Va pcmh study 6 2014[1]

Copyright 2014 American Medical Association. All rights reserved.

Implementation of the Patient-Centered Medical Homein the Veterans Health AdministrationAssociations With Patient Satisfaction, Quality of Care,Staff Burnout, and Hospital and Emergency Department UseKarin M. Nelson, MD, MSHS; Christian Helfrich, MPH, PhD; Haili Sun, PhD; Paul L. Hebert, PhD;Chuan-Fen Liu, MPH, PhD; Emily Dolan, PhD; Leslie Taylor, PhD; Edwin Wong, PhD; Charles Maynard, PhD;Susan E. Hernandez, MPA; William Sanders, AA, AS; Ian Randall, MHSA; Idamay Curtis, BA;Gordon Schectman, MD; Richard Stark, MD; Stephan D. Fihn, MD, MPH

IMPORTANCE In 2010, the Veterans Health Administration (VHA) began implementing thepatient-centered medical home (PCMH) model. The Patient Aligned Care Team (PACT)initiative aims to improve health outcomes through team-based care, improved access, andcare management. To track progress and evaluate outcomes at all VHA primary care clinics,we developed and validated a method to assess PCMH implementation.

OBJECTIVES To create an index that measures the extent of PCMH implementation, describevariation in implementation, and examine the association between the implementation indexand key outcomes.

DESIGN, SETTING, AND PARTICIPANTS We conducted an observational study using data onmore than 5.6 million veterans who received care at 913 VHA hospital-based andcommunity-based primary care clinics and 5404 primary care staff from (1) VHA clinical andadministrative databases, (2) a national patient survey administered to a weighted randomsample of veterans who received outpatient care from June 1 to December 31, 2012, and (3) asurvey of all VHA primary care staff in June 2012. Composite scores were constructed for 8core domains of PACT: access, continuity, care coordination, comprehensiveness,self-management support, patient-centered care and communication, shared decisionmaking, and team-based care.

MAIN OUTCOMES AND MEASURES Patient satisfaction, rates of hospitalization and emergencydepartment use, quality of care, and staff burnout.

RESULTS Fifty-three items were included in the PACT Implementation Progress Index (Pi2).Compared with the 87 clinics in the lowest decile of the Pi2, the 77 sites in the top decileexhibited significantly higher patient satisfaction (9.33 vs 7.53; P < .001), higher performanceon 41 of 48 measures of clinical quality, lower staff burnout (Maslach Burnout Inventoryemotional exhaustion subscale, 2.29 vs 2.80; P = .02), lower hospitalization rates forambulatory care–sensitive conditions (4.42 vs 3.68 quarterly admissions for veterans 65years or older per 1000 patients; P < .001), and lower emergency department use (188 vs245 visits per 1000 patients; P < .001).

CONCLUSIONS AND RELEVANCE The extent of PCMH implementation, as measured by the Pi2,was highly associated with important outcomes for both patients and providers. Thismeasure will be used to track the effectiveness of implementing PACT over time and toelucidate the correlates of desired health outcomes.

JAMA Intern Med. doi:10.1001/jamainternmed.2014.2488Published online June 23, 2014.

Invited Commentary

Supplemental content atjamainternalmedicine.com

Author Affiliations: Authoraffiliations are listed at the end of thisarticle.

Corresponding Author: Karin M.Nelson, MD, MSHS, Seattle Center ofInnovation for Veteran-Centered andValue-Driven Care, VA Puget SoundHealth Care System, 1100 Olive Way,Ste 1400, Seattle, WA 98108 ([email protected]).

Research

Original Investigation

E1


Downloaded From: http://archinte.jamanetwork.com/ by a University of Iowa User on 06/24/2014


A lthough the patient-centered medical home (PCMH) hasbeen endorsed by most major primary care groups asa promising model to strengthen primary care, de-

crease costs, and improve quality,1 early assessment of PCMHimpact have yielded mixed results.2-10 Since 2010, the Veter-ans Health Administration (VHA) has undertaken nationaladoption of a PCMH model, called PACT (Patient Aligned CareTeam).11 The focus of PACT has been to restructure primarycare to provide team-based care that is more comprehensive,coordinated, and patient centered.11

The PACT initiative is a multifaceted and complex inter-vention, creating challenges to measuring implementationacross diverse clinic sites. One of the most widely recognizedPCMH recognition tools is the National Committee for QualityAssurance (NCQA) certification process, which focuses on prac-tice infrastructure and health information technology,12 an areain which the VHA has made considerable past investments.13,14

The VHA has a universally deployed electronic health record,electronic prescribing, patient registries, and a national qual-ity improvement and performance measurement infrastruc-ture for which all clinics in the VHA would receive “credit.” Manynational programs for coordinating care, such as home-basedprimary care, integrated mental health services, and pallia-tive care, were already widely available before PACT was ini-tiated. The focus within the VHA has been on how effectivelythese extensive resources are being applied and coordinatedto fulfill the goals of the PACT initiative.11

Our goal was to derive a comprehensive index from exist-ing data and survey instruments that would have a low re-spondent burden and would reflect processes and attributesthat are essential to effective primary care. Our approach dif-fers from other PCMH measurement tools15 by incorporatingmultiple data sources, including a primary care personnel sur-vey, patient surveys, and administrative data. We sought todevelop a measure to represent areas of focus of the PACT ini-tiative, including continuity through team-based care, patientaccess, care coordination, and patient-centered care.11 Wedesired an instrument that would facilitate comparisonsacross clinical sites within the VHA, assist in identifying sitesthat had most effectively implemented PACT, and determinethe relationship between effective implementation andimportant outcomes, such as patient satisfaction, quality ofcare, provider experience, and use of health care services.

MethodsSurvey Instruments and Data SourcesPatient SurveyWe used data from the previously validated ConsumerAssessment of Health Plans–Patient Centered Medical Home(CAHPS PCMH) survey16 that was administered to a nationallyweighted random sample of veterans who received outpatientcare from June 1 to December 31, 2012. The CAHPS PCMHscales have acceptable internal consistency reliability esti-mates for access (Cronbach α, 0.74), comprehensiveness(0.68), self-management support (0.62), patient-centered careand communication (0.91), and shared decision making

(0.61).16 To test convergent validity, we used information onpatient satisfaction from another sample of veterans from theSurvey of the Health Experiences of Patients, an ongoingnational mailed US Department of Veterans Affairs (VA) sur-vey that assesses the health care experiences of veterans whoreceive care at the VHA and uses a stratified random samplingmethod.17 The evaluation efforts are part of an going qualityimprovement effort at the VHA and are not consideredresearch activity; they are thus not subject to institutionalreview board review or waiver.

Primary Care Personnel SurveyThe PACT Primary Care Personnel Survey was an internallydeveloped instrument designed to measure team functioningin PACT and has been described elsewhere.18 The target popu-lation of the survey was all VHA primary care personnel,including the 4 occupations included in PACT teams: primarycare providers, nurse care managers, medical associates (eg,licensed practical nurses and medical technicians), andadministrative clerks. Data were collected from May 21through June 29, 2012. Team-based care was represented byitems from the primary care personnel survey related to del-egation, staffing, team functioning, and team assignment.18

Administrative DataInformation about demographics, clinical characteristics, anduse of health services was obtained from the VHA CorporateData Warehouse for fiscal year 2012 (n = 5 653 616). Using datafrom the Primary Care Management Module contained withinthe Corporate Data Warehouse, we identified all patients whowere enrolled in primary care and assigned to a primary careprovider.19 We included administrative data for important PACTprogrammatic goals,11 including (1) access to care and use ofnon–face-to-face care, such as telephone clinics and securemessaging; (2) continuity of care; and (3) use of VHA pro-grams to support care coordination (eg, home telemonitor-ing, 2-day posthospital follow-up).

We used data collected by the VHA External Peer ReviewProgram (EPRP) during fiscal year 2012 to assess quality of care.The EPRP is an audit program designed to assess clinical per-formance using standard performance criteria. National dataare collected through manual abstraction of electronic healthrecords by an independent external contractor.20 Previous stud-ies have found high interrater reliability (κ = 0.9) within theEPRP program.14

Construction of the PACT Implementation Progress IndexThe method for developing the PACT Implementation Prog-ress Index (Pi2) and a full description of all items are providedin eTable 1 in the Supplement. Briefly, we mapped data itemsto PACT conceptual domains, calculated domain scores basedon these items, and then generated site-level rankings for eachdomain. Table 1 outlines Pi2 domains and provides examplesof representative variable items. A Pi2 score was assigned toeach clinic based on the number of domains in the top and bot-tom quartiles for the domain scores, ranging from 8 (all do-main scores in the top quartile) to –8 (all domain scores in thebottom quartile). Using these scores, we categorized sites in

Research Original Investigation Patient-Centered Medical Home Implementation

E2 JAMA Internal Medicine Published online June 23, 2014 jamainternalmedicine.com




the top decile of the Pi2 (score, 5 to 8) as having achieved ef-fective implementation and those in the lowest decile of thePi2 (score, –7 to –5) as having been less effective.

Patient- and Provider-Level Outcome MeasuresPatient satisfaction was assessed by using a single item fromthe CAHPS PCMH survey16 as follows: “Using any numberfrom 0 to 10, where 0 is the worst provider possible and 10 isthe best provider possible, what number would you use torate this provider?”

Staff burnout was assessed with both a single-item mea-sure and the emotional exhaustion subscale of the MaslachBurnout Inventory, a widely used measure of burnout.21-23

The single item measure asks: “Overall, based on your defini-tion of burnout, how would you rate your level of burnout,”with 5 ordinal response options.24 We defined burnout as aresponse of 3 or higher, where 3 corresponds to “I am defi-nitely burning out and have one or more symptoms of burn-out, such as physical and emotional exhaustion.”24 We alsoused a 3-item version of the Maslach Burnout Inventorysubscale.25 Items reflecting burnout symptoms are scoredusing a Likert scale ranging from 0 (never) to 6 (every day) andsummed to form a scale score. We defined burnout as a scoreof 10 or higher (range, 0-18).

To assess quality of care, we examined outpatient mea-surements from the EPRP for chronic disease management, be-havioral health screening, and prevention services. These in-

dicators include frequently used measures of the quality ofprevention (eg, vaccinations, screening tests) and outpatientcare of chronic diseases (eg, annual retinal examinations in pa-tients with diabetes mellitus). The performance measure-ment for preventive services and chronic disease cohorts, sam-pling frame, and criterion for meeting the measurement areprovided in eTable 2 in the Supplement.

The EPRP selects a random sample of patient recordsfrom VHA facilities to monitor quality and appropriateness ofmedical care.26 The sample includes veterans who used VHAhealth care at least once in the 2 years before the assessment.Patients who were sampled had at least 1 primary care orspecialty medical visit in the month being sampled. Amongeligible patients, a random sample is drawn with over-sampling of prevalent chronic conditions (eg, diabetes, heartfailure).26

For patients at each primary care site, we determined thenumbers of emergency department or urgent care visits, VAhospital admissions, and hospitalizations for ambulatory care–sensitive conditions (ACSCs), which are postulated to be mostavoidable through provision of effective primary care.27 Hos-pitalizations for ACSCs were based on Agency for HealthcareResearch and Quality Prevention Quality Indicators and wereidentified through standardized protocols using Interna-tional Classification of Diseases, Ninth Revision, diagnoses andCurrent Procedural Terminology codes from inpatient VArecords.27

Table 1. PCMH Concepts and Pi2 Domains

PACT InitiativeGoals Pi2 Domains

TotalItems,

No.

Items From Each DataSource, No.

Example ItemdCAHPSPCMHa CDWb

PersonnelSurveyc

Accessible,continuous, andcoordinated care

Access 11 6 5 0 How often did you get an appointment as soon as you needed?a

When you phoned this provider’s office, how often did you get an answerto your medical question that same day?a

Same-day access to appointments (% of clinics)b

Enhanced access: telephone clinics (%)b

Continuity of care 3 1 2 0 How long have you been going to this provider?a

Proportion of visits to assigned primary care providerb

Care coordination 7 5 2 0 When this provider ordered a test, how often did someone from thisprovider’s office follow up to give you those results?a

Did the provider seem informed and up to date about the care you gotfrom specialists?a

Percentage of patients contacted 2 d after hospital dischargeb

Patient-centeredcare

Comprehensiveness 3 3 0 Did you and anyone in this provider’s office talk about things in your lifethat worry you or cause you stress?a

Self-managementsupport

2 2 0 Did anyone in this provider’s office talk with you about specific goals foryour health?a

Patient-centeredcare andcommunication

6 6 0 How often did this provider explain things in a way that was easy tounderstand?a

Shared decisionmaking

3 3 0 When you talked about starting a prescription medicine, did this providerask you what you thought was best for you?a

Team-based care Delegation, staffing,and teamfunctioning

18 0 0 18 Primary care provider relies on registered nurse care manager for tasksincluding gathering patient preventive services, responding toprescription refillsc

Percentage reporting recommended staffing ratioc

Time spent in team huddles: percentage spending >30 min/dc

Total … 53 26 9 18 …

Abbreviations: CAHPS PCMH, Consumer Assessment of Health Plans–PatientCentered Medical Home; CDW, Corporate Data Warehouse; PACT, PatientAligned Care Team; PCMH, patient-centered medical home; Pi2, PACTImplementation Progress Index.a CAHPS PCMH patient survey.

b CDW data from US Department of Veterans Affairs.c PACT Primary Care Personnel Survey.d Full description in eTable 1 in the Supplement. The questions have been

shortened for the sake of brevity.

Patient-Centered Medical Home Implementation Original Investigation Research

jamainternalmedicine.com JAMA Internal Medicine Published online June 23, 2014 E3




Statistical AnalysisPI2 PropertiesTo test internal consistency reliability, we calculated the Cron-bach α for all items in each domain and all 53 items that makeup the total scale.

Variation in PCMH AdoptionWe evaluated bivariate comparisons of facility characteristicsand level of implementation by using χ2 tests for categoricalvariables and t tests for continuous variables. We comparedsites assessed to have effectively implemented PACT withthose assessed as less effective according to type of facility(hospital or community-based outpatient clinic), number ofpatients, demographic characteristics, and Elixhauser comor-bidity score.28

Associations With Patient and Provider OutcomesWe used a nonparametric test of trend for the ranks across or-dered groups (an extension of the Wilcoxon rank sum test) totest for trends in patient satisfaction and staff burnout by Pi2

scores. We tested differences in the proportions of eligible pa-tients at each VHA clinic fulfilling each of the 48 quality indi-cators according to the success of PACT implementation asmeasured by the Pi2. We calculated rates of services at the fa-cility level by dividing the number of patients who satisfiedthe EPRP quality measure by the number who met inclusioncriteria for each quality measure (eTable 3 in the Supple-ment). For each of the 48 facility-level quality indicators, wetested the trend in proportions of patients fulfilling the EPRPquality guideline by the level of PACT implementation. We usedthe nonparametric test for trend developed by Cuzick, whichis an extension of the Wilcoxon test.29 We adjusted for mul-tiple comparisons using a method described by Benjamini andYekutieli.30 To determine whether more effective implemen-tation (as measured by Pi2) corresponded to higher perfor-mance overall, we included all 48 outcome measures in a lin-ear mixed-effects model that accounted for correlation amongoutcomes from the same facility and estimated an overallimplementation effect. We adjusted for implementation in thismodel as a linear term ranging from 1 to 5, corresponding tothe grouped Pi2 scores. This approach was possible because all48 outcomes were measured on the same scale.

We examined fiscal year 2012 emergency department andurgent care visits and total hospitalizations for sites with moreeffective vs less effective implementation, adjusting for pa-tient age, community-based outreach clinics, and Elixhausercomorbidity scores.28 To account for temporal trends, we mod-eled facility-level trends for hospitalization from 2003 to 2012.The method for examining such trends has been describedelsewhere.31 We estimated interrupted time-series models ofACSC and all-cause hospitalizations from October 1, 2003,through September 20, 2012, for each facility and assessed howthe trends in hospitalizations changed after the start of thePACT initiative in April 2010. All regression models adjustedfor facility-level patient characteristics, unemployment rate inthe VA market area, quarterly dummy variables to capture sea-sonal variation, and a linear time trend. Patient risk was mea-sured using mean facility-level Elixhauser comorbidity scores.

Changes in admissions for ACSC and all-cause hospitaliza-tions after implementation of the PACT initiative were calcu-lated as the difference between the observed rate of admis-sions and the predicted rate had the initiative not beenimplemented during the 2½-year period between April 1, 2010,and September 30, 2012. In this way, we estimated changes inadmissions that might be attributed to the PACT initiative.Trend analyses for hospitalizations were stratified by age (≥65and <65 years) to account for the substantial use of non-VAhealth care by Medicare-eligible veterans.32 We then com-pared the estimated change in admissions among facilities thathad effectively implemented PACT with the change in thosethat had done so less effectively.

ResultsThe final Pi2 consisted of 53 individual items assigned to the 8overarching PACT concepts (Table 1). Detailed descriptions ofall items and descriptive statistics are provided in eTable 1 inthe Supplement. From more than 22 000 primary care per-sonnel at the time of the survey, 5404 (approximately 25% re-sponse rate) from 667 sites of care completed the PACT Pri-mary Care Personnel Survey during the spring of 2012. BetweenJune and December 2012, more than 75 000 veterans who wereenrolled in VA primary care completed the CAHPS PCMH mod-ule included in the Survey of the Health Experiences of Pa-tients (47% response rate).

Psychometric Properties of Pi2

The Pi2 demonstrated satisfactory levels of internal consis-tency for total score (Cronbach α= 0.89), access (0.63), conti-nuity (0.67), comprehensiveness (0.81), self-managementsupport (0.68), patient-centered care and communication(0.95), shared decision making (0.75), and team-based care(0.91). Similar to results reported by Scholle et al,16 the carecoordination composite had lower internal consistency(0.51). With patient satisfaction as measured by the Survey ofthe Health Experiences of Patients used as a measure of con-vergent validity, provider ratings differed between sites withmore vs less effective implementation (mean provider rating,9.05 vs 8.37; P < .001).

Variation in PCMH AdoptionClinical sites that had implemented PACT more effectivelytended to have fewer patients than those that had been lesseffective (Table 2). However, the type of clinic, mean patientage, percentage of male patients, and mean Elixhauser comor-bidity score were similar among all sites irrespective of howwell PACT had been implemented.

Associations With Patient and Provider OutcomesPatient satisfaction was significantly higher among sites thathad effectively implemented PACT than among those thathad not (range of mean rating for satisfaction with provider,9.33-7.53; P < .001) (Table 3 and the eFigure in the Supple-ment). A similarly favorable pattern was observed for staffburnout as measured by the Maslach Burnout Inventory






emotional exhaustion subscale (range, 2.29-2.80; P = .02)(Table 3 and the eFigure in the Supplement) but not for the1-item burnout question.

We observed significant trends in quality of care in rela-tion to the Pi2 score. The 77 sites that achieved the most ef-fective implementation exhibited higher clinical quality out-come measures than less successful sites. There was astatistically significant association (P < .05) between clinicalquality outcomes and the Pi2 score for 19 of 48 measures, andbetter performance was associated with a higher Pi2 score forall but 2. Overall, of all 48 measures, 41 were higher among siteswith higher Pi2 scores. The Figure displays the difference inthe percentage of patients meeting quality criteria between sitesin the highest and lowest quintiles of Pi2 scores. The com-bined effect estimated from the mixed-effects model showeda significant increase in mean outcomes for facilities withhigher Pi2 scores compared with those with lower Pi2 scores(P < .001).

In the comparison of trends across Pi2 scores, veterans withchronic disease who received care at facilities with higher Pi2

scores had small but significant improvements in quality-of-care indicators (eTable 3 in the Supplement). For example, vet-erans with diabetes who received care at sites with the high-est Pi2 scores were more likely to have a hemoglobin A1c valueless than 9% (range at high-implementation vs low-implementation sites, 84.0%-81.8%; P = .04) or a low-density lipoprotein cholesterol (LDL-C) level less than 100mg/dL (to convert to millimoles per liter, multiply by 0.0259)

(range, 70.4%-66.0%; P = .03). Veterans with hypertension weremore likely to have a blood pressure reading less than 140/90mm Hg (range, 80.2%-76.9%; P = .02). Among veterans withischemic heart disease, those at sites with the highest Pi2 scoreswere more likely to have LDL-C measured (range, 97.0%-94%; P < .001), have a measured LDL-C level less than 100mg/dL (range, 70.5%-65.3%; P < .001), and have documenta-tion of a prescription for aspirin at their most recent visit (range,92.9%-89.5%; P = .03). Veterans receiving care at sites that ex-hibited more successful implementation of PACT were morelikely to receive an influenza vaccination (range, 68.5%-64.2%; P < .001 for veterans aged 50-64 years), to be screenedfor cervical cancer (range, 92.8%-86.7%; P = .047 for womenaged 21-64 years), or offered medications for tobacco cessa-tion (range, 96.2%-93.4%; P < .001).

The rate of emergency department visits was signifi-cantly lower in sites with more effective implementation thanin those with less effective implementation (range, 188-245 vis-its per 1000 patients; P < .001; Table 4 and eFigure in theSupplement). Although the total numbers of hospitalizationsin fiscal year 2012 did not differ by level of implementation(Table 4), rates of hospitalization for ACSCs during the 2½-year period after implementation of PACT were lower amongsites that had more effectively implemented PACT than amongthose that were less effective (Table 5). Among sites with Pi2

scores in the highest decile, we estimated that there was a meanreduction of 2.28 admissions for ACSCs per 1000 patientsyounger than 65 years (a 13.4% decrease) attributable to the

Table 2. Site Characteristics by Effectiveness of PACT Implementation

CharacteristicClinic-Level National

Mean (SD)a

Implementation, Mean (95% CI)b

P ValuecMore Effective (Pi2, 5 to 8) Less Effective (Pi2, –7 to −5)Patients served, No. 5917 (6706) 2893 (2250-3535) 5467 (3910-7024) .004

CBOCs, No. (%) 735 (83.0) 72 (94.7) 73 (89.0) .19

Age, y 64.4 (3.9) 65.1 (64.2-66.0) 63.6 (62.7-64.4) .01

Female patients, % 5.4 (0.04) 6.2 (4.0-9.0) 5.9 (5.0-7.0) .77

Elixhauser score 0.76 (0.07) 0.77 (0.75-0.78) 0.76 (0.75-0.77) .43

Abbreviations: CBOCs, community-based outpatient clinics; PACT, Patient Aligned Care Team; Pi2, PACT Implementation Progress Index.a Values represent mean (SD) except where otherwise indicated.b Values represent mean (95% CI) except where otherwise indicated.c P values determined with χ2 or t tests comparing more effective vs less effective implementation sites.

Table 3. Patient Satisfaction, Staff Burnout, and Pi2 Scores

Pi2 ScoreClinics,

No.

Patient Satisfaction, Mean (SD) Staff Burnout, Mean (SD)Provider Rating FromCAHPS PCMH Survey

Provider RatingFrom SHEP

Overall Health CareRating From SHEP MBI EE

Single-ItemMeasure

5 to 8 77 9.33 (0.34) 9.05 (0.28) 8.62 (0.34) 2.29 (1.55) 0.37 (0.36)

2 to 4 213 9.02 (0.44) 8.91 (0.34) 8.49 (0.36) 2.47 (1.29) 0.36 (0.33)

−1 to 1 346 8.67 (0.46) 8.73 (0.34) 8.32 (0.35) 2.56 (1.13) 0.36 (0.29)

−4 to −2 190 8.23 (0.64) 8.55 (0.40) 8.15 (0.40) 2.63 (1.23) 0.37 (0.29)

−7 to −5 87 7.53 (0.88) 8.37 (0.45) 7.87 (0.47) 2.80 (1.42) 0.37 (0.35)

P valuea … <.001 <.001 <.001 .02 .58

Abbreviations: CAHPS PCMH, Consumer Assessment of Health Plans–Patient Centered Medical Home; MBI EE, Maslach Burnout Inventory emotional exhaustionsubscale; Pi2, PACT [Patient Aligned Care Team] Implementation Progress Index; SHEP, Survey of the Health Experiences of Patients.a P values represent test for trend.






Figure. Differences Between Sites With More Effective and Less Effective Implementation in the Percentage of Patients Meeting Quality Criteria

–6.00 0 8.00–2.00 2.00 6.004.00% of Patients

–4.00

SourceCongestive heart failure

LVF documentedLVEF <40% on ACEI or ARB

Diabetes mellitusRenal testing

HypertensionDiagnosis of HTN and BP <140/90 mm Hg

Mental health screening and treatmentAlcohol misuse with timely counseling

Obesity managementEligible patients who participated in MOVE

ImmunizationsImmunization age ≥65 years

Cancer and osteoporosis screeningColorectal cancer screening for those aged 51-75 years

Tobacco use screening and treatmentPatients using tobacco offered referralPatients using tobacco provided with counselingPatients using tobacco offered medicationsScreened for tobacco use

Cervical cancer screening (women aged 21-64 years)Screened for breast cancer (women aged 40-49 years)

Screened for breast cancer (women aged 60-69 years)Osteoporosis screening for women aged ≥65 years

Screened for breast cancer (women aged 50-59 years)

Influenza ages 50-64 yearsPneumococcal

Obese patients screened and offered weight management (MOVE)

Annually screened for alcohol misuse, SUDPatients positive for PTSD with timely dispositionPositive PTSD screen with timely suicide ideation/behavior evaluationPTSD screening using the PC-PTSDCombined scores for timely suicide evaluation if positive for PTSD or MDDTimely evaluation for patients with positive depression screenAnnually screened for depression

Ischemic heart diseaseAMI outpatient, ASA at most recent visitAMI, LVEF <40 on ACEI or ARBAMI discharged who received persistent β–blockerHyperlipidemia screen–overallLDL-C <100 mg/dLLDL-C measured

No diagnosis of HTN and BP <160/100 mm Hg

Diagnosis of HTN and BP <160/100 mm HgNo diagnosis of HTN and BP ≤140/90 mm Hg

LDL-C measured within 1 y of reviewRetinal examination, timely by diseaseBP <160/100 mm HgBP <140/90 mm HgLDL-C <100 mg/dLHbA1c <9%ACEI/ARB in current medicationsASA in current medicationsHbA1c annualFoot sensory examination using monofilamentFoot pedal pulsesFoot inspection

ACEI, angiotensin-converting enzyme inhibitor; AMI, acute myocardialinfarction; ARB, argiotensin II receptor blocker; ASA, aspirin; BP, blood pressure;HbA1c, hemoglobin A1c; HTN, hypertension; LDL-C, low-density lipoproteincholesterol; LVEF, left ventricular ejection fraction; LVF, left ventricular fraction;MDD, major depressive disorder; MOVE, weight management program;

PC-PTSD, primary care posttraumatic stress disorder; PTSD, posttraumaticstress disorder; SUD, substance use disorder.SI conversion factor: To convert LDL-C values to millimoles per liter, multiply by0.0259.






PACT initiative compared with a reduction in admissions ofonly 0.08 for ACSCs (a 3.0% decrease) among sites with lesseffective implementation. Thus, the estimated reduction inhospitalizations for ACSCs was significantly greater at the moreeffective sites, although the absolute number of admissionswas relatively small (2.8 hospitalizations per 1000 patients).For veterans older than 65 years, the projected changes in ad-mission rates for ACSCs were more modest, and the differ-ence among sites was small. Similar trends of smaller magni-tude were noted for all-cause hospitalization.

DiscussionWe constructed the Pi2, a measure to assess progress in imple-menting PCMH in the VHA, using primary care personnel sur-veys, patient surveys, and administrative data. The index wasfavorably and meaningfully associated with important out-comes, including patient satisfaction, staff burnout, quality ofcare, hospitalizations, and emergency department visits. Pa-tient satisfaction was significantly greater (on the order of afull point higher on a 0-10 scale) and staff burnout lower at sitesfor which the Pi2 indicated more effective implementation. De-spite the overall high level of clinical care provided at the VHA,we found measurable differences between clinics by level ofPCMH implementation in terms of the proportion of veteransmeeting criteria for multiple measures of quality. In addition,sites with the highest Pi2 scores exhibited modestly lower ratesof hospital admission for ACSCs and larger projected de-creases in rates of admission after the start of the VHA PACTinitiative. These results are consistent with findings from a re-cent VHA study in which clinic directors’ reports about the

medical home indicated that sites with better care coordina-tion and support for transition had lower rates of hospitaliza-tions for ACSCs.33

Previous studies have found a lack of association withmeasurements of structural and care processes in primarycare and quality of care34,35 or patient experience.36 Thus, indevising a method to assess the degree of PCMH implemen-tation, we adopted an approach that differed substantiallyfrom the widely used NCQA recognition process, includingpatient-reported measures and using administrative data onaccess and continuity. We chose not to use the NCQA recogni-tion process for several reasons: the administrative burden istoo high to determine certification for more than 900 clinicsand much of the process relies on structural changes thathave already been broadly implemented across the VA, there-fore diminishing the ability to discriminate among sites. Pre-vious authors have noted that the NCQA measurement forPCMH may not be able to differentiate on quality-of-caremeasures.10 In contrast, our measure of PCMH implementa-tion detects differences in quality across many clinics. Ourdata support the notion that the assessment of PCMH needsto include both patient-level and practice-level infrastructuremeasures.5,6,37

The VHA has a long-standing investment in infrastruc-ture considered a baseline prerequisite for a functioningPCMH, including a robust quality improvement and perfor-mance system.14 Our data are consistent with previousreports of the high quality of clinical care provided at theVHA.14,38 Paradoxically, this high baseline of quality makes itmore difficult to demonstrate improvements than wouldprobably be the case in other health systems where imple-mentation of PCMH has been evaluated and the baseline

Table 4. Emergency Department Visits, Hospitalizations, and Pi2 Scores for Fiscal Year 2012

Pi2 Score Clinics, No.Emergency Department Visits

per 1000 Patients, No.aHospitalizations per 1000

Patients, No.a

5 to 8 77 188 68

2 to 4 213 227 77

−1 to 1 346 286 87

−4 to −2 190 289 83

−7 to −5 87 245 74

P valueb … <.001 .99

Abbreviation: Pi2, PACT [PatientAligned Care Team] ImplementationProgress Index.a Adjusted for age, Elixhauser

comorbidity score, andcommunity-based outpatientclinics.

b P values represent test for trend.

Table 5. Effects of PACT Initiative Among Sites With More Effective or Less Effective Implementation for ACSC and Total Hospitalizations

Pi2 Scorea

Quarterly Hospital Admissions for ACSCs per 1000 Patients, No. Quarterly Total Hospital Admissions per 1000 Patients, No.Predicted WithPACT Initiative

Predicted WithoutPACT Initiative

Absolute Difference,No. (%)b

Predicted WithPACT Initiative

Predicted WithoutPACT Initiative

Absolute Difference,No. (%)

Patients <65 y

5 to 8 2.28 2.63 −0.35 (−13.4) 26.12 27.43 −1.31 (−4.8)

−7 to −5 2.53 2.61 −0.08 (−3.0) 25.20 24.33 0.87 (3.6)

Patients >65 y

5 to 8 3.68 3.85 −0.17 (−4.6) 21.75 22.21 −0.46 (−2.1)

−7 to −5 4.42 4.33 0.09 (2.1) 23.47 22.37 1.10 (4.9)

Abbreviations: ACSC, ambulatory care sensitive condition; PACT, Patient Aligned Care Team; Pi2, PACT Implementation Progress Index.a Pi2 scores of 5 to 8 indicate more effective implementation; Pi2 scores of −7 to −5, less effective implementation.b All differences were significant at P < .001.






quality of care was substantially lower based on comparablemeasures.2 The favorable findings from our analysis maypartly reflect the extensive improvements in clinical andresource infrastructure that the VHA has made since the1990s, such as the use of electronic prescribing and a univer-sal electronic health record and deployment of pharmacists,dieticians, social workers, and mental health professionals inmany primary care clinics.

These analyses have several limitations. First, several ofthe domain scores rely on self-report, which are subject tobiases, including response bias, framing bias, and others.However, we used well-validated measures, augmentedthese data with important team-based care domains from ourprimary care personnel survey, and used patient report tocapture comprehensiveness of care that other measures maynot include.15,39 Second, our primary care personnel surveyhad a low response rate, but our results were consistent withthose from other surveys of primary care providers40,41 andwith those from a survey of all VA employees. Third, the

cross-sectional design of the study did not permit assessingchange over time, although this is planned in future studies.

ConclusionsWe found that the Pi2 score was favorably associated with pa-tient satisfaction, staff burnout, quality of care, and use of healthcare services. Our results may apply only to large integratedhealth systems that include a robust and integrated electronichealth record and a well-developed quality improvement sys-tem that provides feedback to clinics and providers. All pri-mary care providers in the VHA, for example, have ready ac-cess to detailed information about their patient panels, includingthe likelihood of admission or death (updated weekly), as wellas patients’ use of a range of inpatient, outpatient, and care co-ordination services.42 However, as accountable care organiza-tions evolve, this type of patient-centered measurement couldbe adopted by other large integrated health systems.

ARTICLE INFORMATION

Accepted for Publication: April 25, 2014.

Published Online: June 23, 2014.doi:10.1001/jamainternmed.2014.2488.

Author Affiliations: Seattle Center of Innovationfor Veteran-Centered and Value-Driven Care, VAPuget Sound Health Care System, Seattle,Washington (Nelson, Helfrich, Sun, Hebert, Liu,Dolan, Taylor, Maynard, Hernandez, Randall);General Internal Medicine Service, VA Puget SoundHealth Care System, Seattle, Washington (Nelson,Fihn); Department of Medicine, University ofWashington School of Medicine, Seattle (Nelson,Fihn); Department of Health Services, University ofWashington School of Public Health, Seattle(Helfrich, Sun, Hebert, Liu, Wong, Maynard,Hernandez, Randall); Office of Analytics andBusiness Intelligence, Veterans HealthAdministration, Washington, DC (Sanders, Curtis,Fihn); Office of Patient Care Services, USDepartment of Veterans Affairs, Washington, DC(Schectman); Office of Clinical Operations, USDepartment of Veterans Affairs, Washington, DC(Stark).

Author Contributions: Dr Nelson had full access toall the data in the study and takes responsibility forthe integrity of the data and the accuracy of thedata analysis.Study concept and design: Nelson, Helfrich, Hebert,Dolan, Wong, Hernandez, Schectman, Stark, Fihn.Acquisition, analysis, or interpretation of data: Allauthors.Drafting of the manuscript: Nelson, Helfrich,Sanders, Fihn.Critical revision of the manuscript for importantintellectual content: Nelson, Helfrich, Sun, Hebert,Liu, Dolan, Taylor, Wong, Maynard, Hernandez,Sanders, Randall, Curtis, Schectman, Stark, Fihn.Statistical analysis: Nelson, Sun, Hebert, Liu, Dolan,Taylor, Wong, Maynard, Sanders, Randall.Obtained funding: Fihn.Administrative, technical, or material support:Nelson, Helfrich, Dolan, Hernandez, Sanders,Randall, Curtis, Schectman, Stark, Fihn.Study supervision: Nelson, Schectman, Fihn.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported by theVHA Office of Patient Care Service.

Role of the Sponsors: This study was conducted aspart of VHA Health Care Operations in accordancewith VHA Handbooks 1605.1 and 1605.2. The studyteam had full responsibility for the design andconduct of the study; collection, management,analysis, and interpretation of the data; andpreparation of the manuscript. The study wasreviewed through normal administrative channels.

Additional Contributions: Data for this reportwere developed by the national evaluation team atthe PACT Demonstration Lab Coordinating Centerand the VHA Office of Analytics and BusinessIntelligence. The VHA Office of Primary CareOperations is responsible for PACT implementation,and the VHA Office of Patient Care Services isresponsible for the PACT Demonstration Labprogram. John Messina, BA, was paid to assist withdata acquisition and administrative support.

Previous Presentation: This study was presentedat the Society for General Internal Medicine AnnualMeeting; April 25, 2014; San Diego, California. Thisstudy was also presented at the Academy HealthNational Meeting; June 10, 2014; San Diego,California.

REFERENCES

1. Stange KC, Nutting PA, Miller WL, et al. Definingand measuring the patient-centered medical home.J Gen Intern Med. 2010;25(6):601-612.

2. Werner RM, Duggan M, Duey K, Zhu J, Stuart EA.The patient-centered medical home: an evaluationof a single private payer demonstration in NewJersey. Med Care. 2013;51(6):487-493.

3. Nielsen M, Langner B, Zema C, Hacker T, GrundyP. Benefits of implementing the primary carepatient-centered medical home: a review of cost &quality results, 2012. http://www.pcpcc.org/guide/benefits-implementing-primary-care-medical-home.Patient-Centered Primary Care Collaborative,September 2012. Accessed May 7, 2014.

4. Liss DT, Fishman PA, Rutter CM, et al. Outcomesamong chronically ill adults in a medical homeprototype. Am J Manag Care. 2013;19(10):e348-e358.

5. Day J, Scammon DL, Kim J, et al. Quality,satisfaction, and financial efficiency associated withelements of primary care practice transformation:preliminary findings. Ann Fam Med. 2013;11(suppl 1):S50-S59.

6. Jaén CR, Ferrer RL, Miller WL, et al. Patientoutcomes at 26 months in the patient-centeredmedical home National Demonstration Project. AnnFam Med. 2010;8(suppl 1):S57-S67, S92.

7. Reid RJ, Fishman PA, Yu O, et al.Patient-centered medical home demonstration:a prospective, quasi-experimental, before and afterevaluation. Am J Manag Care. 2009;15(9):e71-e87.

8. Jackson GL, Powers BJ, Chatterjee R, et al. Thepatient-centered medical home: a systematicreview. Ann Intern Med. 2013;158(3):169-178.

9. Solberg LI, Asche SE, Fontaine P, Flottemesch TJ,Anderson LH. Trends in quality during medical hometransformation. Ann Fam Med. 2011;9(6):515-521.

10. Solberg LI, Asche SE, Fontaine P, FlottemeschTJ, Pawlson LG, Scholle SH. Relationship of clinicmedical home scores to quality and patientexperience. J Ambul Care Manage. 2011;34(1):57-66.

11. Rosland AM, Nelson K, Sun H, et al. Thepatient-centered medical home in the VeteransHealth Administration. Am J Manag Care. 2013;19(7):e263-e272.

12. Berenson RA, Hammons T, Gans DN, et al.A house is not a home: keeping patients at thecenter of practice redesign. Health Aff (Millwood).2008;27(5):1219-1230.

13. Kizer KW. The “new VA”: a national laboratoryfor health care quality management. Am J MedQuality. 1999;14(1):3-20.

14. Jha AK, Perlin JB, Kizer KW, Dudley RA. Effectof the transformation of the Veterans Affairs HealthCare System on the quality of care. N Engl J Med.2003;348(22):2218-2227.

15. Burton RA, Devers KJ, Berenson RA.Patient-centered medical home recognition tools:a comparison of ten surveys’ content andoperational details. http://www.urban.org/publications/412338.html. Washington, DC: UrbanInstitute; 2012. Accessed February 1, 2014.






16. Scholle SH, Vuong O, Ding L, et al.Development of and field test results for the CAHPSPCMH Survey. Med Care. 2012;50(suppl):S2-S10.

17. Wright SM, Craig T, Campbell S, Schaefer J,Humble C. Patient satisfaction of female and maleusers of Veterans Health Administration services.J Gen Intern Med. 2006;21(3)(suppl 3):S26-S32.

18. Helfrich CD, Dolan ED, Simonetti J, et al.Elements of team-based care in a patient-centeredmedical home are associated with lower burnoutamong VA primary care employees. J Gen Intern Med.2014. doi:10.1007/s11606-013-2702-z.

19. Shen Y, Hendricks A, Zhang S, Kazis LE. VHAenrollees’ health care coverage and use of care. MedCare Res Rev. 2003;60(2):253-267.

20. Perlin JB, Kolodner RM, Roswell RH. TheVeterans Health Administration: quality, value,accountability, and information as transformingstrategies for patient-centered care. Am J ManagCare. 2004;10(11, pt 2):828-836.

21. Schaufeli WB, Enzmann D, Girault N.Measurement of burnout: a review. In: SchaufeliWB, ed. Professional Burnout: Recent Developmentsin Theory and Research. Philadelphia, PA: Taylor &Francis; 1993:199-215.

22. Wheeler DL, Vassar M, Worley JA, Barnes LL.A reliability generalization meta-analysis ofcoefficient alpha for the Maslach BurnoutInventory. Educ Psychol Meas. 2011;71(1):231-244.doi:10.1177/0013164410391579.

23. Maslach C, Jackson SE. The measurement ofexperienced burnout. J Organiz Behav. 1981;2(2)99-113. doi:10.1002/job.4030020205.

24. Rohland BM, Kruse GR, Rohrer JE. Validation ofa single-item measure of burnout against theMaslach Burnout Inventory among physicians.Stress Health. 2004;20(2):75-79. doi:10.1002/smi.1002.

25. Leiter MP, Shaughnessy K. The areas of worklifemodel of burnout: tests of mediation relationships.Ergonomia. 2006;28:327-341.

26. Goulet JL, Erdos J, Kancir S, et al. Measuringperformance directly using the Veterans HealthAdministration electronic medical record:a comparison with external peer review. Med Care.2007;45(1):73-79.

27. Agency for Healthcare Research and Quality.Prevention quality indicators. http://www.qualityindicators.ahrq.gov/modules/pqi_resources.aspx. Accessed February 28, 2014.

28. Elixhauser A, Steiner C, Harris DR, Coffey RM.Comorbidity measures for use with administrativedata. Med Care. 1998;36(1):8-27.

29. Cuzick J. A Wilcoxon-type test for trend. StatMed. 1985;4(1):87-90.

30. Benjamini Y, Yekutieli D. The control of thefalse discovery rate in multiple testing underdependency. Ann Stat. 2001;29:1165-1188.

31. Hebert PLLC, Wong ES, Hernandez SE, et al.The economic effects and return on investment ofthe Veterans Health Administration’s PatientCentered Home Initiative, 2010 through 2012.Health Aff. In press.

32. Liu CF, Chapko M, Bryson CL, et al. Use ofoutpatient care in Veterans Health Administrationand Medicare among veterans receiving primarycare in community-based and hospital outpatientclinics. Health Serv Res. 2010;45(5, pt 1):1268-1286.

33. Yoon J, Rose DE, Canelo I, et al. Medical homefeatures of VHA primary care clinics and avoidablehospitalizations. J Gen Intern Med. 2013;28(9):1188-1194.

34. Holmboe ES, Arnold GK, Weng W, Lipner R.Current yardsticks may be inadequate for

measuring quality improvements from the medicalhome. Health Aff (Millwood). 2010;29(5):859-866.

35. Friedberg MW, Safran DG, Coltin KL, Dresser M,Schneider EC. Readiness for the patient-centeredmedical home: structural capabilities ofMassachusetts primary care practices. J Gen InternMed. 2009;24(2):162-169.

36. Martsolf GR, Alexander JA, Shi Y, et al. Thepatient-centered medical home and patientexperience. Health Serv Res. 2012;47(6):2273-2295.

37. Gray BM, Weng W, Holmboe ES. An assessmentof patient-based and practice infrastructure-basedmeasures of the patient-centered medical home:do we need to ask the patient? Health Serv Res.2012;47(1, pt 1):4-21.

38. Kerr EA, Gerzoff RB, Krein SL, et al. Diabetescare quality in the Veterans Affairs Health CareSystem and commercial managed care: the TRIADstudy. Ann Intern Med. 2004;141(4):272-281.

39. Birnberg JM, Drum ML, Huang ES, et al.Development of a safety net medical home scale forclinics. J Gen Intern Med. 2011;26(12):1418-1425.

40. Shanafelt TD, Boone S, Tan L, et al. Burnoutand satisfaction with work-life balance among USphysicians relative to the general US population.Arch Intern Med. 2012;172(18):1377-1385.

41. Lewis SE, Nocon RS, Tang H, et al.Patient-centered medical home characteristics andstaff morale in safety net clinics. Arch Intern Med.2012;172(1):23-31.

42. Wang L, Porter B, Maynard C, et al. Predictingrisk of hospitalization or death among patientsreceiving primary care in the Veterans HealthAdministration. Med Care. 2013;51(4):368-373.





Va pcmh study 6 2014[1]

Healthcare

Transcript of Va pcmh study 6 2014[1]