Reporting and replicating trials of exercise-based cardiac rehabilitation: Do we know researchers actually did?
Date of this Version
Background—Complete reporting of all components of complex interventions is essential for translation of research evidence into clinical practice. Previous work has highlighted deficiencies in the reporting of nonpharmacological interventions; however, the reporting quality of exercise-based interventions for coronary heart disease has not been examined. Methods and Results—A systematic search strategy was used to identify randomized controlled trials of exercise-based cardiac rehabilitation published until December 2013. Fifty-seven trials were included, reporting on 74 interventions. Intervention description completeness was assessed using the Template for Intervention Description and Replication checklist. Missing intervention details were then sought from additional published material and also by emailing corresponding authors. Only 6 interventions (8%) sufficiently described all required items within the main publication; this increased to 11 (15%) after searching for additional published material and 32 (43%) after contacting trial authors. Although location/setting and duration were consistently well reported in publications, complete descriptions of the exercise schedule, as well as details about its tailoring and progression, were missing for over half of interventions (complete for 42% and 36% of interventions, respectively). Although some authors (25/61) were able to provide missing intervention details when contacted, others could not be located (20) or did not respond (16). Conclusions—Inadequate reporting of cardiac rehabilitation interventions is a substantial problem, with essential information frequently missing, and for almost half of all interventions, unobtainable after publication. A conscientious effort to address this problem could facilitate an improvement in the quality of cardiac rehabilitation delivered in clinical practice.
This document has been peer reviewed.