Data Safety Monitoring Boards:
Another dimension of protection
The recently-released EMEA guidelines on Data Monitoring Committees have urged the creation of Data Safety Monitoring Boards (DSMBs) to protect the interest of patients enrolled in clinical trials. DSMBs are a must to help improve the safety of volunteer research participants in studies. These boards commonly provide advice on whether trials and/or specific treatments may be safely continued. They are normally comprised of experts external to the trial who review data, blinded or unblinded, as it accumulates.
Most DMSBs will include at least one independent statistician who is familiar with types of analyses as they apply to the therapeutic area or disease indication. One of the most important functions of this individual is to determine, along with other Board members, whether or not to stop a study and/or treatments. Inflation of the type I error (false positive conclusions) as well as possible bias in the future conduct of a clinical trial are the major methodological problems in connection with DSMB activities. Statistical methods exist which control the type I error whilst permitting early stopping of a clinical trial. Also the statistician can assess the potential impact of potential modifications of the study design (e.g. sample size adaption) and provide advice regarding appropriate statistical procedures to avoid an inflation of the type I error.
Biostatisticians who are not officially Board members can also play important roles in the establishment and support1) Assist with the id of DSMBs. They may:
1) Assist with the identification and selection of the DSMB members
2) Help develop the Charter in coordination with, and approval by, the DSMB Chairperson and Sponsor. The DSMB Charter describes the roles, responsibilities, functions, and activities of the DSMB and its members. The contents of a typical charter will include:
• Scope/Remit |
• Maintaining confidentiality |
3) Provide relevant materials to DSMB members prior to the meeting including reports, presentations and other relevant material, including confidential unblinded tables and reports from data files.
Biostatisticians who are involved in preparing unblinded material for DSMB members should have no direct involvement with the study. They may or may not be different than the official independent DSMB statistician.
At ICON, Biostatisticians have extensive experience of being both members of DSMBs and with establishing DSMBs.