Compliance Behavior in Clinical Trials:
Error or Opportunity?
Suzanne Bennett Johnson
University of Florida, Gainesville, FL
Clinical trials are an important component of the National Institutes of Health's (NIH) research mission and are an essential step in the development of new agents by the pharmaceutical industry. Usually biological interventions (e.g., an experimental medication for the treatment of hyperactivity) are the focus of such trials, but occasionally behavioral interventions (e.g., use of behavior therapy for the treatment of hyperactivity) are tested often in conjunction with a biological intervention (e.g., medication plus behavior therapy). Compliance is critical because it influences the investigator's ability to detect true treatment effects. From this perspective, noncompliance is viewed as an annoying source of error-- something that must be controlled or minimized. Historically, two methods have been used to control or minimize this source of error: (a) careful subject selection procedures, and (b) reliance on intent-to-treat analyses.
In this chapter, each of these methods of controlling or minimizing error is carefully examined. I argue that the error perspective has severely restricted the nature of the scientific information we collect, created problems with data interpretation, and limited the significance of our results. In other words, the error perspective has resulted in numerous missed opportunities to collect meaningful scientific information as part of a clinical trial. I argue that an alternative framework, in which compliance behavior is viewed as a central component of clinical trial research rather than solely as an annoying source of error, provides investigators