Should payment for performance depend on mortality?

The introduction of the Hospital Value Based Purchasing (HVBP) programme, as shown recently by Jose F Figueroa and colleagues, did not improve 30 day mortality of Medicare beneficiaries admitted to US hospitals for three incentivised conditions. We agree with the authors' conclusion that an "appropriate mix of quality metrics and incentives to improve patient outcomes" has yet to be identified. The programme was designed to promote better clinical outcomes for hospital patients and to improve their experience of care during hospital stays. However, if reducing mortality was the primary objective, a specific set of indicators should have been selected from relevant practice and organisational guidelines, which was not the case when the programme was designed. The absence of HVBP's impact on mortality is therefore not surprising. In their conclusion Figueroa and colleagues say, "Nations considering similar pay for performance programs may want to consider alternative models." This raises a question: should payment depend on achieving a decrease in mortality rates? The primary objective of pay for performance programmes could still be the development and maintenance of continuous quality improvement (CQI) programmes based on relevant process indicators to assess critical steps in patient management, such as the effective use of checklists in surgical procedures. These process indicators are apt to detect dangerous misconduct, less susceptible to bias, and more actionable. However, a problem remains: such CQI programmes in hospitals rely on a limited number of people repeatedly trying to motivate a large number of health professionals who are insensitive to the benefits of CQI. Instead of diverting the small percentage of budget devoted to pay for performance programmes it could be more profitable, for patients, to think of a better use for this money.