Considerations for Establishing a DSMB
DSMB Composition
A DSMB is an advisory body that reviews available study data at pre-specified periods or in an ad hoc manner to make recommendations to the study sponsor as to whether to continue, modify, or stop a trial due to safety concerns, an early significant benefit of the study treatment, or treatment futility. However, the study sponsor is the final decision-making authority over clinical trial termination, changes, or extensions.
A Clinical Trial Transformation Initiative survey showed that study sponsors accept DSMB recommendations approximately 97% of the time.12 Such a high acceptance rate illustrates the value of DSMB review. It also demonstrates the importance of selecting and appointing the right DSMB members with a wide breadth of expertise, allowing them to make meaningful recommendations based on their safety and efficacy data review.
One of the first decisions sponsors must make when establishing a DSMB is determining how many board members are needed. The primary guidelines do not address this question, and a DSMB may vary in size depending on study needs. However, the consensus is that a DSMB must have at least three members.13 We most commonly work with DSMBs of three or four members. Large studies with complex designs or requiring more detailed monitoring and interim analyses may need larger committees and additional logistics planning for meetings.
Another common question from sponsors is whether to have an even or odd number of committee members to avoid voting impasses. This has been called “a fundamental misunderstanding of how a DMC should operate.”14
Major decisions about the clinical study conduct, such as a continuation or stop of enrollment, must be made by a unanimous consensus, not a vote majority. This approach provides a much more robust basis for these recommendations.14 If the members do not reach a consensus during the meeting, the DSMB may ask for an additional external expert opinion to be arranged soon after the meeting.
Members may still make recommendations for some issues by the majority of votes, and the DSMB Charter should clearly describe the procedure for handling an equal split of votes. We concur with this opinion and frequently manage DSMBs with an even number of members.
Due to the growing complexity of study designs and statistical methods, a DSMB should include clinicians and a biostatistician (who is usually a non-voting member).15 The biostatistician should also have experience in clinical trials, preferably within the same therapeutic area. For example, for a PSI-managed clinical trial enrolling pediatric cancer patients, the study DSMB included a biostatistician who had previously participated in pediatric clinical studies.
DSMB members should also have experience within the study’s indication. This is different from experience in the therapeutic area; as trials become increasingly complex, they require specialists with specific clinical expertise to be able to interpret the study data. For example, in a radiopharmaceutical trial in prostate cancer, an ideal board member would be a medical oncologist who is specifically familiar with radiopharmaceuticals in this indication. In addition, members with experience in causality assessments of adverse events, which may fall outside of the disease-specific expertise, can benefit the overall data quality of a trial.
Large, international trials usually require an international DSMB, with members representing each geographical region where the trial will occur. Board members don’t need to have previous experience with DSMBs or other types of safety review committees. While this kind of experience is helpful, it severely limits the potential candidates. However, the DSMB Chairperson position does require prior DSMB experience due to the nature of the role, which includes leading the meeting discussions, signing the meetings’ minutes and recommendations, and being responsible for communications with the sponsor. A potential solution to the challenge of eligible candidates with no prior experience is a bespoke training plan for all DSMB members, including a review of DSMB goals, responsibilities, processes, and reporting.
While choosing expert individuals for a DSMB is important, those with more professional experience may have conflicts or constraints to consider. Any member serving on a DSMB board must be independent of the study sponsor and uninterested in the trial’s outcome. Members should also be free of both financial and non-financial conflicts of interest, such as stock ownership in pharmaceutical companies or having served as an investigator for the study of interest. Finally, DSMB members should not have relationships that would result in significant financial, academic, intellectual, career, professional advancement, or other gains for themselves, their family members, or other close personal relationships based on the trial outcome.16
The general responsibility for establishing a DSMB, including appointing DSMB members, lies with the sponsoring company. They may either do it themselves or delegate to a CRO – partly or entirely. For instance, the sponsor may suggest desired experts but leave the logistics and additional member selection up to the CRO. Either method is acceptable from a regulatory or operational standpoint. However, instances where DSMB member selection is not completely outsourced may lead to potential recruitment biases.
To maintain DSMB independence, PSI recommends keeping all DSMB members entirely third-party to avoid potential conflicts of interest. The most common potential bias is financial – any DSMB member must be financially independent of the sponsor, meaning that each member should not receive any payment except remuneration for their participation or service working on a particular DSMB, including stocks or other investments with the sponsor company. This bias may even extend to any prior or current work for which a potential member has received payment.
Another common bias may occur if an expert is not involved in the referred study but is an investigator in another study within the program. This may mean that the potential candidate may be interested in the results of the particular study for which they may be a DSMB candidate.
To address these potential biases, PSI developed a database of DSMB members to access a pool of experts in various therapeutic indications from which we can select. Shifting the burden of the DSMB member selection process from the sponsor to a CRO with a comprehensive database of eligible members can reduce the timeline and the risk of undue influence over member selection.
The contracting process is an essential part of DSMB members’ appointments. These contracts outline the particular goals and responsibilities of a given DSMB and are different from investigator contracts. Typically, individual DSMB member contracts include a reference to the DSMB Charter, which describes in detail the members’ responsibilities and the procedural rules of the DSMB. They also contain provisions on confidentiality, conflict of interest, non-debarment, intellectual property, and data protection and define the honorarium amount per meeting or hour.
The DSMB contract template usually does not include indemnification provisions; if the DSMB member requests such clauses, the study sponsor issues a Letter of Indemnification. Because many DSMBs have a global role, the contracting party is the CRO responsible or the study sponsor unless otherwise recommended on a case-by-case basis. The PSI Legal Department prepares the DSMB contract templates for each study, obtains approval from the sponsor, and supports the legal negotiations of the individual contracts.
For a successful DSMB selection process, the following should be in place:
A search process for potential DSMB members, with a clear distribution of responsibilities and templates of pertinent forms readily available.
Established selection criteria tailored to particular study needs.
A database of potential DSMB members (including independent biostatisticians) that can be matched for the particular study.
A well-established process of contracting DSMB members, including contract templates covering all required areas.
All four are part of the PSI DSMB management process and are routinely used to select and contract DSMB members.
Next up: Considerations for Establishing a DSMB DSMB Charter and Meeting Schedule
