Evaluating the policy role of the small area variations and physician practice style hypotheses.

The primary purpose of this article is to develop a framework for reinterpreting the role of physician practice style in the small area variations phenomenon. This phenomenon deals with the wide interarea variations in per capita use rates which have been found for many medical and surgical procedures. The variations have been interpreted by many to suggest that large amounts of unnecessary care are being provided. The variations and corresponding perceptions of unnecessary care have also led to a US health policy which is increasingly emphasizing patient outcomes research. I show, however, that most of the empirical studies of the variations phenomenon have inappropriately aggregated either across procedures or across market areas so as to obscure the role of practice style. Its role has also been obscured by the common failure to distinguish practice style from other determinants of utilization. As a result, small area methods can lead to substantial error in identifying procedures associated either with major differences in practice style or with substantial amounts of unnecessary care if all variation is attributed to practice style.