Data Monitoring in Clinical Trials
Investigators commonly have questions regarding data and safety monitoring plans (DSMPs) in clinical trials. The Committee on Human Research (CHR) website explains when a DSMP is required at:
http://www.research.ucsf.edu/chr/Guide/hsppGP.asp.
The Regulatory Core offers advice on the "nuts and bolts" of setting up a DSMP, including sample plans.
On this site, we explain the ethical rationale for DSMPs in clinical trials.
What is the Purpose of a DSMP?
As interim data accrues in a clinical trial, they should be monitored to address the following questions: 1. Are the risks and benefits substantially different than expected? If so, the balance of risks and benefits in the trial needs to be reassessed. Keep in mind that the CHR requires that the risks be acceptable and minimized.
2. Will the trial's findings be valid? If the trial will not produce valid and generalizable findings, no risks to participants can be justified.
3. Has the research question already been answered? If so, it is ethically unacceptable to withhold the superior intervention from participants in the other arm(s) of the trial.
What Might Happen After Review of Interim Data?
In turn, if the data raise concerns about these issues, the investigators and the IRB need to consider several practical questions:
1.Should the informed consent process be modified? Additional information needs to be discussed with participants if it would alter the willingness of persons to enroll in or to continue to participate in the study.
2.Should the protocol be modified? This might be required to reduce risk to participants or to increase the validity the data.
3.Should the trial be continued? If interim data show conclusively that the risks are unacceptable, that one arm is more effective, or that the trial will not answer the research question, it would be unethical to continue the study.
Who will monitor data from the trial? The most rigorous way to carry out interim analyses of the data as it accrues is through a data and safety monitoring board (DSMB) that is independent of the investigators and the sponsor [Slutsky, 2004 #4223]. The advantage of a DSMB is that it can be unblinded without unblinding the investigators. Thus, the DSMB can examine interim statistical analyses comparing the arms of the trial, to determine if stopping rules have been met. However, the investigators remain blinded, which minimizes the potential for bias as the study continues.
The arrangements for monitoring interim data will depend on the type of trial. In Phase I studies, blinding may not be necessary or possible. The principal investigator, a colleague of the PI, or the study sponsor may monitor the data for unexpected serious harm.
Useful Links
NIH has general guidance on DSMBs on its website at http://grants.nih.gov/grants/guide/notice-files/not98-084.html.
The National Cancer Institute (NCI) has more detailed guidance regarding DSMBs, including examples of plans for DSMBs, at the website: http://www.cancer.gov/clinicaltrials/conducting/dsm-guidelines/page1.
Institute-Specific Guidance
Each institute at NIH has its own policies regarding DSMBs.
- At UCSF:
- [Add guidance at other institute]
- [Add guidance at other institute]
An alternative is to search the specific institute's webpage.
Data Monitoring Committees in Clinical Trials: A Practical Perspective by Susan S. Ellenberg, Thomas R. Fleming, and David L. DeMets
Data Monitoring in Clinical Trials: A Case Studies Approach by David L. DeMets, Curt D. Furberg, and Lawrence M. Friedman)
Chapter 12 in Fundamentals of Clinical Trials by Lawrence M. Friedman, Curt D. Furberg, and David L. DeMets
Chapter 11 of Designing Clinical Research by Stephen B. Hulley et al.
Bernard Lo, M.D.
Share your own advice about data safety monitoring plans by adding a comment to this article
%COMMENT%