A study recently released by the Harvard School of Public Health indicates that hospitals’ participation in pay-for-performance (P4P) incentive programs is unlikely to produce meaningful improvement in patient outcomes.¹ The 252 hospitals that participated in Premier Inc.’s Hospital Quality Incentive Demonstration (HQID) and treated more than six million patients over the course of six years failed to achieve lower 30-day mortality rates when compared to their non-participating counterparts.² In addition, the Harvard study did not reveal any significant differences in mortality trends between the conditions whose outcomes were explicitly linked to payment incentives and the outcomes for those conditions that were not.³ Although the study did indicate some impact of financial incentives on process improvement, it may cast doubt on current value-based purchasing strategies.

Despite the studies negative findings, Dr. Ashish K. Jha, M.D., the primary researcher on the study, did point out that public reporting requirements may have incentivized all hospitals to improve their performance, regardless of whether they participated in Premier’s program.⁴ Furthermore, Dr. Jha suggested that clinicians have improved their treatment methodologies over time, a factor unrelated to incentive payments, but one which also contributed to a reduction in mortality rates.⁵ Premier contended that the HQID achieved its goal, as the program was intended “to determine whether incentives would improve care processes in hospitals,” and most of the quality measures focused on whether hospital staff followed certain procedures or “process measures,” such as providing heart attack patients with beta-blockers when being admitted and discharged.⁶  Although the HQID demonstrated the success of financial incentives on improving process measures, Premier acknowledged these metrics may not be the best method for improving patient outcomes, which may leave many providers confused as to how to approach quality improvement within their organizations given the HQID’s demonstrated lack of improvement on mortality rates.⁷

The results of Premier’s pay-for-performance program may exemplify a recurring theme in government-sponsored programs aimed at reducing healthcare spending. According to the Congressional Budget Office (CBO) evaluations appear to indicate that, in nearly every program involving disease management or care coordination, spending either increased or remained unchanged when incentive costs were included.⁸ The CBO recommends that future policies designed to curb spending should focus on data gathering, particularly with respect to hospital admissions; smoothing patient transitions among different healthcare facilities; using team-based care; targeting interventions towards high-cost enrollees; and, limiting the costs of interventions.⁹

Significantly, other recent research suggests the logic behind tying together cost savings and quality improvement measures may be flawed. In a recent study by the National Bureau of Economic Research, Professor Joseph Doyle suggests that reducing healthcare costs may have a negative impact on patients, and that the approach may actually lead to worse patient outcomes.¹⁰ In Doyle’s study, the one-year mortality rates for patients treated in higher-spending hospital emergency rooms were “as much as 30 percent lower” than patients treated at lower-cost facilities.¹¹ A recent Canadian study also observed an association between higher spending, lower mortality, and fewer readmissions.¹² It is possible that, due to limited emphases, both studies may fail to address quality and cost in the larger scope of the U.S. healthcare system.  Jon Skinner of the Dartmouth Atlas of Health Care made this argument in defense of his findings, which contend that greater spending does not correlate with better quality care.¹³

Despite opposing theories on the relationship between healthcare spending and quality, the results of the Harvard study will likely prompt reevaluation of the Center for Medicare and Medicaid Services’ (CMS) Hospital Value-Based Purchasing (HVBP) program.¹⁴ As part of the Patient Protection and Affordable Care Act, CMS was required to adopt a nationwide pay-for-performance program, which resulted in the creation of its HVBP program.¹⁵ CMS’s program was modeled after Premier’s HQID, and although the HVBP program’s current incentives focus on process measures and patient satisfaction data, it will soon be expanded to include 30-day mortality, a measure which promises little improvement according to the Harvard study’s key findings.¹⁶

Beginning in FY2013, CMS will withhold one percent of hospitals’ Medicare payments in order to fund its HVBP program and hospitals will have to perform well on a number of quality measures in order to receive incentive payments under the program and recover a portion of their withheld Medicare payments.¹⁷ This task may prove difficult for many hospitals to accomplish given the recent research findings, and the hospitals with the lowest scores will not recover any of their withheld payments.¹⁸ Despite his study’s findings, Dr. Ashish K. Jha still supports the pay-for-performance model and believes success can be achieved through reworking both the “pay” and the “performance” components of the model.¹⁹ In the interim, however, providers will have to reconcile the possibility of poor performance on mortality measures with the resulting effect it may have on their Medicare reimbursements in the near future.