HSR&D Citation Abstract
Search | Search by Center | Search by Source | Keywords in Title
Implementation of the Veterans Health Administration Patient Centered Medical Home Initiative and Primary Care Use: Differences by Clinic Comorbidity Burden
Wong ES, Rosland AM, Fihn SD, Nelson KM. Implementation of the Veterans Health Administration Patient Centered Medical Home Initiative and Primary Care Use: Differences by Clinic Comorbidity Burden. Poster session presented at: AcademyHealth Annual Research Meeting; 2015 Jun 15; Minneapolis, MN.
1Center of Innovation for Veteran-Centered and Value-Driven Care, VHA Puget Sound Health Care System, Seattle, WA; 2Department of Health Services, University of Washington, Seattle, WA; 3Center for Clinical Management Research, VHA Ann Arbor Healthcare System, Ann Arbor, MI; 4Department of Internal Medicine, University of Michigan, Ann Arbor, MI; 5Office of Analytics and Business Intelligence, Veterans Health Administration, Seattle, WA; 6Department of Medicine, University of Washington, Seattle, WA;
Objective: In April 2010, the Veterans Health Administration (VHA) began to establish patient-centered medical homes at all primary care (PC) clinics nationwide as part of the Patient Aligned Care Teams (PACT) initiative. Earlier research indicated the adoption of PACT was associated with an increase in PC use among VHA users age 65. We sought to determine whether clinics with a higher proportion of complex patients had larger increases in PC visits attributable to PACT implementation relative to clinics with a lower proportion of such patients.
Study Design: Using encounter level data from the VHA Corporate Data Warehouse, we applied an interrupted time-series design for which VHA PC use was defined as the total number of PC visits delivered to all veterans assigned to a clinic in a given quarter. Clinic-level comorbidity burden was defined as the proportion of clinics' patients classified as having high comorbidity (Gagne score 2). We modeled time trends in quarterly PC use using mixed effects negative binomial models, adjusting for demographics of clinics' patient population, time-invariant facility characteristics and local economic conditions. Using these trends to project estimated PC use subsequent to April 2010, we estimated the expected difference in quarterly PC visits with (i.e., observed) and without PACT (i.e., projected) for each clinic and assessed the differences by quartiles of clinic-level comorbidity. The unit of analysis for all models was clinic-quarter.
Population Studied: 5.3 million VHA enrollees age 65+ receiving primary care from 972 VA clinics at any time between October 2003 and September 2013.
Principal Findings: The number of age 65 VA primary care users increased from 2.1 million in fiscal year 2004, quarter 1 (FY2004Q1) to 2.9 million in FY2013Q4. On average, 18.2% (interquartile range = 14.5% to 21.2%) of clinics' patients were considered high-comorbidity. At the time of PACT implementation, mean visits per 1,000 veterans per quarter were 759 and 899 for clinics with the lowest and highest proportion of high-comorbidity patients, respectively. PC use increased from 815 visits per 1,000 patients per quarter prior to implementation of PACT to 845 visits per 1,000 patients per quarter after implementation of PACT. Overall, PACT was associated with a significant shift in the long-run trend for clinic-level PC use (49 visit increase per 1,000 veterans, p < 0.001). This positive association was strongest among clinics in the lowest (192 visits per 1,000 veterans per quarter, p < 0.001) and 2nd lowest quartiles (53 visits per 1,000 veterans per quarter, p < 0.001) of clinic-level comorbidity.
Conclusions: Implementation of PACT within VA was associated with modest (~5%) increase in visits among age 65+ veterans that was more concentrated at clinics with relatively fewer high-comorbidity patients.