Features of registered clinical trials and associated papers on physical activity involving people with spinal cord injury.

Clinical recommendations are typically developed based on published data. However, if not all data are published this may influence the accuracy and effectiveness of recommendations. We examined whether failure to publish clinical trial data is a problem in physical activity (PA) research involving people with spinal cord injury (SCI). This article reports features of 102 PA trials registered on ClinicalTrials.gov including participants with SCI, of which 63% had associated papers. Target enrolment was achieved in 41% of published trials and 85% of participants enrolled completed trials. Only 60% of papers stated hypotheses, of which 76% were supported. We discuss why ~2/5 trials remain unpublished, whether researcher prefer to 'file-drawer' negative findings, and/or editors/reviewers look favourably on positive findings. We stress the importance of transparently reporting all results from SCI-PA studies regardless of positive/negative directionality and raise concern regarding the influence of publication biases and selective data reporting on clinical recommendations.