Data & Safety Monitoring for Clinical Trials

Data and safety monitoring is intended to protect the safety of research participants and ensure the integrity of the research data. All studies involving human subjects require some level of monitoring.

Federal regulations require appropriate provisions be in place to adequately monitor the data collected and ensure the safety of research subjects.1 The NIH requires data and safety monitoring for all types of clinical trials, including physiologic, toxicity, and dose-finding studies (phase I); efficacy studies (phase II); and efficacy, effectiveness, and comparative trials (phase III).2 The method and level of data and safety monitoring should be commensurate with the degree of risk of harm to subjects and the size and complexity of the study.

Data Safety Monitoring Plans (DSMP)

A data safety monitoring plan (DSMP) is a written plan for monitoring research data and the safety of study participants. DSMPs are required for clinical studies involving more than minimal risk. A DSMP ordinarily appears in the study protocol and assures there is a plan and system for appropriate oversight of the study. The DSMP should delineate who is responsible for reviewing and reporting adverse events, to whom they will be reported, and on what schedule this monitoring and reporting will occur. The DSMP should also explicitly indicate whether or not there will be a formal data safety monitoring board (DSMB) or data monitoring committee (DMC). Though every study involving more than minimal risk will have a DSMP, only a subset of studies will have a DSMB or DMC.

Data Safety Monitoring Boards (DSMB) and Data Monitoring Committees (DMC)

A DSMB or DMC is a formally convened group charged with reviewing data and endpoints on a timeline set forth in a protocol’s DSMP. These groups are made up of members from different disciplines who are knowledgeable and experienced in the responsible conduct of research. Membership must include experts in biostatistics, experimental design, bioethics, and the medical field under study.

DSMBs and DMCs are typically required for studies that involve more than minimal risk and include at least one of the following:

  • Multiple study sites
  • Vulnerable populations
  • Blinded study design
  • New therapies or science
  • High-risk interventions

To request the above document be made accessible please email communications@catalyst.harvard.edu.

1 45 CFR 46.111(a)(6); 21 CFR 56.111(a)(6)
2 NIH Policy for Data and Safety Monitoring released June 10, 1998