Will Pay-for-Performance Pay for Itself?

Over the last two decades, skyrocketing medical costs have forced researchers and policy makers to investigate ways to restrain medical inflation without sacrificing health care quality. Pay-for-Performance (P4P) schemes that reward care providers for positive outcomes, rather than services provided or procedures performed, promise to reward quality rather than quantity and shift the medical cost-curve. The federal Medicaid and Medicare  programs, as well as individual states, have implemented P4P plans in various health care arenas, and the Affordable Care Act plans to expand P4P programs even further. However, despite the promise of using the market to incentivize better and more efficient patient care, academic research on P4P has revealed at best mixed results. In The Effect of Pay-for-Performance in Nursing Homes, Rachel Werner, Tamara Konetzka, and Daniel Polsky compare changes in quality of care between states that did and did not implement P4P plans for nursing homes using Medicaid funding from 2001 to 2009.

In 2007, nursing homes provided care to over 1.5 million residents at a cost of over $120 billion. They represent a rich object of study, as poor quality in nursing homes motivated the 1987 Nursing Home Reform Act. The act mandated quarterly reports on each resident, and inspections of nursing homes at least every 15 months. The authors used this database to study the implementation of P4P schemes in eight states over the past decade. These states implemented similar programs that rewarded nursing homes for meeting requirements of care, including staffing levels, regulatory compliance, and in some cases specific clinical care measures such as rate of pressure sores or numbers of residents who were physically restrained. Rewards ranged from a 0.5 to five percent bonus on the per diem reimbursement per resident, and cost between $1 and $18 million per state in 2008. The eight states implemented their programs at different times between 2002 and 2009, allowing the researchers to see how quality varied with different times of implementation. The study included 17,579 nursing homes, of which 3,513 were in a state with P4P plan. The study included 44.1 million individual observations. Researchers used a difference-in-difference approach to study P4P’s impact on nursing home quality, comparing changes in quality measures before and after implementation between states with P4P programs and the 42 without. The study also controlled for characteristics of nursing homes and residents, such as number of beds or race and gender, which could affect care quality. The study examined changes at both one and two years before and after the start of P4P.

The study could not find systemic improvement in care due to P4P. Regulatory deficiencies increased by an average of 1.12 percent in P4P states. P4P was associated with a decline in pressure sores of 0.3 percent from a base of 12.3 percent, and incidents of severe pain decreased 0.5 percent from 14 percent, but nursing homes in P4P states experienced relatively higher rates of catheter use and patient weight loss. The researchers also looked at effects of P4P by state, but only Georgia saw significant improvement in all quality measures of interest.

The authors suggested several reasons for the incentive pay schemes’ disappointing results. The incentives, with a maximum payout of an additional five percent per resident, may have been less than the additional cost of providing better care. Additionally, incentives were targeted at the nursing home, rather than individual nurses and aides who actually provided care. The authors also suggested paying bonuses on a monthly rather than annual basis to better motivate nursing homes and more tightly link benefits to performance.

This study adds to the literature finding that pay-for-performance in health care has only mixed results at best. Incentive programs could possibly produce positive results by incorporating better-targeted rewards and punishments, but more research would be beneficial before pay-for-performance is rolled out on a larger scale.