Improving the Quality of Measurement and Evaluation in Quality Improvement Efforts

short of the desired outcome. Efforts to measure, publicize, and reward clinicians and health care organizations for their quality of care have grown exponentially over the last decade. The underlying goal of these efforts is to increase the number of patients treated by high-quality health care providers. Two methods to achieve this goal have received widespread attention. First, convert lowquality providers to high quality; pay-for-performance and quality improvement (QI) efforts are strategies to accomplish this conversion. Second, move market share from low-quality to high-quality providers; public reporting of performance measures is a strategy to achieve this shift. Underpinning all of these efforts is the need for valid quality of care measures. The science of quality measurement is still immature, despite the near exponential growth in quality of care measures. Most quality measurement efforts struggle to find a balance between measures that are scientifically sound yet feasible to implement with the limited resources allocated. Unfortunately, feasibility generally trumps sound science. Ideally, risks and benefits of more versus less rigorous measurement would be explicit and used to guide measurement in QI projects. Unfortunately, the relative risks and benefits are implicit or completely opaque. This contrasts with the more rigorous measurement often used in clinical research. Historically, QI measurement has been less scientifically rigorous than measurement for clinical research. Public reporting of quality of care, both voluntary and mandatory, has made this difference more relevant. Regardless of the label, patients, providers, and payers should have confidence that public statements regarding the quality of care provided are true. In support of QI efforts, we lack explicit methods to evaluate the strength and amount of evidence required for a measure to be included in pay for performance or public reporting and to evaluate the validity of a measure.This article discusses how quality measurement could be informed by the measurement principles for clinical research, explores the strength of evidence required to use a measure in pay for performance, and considers a model to evaluate the validity of quality measures.