Data Monitoring Committees (DMCs) are groups of independent experts who periodically receive (by-arm) reports created by an independent Statistical Data Analysis Center (SDAC) using interim data from ongoing studies. The primary role of the DMC is to help protect the safety of the current and prospective patients in the study through assessment of risk\/benefit. They do so by making recommendations about the continuation of the study they are overseeing based on their best judgment and in accordance with guidelines specified in a DMC Charter.<\/p>\n
<\/p>\n
DMC members are typically independent from the sponsor of the study. They are generally practicing physicians at academic centers or professors in statistics or biostatistics departments. They may be retired, and no longer licensed to practice (but they should not be debarred from clinical practice or participation in clinical research). \u00a0As a group, they have experience in the disease being studied, available standard of care, expected side effects from the experimental treatment, clinical trials, responsible health authority guidelines and practices in the relevant regions, and the statistical methods to be used in the study.<\/p>\n
<\/p>\n
While DMC members will have a range of expertise, here we highlight 10 key qualifications important for all DMC<\/a> members.<\/p>\n <\/p>\n Clinicians on the DMC are expected to have many years of experience treating patients with the disease and its complications. The statistician on the DMC is expected to know the statistical approaches used in the study, both for the final analysis as well as any interim analyses. All DMC members should have at least some experience with the clinical trials process, perhaps by being part of an investigative site for previous studies or doing the statistical analysis of other clinical studies.<\/p>\n There can be \u2014 and should be \u2014 a range of DMC experience. There should be experienced DMC members (with one of them being the DMC chair), but there is also a need to mentor and grow future DMC members. Therefore, it is also reasonable to have a member with very limited (or no) DMC experience.<\/p>\n <\/p>\n Many clinical studies are conducted worldwide, or at least in multiple geographic regions. It is valuable that local knowledge is represented in the DMC<\/a>. For example, if a study is enrolling half the subjects in the Asia Pacific region, it would be advisable to have a DMC member located in that region or who is at least knowledgeable of the standard of care in that region and the possible impact of region-specific baseline characteristics and physiology on treatment effect.<\/p>\n <\/p>\n DMC members should be independent. That is not just financially independent, but they should also be intellectually, professionally, and regulatory independent. This may be difficult to assess. For example, one might question the independence of a DMC member who went to medical school with the CMO of the sponsor of the study and is still meeting socially, or who is working at an institution that is a participating site for the study.<\/p>\n DMC members should be open to disclosing potential conflicts of interest when being selected as a DMC member and any that arise during the study. These could be other DMC activities (perhaps with the same company, or a competitor), consulting activities (again, perhaps with the same company, or a competitor), and changes in employment. Note that not all potential conflicts will be significant enough to cause the removal of a DMC member.<\/p>\n The combination of both being independent and openly disclosing are important aspects of reassuring external groups (e.g., IRBs and regulatory agencies) of the integrity of the DMC<\/a> process.<\/p>\n <\/p>\n When discussing topics with the study team, DMC members need to maintain a neutral stance and not give away their inclinations by way of their body language or how questions are asked or answered. It is essential that the DMC does not convey any sense of by-arm results unintentionally.<\/p>\n Confidentiality of the DMC is key. No other group has this highly important by-arm data. The DMC must make sure that any files received are kept confidential. And the DMC must remain tight-lipped if chatting formally or informally with anyone, for example at scientific conferences or with work colleagues. Even side discussions between individual DMC members are discouraged. It is preferred that all DMC discussions take place formally in the DMC meeting with quorum established and where formal minutes are taken.<\/p>\n <\/p>\n DMC members must understand how to effectively participate in virtual meetings using various platforms (e.g., using video, muting themselves, \u201craising hand,\u201d etc.). The SDAC typically will also provide a portal to access documents for review for DMC<\/a> meetings and the DMC members should be comfortable accessing, editing, and uploading. Similarly, using an electronic signature may be needed by the DMC Chair or the full DMC. In the future, there could be more use of interactive tools for data review provided directly to the DMC or facilitated by the SDAC reporting statistician that could require other technical skills.<\/p>\n <\/p>\n The SDAC will send documents for review (e.g., DMC charter, meeting minutes) and expect relatively quick (several business day) turnaround from the DMC. Similarly, the SDAC typically will send scheduling polls and expect quick turnaround on those. Sometimes potential DMC members who otherwise would be excellent DMC members (perhaps key opinion leaders) are too busy to attend to these mundane but important logistic requests.<\/p>\n More critical, if an emergency comes up in the study (for example, a death in a study where no deaths were expected) there might be a need for an immediate review by the DMC and perhaps an ad hoc DMC meeting within a week. If it is generally impossible to get time on a DMC<\/a> member\u2019s calendar with less than a month of advance lead time, then perhaps that person would not be appropriate to have on the DMC.<\/p>\n <\/p>\n It is expected that the DMC members will take their role seriously and invest the proper time in it. When materials are provided by the SDAC a week in advance of the meeting, the DMC members should make sure that they dedicate sufficient time in advance to be prepared for discussion. It is important that each DMC member attends the full duration of the DMC meeting without last moment cancellation or early departure unless truly an emergency arises.<\/p>\n The DMC must take the discussions seriously \u2014 weighing risk\/benefit of the current and prospective patients. The DMC<\/a> also may enter into serious discussion to assess over- vs. under-reacting to imbalances in events. It requires deliberate thought to do this \u2014 assessing risk\/benefit, while acknowledging the uncertainty of data that is developing in the ongoing study.<\/p>\n There is a limited time for DMC meetings. It is important that the DMC not get overly off topic on matters that are not core to their responsibility to form a recommendation that will help protect the ongoing safety of current and potential patients. A DMC member might ponder \u201cwouldn\u2019t it be interesting to get this new table at the next meeting\u201d or \u201cwhat if the final analysis used this new statistical method I developed,\u201d but that could be distracting from the primary purpose of the DMC.<\/p>\n <\/p>\n Some less experienced DMC members fall into a trap of envisioning themselves as the treating physician in individual cases \u2014 for example, asking why certain tests were or were not done, or why it was deemed related or not to the experimental treatment. Looking at individual cases in detail can sometimes be appropriate in smaller, earlier phase studies, or when the event (e.g., SAE, death, or SUSAR) is particularly surprising or severe. But for most later phase studies, the value of the DMC is looking at group differences and the DMC is encouraged to focus on those results which are uniquely available to the DMC via the SDAC.<\/p>\n The DMC should certainly look carefully at the group comparisons. But the DMC must also be aware \u2014 perhaps aided by thoughts from the DMC statistician and the SDAC reporting statistician \u2014 of the hazards of multiple comparisons. If the DMC report has 200 pages, with thousands of comparisons in total, there will assuredly be some \u201cinteresting\u201d imbalances that are the result of chance alone. It is one of the trickiest aspects of DMC review to decide how to evaluate imbalances knowing that some might be a result from repeated looks at many, many comparisons. Similar caution is needed to assess the credibility of heterogeneous subgroup results.<\/p>\n The DMC must understand that interim data is used for reports, and it is unrealistic to assume perfectly clean data and perfectly clean outputs will be available. This adds to the difficulty of reviewing these outputs, and again shows the need for an experienced DMC member to be present.<\/p>\n <\/p>\n Even if the DMC<\/a> member is responsive and has reviewed material in advance, it is important that the DMC member share their thoughts during the meeting. DMC membership is usually 3\u20135 members, and assuredly each member will bring unique experiences and reactions, but that is useless unless the rest of the DMC hears those thoughts. Hopefully the DMC chair or the SDAC reporting statistician can help elicit feedback from all DMC members, even if hesitant to speak up. Even DMC members that speak up must also be good listeners to what the quieter members have to say.<\/p>\n More questions than answers may arise in the DMC\u2019s deliberations. Collectively, the DMC needs to be able to reach a consensus as to whether it has adequate information to make a recommendation, or whether additional data or additional analyses of existing data are needed.<\/p>\n <\/p>\n The DMC should (collectively) set the tone for what materials they receive and how the DMC meetings proceed. For example, less experienced sponsor study teams might go into detail in an open session on every SAE, which has occurred (using slides that were available prior to the meeting), thinking that is helpful to the DMC. The DMC should feel empowered to redirect the sponsor\u2019s medical monitor and emphasize that the limited time of the DMC meeting would be better spent in closed session and just check to see if the DMC has any questions on particular SAEs based on their pre-review of the materials.<\/p>\n Similarly, if the DMC is not getting statistical outputs in the way they want, they should feel emboldened to request new or improved outputs from the SDAC. That is particularly true if the DMC<\/a> is being provided outputs that are initially more suited for final analysis of the study, but not fine-tuned for the purposes of DMC review (e.g., initially thousands of pages of outputs but lacking any useful figures).<\/p>\n The DMC charter will specify the DMC\u2019s responsibility on many matters. It is important that the DMC advocate at an organizational meeting on aspects they feel are important, such as having full access to treatment assignment from the outset (Active vs. Placebo, instead of A vs. B), having access to efficacy data upon request even outside of formal interim analyses, and the use of non-binding assessments based on totality of data rather than binding assessments for those formal interim analyses.<\/p>\n <\/p>\n1. Range in DMC experience<\/h3>\n
2. Geographical representation<\/h3>\n
3. Independent and open to disclosure<\/h3>\n
4. Discrete<\/h3>\n
5. Tech savvy<\/h3>\n
6. Responsive<\/h3>\n
7. Serious and focused<\/h3>\n
8. Thoughtful review of groups vs. individuals<\/h3>\n
9. Actively participating (and listening)<\/h3>\n
10. Authoritative (but collaborative)<\/h3>\n