Gray B, Vandergrift J, McAllister K. — American Board of Internal Medicine
Presented: AcademyHealth Annual Research Meeting, June 2015
To assess the association between physician Medicaid fee generosity for primary care services and access to board certified primary care internists of varies skill levels.
The physician sample included board certified internal medicine physicians about 10 years past their residency who reported a majority of time in primary care and were employed in group or solo office-based practice. These practice types may be particularly responsive to fee changes because of how they are generally compensated. The skill measures were residency evaluation ratings and initial board certifying examination performance. Share of patients Medicaid insured, practice location and other practice characteristics were based on surveys completed by nearly all internists as part of their ten-year Maintenance of Certification (MOC) requirement ( 85% response rate). The access measure was share Medicaid insured. The primary care fee measure was from 2008 and 2012 surveys of state Medicaid physician fees published by Kaiser. These fees were divided by the Medicare equivalent applicable to each state to account for regional difference in costs. County level socio-demographic measures were drawn from the Area Health Resource File (AHRF).
Applying a two-part state fixed effects model (having any Medicaid insured, conditional on any, the share Medicaid insured), the relationship estimated was between share Medicaid insured, the fee measure, and two physician skill indicators (i.e., top residency evaluation [score 9/9]; top quartile examination score). To measure the differential association between fees and share Medicaid insured across skill levels, the model included an interaction between a skill indicator and the fee measure. State indicator variables were applied to account for state characteristics that may be correlated with our fee measures but are constant over time, a 2008 indicator to account for factors that vary over time, and county-level Medicaid demand measures (Medicaid share eligible for 2008 and other socio-demographic measures from 2008 and 2012). State-level robust cluster adjustment was used to account for correlated errors.
Mid-career primary care office based internists.
On average, physicians reported that 8% of their patients had Medicaid. An increase in the primary care Medicaid/Medicare fee ratio (mean=.60) of 25 percentage points over time was associated with a 1.4 percentage point (P=.043) increase in the share of Medicaid insured in a physician’s practice (18% increase). The skill-fee interaction term coefficients indicated that this association was 2.1 percentage points larger for physicians with top residency evaluations than poorer evaluations, and 1.1 percentage point larger for physicians whose initial certification examination score was in the top quartile than those in a lower quartile (Ps<.05). These figures represent 26% and 14% of the share of patients with Medicaid insurance respectively.
Increased Medicaid primary care fees were associated with more access to our sample of overall high quality primary care physicians (all maintaining board certification) and this association was significantly larger among the highest skilled physicians in this group.
Implications for Policy or Practice:
Medicaid physician fees matter. A policy maker whose aim it is to increase access for underserved populations to high skilled physicians should consider fees as an important policy tool.
For more information about this presentation, please contact Research@abim.org.