Building a More Connected DSMB: Better Integrating Ethics Review and Safety Monitoring

Abstract

Data and Safety Monitoring Boards (DSMBs) have become an increasingly common feature of clinical trial oversight, yet a paucity of legal or ethical frameworks govern these Boards’ composition or operation, or their relationship with other actors with monitoring responsibilities. This paper argues that prevailing structural gaps are impeding harmonized systems for monitoring the ongoing ethical acceptability of clinical trials. Particular tensions stem from DSMBs’ sweeping discretion in deciding whether and when to recommend that a trial should be terminated or amended based on safety and efficacy information. This discretion becomes especially challenging in light of DSMBs’ monopoly over emerging trial data, which prevents Institutional Review Boards, sponsors, and investigators from participating in certain pivotal and ethically charged decisions. To address these disconnects, I advocate for strengthened pre-trial and post-trial communication in addition to innovative strategies to support DSMB decision making through the life of a trial.

Keywords:

• clinical trials
• data monitoring
• Data Safety and Monitoring Boards
• human subjects protection
• Institutional Review Board
• research ethics

Notes

1. A more cynical interpretation would also reflect potential conflicts of interest a DSMB faces; in particular, a Board constituted by a commercial sponsor eager to market a new drug or indication for a drug. While recognizing the potential role that such conflicts of interest play in DSMB decision-making, this paper focuses, instead, on the structural disconnects that limit transparency even in the most well-intentioned and conflict-free Board.

2. At the second interim review, a total of 134 incident invasive breast cancers had been reported, including 89 in those randomized to placebo versus 45 among those randomized to tamoxifen. At the fourth and final review, a total of 239 incident breast cancers had been reported, including 154 in the placebo group versus 85 in the tamoxifen group. Presumably at least some of these women would have been diagnosed with breast cancer even if they had received tamoxifen; however, it is fair to suppose that at least some of them could have been prevented.

3. Essentially, ethical walls work by limiting disclosure of information to certain individuals who may be subject to a conflict of interest. These have been upheld as means of rebutting the presumption that an attorney has shared confidential information his or her law firm, thereby mitigating the effects of certain conflicts of interest in law firm representation.