Gray BM, Vandergrift JL. — American Board of Internal Medicine
Presented: AcademyHealth Annual Research Meeting, June 2016
Objective: Patient-centered health care is central to ongoing healthcare reform initiatives. Providers across the nation are attempting to transform organizations with the expectation that patient-centered care will translate into higher quality and more efficient healthcare. Organizations, such as the NCQA, have facilitated this transformation by providing a framework for patient-centered reorganization.
In our study, we examined whether structural characteristics that are the basis for NCQA’s PCMH designation, as measured by their practice infrastructure survey score (PCMH-PSC), were associated care costs and quality.
Population Studied: Medicare beneficiaries treated by mid-career general internists.
Study Design: We used data from 2,293 mid-career (10 years past certification) office-based general internists who completed a Practice Improvement Module (PIM) between 2009 and 2012 as part of the American Board of Internal Medicine's Maintenance of Certification requirement. PIMs include a practice infrastructure survey mirroring NCQA’s PCMH qualification survey (PPC-Readiness Survey v2). We merged these data with Medicare claims, attributing a beneficiary to an internist if that internist accounted for a plurality of outpatient visits over the year the physician completed the PMHC-PSC survey and the prior two years. We attributed 322,134 beneficiaries to 1,917 physicians.
We used healthcare costs (inpatient and outpatient) and five HEDIS process quality measures as dependent variables: biennial mammography (69,967 woman, 1,807 physicians); two annual A1c tests, annual LDL testing, and biennial eye examinations for diabetics (38,533 beneficiaries, 1,746 physicians); and annual LDL testing for beneficiaries with cardiovascular disease (47,800 beneficiaries, 1,780 physicians).
To estimate associations, we applied GEE regression models that accounted for correlated errors associated with physicians seeing multiple patients (logistic for HEDIS measures and log-gamma for cost measures). As controls, we included patient demographics, 27 chronic condition indicators, CMS's HCC risk adjuster, 10 regional indicators, year indicators, a Medicaid eligibility indicator, and indicators describing physician characteristics.
Principal Findings: The overall PCMH-PSC was positively associated with meeting all five HEDIS process quality measures (Ps<.01) with a 25-point PCMH-PSC increase (25% of the maximum score) associated with between a 3.5 percentage point (diabetic eye examinations) to a 1.4 percentage point (LDL testing for patients with cardiovascular disease) regression adjusted increase in process quality measures. Higher infrastructure scores were significantly associated with reductions in outpatient costs (P=0.002) with a 25-point increased score associated with a regression adjusted $98 per year beneficiary cost reduction or $16,485 per year savings for each physician in our sample. Three PCMH-PSC sub-scores (institutional processes for quality improvement, using patient data to follow patients with important conditions, and systematic processes such as E-prescribing and test/referral tracking) were significantly associated with most process measures, with the first two significantly associated with a reduction in outpatient costs.
Conclusions: Higher PCMH-PSC scores were associated with meeting HEDIS process quality standards and lower outpatient costs.
Implication for Policy or Practice We identified two benefits, reduced cost and increased quality, that support the implantation of the kind of practice infrastructure, policies, and procedures encouraged by NCQA's PCMH standards. Furthermore, cost saving we found might partially offset the cost of this implementation.
For more information about this presentation, please contact Research@abim.org.