The Ethics of Withdrawal from Study Participation

Abstract

We investigated whether consent forms adhere to Federal Common Rule regulations pertaining to withdrawal from research, described the language of withdrawal provisions, and assessed differences in studies by withdrawal provisions. A random sample of 114 consent forms from a midwestern, academic medical center were examined for descriptive content of withdrawal parameters stated within consent forms. All consent forms included the required statement about withdrawal pursuant to the Code of Federal Regulations (CFR) 45 CFR (§46.116(a)(8)), and all adhered to regulation 45 CFR (§46.116(b)(4)) by including a statement that withdrawal will have no affect on care provided. Of 114 studies, thirty (26%) studies explicitly requested subjects/participants to engage in a further behavior before withdrawing from the study. Safety was mentioned in only 4 (13%) instances as the reason for an additional visit or test. None of the consent forms provided information about the consequences to the subject's health or well being by withdrawing from study participation. Consent forms generally conform to current regulations. Future research should examine subjects' experiences of withdrawing from research in order to help clinical investigators and Institutional Review Boards (IRBs) assess the extent to which consent forms indicate barriers to withdrawal and for compliance with Federal Common Rule regulations.

Keywords:

• consent form
• withdrawal
• policy
• clinical research
• Federal Common Rule
• content analysis
• voluntariness
• undue influence

Acknowledgments

Supported by grant DK063953 from the National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD. Thanks go to Nancy King, Rick Kodish, and Mark Kuczewski for helpful insights on earlier drafts. An earlier draft of this article was presented at the 7^th Annual Meeting of the American Society of Bioethics and Humanities, October 21, 2005, Washington, DC.

Notes

∗Some PIs sponsored more than one protocol.

†Since 1 protocol was not a clinical trial, total for analysis was one less than N for notify the PI and further testing.

‡Chi-Square analyses were only performed on ‘Notify the PI’ and the set of ‘All Withdrawal Stipulations Combined’, and not performed on ‘Return for another visit’ or ‘Return for further testing’ because of small numbers.

§P < 0.05, °P < 0.005, ^││P < 0.0005.

21 C.F.R. 50 (1991).

45 C.F.R. 46 (2005).