Arkansas Department of Health and MONAHRQ Presented by: Connie Marie Ardwin.

Arkansas Department of Health and MONAHRQ

Presented by:Connie Marie Ardwin

Introduction

• Hospital Discharge Data Section, Arkansas Department of Health

• Collecting inpatient discharge data since 1997 • Includes many additional fields; refined over

time to improve data quality and annual report.

• In the past, had “Request for Information” form when specific or aggregate data was needed

• Today, take the next step to make our county and area level aggregate data publicly available

Who will use the site?

• Hospital Discharge Data Section

• Arkansas Hospital

• Public Health Employee• Local Community• Future Inpatient• Researcher

In house

Contributors

Service usersand analysts

What makes Arkansas different?

• Arkansas has 75 counties– 41 of these are serviced by a single hospital– 20 do not have access to a hospital within county borders

• Arkansas has 108 hospitals– 52 of these are Acute Care and 29 are CAH

• Arkansas State law and policy– Centralized Health Department tasked with data collection– Requires that hospital and provider level information not be

disclosed publicly without hospital consent– Requires that our Data Submittal Guide (rules for formatting and

sending data) go through legislative process– Currently only collect inpatient data, however we have begun the

legislative process for collecting ED data starting in 2012.

How has AR been reporting?

• Publicly– Annual Report– http://www.healthy.arkansas.gov/programsServic

es/healthStatistics/Documents/HospitalDischarge/2009HospitalDischargeAnnualReport.pdf

• Privately / In House– Requests for Information– Detailed summary of hospital utilization provided

for each hospital upon request– Datasets or aggregate data summaries for

researchers and other State departments

http://www.healthy.arkansas.gov/programsServices/healthStatistics/Documents/HospitalDischarge/2009HospitalDischargeAnnualReport.pdf



How will we use MONAHRQ?

• Internally– To check our analyses done by SAS– To provide hospitals information on their facility

• Publicly– To report county level rates and quality measures– To enhance our Annual Report

• Password Protected– We had originally considered asking hospitals if they

would like their data to be included in a masked ‘hospital level, utilization’ website.

How did AR implement MONAHRQ?

• Had to have staff available– Look at the uses of MONAHRQ and which aspects

would benefit us.– Learn how to use MONAHRQ

• Had to have the right political climate– Public reporting is a hot topic– Work with our Hospital Association to get word

out about our publications (Annual Report)

Software and IT Issues

• Software is extremely easy to set up and use.

• AR data required lots of prep work.– Mapping the dataset– Easier to work qualifying

logic before loading than after

– Double and triple check the #s

Mapping File was made using the HCUP resources for mapping.

Microsoft Office Word 97 - 2003 Document

SAS to MQ map.doc



• Data requires lots of prep work.– Mapping the dataset– Easier to work qualifying



Our data is collected as a flat file from the hospitals, then all hospitals are combined into a Master dataset for the year and stored as a SAS dataset.

We use a SAS program to roll up / map the data elements, drop unnecessary elements, and use the MONAHRQ logic to delete non-qualifying records.



• Data requires lots of prep work.– Mapping the dataset– Easier to work qualifying



Never assume that because it worked last time that the numbers are good this time. Double check all numbers.

Example: I changed from using the total charges stored as a character to the total charges stored as a number. I removed the decimal point, and got a surprised question: “We could save three billion dollars if we reduced hospitalizations for Congestive Heart Failure by 50%?

Hosting and HTML Issues

• Generated pages can be linked to from our SharePoint site, but not incorporated into them.

• The company hosting our department site had trouble understanding what we needed to add, as they did not know what MONAHRQ generated.

They can also be converted to a SharePoint site, but that would be costly.

The static IP address assigned to the pages where they are to be stored is very user un-friendly. So, we are obtaining a .gov address through our DIS department that will redirect to the static IP address.

Hosting and HTML Issues

• Generated pages can be linked to from our SharePoint site, but not incorporated into them.

• The company hosting our department site had trouble understanding what we needed to add, as they did not know what MONAHRQ generated.

After many emails, file transfers, phone calls, and sit downs we figured out how to describe what we need.

“We need a link from our section of the ADH website to the homepage (index) where our generated pages are stored. The web pages we generate are all relative. We will need the .gov address to redirect to this address.

Hosting and HTML Issues• HTML text changes– Some definitions, such as Cost Savings, need further

clarification.– Add Disclaimers to certain pages, such as how this data

does not include Arkansas patients that were treated in an out of state facility.

• Apply the “Look and Feel” of the ADH site to the MONAHRQ pages; this cannot be easily done without the help of Web Design personnel.

• We have been told that we can add visitor counters and a link for contacting us if there are questions, however we have not been informed how to go about adding these to the pages.

Home Page - Generated

Home Page - Altered

How the ADH uses MONAHRQ

• MONAHRQ is used internally by our Hospital Discharge Data Section.

• Error check our data.• Error check our

mappings.• Compare utilization

statistics.

Data Load Summary Data have been loaded from your input file and are ready for analysis. The following shows descriptive statistics for the loaded data. You may go back and change any of your data mapping and crosswalk options and reload the file to correct any errors. Total Rows Loaded 418,562 Total Rows Excluded 0 Number of variables per record 73 Records with extra values (more than 73) 418,562

Record Warnings Records Column Affected Message Race 666 Value mapped to null based on crosswalk (info) Primary Payer 648 Value mapped to null based on crosswalk (info) Discharge Disposition 945 Value mapped to null based on crosswalk (info) Admission Type 2 Value mapped to null based on crosswalk (info) Admission Source 5,467 Value mapped to null based on crosswalk (info) APR-DRG 84 Invalid discharge status (Grouper) 2 Record does not meet criteria for any DRG (Grouper) 30 Gestational age/birth weight conflict (Grouper) 153 Invalid principal diagnosis (Grouper) Principal Diagnosis 1 String value over than 5 characters has been shortened 4 String value less than 3 characters Diagnosis Code 8 3 String value less than 3 characters Diagnosis Code 9 2 String value less than 3 characters Diagnosis Code 10 39 String value less than 3 characters Diagnosis Code 15 1 String value over than 5 characters has been shortened Diagnosis Code 31 6 String value over than 5 characters has been shortened Principal Procedure 366 String value over than 4 characters has been shortened Procedure Code 2 4,028 String value over than 4 characters has been shortened Procedure Code 3 2,429 String value over than 4 characters has been shortened Procedure Code 4 2,750 String value over than 4 characters has been shortened Procedure Code 5 1,979 String value over than 4 characters has been shortened Procedure Code 6 1,410 String value over than 4 characters has been shortened Procedure Code 7 778 String value over than 4 characters has been shortened Procedure Code 8 633 String value over than 4 characters has been shortened

File Warnings Column Message Diagnosis Code 31 Column of ICD-9-CM codes does not have any leading zeros (warning) Diagnosis Code 32 Column of ICD-9-CM codes does not have any leading zeros (warning) Diagnosis Code 33 Column of ICD-9-CM codes does not have any leading zeros (warning)








statistics.

Invalid values: diagnosis codes and procedure codes

Incorrect line ups: Present on Admissions and Days to Procedure present without a ICD-9 code or missing when there was.








statistics.

Invalid mappings: race/ethnicity, days to procedure, source of payment, patient disposition, source and type of admission.








statistics.

Discharges, Length of Stay, Charges, Cost, and means

By

Status, Patient demographics, Hospital demographics


• A section of MONAHRQ has been approved to be used publicly within an online published document.

• Produce the PQI section of our Annual Report.

• This information is now online and available for everyone to view!

Hospital Discharge Page

http://www.healthy.arkansas.gov/programsServices/healthStatistics/Pages/HospitalDischarge.aspx

Help with setting up MONAHRQ

• Email [email protected]

• AHRQ Quality Indicators Web Sitehttp://www.qualityindicators.ahrq.gov/

• HCUP Data Elementshttp://www.hcup-us.ahrq.gov/db/state/siddist/siddist_ddeavailbyyear.jsp

mailto:[email protected]

http://www.qualityindicators.ahrq.gov/

http://www.hcup-us.ahrq.gov/db/state/siddist/siddist_ddeavailbyyear.jsp

http://www.hcup-us.ahrq.gov/db/state/siddist/siddist_ddeavailbyyear.jsp

Public Reporting

Challenges• Collecting hospital data

• Getting accurate data

• Border counties have a high population % going to out of state hospital

• Protecting patient information, particularly in low population counties

Solutions• Law requires hospitals to

report• Get corrections on records

from hospitals until less than 1% error by our checks

• Requesting out of state hospital data and using the Exchange

• Make sure all counties reported on have a high enough denominator – currently done by hand

Public Reporting

Challenges• Law does not allow to publish

hospital level or identifying information

• Numerous (75) counties, the majority with a small population, cause many of the county rates to be suppressed.

Solutions• Do not publish the hospital

utilization and rates sections of the website

• Group hospitals up into a peer group and report on the group

• Considered rolling our counties up into larger reporting regions in future, but we are keeping it at the county level for now.

Impact of MONAHRQ

• Internal exposure to MONAHRQ within the Health Statistics Branch has generated a lot of interest in both the software and reporting elements.

• Hospital Discharge Data Section

• Other Departments

Impact of MONAHRQ


Hospital Discharge Data Section

• Will ease reporting to and on hospitals, including some requests for information

• Paved way for adding estimated costs

• Added reliable Quality Indicators by county

Impact of MONAHRQ


Other Departments / Reporting

• Critical Access Hospitals (special reporting)

• Emergency Department Data (hope to start collecting in 2012)

• Epidemiology (like to see a future version with Injury Indicators – CDC)

Future of MONAHRQ use in AR

• We hope to have our first hospital inpatient public reporting site go live as early as October this year. Technical issues could push this launch back to November.

• We plan on speaking with our Hospital Association about options for making the Utilization and Quality Ratings available to the hospitals.

Questions?

If you have any additional questions you can contact me at:

[email protected] # 501-280-4064

Thank you.

mailto:[email protected]

Arkansas Department of Health and MONAHRQ Presented by: Connie Marie Ardwin.

Documents

Transcript of Arkansas Department of Health and MONAHRQ Presented by: Connie Marie Ardwin.