Objective: To quantify data sharing policy compliance at the BMJ by analysing the rate of data sharing practices, and investigate attitudes and examine barriers towards data sharing. Design: Observational study. Setting: The BMJ research archive. Participants: 160 randomly sampled BMJ research articles, excluding meta-analysis and systematic reviews. Main outcome measures: Percentages of research articles that indicated the availability of their raw datasets in their data sharing statements and those that provided their datasets upon request. Results: Fifty out of 160 (31%) research articles indicated the availability of their datasets. Twelve used publicly available data and the remaining 38 were sent email requests to access their datasets. Only 1 publicly available dataset could be accessed and only 6 out of 38 shared their data via e-mail. So only 7/160 research articles shared their datasets, 4.4% (95% confidence interval: 1.8% to 8.8%). Conclusions: Despite the BMJ’s strong data sharing policy, sharing rates are low. Possible explanations for low data sharing rates could be: the wording of the BMJ data sharing policy, which leaves room for individual interpretation and possible loopholes; that our email requests ended up in researchers spam folders; and, that researchers are not rewarded in the scientific community for sharing their data. It might be time for a more effective data sharing policy and better incentives for health and medical researchers to share their data.