The University of Alabama Interregional Headache School Study: Anatomy of a Failure
D. Andress‐Rothrock,J. Rothrock
DOI: https://doi.org/10.1111/j.1526-4610.2012.02120.x
2012-04-01
Abstract:Why do research studies “fail”? More specifically, why do clinical research studies involving a therapeutic intervention fail? It has long been said that “success is born of many fathers; failure of but one.” While appropriate in many situations, this maxim does not extend to clinical research, an enterprise where “failure” goes by many definitions and may result from a myriad of causes. For example, if the study involves a therapeutic intervention, was the treatment under investigation simply ineffective? or unsafe? Is the “failed” therapy, in fact, effective for the intended indication . . . but unable to achieve a significant P value because the study was underpowered (the so-called type II error)? Was the study population poorly chosen? Was the study poorly conducted? The list goes on. Many fathers. What follows is a description of one such failed study, an analysis of those factors most relevant to its lack of success and suggestions as to how one might avoid repeating these same mistakes. Consider the last . . . advice from a failure. In 2005-2006, my colleagues and I at the University of South Alabama in Mobile sought to investigate what effect an intensive patient-directed educational intervention might have on various parameters related to the management and clinical outcome in the migraine patient population. Specifically, we randomized patients to receive either our “standard” educational intervention (ie, in-clinic physiciangenerated verbal education and written materials) vs this same intervention augmented by 3 after-hours classes conducted by trained lay instructors and concerning the topics of migraine’s epidemiology and pathophysiology, acute migraine treatment, and migraine prevention. The treating neurologist was blinded to the results of randomization and managed all participating subjects on a nonuniform, case-bycase basis. For literally all of the outcome variables prospectively chosen and then analyzed, the group randomized to this “headache school” exhibited a significantly more favorable response. Recognizing that greater Mobile represented a somewhat unique locale in terms of its social and clinical demography, we sought to investigate whether the favorable results from our initial experience with “headache school” could be extrapolated to other geographic regions and medical practice environments. We specifically chose for this interregional study a university-based headache clinic in New England, a university-based headache clinic in Birmingham, Alabama, a private sector headache clinic in Montgomery, Alabama, and a private sector headache clinic in southern California. The University of Alabama (Birmingham) served as the coordinating center, and the authors made site visits to each of the other 3 participating clinics to familiarize the physician coinvestigators with the study protocol and to train their designated lay instructors. The study was conducted under the auspices of each site’s respective investigational review board (IRB). All study-related data were recorded electronically and transmitted to a centralized registry at the coordinating center. The study was supported via an investigator-initiated grant provided by Merck. The study protocol called for each site to enroll a total of 100 subjects over a period of 6 months and to follow each subject for 6 additional months. The ISSN 0017-8748 doi: 10.1111/j.1526-4610.2012.02120.x Published by Wiley Periodicals, Inc. Headache © 2012 American Headache Society