it results in a net improvement in quality. Thus, the direct benefits of implementing a particular measure cannot be outweighed by the indirect harms, e.g., resource and opportunity costs, antagonizing providers, incentivizing perverse behaviors, or negatively affecting other domains of quality. Although simple in concept, measuring these benefits and harm is often difficult and heavily influenced by which group—patients, payers, or providers—is doing the accounting.

Expanding on other recent reviews of performance measurement (Bird et al., 2005; Birkmeyer et al., 2004; Landon et al., 2003), this paper provides an overview of measures most commonly used to profile the quality of physicians, hospitals, or systems and their main limitations. We describe the trade-offs associated with structure, process, and outcome measures (see Table F-1). We address the question of “how good is good enough?” and make the case that the answer depends on the purpose of measurement—quality improvement or selective referral. Finally, we consider which measures are ready (or almost ready) for implementation right now and a research agenda aimed at improving performance measurement for the future.


A large number of performance measures has been developed for assessing health care quality. Tables F-2 and F-3 include a representative list of commonly used quality indicators and measurement sets. Almost all of these measures have been either endorsed by leading organizations in quality measurement and/or already applied in hospital accreditation, pay for performance, or public reporting efforts. A more exhaustive list of performance measures can be found on the Agency for Healthcare Research and Quality’s (AHRQ’s) National Quality Measures Clearinghouse Web site ( Although the measures could be sorted on other dimensions, we consider them below according to whether they focus on ambulatory care (preventive care and chronic disease management) or hospital-based care (including surgery).

Ambulatory Care

Although not the only measurement set in ambulatory care, the Health Plan Employer Data and Information Set (HEDIS), developed by the National Committee on Quality Assurance is by far the most familiar (Table F-2). HEDIS measures focus largely on processes of care relating to preventive and other primary care services, but they also include measures of health plan stability, access to care, and use of services. The National Quality Forum (NQF) is endorsing a set of ambulatory care quality indicators devel-

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement