1. High risk or vulnerable populations
2. Trials investigating new interventions with few or no safety data available
3. Trials with primary endpoints such as mortality or major morbidity
4. Randomized multicenter or international trials
5. Planned interim analyses with the potential for early termination
6. Trials with a large sample size
5.3 Organizing a DSMB
A DSMB typically consists of one or more clinicians knowledgeable in the field of investigation as well as biostatisticians or clinical trial methodologists . The expertise required for each DSMB varies according to the design and complexity of the trial. The clinicians could be physicians, nurses or other allied health specialists with expertise in the particular disease under study. They are responsible for reviewing the adverse event monitoring plans and serious adverse events. These experts also determine the appropriateness of measurements used for data reporting and can advise on quality control and/or assurance associated with laboratory or imaging measurements and instrumentation. Biostatisticians with experience in clinical trial design or clinical trial methodologists should be included in DSMB composition and tasked with review of the analysis plan and sequential analysis of interim trial data. Other specialists that may be valuable for some trials include clinical pharmacologists, bioethicists, and patient advocacy experts. Bioethicists on DSMBs provide expertise on informed consent and to ensure protection of vulnerable populations. A consumer or community advocate (often a current or former patient, a parent, or family member of someone with the disease) may also provide a helpful perspective .
Although there is no set number of members required for a DSMB, a general principle is that it should be as small as possible while encompassing all relevant expertise (odd numbers are preferred to avoid tie votes) . One of the individuals is selected to serve as DSMB chairperson; he/she is the point person for the DSMB and is responsible for overseeing the meetings. Although there is no official requirement for the selection of the DSMB chairman, it is desirable to have someone with previous DSMB experience who is an accomplished group leader with good interpersonal and organizational skills.
DSMB members are usually appointed either by principal investigators or trial sponsors, depending on the type of trial, the phase of trial, and sponsoring agency. For early-phase NIH-sponsored trials, the principal investigator typically appoints the members of the DSMB. For large, multicenter phase III and IV trials, the funding institute often nominates the board. For industry-sponsored trials, the pattern varies from either the company appointing the members or it adopts a “hands-off” approach and lets an external group do it. The members of DSMB should be independent of the sponsor of the study and manufacturer of the product evaluated. They should also not be related to the study’s investigative group. The DSMB members should have no financial, scientific, or any other conflict of interest with the trial. Absence of major conflict of interest is an essential requirement for all DSMB members to ensure the independence of the board. Financial conflicts (e.g., substantial stock holdings or yearly stipend from a sponsor) have received the most attention in the media, but other conflicts (e.g., strong intellectual investment or personal relationships) can also be important. For example, an individual whose initial concept or early work is being tested in the study could gain tremendously if the original research concept is validated and therefore may be hesitant to consider early termination of a trial in case of severe adverse events, compared to an individual who is completely independent. Similarly, the members of DSMB should not receive coauthorship on articles or promotions based on trial results because it creates conflict of interest.
5.4 Responsibilities of DSMB
The overarching purpose of a DSMB is to protect the safety of study participants and to give credibility and validity to the results of the study. The fundamental responsibility of every DSMB is to make recommendations to the trial’s sponsor and/or steering committee concerning the appropriateness of trial continuation. The DSMB has broad responsibilities which are summarized in Table 5.2.
Responsibilities of a data safety monitoring board
1. Review research protocol, informed consent documents, and plans for data safety and monitoring
2. Give priority to monitoring harm or adverse effects on study participants
3. Evaluate performance of individual centers including periodic assessments of data quality, participants recruitment, accrual, and retention
4. Analyze interim results and evidence of efficacy or adverse events
5. Consider any newly available external evidence such as scientific or therapeutic developments that can impact the ethics of the trial or safety of trial participants
6. Maintain confidentiality of the trial data
7. Report on the safety and progress of the trial
8. Make recommendations to the sponsor and investigators regarding continuation, modification, or
9. Assist in resolution of the problems reported by investigators
Based on the accumulating data and interim analysis, DSMBs typically make one of the following recommendations: continuation of the trial as planned; continuation of trial with modifications; early termination of the trial for either futility or adverse events; early termination of the trial based on unequivocal efficacy; or extension of the trial.
Commonly, DSMB is asked to review and approve the study protocol before initiation. Although DSMB is not responsible for study approval (which is done by an IRB), the purpose of this practice is to ensure that the board members are in agreement with the study investigators and sponsors about the acceptability of study design, mainly the statistical monitoring plan and any criteria for early termination. It is a delicate balance between protecting study participants and ensuring the reliability of the trial results when selecting criteria for early termination. For example, in a trial evaluating a potentially lifesaving treatment of a seriously ill patient, a less stringent criterion might be used to allow quick availability of a superior treatment. On the other hand, in evaluating a new vaccine, it is important to collect data over a longer duration to ensure that the vaccine can be safely administered to millions of healthy individuals and in such a scenario early termination is extremely rare. Occasionally, sponsors or investigators may ask DSMB to release interim data because of safety concerns arising in other investigational treatment trials or requests by regulatory authorities. Since DSMB protects the confidentiality of the interim data, it determines on an individual basis whether to grant such a request while safeguarding the integrity of the trial.
5.5 Workings of a DSMB
The DSMB functions with written operating guidelines which are typically prepared by the sponsor and approved by the DSMB during the first meeting [4, 9]. This document outlines the conduct and frequency of meetings, meeting formats, documentation and reporting procedures of all meetings, discussions and decisions, and ensures that IRBs are promptly informed of meeting recommendations. The number of interim analyses should be adequately planned and should be conducted on predefined outcomes. DSMB meetings can be divided into “open” or “closed” sessions, depending on the phase of the study. During the open session, study progress is reviewed, and therefore it can include study investigators, sponsors, monitors, and regulators who can answer questions raised by the DSMB related to the trial’s conduct. During the closed session, study efficacy and safety data is reviewed and typically includes only DSMB members. The objective is to avoid unblinding the investigators or industry sponsors and to prevent them from developing biases.
The DSMB receives two kinds of reports to evaluate, one regarding process and the other relating to study outcomes. The study outcome report consists of primary and secondary outcome variables as well as adverse events. This outcome report is provided to the DSMB members only. The process reports usually consist of study participant recruitment, accrual and retention status, data quality and timeliness and performance of trial sites. These process reports, including DSMB recommendations, are shared with trial investigators, sponsors, and regulatory authorities.
5.6 Statistical Issues Related to DSMBs
The DSMB must review accumulating data periodically in order to assess whether an important safety issue has arisen (e.g., severe side effects), or whether the intervention under study is showing a substantial beneficial effect earlier than expected. The required frequency of these reviews depends on the disease and the specific interventions. Most DSMBs hold meetings two to four times a year. While these interim reviews are conducted, the repeated statistical evaluation of data must be performed with caution, especially in the early stages of a trial when the number of participants and the number of events are relatively small. In addition to clinicians with expertise in relevant clinical specialties, most DSMBs have at least one biostatistician who is knowledgeable about statistical methods for clinical trials and sequential analysis of trial data.
5.6.1 Interim Analysis
An interim analysis is planned that intends to compare treatment arms at any time prior to formal completion of a trial. When interim analyses are necessary, they should be preplanned for early stopping (either for futility or for positive efficacy), or in the case of adaptive designs where a possible modification of the study design is based on unblinded interim data. Interim data, whether blinded or unblinded, should be analyzed by a statistical group that is independent of the sponsor and investigators—that is, the group is not otherwise involved in the trial design or conduct and has no financial connections to the sponsor or other trial organizers.