Data and Safety Monitoring Board Guidelines for NCCAM-Supported Clinical Trials
On this page:
- Board Activities
- Board Process
- Mailings to the DSMB
- Access to Interim Data
It is National Institutes of Health (NIH) policy that all clinical trials require data and safety monitoring, with the method and degree of monitoring being commensurate with the risks (NIH Guide, June 12, 1998, Notice 98-084: NIH POLICY FOR DATA AND SAFETY MONITORING). The following document summarizes National Center for Complementary and Alternative Medicine (NCCAM) guidelines concerning the establishment, responsibilities, and operating procedure of data safety and monitoring boards (DSMB).
For Phase III trials, as well as for some masked Phase II therapeutic trials, NCCAM requires that a DSMB perform monitoring functions. The PI should confer with the NCCAM Program Director prior to funding to discuss whether a DSMB will be required. This DSMB would be composed of experts relevant to the study who would regularly assess the trial and offer recommendations to the NCCAM concerning trial continuation. The remainder of this document addresses DSMB activities, including its establishment, responsibilities, and operating procedures.
The independent Data and Safety Monitoring Board (DSMB) plays a crucial role in ensuring the safety and welfare of patients enrolled in clinical trials. Funds to support the functions and operations of the DSMB will be provided by NCCAM in a fashion appropriate for the trial award mechanism (e.g., grant, contract, cooperative agreement). In order to increase efficiency and cost-effectiveness, a single DSMB will often monitor multiple trials.
At periodic intervals during the course of the trial, the DSMB responsibilities are to
- Review the research protocol, informed consent documents, and plans for data safety and monitoring;
- Evaluate the progress of intervention trial(s), including periodic assessments of data quality and timeliness, participant recruitment, accrual and retention, participant risk versus benefit, performance of the trial site(s), and other factors that can affect study outcome;
- Consider factors external to the study when relevant, such as scientific or therapeutic developments that may have an impact on the safety of the participants or the ethics of the trial;
- Protect the safety of the study participants;
- Report on the safety and scientific progress of the trial;
- Make recommendations to the PI, NCCAM, and, if required, to the FDA concerning continuation, termination, or other modifications of the trial based on the observed beneficial or adverse effects of the treatment under study;
- If appropriate, conduct interim analysis of efficacy in accordance with stopping rules that are clearly defined in advance of data analysis. In most cases, these stopping rules are proposed by the PI and approved by the DSMB with modification as necessary to ensure patient safety.
- Ensure the confidentiality of the trial data and the results of monitoring; and
- Assist NCCAM by commenting on any problems with study conduct, enrollment, sample size, and/or data collection.
Members of the DSMB are appointed by, and report to, the Director of NCCAM, after consultation with the PI. Members should be completely independent of the trial investigators and have no financial, scientific, or other conflict of interest with the trial as described in the NIH Grants Policy Statement, Part II, section 12, and in the Combined Federal Register announcement - 45CFR94. Written documentation attesting to absence of conflict of interest is required; potential conflicts that develop during a member's tenure on a DSMB must be disclosed to the NCCAM in a timely fashion.
Current or past collaborators or associates of the PI are not eligible to serve in the DSMB. Members may be from within either the awardee institution or collaborating institutions, but the majority of voting members should not be affiliated with these institutions.
Disciplines represented on the DSMB should include experts in or representatives of the relevant clinical practice, the relevant complementary and alternative medicine modality, clinical trials methodology, and biostatistics. Additional DSMB members should include a medical ethicist and a public ombudsman. An NCCAM official(s) may serve as an ex-officio member(s) of the DSMB. NCCAM Program Directors may fill this role if they are not in conflict with the trial (e.g., cooperative agreements and contracts).
A Chairperson will be selected prior to the first meeting. While the Chair will lead all meetings of the DSMB, the Chair and the NCCAM-appointed DSMB Executive Secretary will cooperate on development and execution of all logistic activities associated with the DSMB including 1) oversight and coordination of all meetings and communications; 2) development of the agenda; 3) identification and review of substantive materials; and 4) production of meeting minutes and reports. In addition, the Executive Secretary ensures that the DSMB complies with all NIH and PHS policies and procedures concerning DSMB activities.
The first meeting will take place face-to-face before initiation of the trial to discuss the protocol and to establish guidelines to monitor the study. The Chair, with input from the PI and NCCAM staff, will prepare the initial agenda, which should address issues concerning commencement of the trial. The initial tasks of the DSMB are to review the entire protocol and informed consent forms with regard to subject safety; identify needs for protocol modification; and, after receipt of a satisfactory protocol, recommend to the NCCAM initiation of expenditure of project funds. The DSMB will also identify the relevant data parameters to be reported by the PI to the DSMB and the format of these data.
Following the initial meeting, the DSMB should convene at least twice a year to review accumulated data on safety and, if appropriate, conduct interim analysis. Meetings may be convened as conference calls, as well as in person, although the initial meeting and meetings to discuss interim analysis will be face-to-face. An emergency meeting of the Board may be called at any time by the Chairperson or by NCCAM should a question of patient safety arise.
An appropriate format for DSMB meetings consists of a briefing session, a closed session, and, if appropriate, an executive session. The briefing sessions may be attended by the PI and other trial investigators, by institution staff, and by NCCAM staff and other appropriate federal officials, but should always include the study biostatistician. Issues discussed at briefing sessions usually include conduct and progress of the study, including patient accrual, compliance with protocol, general safety issues, and any other problems encountered. Patient-specific data and treatment group data may not be presented in the briefing session.
The closed session is normally attended only by DSMB members and, if required, by the trial biostatistician; other NCCAM staff may attend as appropriate. All safety and efficacy data are presented at this session. The DSMB may choose to retain the study blind by assigning study accrual data to groups A, B, etc. Alternatively the DSMB may choose to review unblinded data. Generally, these data would be prepared by and presented to the DSMB by the trial biostatistician. The discussion at the closed session is completely confidential.
A final executive session involving only DSMB members is held to discuss the general conduct of the trial and all outcome results, including adverse events. Recommendations are developed and votes taken as necessary.
Should the Board recommend termination of the trial, a full vote of the Board will be required. In the event of a split vote, a majority vote will rule and a minority report should be appended.
Interim Reports: Interim reports are generally prepared by the study biostatistician(s) and distributed to the DSMB, preferably at least 5 days prior to a scheduled meeting. These interim reports are numbered and provided in sealed envelopes within an express mailing package. The contents of the report are determined by the DSMB. Additions and other modifications to these reports may be directed by the DSMB on a one-time or continuing basis. Interim data reports generally consist of two parts: Part 1 (Open Session Report) provides information on study aspects such as accrual, baseline characteristics, and other general information on study status. Part 2 (Closed Session/Executive Session Report) may contain data on study outcomes, including safety data and, depending on the study, perhaps efficacy data. The Closed Session Report is considered confidential.
Copies distributed prior to and during a meeting are collected by the study biostatistician(s) following the meeting. Data files to be used for interim analyses should have undergone established editing procedures to the extent possible. Interim analyses of efficacy data are performed only if they are specified and approved in advance and criteria for possible stopping is clearly defined.
Reports from the DSMB: A formal report from the Chair and approved by the DSMB should be supplied to the PI, the grantee institution, the NCCAM, and, if appropriate, the FDA within 6 weeks of each meeting. Each report should conclude with a recommendation to either 1) continue the study without changes; 2) continue the study with changes; or 3) terminate the study. This recommendation should be made by formal majority vote and fully justified by the preceding report. These recommendations should be transmitted to the PI, a business official of the grantee institution, the NCCAM, and the FDA (if appropriate) as rapidly as possible by the Executive Secretary, by immediate telephone and telefax if sufficiently urgent (such as the case of trial termination). In the event of a split vote in favor of continuation, a minority report should be contained within the regular DSMB report. The report should not include unmasked data, discussion of the unmasked data, etc.
In the unlikely situation that the trial PI does not concur with the DSMB, then the Director, NCCAM must be informed of the reason for the disagreement. The trial PI, DSMB Chair, and appropriate NCCAM staff will be responsible for reaching a mutually acceptable decision.
Mailings to the DSMB
On a scheduled basis (as agreed upon by the DSMB), masked safety data should be communicated to all DSMB members or to the one member who serves as the designated safety officer. Any concerns noted should be brought to the attention of the Chair or designated safety officer, who will take appropriate action.
Access to Interim Data
Access to the accumulating endpoint data should be limited to as small a group as possible. Limiting access of interim data to the DSMB relieves the investigators of the burden of deciding whether it is ethical to continue to randomize patients and helps protect the study from bias in patient entry and/or evaluation.
All materials, discussions, and proceedings of the DSMB are completely confidential. Members and other participants in DSMB meetings are expected to maintain confidentiality. Each member of the DSMB, including nonvoting members, must sign a statement of confidentiality.