High-risk studies are influenced by indirect range restriction.

Abstract

High-risk studies select subjects who are at high risk for existing or future disease. Therefore, the range of disease is restricted in high-risk studies. This paper shows that high-risk studies are vulnerable to a particular type of range restriction referred to as indirect range restriction. A simulation study is used to illustrate the effects of indirect range restriction on high-risk studies. The results suggest that indirect range restriction can have a substantial impact on the results of high-risk studies. In addition, a review of several areas of behavioral medicine research suggests that high-risk studies have produced many misleading findings. The range restriction approach can be used to estimate statistical power in high-risk studies, interpret the results of high-risk studies, and design future high-risk studies.