Threats to validity in randomized clinical trials.

The purposes of this article are to present an overview of randomized clinical trials (RCTs) and describe some of the methodological problems inherent in using RCTs in nursing research. Many nursing intervention studies are fraught with problems that defy the stringent control criteria required for RCTs, leading to biased estimates of intervention efficacy. Five threats to validity in RCTs are presented, including problems related to (a) differential dropout, (b) random assignment, (c) identifying and maintaining an adequate control condition, (d) nonadherence to research protocols, and (e) assessment of clinically meaningful change. Three strategies are recommended for addressing some of the problems posed by RCTs and improving inference. Copyright 2000 John Wiley & Sons, Inc. 23:79-87, 2000