We will discuss the methods, results, and implications of a series of scoping and systematic reviews that synthesized evidence on the reproducibility of health science research. Specific aims of the projects included describing and categorizing methods and metrics used to quantify reproducibility, synthesizing estimates of the prevalence of reproducible research, and characterizing factors associated with reproducibility. We will describe our dataset of 1,340 articles with a focus on the sub-set with empirical data. We will report on the seven methods used for conducting reproducibility research, 38 metrics to quantify reproducibility, 180 estimates of the prevalence of reproducibility, and 234 factors associated with reproducibility which we grouped into modifiable and non-modifiable categories. Finally, we will discuss the significant heterogeneity in the evidence and challenges in synthesizing this data.
Speakers:
Niklas Bobrovitz, MD, MSc, DPhil
third-year emergency medicine resident
Centre for Health Informatics
University of Calgary
Stephana Moss, MSc, PhD
Assistant Professor
Department of Pediatrics
Dalhousie University