Penn State Translational Science Seminar: Reducing risk of bias in study design and reporting in human physiology research

Presented by Nathalie Kirby, PhD

Scientists and decision-makers rely on accurate scientific evidence to advance knowledge, reach consensus, and develop and enact policy. A key consideration in the design of studies for this purpose is minimizing bias, which, in this context, describes a systematic tendency that causes differences between the study’s findings and the “truth”. Such bias can be introduced in any step of the research pipeline, including study design (e.g., by not controlling factors unrelated to the study’s goal), analysis (e.g., through researchers’ preconceived notions), and reporting (e.g., by emphasizing statistically significant findings over non-significant ones). While the introduction of bias is often unintentional, it nonetheless undermines confidence in a study’s findings. Thus, identifying and minimizing sources of bias in our experiments is critical to the scientific integrity of our field. To provide a snapshot of common areas of potential bias in health sciences, this presentation will dsicuss the findings of a risk of bias analysis (Cochrane Risk of Bias 2 [RoB2] tool) from an ongoing meta-analysis of >400 environmental physiology studies investigating passive heat exposure. Studies were evaluated within the six RoB2 domains: randomization, deviation from intended interventions, missing data, measurement of the outcome, selection of reported results, and carryover effects. Further, while the RoB2 tool is useful to evaluate bias in clinical trials in medicine and adjacent fields, differences in common practice between research areas make it difficult to employ when assessing physiology studies. Leveraging this unique data set, the ongoing project aims to contextualize this tool for use across kinesiology and physiology fields. Throughout this presentation, Dr. Kirby will provide easily actionable examples and strategies to assist researchers in identifying and reducing sources of unintentional bias in their experiments and showcasing the rigor of their work in their manuscripts.