Summary Background: Increasing regulation of medical

research, in particular the requirement for explicit consent, may reduce the quantity and quality of clinical epidemiological research. Aim: To assess the potential biases arising from the need for explicit consent in our hospital-based stroke research register. Design: Comparison of patients enrolled into our stroke research register with those included in a concurrent clinical stroke audit that targeted the same population but did not require explicit consent. Methods: We obtained the numbers of consenters, refusers and those from whom consent was not sought for various logistical reasons. We compared characteristics of participants (those eventually included in the research register) vs. non-participants. Results: Of 1228 patients included in the stroke audit during an 18-month period, 1075 (88%) were also included in the research register, with higher participation among outpatients than inpatients. Only 1 % of eligible patients refused involvement in any aspect of the research register. By far the largest number of non-participants was those from whom we could not seek consent for practical reasons. Comparison of baseline characteristics showed important differences between participants and non-participants that could affect outcome. Conclusions: Very few patients refused inclusion in our research register, but the need for explicit consent reduced participation and introduced bias. An opt-out system avoiding the need for explicit patient consent for minimally intrusive clinical epidemiological studies would minimize bias and reduce the considerable time and costs associated with the consent process