The Health Plan Employer Data and Information Set (HEDIS) is now almost surely the most commonly used system for gauging health care performance. However, despite its broad use, HEDIS has many critics. The quality measures included in HEDIS do not asses many important aspects of health care delivery and the individual measures could be improved. In the latter part of July 1996, NCQA will release a new version of HEDIS 3.0 for public review and comment. That draft will be revised and released in final form in the latter part of 1996. HEDIS 3.0 will identify two types of measures, a reporting set and a testing set. Health plans nationally will be expected to collect data on 'reporting set' measures to meet purchaser and consumer requirements for information. The 'testing set' will identify a series of measures that are particularly promising, but that are less well developed and/or have been less thoroughly evaluated. The purpose of identifying a 'testing set' is to provide a signal about measures that may be incorporated into HEDIS (after appropriate revision) in the future, and to stimulate their refinement. Because of the widespread use of HEDIS, it is critically important to evaluate its components and strengthen them where needed. If there are problems with the new measures, it is important to identify those problems as soon as possible so that appropriate changes can be made. This is particularly true of the 'reporting set' measures because they will be so widely used. Further development and evaluation of the 'testing set' also is needed. As a result, we propose a rigorous and broad evaluation of HEDIS 3.0. Specifically, we propose to: 1) Evaluate the new 'reporting set' measures in HEDIS 3.0 and a subset of the original 'reporting set' measures with respect to their relevance for users, the soundness of the science that underlies them, and the feasibility of implementing them; 2) Develop complete operational specifications for a subset of 'testing set' measures that are particularly strong candidates for the next version of HEDIS; and 3) Evaluate the 'testing set' measures that might be used in the next version of HEDIS with respect to their relevance, scientific soundness and logistic feasibility. Based on these analyses we will suggest refinement of specific measures and identify important problems with individual indicators that can guide decisions about whether to include them in subsequent versions of HEDIS. We propose this work as part of a general strategy and method for developing and refining measures such as HEDIS in the future.
