Cost Analysis of the Medicare Physician Group Practice Demonstration

Accountable care organizations (ACOs) are increasingly being created as payment models designed to improve care and slow the growth of healthcare costs. The Centers for Medicare & Medicaid Services (CMS) has created 3 ACOs that differ in details but employ a shared approach. In all 3 programs—Pioneer, the Medicare Shared Savings Program, and the Advance Payment Model—participating organizations can share in savings if they meet quality and cost targets for their assigned beneficiaries.

The Patient Protection and Affordable Care Act included ACOs because simulations demonstrated that CMS could realize savings with the ACO model. The ACOs developed by CMS were built on a pilot program, the Medicare Physician Group Practice Demonstration (PGPD). In the PGPD, physicians received bonus payments if they achieved lower growth in costs compared with local controls at the same time they met quality targets.

Noting that while evidence from the PGPD demonstrated improvements in quality, the effect of the program on cost remains unclear. Researchers recently conducted an analysis to estimate cost savings associated with the PGPD overall and for dual-eligible Medicare and Medicaid beneficiaries. They reported results of the analysis in the Journal of the American Medical Association [2012;308(10):1015-1023].

The researchers utilized Medicare administrative data to analyze changes in coding for beneficiaries assigned to each of the 10 PGPD participants and their local groups. The primary outcome was annual spending per Medicare fee-for-service beneficiary.

The intervention group included fee-for-service Medicare beneficiaries who received care primarily from physicians in the PGPD program (n=990,177); the control group included Medicare beneficiaries from the same region who received care primarily from non-PGPD physicians (n=7,514,453). Overall, approximately 15% of beneficiaries were dual eligible for Medicare and Medicaid.

Characteristics in the 2 groups were similar at baseline, including mean age, proportion disabled, proportion dying annually, average number of comorbidities, and prevalence of each comorbidity. Beneficiaries in the control group were more likely to be women, eligible for Medicaid, and black.

The researchers noted that the demographics of the 2 groups did not change significantly between the preintervention and postintervention periods, “suggesting PGPD participants did not systematically target specific demographic groups for either enrollment or disenrollment.”

After adjustments, per capita annual savings estimates for the intervention group compared with the control group were modest ($114; 95% confidence interval [CI], $12-$216; P=.03), but not statistically significant among beneficiaries who were not eligible for both Medicare and Medicaid ($59; 95% CI, $166 in savings to $47 in additional spending). The savings among dual-eligible beneficiaries were significant ($532; 95% CI, $277-$786; P<.001).

Overall, the adjusted mean reductions in spending were concentrated in acute care ($118; 95% CI, $65-$170). Among single-eligible beneficiaries, the adjusted mean reductions in savings were $85 (95% CI, $32-$138); among dual-eligible beneficiaries the reductions in savings were $381 (95% CI, $247-$515).

The analysis found significant variation in savings across practice groups; the savings ranged from an overall mean per capita annual savings of $866 (95% CI, $815-$918) to an increase in expenditures of $749 (95% CI, $698-$799). Overall, 30-day medical readmissions decreased by 0.67% (95% CI, -1.11% to -0.23%); in the dual-eligible population, the decrease was 1.07% (95% CI, -1.73% to -0.41%).

In conclusion, the researchers said, “substantial PGPD savings achieved by some participating institutions were offset by a lack of savings at other participating institutions. Most of the savings were concentrated among dual-eligible beneficiaries.”