term opioid therapy: Unmet needs, risks, and solutions. Mayo Clinic

Proceedings 84 (7): 593–601.

Patrick, S., C. Fry, T. Jones, and M. Buntin. 2016. Implementation of prescription drug monitoring programs

associated with reductions in opioid-related death rates. Health Affairs. Published online before print

June 2016, DOI:10.1377/hlthaff.2015.1496.

The Pew Charitable Trusts. 2016. Prescription drug monitoring programs: Evidence-based practices to optimize

prescriber use. Philadelphia, PA.

Pomerleau, A., L. Nelson, J. Hoppe, M. Salsman, P. Weiss, and J. Perrone. 2016. The impact of prescription



91

drug monitoring programs and prescribing guidelines on emergency department opioid prescribing:

A multi-center survey. Pain Medicine 17 (3). First published online: 19 March 2016. Oxford

University Press.

Prescription Drug Monitoring Program Center of Excellence at Brandeis University (PDMP COE). 2014a.

Briefing on PDMP effectiveness. Brandeis University: Waltham, MA.

———. 2014b. Mandating PDMP participation by medical providers: Current status and experience in selected

states. Brandeis University: Waltham, MA.

———. 2016. PDMP prescriber use mandates: Characteristics, current status, and outcomes in selected states.

Waltham, MA. Retrieved from

http://www.pdmpassist.org/pdf/COE_documents/Add_to_TTAC/COE%20briefing%20on%20man

dates%203rd%20revision.pdf (accessed March 27, 2017).

Prescription Drug Monitoring Program (PDMP) Training and Technical Assistance Center. n.d. State profiles.

Retrieved from http://www.pdmpassist.org/content/state-profiles (accessed February 23, 2016).

———. 2015. Status of prescription drug monitoring programs (PDMPs). Retrieved from

http://www.pdmpassist.org/pdf/PDMPProgramStatus2015_v5.pdf (accessed February 23, 2016).

Reifler, L., D. Droz, J. Bailey, S. Schnoll, R. Fant, R. Dart, and B. Bartelson. 2012. Do prescription monitoring

programs impact state trends in opioid abuse/misuse? Pain Medicine 3: 434–442.

Rutkow, L., H. Chang, M. Daubresse, G. Alexander. 2015. Effect of Florida’s prescription drug monitoring

program and pill mill laws on opioid prescribing and use. JAMA Internal Medicine 175 (10). DOI:

10.1001/jamainternmed.2015.3931.

Silverman, S. 2009. Opioid induced hyperalgesia: Clinical implications for the pain practitioner. Pain

Physician 12: 679–684.

Simeone, R., and L. Holland. 2006. An evaluation of prescription drug monitoring programs. Albany, NY:

Simeone Associates, Inc.

Smith, V., K. Gifford, E. Ellis, B. Edwards, R. Rudowitz, E. Hinton, L. Antonisse, and A. Valentine. 2016.

Implementing coverage and payment initiatives results from a 50-state Medicaid budget survey for state

fiscal years 2016 and 2017. The Henry J. Kaiser Family Foundation and National Association of

Medicaid Directors.

Stapleton, D., G. Livermore, J. Cores, M. Nowak, and E. Eiseman. 2001. Evaluation of the New York State

ONECARD Rx workers’ compensation benefit: Final report. Prepared for State of New York,

Department of Civil Service. Project Officer: Stephanie Washington. Prepared by the Lewin Group,

Inc. and Cornell University.

Stover, B., J. Turner, G. Franklin, J. Gluck, D. Fulton-Kehoe, L. Sheppard, T. Wickizer, J. Kaufman, and K.

Egan. 2006. Factors associated with early opioid prescription among workers with low back injuries.

The Journal of Pain 7: 718–725.

Sullivan, M., M. Edlund, M. Fan, A. DeVries, J. Braden, and B. Martin. 2008. Trends in use of opioids for

non-cancer pain conditions 2000–2005 in commercial and Medicaid insurance plans: The TROUP

Study. Pain 138 (2): 440–449.

Swedlow, A., L. Gardner, J. Ireland, and E. Genovese. 2008. Pain management and the use of opioids in the

treatment of back conditions in the California workers’ compensation system. Oakland, CA: California

Workers’ Compensation Institute.

Swedlow, A., S. Hayes, and R. David. Are formularies a viable solution for controlling prescription drug

utilization and cost in California workers’ compensation? Oakland, CA: California Workers’

Compensation Institute.



92

http://www.pdmpassist.org/pdf/COE_documents/Add_to_TTAC/COE%20briefing%20on%20mandates%203rd%20revision.pdf

Swedlow, A., J. Ireland, and G. Johnson. 2011. Prescribing patterns of Schedule II opioids in California workers’

compensation. Oakland, CA: California Workers’ Compensation Institute.

Swedlow, A., and B. Young. 2012. Drug testing utilization and cost trends in California workers’ compensation.

Oakland, CA: California Workers’ Compensation Institute.

Tanabe, R. 2015. Workers’ compensation medical cost containment: A national inventory, 2015. Cambridge,

MA: Workers Compensation Research Institute.

Telles, C., and R. Tanabe. 2014. Monitoring trends in the New York workers’ compensation system. Cambridge,

MA: Workers Compensation Research Institute.

Tennessee Department of Health. Tennessee chronic pain guidelines: Clinical practice guidelines for outpatient

management of chronic non-malignant pain. Nashville, TN.

Texas Department of Insurance (TDI), Texas Workers’ Compensation Research and Evaluation Group. 2013.

Impact of the Texas pharmacy closed formulary: A preliminary report based on 12-month injuries with

9-month services. Retrieved from http://www.tdi.texas.gov/reports/wcreg/documents/

Pharma_070913.pdf (accessed February 23, 2016).

Thumula, V., D. Wang, and T. Liu. 2016. Interstate variations in use of opioids, 3rd edition. Cambridge, MA:

Workers Compensation Research Institute.

———. 2017. Interstate variations in use of opioids, 4th edition. Cambridge, MA: Workers Compensation

Research Institute.

Trust for America’s Health. 2013. Prescription drug abuse: Strategies to stop the epidemic. Washington, DC.

United States General Accounting Office (GAO). 2002. Prescription drugs: State monitoring programs provide

useful tool to reduce diversion. GAO-02-634. Retrieved from

http://www.gao.gov/new.items/d02634.pdf (accessed February 23, 2016).

Utah Department of Health. 2009. Utah clinical guidelines on prescribing opioids for treatment of pain.

Retrieved from http://www.dopl.utah.gov/licensing/forms/OpioidGuidlines.pdf (accessed February

23, 2016).

Victor, R., and P. Petrova. 2006a. State policies affecting the cost and use of pharmaceuticals in workers’

compensation: A national inventory. Cambridge, MA: Workers Compensation Research Institute.

———. 2006b. The cost and use of pharmaceuticals in workers’ compensation: A guide for policymakers.

Cambridge, MA: Workers Compensation Research Institute.

Volinn, E., J. Fargo, and P. Fine. 2009. Opioid therapy for nonspecific low back pain and the outcome of

chronic work loss. Pain 142: 194–201.

Wang, D. 2014. Longer-term use of opioids, 2nd edition. Cambridge, MA: Workers Compensation Research

Institute.

———. 2016. Longer-term use of opioids, 3rd edition. Cambridge, MA: Workers Compensation Research

Institute.

Wang, D., D. Hashimoto, and K. Mueller. 2012. Longer-term use of opioids. Cambridge, MA: Workers


Wang, D., and T. Liu. 2011. Prescription benchmarks, 2nd edition: Trends and interstate comparisons.

Cambridge, MA: Workers Compensation Research Institute.

Wang, D., K. Mueller, and D. Hashimoto. 2011. Interstate variations in use of narcotics. Cambridge, MA:

Workers Compensation Research Institute.

Wang, D., and R. Victor. 2010. Prescription benchmarks, first edition. 13 volumes. Cambridge, MA: Workers


Washington State Agency Medical Directors Group. 2010. Interagency guideline on opioid dosing for chronic



93

http://www.tdi.texas.gov/reports/wcreg/documents/Pharma_070913.pdf

non-cancer pain: An educational aid to improve care and safety with opioid therapy, 2010 update.

Washington State Department of Labor & Industries. 2013. Guideline for prescribing opioids to treat pain in

injured workers. Office of the Medical Director. Washington State Department of Labor & Industries.

Retrieved from http://lni.wa.gov/ClaimsIns/Files/OMD/MedTreat/

FINALOpioidGuideline010713.pdf (accessed February 23, 2016).

Webster, B., S. Verma, and R. Gatchel. 2007. Relationship between early opioid prescribing for acute

occupational low back pain and disability duration, medical costs, subsequent surgery and late

opioid use. Spine 32: 2,127–2,132.

Work Loss Data Institute. 2011. Pain (chronic). Encinitas, CA: Work Loss Data Institute.

Worley, J. 2012. Prescription drug monitoring programs, a response to doctor shopping: Purpose,

effectiveness, and directions for future research. Issues in Mental Health Nursing 33: 319–328.

Wynn, B., C. Buttorff, E. Meza, E. Taylor, A. Mulcahy. 2016. Implementing a drug formulary for California's

workers’ compensation program. Santa Monica, CA: RAND Corporation.

Yang, R., N. Coomer, S. Landes, E. Radeva, C. Telles, and R. Tanabe. 2009. CompScope™ benchmarks:

Technical appendix, 10th edition. Cambridge, MA: Workers Compensation Research Institute.

Yee, C., S. Pizer, and O. Fomenko. 2015. Why surgery rates vary. Cambridge, MA: Workers Compensation

Research Institute.



94

http://lni.wa.gov/ClaimsIns/Files/OMD/MedTreat/FINALOpioidGuideline010713.pdf

WCRI Publications

Medical Costs, Utilization, and Health Care Delivery wcri medical price index for workers’ compensation, ninth edition (mpi-wc). Rui Yang and Olesya

Fomenko. July 2017. wc-17-33. a multistate perspective on physician dispensing, 2011–2014. Dongchun Wang, Vennela Thumula, and Te-

Chun Liu. July 2017. wc-17-30. hospital outpatient payment index: interstate variations and policy analysis, 6th edition. Olesya

Fomenko and Rui Yang. June 2017. wc-17-32. interstate variations in use of opioids, 4th edition. Vennela Thumula, Dongchun Wang, and Te-Chun Liu.

June 2017. wc-17-28. the effects of provider choice policies on workers’ compensation costs. David Neumark and Bogdan

Savych. April 2017. wc-17-21. evaluation of the 2015 and 2016 fee schedule changes in delaware. Olesya Fomenko, Rui Yang, and Te-

Chun Liu. February 2017. wc-17-18. wcri medical price index for workers’ compensation, eighth edition (mpi-wc). Rui Yang and Olesya

Fomenko. November 2016. wc-16-74. designing workers’ compensation medical fee schedules, 2016. Olesya Fomenko and Te-Chun Liu. November

2016. wc-16-71. compscope™ medical benchmarks for california, 17th edition. Rui Yang and Karen Rothkin. October 2016.

wc-16-53. compscope™ medical benchmarks for florida, 17th edition. Rui Yang and Roman Dolinschi. October 2016.

wc-16-54. compscope™ medical benchmarks for georgia, 17th edition. Rui Yang. October 2016. wc-16-55. compscope™ medical benchmarks for illinois, 17th edition. Evelina Radeva. October 2016. wc-16-56. compscope™ medical benchmarks for indiana, 17th edition. Carol A. Telles. October 2016. wc-16-57. compscope™ medical benchmarks for kentucky, 17th edition. Carol A. Telles. October 2016. wc-16-58. compscope™ medical benchmarks for louisiana, 17th edition. Carol A. Telles. October 2016. wc-16-59. compscope™ medical benchmarks for massachusetts, 17th edition. Evelina Radeva. October 2016. wc-16-60. compscope™ medical benchmarks for michigan, 17th edition. Evelina Radeva. October 2016. wc-16-61. compscope™ medical benchmarks for minnesota, 17th edition. Sharon E. Belton. October 2016. wc-16-62. compscope™ medical benchmarks for new jersey, 17th edition. Carol A. Telles. October 2016. wc-16-63. compscope™ medical benchmarks for north carolina, 17th edition. Carol A. Telles. October 2016. wc-16-64. compscope™ medical benchmarks for pennsylvania, 17th edition. Evelina Radeva. October 2016. wc-16-65. compscope™ medical benchmarks for texas, 17th edition. Carol A. Telles. October 2016. wc-16-66. compscope™ medical benchmarks for virginia, 17th edition. Bogdan Savych. October 2016. wc-16-67. compscope™ medical benchmarks for wisconsin, 17th edition. Sharon E. Belton. October 2016. wc-16-68. hospital outpatient payment index: interstate variations and policy analysis, 5th edition. Olesya

Fomenko and Rui Yang. August 2016. wc-16-72. monitoring connecticut reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun

Liu. July 2016. wc-16-44. early impact of florida reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun

Liu. July 2016. wc-16-45. impact of georgia reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun Liu.

July 2016. wc-16-46. monitoring illinois reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun Liu.

July 2016. wc-16-47. monitoring indiana reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun Liu.

July 2016. wc-16-48.


monitoring michigan reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun Liu. July 2016. wc-16-49.

impact of south carolina reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun Liu. July 2016. wc-16-50.

monitoring tennessee reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-Chun Liu. July 2016. wc-16-51.

longer-term use of opioids, 3rd edition. Dongchun Wang. June 2016. wc-16-42. interstate variations in use of opioids, 3rd edition. Vennela Thumula, Dongchun Wang, and Te-Chun Liu.

June 2016. wc-16-22. payments to ambulatory surgery centers, 2nd edition. Bogdan Savych. May 2016. wc-16-39. comparing payments to ambulatory surgery centers and hospital outpatient departments, 2nd edition.

Bogdan Savych. May 2016. wc-16-40. crossing state lines for ambulatory surgical care: exploring claims from new york. Bogdan Savych. May

2016. wc-16-41. do higher fee schedules increase the number of workers’ compensation cases? Olesya Fomenko and

Jonathan Gruber. April 2016. wc-16-21. physician dispensing of higher-priced new drug strengths and formulation. Dongchun Wang, Vennela

Thumula, and Te-Chun Liu. April 2016. wc-16-18. texas-like formulary for north carolina state employees. Vennela Thumula and Te-Chun Liu. March 2016.

wc-16-19. evaluation of the 2015 fee schedule rates in illinois. Olesya Fomenko. February 2016. wc-16-20. wcri medical price index for workers’ compensation, seventh edition (mpi-wc). Rui Yang and Olesya

Fomenko. November 2015. wc-15-47. wcri flashreport: evaluation of the 2015 professional fee schedule update for florida. Olesya Fomenko.

November 2015. fr-15-01. will the affordable care act shift claims to workers’ compensation payors? Richard A. Victor, Olesya

Fomenko, and Jonathan Gruber. September 2015. wc-15-26. why surgery rates vary. Christine A. Yee, Steve Pizer, and Olesya Fomenko. June 2015. wc-15-24. workers’ compensation medical cost containment: a national inventory, 2015. Ramona P. Tanabe. April

2015. wc-15-27. hospital outpatient cost index for workers’ compensation, 4th edition. Olesya Fomenko and Rui Yang.

February 2015. wc-15-23. are physician dispensing reforms sustainable? Dongchun Wang, Vennela Thumula, and Te-Chun Liu. January

2015. wc-15-01. hospital outpatient cost index for workers’ compensation, 3rd edition. Olesya Fomenko and Rui Yang.

December 2014. wc-14-66. the impact of physician dispensing on opioid use. Vennela Thumula. December 2014. wc-14-56. early impact of tennessee reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-

Chun Liu. December 2014. wc-14-55. early impact of south carolina reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and

Te-Chun Liu. November 2014. wc-14-54. early impact of connecticut reforms on physician dispensing. Dongchun Wang, Vennela Thumula, and Te-

Chun Liu. November 2014. wc-14-53. estimating the effect of california’s fee schedule changes: lessons from WCRI studies. Rui Yang. October

2014. wc-14-49. impact of physician dispensing reform in georgia, 2nd edition. Dongchun Wang, Te-Chun Liu, and Vennela

Thumula. September 2014. wc-14-50. physician dispensing in pennsylvania, 2nd edition. Dongchun Wang, Te-Chun Liu, and Vennela Thumula.

September 2014. wc-14-51. wcri medical price index for workers’ compensation, sixth edition (mpi-wc). Rui Yang and Olesya

Fomenko. July 2014. wc-14-34. impact of a texas-like formulary in other states. Vennela Thumula and Te-Chun Liu. June 2014. wc-14-31. comparing payments to ambulatory surgery centers and hospital outpatient departments. Bogdan

Savych. June 2014. wc-14-30. payments to ambulatory surgery centers. Bogdan Savych. June 2014. wc-14-29.


interstate variations in use of narcotics, 2nd edition. Vennela Thumula, Dongchun Wang, and Te-Chun Liu. May 2014. wc-14-18.

longer-term use of opioids, 2nd edition. Dongchun Wang. May 2014. wc-14-19. the effect of reducing the illinois fee schedule. Rui Yang and Olesya Fomenko. January 2014. wc-14-01. the prevalence and costs of physician-dispensed drugs. Dongchun Wang, Te-Chun Liu, and Vennela

Thumula. September 2013. wc-13-39. physician dispensing in the pennsylvania workers’ compensation system. Dongchun Wang, Te-Chun Liu,

and Vennela Thumula. September 2013. wc-13-23. physician dispensing in the maryland workers’ compensation system. Dongchun Wang, Te-Chun Liu, and

Vennela Thumula. September 2013. wc-13-22. impact of banning physician dispensing of opioids in florida. Vennela Thumula. July 2013. wc-13-20. impact of reform on physician dispensing and prescription prices in georgia. Dongchun Wang, Te-Chun

Liu, and Vennela Thumula. July 2013. wc-13-21. a new benchmark for workers' compensation fee schedules: prices paid by commercial insurers?. Olesya

Fomenko and Richard A. Victor. June 2013. wc-13-17. comparing workers’ compensation and group health hospital outpatient payments. Olesya Fomenko. June

2013. wc-13-18. wcri medical price index for workers’ compensation, fifth edition (mpi-wc). Rui Yang and Olesya

Fomenko. June 2013. wc-13-19. workers’ compensation medical cost containment: a national inventory, 2013. Ramona P. Tanabe.

February 2013. wc-13-02. hospital outpatient cost index for workers’ compensation, 2nd edition. Olesya Fomenko and Rui Yang.

January 2013. wc-13-01. longer-term use of opioids. Dongchun Wang, Dean Hashimoto, and Kathryn Mueller. October 2012. wc-12-39. impact of treatment guidelines in texas. Philip S. Borba and Christine A. Yee. September 2012. wc-12-23. physician dispensing in workers’ compensation. Dongchun Wang. July 2012. wc-12-24. designing workers’ compensation medical fee schedules. Olesya Fomenko and Te-Chun Liu. June 2012.

wc-12-19. compscope™ medical benchmarks for maryland, 12th edition. Rui Yang. May 2012. wc-12-06. why surgeon owners of ambulatory surgical centers do more surgery than non-owners. Christine A. Yee.

May 2012. wc-12-17. wcri medical price index for workers’ compensation, fourth edition (mpi-wc). Rui Yang and Olesya

Fomenko. March 2012. wc-12-20. hospital outpatient cost index for workers’ compensation. Rui Yang and Olesya Fomenko. January 2012.

wc-12-01. wcri medical price index for workers’ compensation, third edition (mpi-wc). Rui Yang. August 2011. wc-

11-37. prescription benchmarks, 2nd edition: trends and interstate comparisons. Dongchun Wang and Te-Chun

Liu. July 2011. wc-11-31. prescription benchmarks for florida, 2nd edition. Dongchun Wang and Te-Chun Liu. July 2011. wc-11-32. prescription benchmarks for washington. Dongchun Wang and Te-Chun Liu. July 2011. wc-11-33. interstate variations in use of narcotics. Dongchun Wang, Kathryn Mueller, and Dean Hashimoto. July 2011.

wc-11-01. impact of preauthorization on medical care in texas. Christine A. Yee, Philip S. Borba, and Nicole Coomer.

June 2011. wc-11-34. workers' compensation medical cost containment: a national inventory, 2011. April 2011. wc-11-35. prescription benchmarks for minnesota. Dongchun Wang and Richard A. Victor. October 2010. wc-10-53. prescription benchmarks for florida. Dongchun Wang and Richard A. Victor. March 2010. wc-10-06. prescription benchmarks for illinois. Dongchun Wang and Richard A. Victor. March 2010. wc-10-05. prescription benchmarks for louisiana. Dongchun Wang and Richard A. Victor. March 2010. wc-10-10. prescription benchmarks for maryland. Dongchun Wang and Richard A. Victor. March 2010. wc-10-08. prescription benchmarks for massachusetts. Dongchun Wang and Richard A. Victor. March 2010. wc-10-07. prescription benchmarks for michigan. Dongchun Wang and Richard A. Victor. March 2010. wc-10-09. prescription benchmarks for north carolina. Dongchun Wang and Richard A. Victor. March 2010. wc-10-16.


prescription benchmarks for new jersey. Dongchun Wang and Richard A. Victor. March 2010. wc-10-15. prescription benchmarks for pennsylvania. Dongchun Wang and Richard A. Victor. March 2010. wc-10-11. prescription benchmarks for tennessee. Dongchun Wang and Richard A. Victor. March 2010. wc-10-13. prescription benchmarks for texas. Dongchun Wang and Richard A. Victor. March 2010. wc-10-12. prescription benchmarks for wisconsin. Dongchun Wang and Richard A. Victor. March 2010. wc-10-14. fee schedules for hospitals and ambulatory surgical centers: a guide for policymakers. Nicole M.

Coomer. February 2010. wc-10-01. national inventory of workers’ compensation fee schedules for hospitals and ambulatory surgical

centers. Nicole M. Coomer. February 2010. wc-10-02. workers’ compensation medical cost containment: a national inventory. August 2009. wc-09-15. wcri flashreport: information requested by medicare to support decision-making on medicare

secondary payer regulations. Ramona P. Tanabe. April 2009. fr-09-01. wcri medical price index for workers’ compensation, second edition (mpi-wc). Stacey M. Eccleston with the

assistance of Juxiang Liu. June 2008. wc-08-29. wcri flashreport: connecticut fee schedule rates compared to state medicare rates: common medical

services delivered to injured workers by nonhospital providers. Stacey M. Eccleston. December 2007. fr-07-04.

wcri flashreport: what are the most important medical conditions in workers’ compensation. August 2007. fr-07-03.

wcri flashreport: what are the most important medical conditions in new york workers’ compensation. July 2007. fr-07-02.

wcri flashreport: analysis of illustrative medical fee schedules in wisconsin. Stacey M. Eccleston, Te-Chun Liu, and Richard A. Victor. March 2007. fr-07-01.

wcri medical price index for workers’ compensation: the mpi-wc, first edition. Stacey M. Eccleston. February 2007. wc-07-33.

benchmarks for designing workers’ compensation medical fee schedules: 2006. Stacey M. Eccleston and Te-Chun Liu. October 2006. wc-06-14.

analysis of the workers’ compensation medical fee schedules in illinois. Stacey M. Eccleston. July 2006. wc-06-28.

state policies affecting the cost and use of pharmaceuticals in workers’ compensation: a national inventory. Richard A. Victor and Petia Petrova. June 2006. wc-06-30.

the cost and use of pharmaceuticals in workers’ compensation: a guide for policymakers. Richard A. Victor and Petia Petrova. June 2006. wc-06-13.

how does the massachusetts medical fee schedule compare to prices actually paid in workers’ compensation? Stacey M. Eccleston. April 2006. wc-06-27.

the impact of provider choice on workers’ compensation costs and outcomes. Richard A. Victor, Peter S. Barth, and David Neumark, with the assistance of Te-Chun Liu. November 2005. wc-05-14.

adverse surprises in workers’ compensation: cases with significant unanticipated medical care and costs. Richard A. Victor. June 2005. wc-05-16.

wcri flashreport: analysis of the proposed workers’ compensation fee schedule in tennessee. Stacey M. Eccleston and Xiaoping Zhao. January 2005. fr-05-01.

wcri flashreport: analysis of services delivered at chiropractic visits in texas compared to other states. Stacey M. Eccleston and Xiaoping Zhao. July 2004. fr-04-07.

wcri flashreport: supplement to benchmarking the 2004 pennsylvania workers’ compensation medical fee schedule. Stacey M. Eccleston and Xiaoping Zhao. May 2004. fr-04-06.

wcri flashreport: is chiropractic care a cost driver in texas? reconciling studies by wcri and mgt/texas chiropractic association. April 2004. fr-04-05.

wcri flashreport: potential impact of a limit on chiropractic visits in texas. Stacey M. Eccleston. April 2004. fr-04-04.

wcri flashreport: are higher chiropractic visits per claim driven by “outlier” providers? Richard A. Victor. April 2004. fr-04-03.

wcri flashreport: benchmarking the 2004 pennsylvania workers’ compensation medical fee schedule. Stacey M. Eccleston and Xiaoping Zhao. March 2004. fr-04-01.

evidence of effectiveness of policy levers to contain medical costs in workers’ compensation. Richard A. Victor. November 2003. wc-03-08.


wcri medical price index for workers’ compensation. Dongchun Wang and Xiaoping Zhao. October 2003. wc-03-05.

wcri flashreport: where the medical dollar goes? how california compares to other states. Richard A. Victor and Stacey M. Eccleston. March 2003. fr-03-03.

patterns and costs of physical medicine: comparison of chiropractic and physician-directed care. Richard A. Victor and Dongchun Wang. December 2002. wc-02-07.

provider choice laws, network involvement, and medical costs. Richard A. Victor, Dongchun Wang, and Philip Borba. December 2002. wc-02-05.

wcri flashreport: analysis of payments to hospitals and surgery centers in florida workers’ compensation. Stacey M. Eccleston and Xiaoping Zhao. December 2002. fr-02-03.

benchmarks for designing workers’ compensation medical fee schedules: 2001–2002. Stacey M. Eccleston, Aniko Laszlo, Xiaoping Zhao, and Michael Watson. August 2002. wc-02-02.

wcri flashreport: changes in michigan’s workers’ compensation medical fee schedule: 1996–2002. Stacey M. Eccleston. December 2002. fr-02-02.

targeting more costly care: area variation in texas medical costs and utilization. Richard A. Victor and N. Michael Helvacian. May 2002. wc-02-03.

wcri flashreport: comparing the pennsylvania workers’ compensation fee schedule with medicare rates: evidence from 160 important medical procedures. Richard A. Victor, Stacey M. Eccleston, and Xiaoping Zhao. November 2001. fr-01-07.

wcri flashreport: benchmarking pennsylvania’s workers’ compensation medical fee schedule. Stacey M. Eccleston and Xiaoping Zhao. October 2001. fr-01-06.

wcri flashreport: benchmarking california’s workers’ compensation medical fee schedules. Stacey M. Eccleston. August 2001. fr-01-04.

managed care and medical cost containment in workers’ compensation: a national inventory, 2001–2002. Ramona P. Tanabe and Susan M. Murray. December 2001. wc-01-04.

wcri flashreport: benchmarking florida’s workers’ compensation medical fee schedules. Stacey M. Eccleston and Aniko Laszlo. August 2001. fr-01-03.

the impact of initial treatment by network providers on workers’ compensation medical costs and disability payments. Sharon E. Fox, Richard A. Victor, Xiaoping Zhao, and Igor Polevoy. August 2001. dm-01-01.

the impact of workers’ compensation networks on medical and disability payments. William G. Johnson, Marjorie L. Baldwin, and Steven C. Marcus. November 1999. wc-99-5.

fee schedule benchmark analysis: ohio. Philip L. Burstein. December 1996. fs-96-1. the rbrvs as a model for workers’ compensation medical fee schedules: pros and cons. Philip L. Burstein.

July 1996. wc-96-5. benchmarks for designing workers’ compensation medical fee schedules: 1995–1996. Philip L. Burstein.

May 1996. wc-96-2. fee schedule benchmark analysis: north carolina. Philip L. Burstein. December 1995. fs-95-2. fee schedule benchmark analysis: colorado. Philip L. Burstein. August 1995. fs-95-1. benchmarks for designing workers’ compensation medical fee schedules: 1994–1995. Philip L. Burstein.

December 1994. wc-94-7. review, regulate, or reform: what works to control workers’ compensation medical costs. Thomas W.

Grannemann, ed. September 1994. wc-94-5. fee schedule benchmark analysis: michigan. Philip L. Burstein. September 1994. fs-94-1. medicolegal fees in california: an assessment. Leslie I. Boden. March 1994. wc-94-1. benchmarks for designing workers’ compensation medical fee schedules. Stacey M. Eccleston, Thomas W.

Grannemann, and James F. Dunleavy. December 1993. wc-93-4. how choice of provider and recessions affect medical costs in workers’ compensation. Richard B. Victor

and Charles A. Fleischman. June 1990. wc-90-2. medical costs in workers’ compensation: trends & interstate comparisons. Leslie I. Boden and Charles A.

Fleischman. December 1989. wc-89-5.


Worker Outcomes comparing outcomes for injured workers in indiana, 2016 interviews. Bogdan Savych and Vennela Thumula.

June 2017. wc-17-22. comparing outcomes for injured workers in massachusetts, 2016 interviews. Bogdan Savych and Vennela

Thumula. June 2017. wc-17-23. comparing outcomes for injured workers in michigan, 2016 interviews. Bogdan Savych and Vennela

Thumula. June 2017. wc-17-24. comparing outcomes for injured workers in north carolina, 2016 interviews. Bogdan Savych and Vennela

Thumula. June 2017. wc-17-25. comparing outcomes for injured workers in virginia, 2016 interviews. Bogdan Savych and Vennela

Thumula. June 2017. wc-17-26. comparing outcomes for injured workers in wisconsin, 2016 interviews. Bogdan Savych and Vennela

Thumula. June 2017. wc-17-27. comparing outcomes for injured workers in arkansas. Bogdan Savych and Vennela Thumula. May 2016. wc-

16-23. comparing outcomes for injured workers in connecticut. Bogdan Savych and Vennela Thumula. May 2016.

wc-16-24. comparing outcomes for injured workers in florida. Bogdan Savych and Vennela Thumula. May 2016. wc-16-

25. comparing outcomes for injured workers in georgia. Bogdan Savych and Vennela Thumula. May 2016. wc-16-

26. comparing outcomes for injured workers in indiana. Bogdan Savych and Vennela Thumula. May 2016. wc-16-

27. comparing outcomes for injured workers in iowa. Bogdan Savych and Vennela Thumula. May 2016. wc-16-28. comparing outcomes for injured workers in kentucky. Bogdan Savych and Vennela Thumula. May 2016. wc-

16-29. comparing outcomes for injured workers in massachusetts. Bogdan Savych and Vennela Thumula. May

2016. wc-16-30. comparing outcomes for injured workers in michigan. Bogdan Savych and Vennela Thumula. May 2016. wc-

16-31. comparing outcomes for injured workers in minnesota. Bogdan Savych and Vennela Thumula. May 2016. wc-

16-32. comparing outcomes for injured workers in north carolina. Bogdan Savych and Vennela Thumula. May

2016. wc-16-33. comparing outcomes for injured workers in pennsylvania. Bogdan Savych and Vennela Thumula. May 2016.

wc-16-34. comparing outcomes for injured workers in tennessee. Bogdan Savych and Vennela Thumula. May 2016. wc-

16-35. comparing outcomes for injured workers in virginia. Bogdan Savych and Vennela Thumula. May 2016. wc-

16-36. comparing outcomes for injured workers in wisconsin. Bogdan Savych and Vennela Thumula. May 2016. wc-

16-37. predictors of worker outcomes in arkansas. Bogdan Savych, Vennela Thumula, and Richard A. Victor. January

2015. wc-15-02. predictors of worker outcomes in connecticut. Bogdan Savych, Vennela Thumula, and Richard A. Victor.

January 2015. wc-15-03. predictors of worker outcomes in iowa. Bogdan Savych, Vennela Thumula, and Richard A. Victor. January

2015. wc-15-04. predictors of worker outcomes in tennessee. Bogdan Savych, Vennela Thumula, and Richard A. Victor.

January 2015. wc-15-05. predictors of worker outcomes in indiana. Bogdan Savych, Vennela Thumula, and Richard A. Victor. June

2014. wc-14-20. predictors of worker outcomes in massachusetts. Bogdan Savych, Vennela Thumula, and Richard A. Victor.

June 2014. wc-14-21.


predictors of worker outcomes in michigan. Bogdan Savych, Vennela Thumula, and Richard A. Victor. June 2014. wc-14-22.

predictors of worker outcomes in minnesota. Bogdan Savych, Vennela Thumula, and Richard A. Victor. June 2014. wc-14-23.

predictors of worker outcomes in north carolina. Vennela Thumula, Bogdan Savych, and Richard A. Victor. June 2014. wc-14-24.

predictors of worker outcomes in pennsylvania. Vennela Thumula, Bogdan Savych, and Richard A. Victor. June 2014. wc-14-25.

predictors of worker outcomes in virginia. Vennela Thumula, Bogdan Savych, and Richard A. Victor. June 2014. wc-14-26.

predictors of worker outcomes in wisconsin. Vennela Thumula, Bogdan Savych, and Richard A. Victor. June 2014. wc-14-27.

how have worker outcomes and medical costs changed in wisconsin? Sharon E. Belton and Te-Chun Liu. May 2010. wc-10-04.

comparing outcomes for injured workers in michigan. Sharon E. Belton and Te-Chun Liu. June 2009. wc-09-31.

comparing outcomes for injured workers in maryland. Sharon E. Belton and Te-Chun Liu. June 2008. wc-08-15.

comparing outcomes for injured workers in nine large states. Sharon E. Belton, Richard A. Victor, and Te-Chun Liu, with the assistance of Pinghui Li. May 2007. wc-07-14.

comparing outcomes for injured workers in seven large states. Sharon E. Fox, Richard A. Victor, and Te-Chun Liu, with the assistance of Pinghui Li. February 2006. wc-06-01.

wcri flashreport: worker outcomes in texas by type of injury. Richard A. Victor. February 2005. fr-05-02. outcomes for injured workers in california, massachusetts, pennsylvania, and texas. Richard A. Victor,

Peter S. Barth, and Te-Chun Liu, with the assistance of Pinghui Li. December 2003. wc-03-07. outcomes for injured workers in texas. Peter S. Barth and Richard A. Victor, with the assistance of Pinghui Li

and Te-Chun Liu. July 2003. wc-03-02. the workers’ story: results of a survey of workers injured in wisconsin. Monica Galizzi, Leslie I. Boden,

and Te-Chun Liu. December 1998. wc-98-5. workers’ compensation medical care: effective measurement of outcomes. Kate Kimpan. October 1996.

wc-96-7.

Benefits and Return to Work adequacy of workers’ compensation income benefits in michigan. Bogdan Savych and H. Allan Hunt. June

2017. wc-17-20. return to work after a lump-sum settlement. Bogdan Savych. July 2012. wc-12-21. factors influencing return to work for injured workers: lessons from pennsylvania and wisconsin.

Sharon E. Belton. November 2011. wc-11-39. the impact of the 2004 ppd reforms in tennessee: early evidence. Evelina Radeva and Carol Telles. May 2008.

fr-08-02. factors that influence the amount and probability of permanent partial disability benefits. Philip S.

Borba and Mike Helvacian. June 2006. wc-06-16. return-to-work outcomes of injured workers: evidence from california, massachusetts, pennsylvania,

and texas. Sharon E. Fox, Philip S. Borba, and Te-Chun Liu. May 2005. wc-05-15. who obtains permanent partial disability benefits: a six state analysis. Peter S. Barth, N. Michael

Helvacian, and Te-Chun Liu. December 2002. wc-02-04. wcri flashreport: benchmarking oregon’s permanent partial disability benefits. Duncan S. Ballantyne and

Michael Manley. July 2002. fr-02-01. wcri flashreport: benchmarking florida’s permanent impairment benefits. Richard A. Victor and Duncan

S. Ballantyne. September 2001. fr-01-05. permanent partial disability benefits: interstate differences. Peter S. Barth and Michael Niss. September

1999. wc-99-2. measuring income losses of injured workers: a study of the wisconsin system—A WCRI Technical Paper.

Leslie I. Boden and Monica Galizzi. November 1998.


permanent partial disability in tennessee: similar benefits for similar injuries? Leslie I. Boden. November 1997. wc-97-5.

what are the most important factors shaping return to work? evidence from wisconsin. Monica Galizzi and Leslie I. Boden. October 1996. wc-96-6.

do low ttd maximums encourage high ppd utilization: re-examining the conventional wisdom. John A. Gardner. January 1992. wc-92-2.

benefit increases and system utilization: the connecticut experience. John A. Gardner. December 1991. wc-91-5.

designing benefit structures for temporary disability: a guide for policymakers—Two-Volume Publication. Richard B. Victor and Charles A. Fleischman. December 1989. wc-89-4.

return to work incentives: lessons for policymakers from economic studies. John A. Gardner. June 1989. wc-89-2.

income replacement for long-term disability: the role of workers’ compensation and ssdi. Karen R. DeVol. December 1986. sp-86-2.

Cost Drivers and Benchmarks of System Performance compscope™ benchmarks for california, 17th edition. Rui Yang. April 2017. wc-17-01. compscope™ benchmarks for florida, 17th edition. Rui Yang. April 2017. wc-17-02. compscope™ benchmarks for georgia, 17th edition. Rui Yang and William-Monnin Browder. April 2017. wc-

17-03. compscope™ benchmarks for illinois, 17th edition. Evelina Radeva. April 2017. wc-17-04. compscope™ benchmarks for indiana, 17th edition. Carol A. Telles. April 2017. wc-17-05. compscope™ benchmarks for kentucky, 17th edition. William-Monnin Browder. April 2017. wc-17-06. compscope™ benchmarks for louisiana, 17th edition. Carol A. Telles. April 2017. wc-17-07. compscope™ benchmarks for massachusetts, 17th edition. Evelina Radeva. April 2017. wc-17-08. compscope™ benchmarks for michigan, 17th edition. Evelina Radeva. April 2017. wc-17-09. compscope™ benchmarks for minnesota, 17th edition. Sharon E. Belton. April 2017. wc-17-10. compscope™ benchmarks for new jersey, 17th edition. Evelina Radeva. April 2017. wc-17-11. monitoring the north carolina system: compscope™ benchmarks, 17th edition. Carol A. Telles. April 2017.

wc-17-12. compscope™ benchmarks for pennsylvania, 17th edition. Evelina Radeva. April 2017. wc-17-13. compscope™ benchmarks for texas, 17th edition. Carol A. Telles. April 2017. wc-17-14. compscope™ benchmarks for virginia, 17th edition. Bogdan Savych. April 2017. wc-17-15. compscope™ benchmarks for wisconsin, 17th edition. Sharon E. Belton. April 2017. wc-17-16. monitoring trends in the new york workers’ compensation system, 2005–2014. Carol A. Telles and William

Monnin-Browder. February 2017. wc-17-19. monitoring trends in the new york workers’ compensation system, 2005–2013. Carol A. Telles and Ramona

P. Tanabe. February 2016. wc-16-38. monitoring trends in the new york workers’ compensation system. Carol A. Telles and Ramona P. Tanabe.

September 2014. wc-14-33. monitoring changes in new york after the 2007 reforms. Carol A. Telles and Ramona P. Tanabe. October

2013. wc-13-24. monitoring the impact of the 2007 reforms in new york. Carol A. Telles and Ramona P. Tanabe. October 2012.

wc-12-22. compscope™ benchmarks for maryland, 12th edition. Rui Yang, with the assistance of Syd Allan. December

2011. wc-11-45. early impact of 2007 reforms in new york. Carol A. Telles and Ramona P. Tanabe. December 2011. wc-11-38. compscopeTM benchmarks for tennessee, 11th edition. Evelina Radeva, Nicole M. Coomer, Bogdan Savych,

Carol A. Telles, Rui Yang, and Ramona P. Tanabe, with the assistance of Syd Allan. January 2011. wc-11-13. baseline trends for evaluating the impact of the 2007 reforms in new york. Ramona P. Tanabe and Carol

A. Telles. November 2010. wc-10-36. updated baseline for evaluating the impact of the 2007 reforms in new york. Ramona P. Tanabe, Stacey

Eccleston, and Carol A. Telles. April 2009. wc-09-14.


interstate variations in medical practice patterns for low back conditions. Dongchun Wang, Kathryn Meuller, Dean Hashimoto, Sharon Belton, and Xiaoping Zhao. June 2008. wc-08-28.

wcri flashreport: timeliness of injury reporting and first indemnity payment in new york: a comparison with 14 states. Carol A. Telles and Ramona P. Tanabe. March 2008. fr-08-01.

baseline for evaluating the impact of the 2007 reforms in new york. Ramona P. Tanabe, Stacey Eccleston, and Carol A. Telles. March 2008. wc-08-14.

why are benefit delivery expenses higher in california and florida? Duncan S. Ballantyne and Carol A. Telles. December 2002. wc-02-06.

compscopeTM benchmarks: massachusetts, 1994–1999. Carol A. Telles, Aniko Laszlo, and Te-Chun Liu. January

2002. cs-01-03. compscopeTM

benchmarks: florida, 1994–1999. N. Michael Helvacian and Seth A. Read. September 2001. cs-01-1. wcri flashreport: where the workers’ compensation dollar goes. Richard A. Victor and Carol A. Telles.

August 2001. fr-01-01. predictors of multiple workers’ compensation claims in wisconsin. Glenn A. Gotz, Te-Chun Liu, and

Monica Galizzi. November 2000. wc-00-7. area variations in texas benefit payments and claim expenses. Glenn A. Gotz, Te-Chun Liu, Christopher J.

Mazingo, and Douglas J. Tattrie. May 2000. wc-00-3. area variations in california benefit payments and claim expenses. Glenn A. Gotz, Te-Chun Liu, and

Christopher J. Mazingo. May 2000. wc-00-2. area variations in pennsylvania benefit payments and claim expenses. Glenn A. Gotz, Te-Chun Liu, and

Christopher J. Mazingo. May 2000. wc-00-1. benchmarking the performance of workers’ compensation systems: compscopeTM

measures for minnesota. H. Brandon Haller and Seth A. Read. June 2000. cs-00-2.

benchmarking the performance of workers’ compensation systems: compscopeTM measures for

massachusetts. Carol A. Telles and Tara L. Nells. December 1999. cs-99-3. benchmarking the performance of workers’ compensation systems: compscopeTM

measures for california. Sharon E. Fox and Tara L. Nells. December 1999. cs-99-2.

benchmarking the performance of workers’ compensation systems: compscopeTM measures for

pennsylvania. Sharon E. Fox and Tara L. Nells. November 1999. cs-99-1. cost drivers and system performance in a court-based system: tennessee. John A. Gardner, Carol A. Telles,

and Gretchen A. Moss. June 1996. wc-96-4. the 1991 reforms in massachusetts: an assessment of impact. John A. Gardner, Carol A. Telles, and Gretchen

A. Moss. May 1996. wc-96-3. the impact of oregon’s cost containment reforms. John A. Gardner, Carol A. Telles, and Gretchen A. Moss.

February 1996. wc-96-1. cost drivers and system change in georgia, 1984–1994. John A. Gardner, Carol A. Telles, and Gretchen A. Moss.

November 1995. wc-95-3. cost drivers in missouri. John A. Gardner, Richard A. Victor, Carol A. Telles, and Gretchen A. Moss. December

1994. wc-94-6. cost drivers in new jersey. John A. Gardner, Richard A. Victor, Carol A. Telles, and Gretchen A. Moss. September

1994. wc-94-4. cost drivers in six states. Richard A. Victor, John A. Gardner, Daniel Sweeney, and Carol A. Telles. December

1992. wc-92-9. performance indicators for permanent disability: low-back injuries in texas. Sara R. Pease. August 1988.

wc-88-4. performance indicators for permanent disability: low-back injuries in new jersey. Sara R. Pease.

December 1987. wc-87-5. performance indicators for permanent disability: low-back injuries in wisconsin. Sara R. Pease. December

1987. wc-87-4.

Administration/Litigation wcri flashreport – worker attorney involvement: a new measure. Rui Yang, Karen Rothkin, and Roman

Dolinschi. May 2017. FR-17-01.


workers’ compensation laws as of january 1, 2016. Joint publication of IAIABC and WCRI. May 2016. wc-16-43.

workers’ compensation laws as of january 1, 2014. Joint publication of IAIABC and WCRI. April 2014. wc-14-28.

workers’ compensation laws as of january 2012. Joint publication of IAIABC and WCRI. Ramona P. Tanabe. March 2012. wc-12-18.

workers’ compensation laws, 3rd edition. Joint publication of IAIABC and WCRI. Ramona P. Tanabe. October 2010. wc-10-52.

avoiding litigation: what can employers, insurers, and state workers’ compensation agencies do?. Richard A. Victor and Bogdan Savych. July 2010. wc-10-18.

workers’ compensation laws, 2nd edition. Joint publication of IAIABC and WCRI. June 2009. wc-09-30. did the florida reforms reduce attorney involvement? Bogdan Savych and Richard A. Victor. June 2009.

wc-09-16. lessons from the oregon workers’ compensation system. Duncan S. Ballantyne. March 2008. wc-08-13. workers’ compensation in montana: administrative inventory. Duncan S. Ballantyne. March 2007. wc-07-12. workers’ compensation in nevada: administrative inventory. Duncan S. Ballantyne. December 2006.

wc-06-15. workers’ compensation in hawaii: administrative inventory. Duncan S. Ballantyne. April 2006. wc-06-12. workers’ compensation in arkansas: administrative inventory. Duncan S. Ballantyne. August 2005.

wc-05-18. workers’ compensation in mississippi: administrative inventory. Duncan S. Ballantyne. May 2005. wc-05-13. workers’ compensation in arizona: administrative inventory. Duncan S. Ballantyne. September 2004.

wc-04-05. workers’ compensation in iowa: administrative inventory. Duncan S. Ballantyne. April 2004. wc-04-02. wcri flashreport: measuring the complexity of the workers’ compensation dispute resolution processes

in tennessee. Richard A. Victor. April 2004. fr-04-02. revisiting workers’ compensation in missouri: administrative inventory. Duncan S. Ballantyne. December

2003. wc-03-06. workers’ compensation in tennessee: administrative inventory. Duncan S. Ballantyne. April 2003. wc-03-01. revisiting workers’ compensation in new york: administrative inventory. Duncan S. Ballantyne. January

2002. wc-01-05. workers’ compensation in kentucky: administrative inventory. Duncan S. Ballantyne. June 2001. wc-01-01. workers’ compensation in ohio: administrative inventory. Duncan S. Ballantyne. October 2000. wc-00-5. workers’ compensation in louisiana: administrative inventory. Duncan S. Ballantyne. November 1999.

wc-99-4. workers’ compensation in florida: administrative inventory. Peter S. Barth. August 1999. wc-99-3. measuring dispute resolution outcomes: a literature review with implications for workers’

compensation. Duncan S. Ballantyne and Christopher J. Mazingo. April 1999. wc-99-1. revisiting workers’ compensation in connecticut: administrative inventory. Duncan S. Ballantyne.

September 1998. wc-98-4. dispute prevention and resolution in workers’ compensation: a national inventory, 1997–1998. Duncan S.

Ballantyne. May 1998. wc-98-3. workers’ compensation in oklahoma: administrative inventory. Michael Niss. April 1998. wc-98-2. workers’ compensation advisory councils: a national inventory, 1997–1998. Sharon E. Fox. March 1998.

wc-98-1. the role of advisory councils in workers’ compensation systems: observations from wisconsin. Sharon E.

Fox. November 1997. revisiting workers’ compensation in michigan: administrative inventory. Duncan S. Ballantyne and

Lawrence Shiman. October 1997. wc-97-4. revisiting workers’ compensation in minnesota: administrative inventory. Carol A. Telles and Lawrence

Shiman. September 1997. wc-97-3. revisiting workers’ compensation in california: administrative inventory. Carol A. Telles and Sharon E.

Fox. June 1997. wc-97-2. revisiting workers’ compensation in pennsylvania: administrative inventory. Duncan S. Ballantyne. March

1997. wc-97-1.


revisiting workers’ compensation in washington: administrative inventory. Carol A. Telles and Sharon E. Fox. December 1996. wc-96-10.

workers’ compensation in illinois: administrative inventory. Duncan S. Ballantyne and Karen M. Joyce. November 1996. wc-96-9.

workers’ compensation in colorado: administrative inventory. Carol A. Telles and Sharon E. Fox. October 1996. wc-96-8.

workers’ compensation in oregon: administrative inventory. Duncan S. Ballantyne and James F. Dunleavy. December 1995. wc-95-2.

revisiting workers’ compensation in texas: administrative inventory. Peter S. Barth and Stacey M. Eccleston. April 1995. wc-95-1.

workers’ compensation in virginia: administrative inventory. Carol A. Telles and Duncan S. Ballantyne. April 1994. wc-94-3.

workers’ compensation in new jersey: administrative inventory. Duncan S. Ballantyne and James F. Dunleavy. April 1994. wc-94-2.

workers’ compensation in north carolina: administrative inventory. Duncan S. Ballantyne. December 1993. wc-93-5.

workers’ compensation in missouri: administrative inventory. Duncan S. Ballantyne and Carol A. Telles. May 1993. wc-93-1.

workers’ compensation in california: administrative inventory. Peter S. Barth and Carol A. Telles. December 1992. wc-92-8.

workers’ compensation in wisconsin: administrative inventory. Duncan S. Ballantyne and Carol A. Telles. November 1992. wc-92-7.

workers’ compensation in new york: administrative inventory. Duncan S. Ballantyne and Carol A. Telles. October 1992. wc-92-6.

the ama guides in maryland: an assessment. Leslie I. Boden. September 1992. wc-92-5. workers’ compensation in georgia: administrative inventory. Duncan S. Ballantyne and Stacey M. Eccleston.

September 1992. wc-92-4. workers’ compensation in pennsylvania: administrative inventory. Duncan S. Ballantyne and Carol A.

Telles. December 1991. wc-91-4. reducing litigation: using disability guidelines and state evaluators in oregon. Leslie I. Boden, Daniel E.

Kern, and John A. Gardner. October 1991. wc-91-3. workers’ compensation in minnesota: administrative inventory. Duncan S. Ballantyne and Carol A. Telles.

June 1991. wc-91-1. workers’ compensation in maine: administrative inventory. Duncan S. Ballantyne and Stacey M. Eccleston.

December 1990. wc-90-5. workers’ compensation in michigan: administrative inventory. H. Allan Hunt and Stacey M. Eccleston.

January 1990. wc-90-1. workers’ compensation in washington: administrative inventory. Sara R. Pease. November 1989. wc-89-3. workers’ compensation in texas: administrative inventory. Peter S. Barth, Richard B. Victor, and Stacey M.

Eccleston. March 1989. wc-89-1. reducing litigation: evidence from wisconsin. Leslie I. Boden. December 1988. wc-88-7. workers’ compensation in connecticut: administrative inventory. Peter S. Barth. December 1987. wc-87-3. use of medical evidence: low-back permanent partial disability claims in new jersey. Leslie I. Boden.

December 1987. wc-87-2. use of medical evidence: low-back permanent partial disability claims in maryland. Leslie I. Boden.

September 1986. sp-86-1.

Vocational Rehabilitation

improving vocational rehabilitation outcomes: opportunities for early intervention. John A. Gardner. August 1988. wc-88-3.

appropriateness and effectiveness of vocational rehabilitation in florida: costs, referrals, services, and outcomes. John A. Gardner. February 1988. wc-88-2.

vocational rehabilitation in florida workers’ compensation: rehabilitants, services, costs, and outcomes. John A. Gardner. February 1988. wc-88-1.


vocational rehabilitation outcomes: evidence from new york. John A. Gardner. December 1986. wc-86-1. vocational rehabilitation in workers’ compensation: issues and evidence. John A. Gardner. June 1985.

s-85-1.

Occupational Disease

liability for employee grievances: mental stress and wrongful termination. Richard B. Victor, ed. October 1988. wc-88-6.

asbestos claims: the decision to use workers’ compensation and tort. Robert I. Field and Richard B. Victor. September 1988. wc-88-5.

Other

workers’ compensation: where have we come from? where are we going?. Richard A. Victor and Linda L. Carrubba, eds. November 2010. wc-10-33.

recession, fear of job loss, and return to work. Richard A. Victor and Bogdan Savych. April 2010. wc-10-03. wcri flashreport: what are the prevalence and size of lump-sum payments in workers’ compensation:

estimates relevant for medicare set-asides. Richard A. Victor, Carol A. Telles, and Rui Yang. November 2006. fr-06-01.

the future of workers’ compensation: opportunities and challenges. Richard A. Victor, ed. April 2004. wc-04-03.

managing catastrophic events in workers’ compensation: lessons from 9/11. Ramona P. Tanabe, ed. March 2003. wc-03-03.

wcri flashreport: workers’ compensation in california: lessons from recent wcri studies. Richard A. Victor. March 2003. fr-03-02.

wcri flashreport: workers’ compensation in florida: lessons from recent wcri studies. Richard A. Victor. February 2003. fr-03-01.

workers’ compensation and the changing age of the workforce. Douglas J. Tattrie, Glenn A. Gotz, and Te-Chun Liu. December 2000. wc-00-6.

medical privacy legislation: implications for workers’ compensation. Ramona P. Tanabe, ed. November 2000. wc-00-4.

the implications of changing employment relations for workers’ compensation. Glenn A. Gotz, ed. December 1999. wc-99-6.

workers’ compensation success stories. Richard A. Victor, ed. July 1993. wc-93-3. the americans with disabilities act: implications for workers’ compensation. Stacey M. Eccleston, ed. July

1992. wc-92-3. twenty-four-hour coverage. Richard A. Victor, ed. June 1991. wc-91-2.

These publications can be obtained by visiting our web site at www.wcrinet.org or by sending a written request by fax to (617) 661-9284, or by mail to

Publications Department Workers Compensation Research Institute 955 Massachusetts Avenue Cambridge, MA 02139


About the Institute

Our Mission:

To be a catalyst for significant improvements in workers’ compensation

systems, providing the public with objective, credible, high-quality

research on important public policy issues.

The Institute:

Founded in 1983, the Workers Compensation Research Institute (WCRI)

is an independent, not-for-profit research organization which strives to

help those interested in making improvements to the workers’

compensation system by providing highly-regarded, objective data

and analysis.

The Institute does not take positions on the issues it researches; rather,

it provides information obtained through studies and data collection

efforts, which conform to recognized scientific methods. Objectivity is

further ensured through rigorous, unbiased peer review procedures.

The Institute’s work includes:

Original research studies of major issues confronting workers’

compensation systems (for example, outcomes for injured

workers).

Studies of individual state systems where policymakers have

shown an interest in change and where there is an unmet need

for objective information.

Studies of states that have undergone major legislative changes

to measure the impact of those reforms and draw possible lessons

for other states.

Presentations on research findings to legislators, workers’

compensation administrators, industry groups, and other

stakeholders.

With WCRI’s research, policymakers and other system stakeholders—

employers, insurers, and labor unions—can monitor state systems on a

regular basis and identify incremental changes to improve system

performance. This results in a more enduring, efficient, and equitable

system that better serves the needs of workers and employers.


WCRI Research Brief · 2017-08-03 · Page 1 of 2 WCRI Research Brief: Longer-Term Dispensing of...

Documents

Transcript of WCRI Research Brief · 2017-08-03 · Page 1 of 2 WCRI Research Brief: Longer-Term Dispensing of...