Hospital-physician integration and clinical volume in traditional Medicare

Abstract

Objective: To test the effect of hospital-physician integration on primary care physicians' (PCP) clinical volume in traditional Medicare.

Data sources and study setting: Nationwide retrospective longitudinal study using Medicare claims and other data sources from 2010 to 2016.

Study design: We identified 70,000 PCPs, some of whom remained non-integrated and some who became hospital-integrated during this study period. We used an event study design to identify the effect of integration on key measures of physicians' clinical volume, including the number of claims, work-relative value units (RVUs), professional revenue generated, number of patients treated, and facility fee revenue generated.

Principal findings: Per-physician clinical volume declined by statistically and economically significant margins. Relative to the comparison group who remained non-integrated, work RVUs fell by 7% (95% confidence interval [CI]: -8.6% to -5.5%); the number of patients treated fell by 4% (95% CI: -5.8% to -2.6%); and claims volume among PCPs who became hospital-integrated fell by over 15% (95% CI: -16.8% to -13.5%). Though professional revenue declined by $29,165 (95% CI: -$32,286 to -$26,044), this loss was almost entirely offset by increased facility fee revenue of $28,556 (95% CI: 26,909 to 30,203).

Conclusions: Hospital-physician integration may affect the quantity of clinical services delivered by PCPs to traditional Medicare beneficiaries. Reductions in clinical volume associated with integration may have long-term consequences for the supply of physician services and patient access to primary care. Future research on physician time use and patient access following hospital integration would further add to the evidence base.

Keywords: Medicare; delivery system organization; hospital workforce; outpatient care delivery; physician employment; physician practice organization.

FIGURE 1

Physician clinical volume before and after integration. Each panel displays the event study specification of the effect of integration on the dependent variables of the number of claims, professional revenue (dollars), work‐relative value units, and the number of patients. The y‐axis displays the regression coefficients for time dummies. The x‐axis displays the time relative to when the physician became hospital‐integrated (t = −1 represents the last year pre‐integration). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

Effects of integration on work‐relative value units (RVUs), by practice size. Each panel displays the event study specification of the effect of integration on work RVUs. The y‐axis displays the regression coefficients for time dummies. The x‐axis displays the time relative to when the physician became hospital‐integrated (t = −1 represents the last year pre‐integration). The left and right panes show the effects among physicians coming from small (2 physicians) and large (32 or more physicians) practices. [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 3

Facility fees for physician services before and after integration. This figure displays the event study specification of the effect of integration on the value of Medicare facility fees associated with services performed by integrated physicians. The y‐axis displays the regression coefficients for time dummies. The x‐axis displays the time relative to when the physician became hospital‐integrated (t = −1 represents the last year pre‐integration). [Color figure can be viewed at wileyonlinelibrary.com]