In June 1998, the US National Institutes of Health (NIH) issued a policy for data and safety monitoring, which stated ‘It is the policy of the NIH that each Institute and Center (IC) should have a system for the appropriate oversight and monitoring of the conduct of clinical trials to ensure the safety of participants and the validity and integrity of data for all NIH-supported or conducted clinical trials.’ The policy provides flexibility so that data and safety monitoring can be implemented in the manner appropriate for different clinical research activities. Monitoring exists in a continuum from ongoing review by the principal investigator or NIH programme staff in a small phase I study to the establishment of an independent data safety and monitoring board (DSMB) for a large phase III clinical trial.
Although studies may be monitored by a variety of mechanisms and methods, the NIH has mandated that every randomized clinical trial must have an independent DSMB. This short report describes the composition and role of the DSMB for the NIMH Collaborative HIV/STD Prevention Trial.
Components of data safety and monitoring
Data and safety monitoring includes the following four important components:
The principal investigator of each study must include a data and safety monitoring plan in each new protocol.
The institutional review board must approve the plan and determine what types of safety monitoring processes are required (e.g. principal investigator monitoring only, a single independent monitor, a group appointed by the principal investigator or an independent DSMB).
The NIH Institute or Center is responsible for appointing an independent monitor or convening a DSMB.
The study principal investigator is responsible for providing all required data to the individual monitor or to the DSMB and for acting on the findings of any monitor or the DSMB.
This Trial has engaged in all four of these components.
Membership of the data safety and monitoring board
The first step in establishing the DSMB for this Trial was to identify the areas of expertise needed to monitor all aspects of the Trial effectively. The following expertise was deemed critical: prevention science, ethnography, infectious diseases (especially HIV and sexually transmitted diseases), laboratory diagnostics, clinical practice, methodology, international trial experience, statistics, and ethics.
National Institute of Mental Health (NIMH) staff identified and appointed to the DSMB members who represent the areas of expertise listed above. Some members have expertise in more than one of the areas, and no member of the DSMB has direct involvement in the study or any conflict of interest with the investigators. The nine members of the DSMB are acknowledged in this issue.
Role of the data safety and monitoring board
As specified by NIH policy, the DSMB assessed the overall study, including the protocol and the performance of the staff implementing the protocol, for any concern about plans or implementation. The DSMB also reviewed cumulative study data to evaluate the safety of study participants, the ongoing conduct of the study, and the scientific validity and integrity of the Trial. Because the Trial was international in scope, the DSMB was especially concerned with the effects of cultural differences on study implementation and fidelity.
The members of the DSMB closely reviewed the quality assurance of the biospecimens (from the collection of the specimens to analysis), the reporting of results to participants, and the treatment of participants with positive laboratory results. As is common in international studies, this was of special importance because some of the sites where laboratory capacity was not available were establishing or upgrading laboratories for the Trial. Also, members realized that, in addition to the normal delays encountered with domestic shipments, shipping specimens halfway round the world for quality assurance involves special difficulties related to distance, various customs clearances, and permission for shipping from foreign governments.
In particular, the DSMB reviewed: Trial protocols for appropriateness in achieving goals, including sample size assumptions and randomization procedures; epidemiological data to verify the selection of appropriate study populations and venues; adherence to the Trial protocol; data for evidence of study-related adverse events; data for evidence of efficacy in accordance with the approved statistical plan; data quality, completeness, and timeliness; quality assurance data that might identify problems with Trial procedures; and factors that might compromise the confidentiality of the Trial data.
As the Trial progressed, the data coordinating center (DCC) provided data to the DSMB in the following categories: epidemiological data; baseline data; 12-month follow-up data; 24-month follow-up data; process data; adverse events; and hypothesis testing (interim and final). See Table 1 for a description of the types of data that were presented over time. These data were used by the DSMB to evaluate the likelihood of the trial having significant results on the basis of the recruitment and retention rates and the data safety procedures.
The safety of the research subjects was evaluated by multiple methods and multiple groups including site staff, the Trial steering committee, the DSMB, and institutional review boards. The DSMB evaluated safety by carefully evaluating the protocol that described every planned activity in detail. This group also had access to the reports from the central quality assurance monitors from the DCC that evaluated each component of the Trial (human subjects, data collection and assessment, biological specimen collection, intervention) using structured evaluation forms during periodic site visits. Finally, the DSMB had access to listings of both protocol violations and reported adverse events. Of the eight adverse events reported to date, four involved robbery of a field office or a staff member, two involved incorrectly giving test results to participants, and two involved research participants being harassed or assaulted.
Meetings of the data safety and monitoring board
Based on progress of the Trial, the frequency of the DSMB meetings varied. The DSMB first met shortly after the first draft of the protocol was available. While pilot studies were being conducted and the Trial protocol was being revised, the DSMB met approximately three times a year to provide timely feedback. Once the Trial protocol was approved and the study was in the field, the DSMB met approximately once a year, sometimes by conference call.
In advance of each meeting or call, the NIMH staff in consultation with the DSMB chair prepared a set of questions to be addressed. The specific questions being addressed varied depending on the phase of the study being evaluated. As an example, the questions contemplated during the DSMB meeting that reviewed the full protocol on 17 March 2000 are presented in Table 2.
The DSMB meetings generally lasted approximately a day and included both closed and open sessions within the day. Materials required for the discussion were provided to the members of the DSMB approximately a week before the meeting. Meetings began with a closed session attended by the DSMB members and NIMH staff not involved in the study. During this portion of the meeting, the members discussed major areas of concern and developed the questions they wanted to ask members of the trial steering committee.
This closed session was often followed by an open session during which members of the steering committee met with the DSMB, in person or by telephone, to make brief presentations or to respond to questions. Finally, the DSMB reconvened in a closed session to review both the interim data provided by the DCC and information provided by the steering committee and to formulate a report to the NIMH. Confidential study documents were collected from members at the end of the meeting and were subsequently destroyed.
Confidentiality was an important aspect of these meetings. Voting members of the DSMB viewed interim analyses and individual adverse event data. Therefore, discussions from the meeting were not revealed outside of the meeting room.
Reports of the data safety and monitoring board
After each meeting, the DSMB provided the NIMH with a written report responding to the questions posed and recommending whether the study should continue without change, be modified, or be terminated. Recommendations regarding design modification and study conduct were forwarded to the Trial steering committee, which carefully considered each recommendation, made necessary modifications to the protocol, and provided a written response to the NIMH and the DSMB describing how each issue had been addressed.
In the course of these reports, the DSMB made a number of recommendations that strengthened the Trial. One of the most striking was the recommendation to conduct initial epidemiological studies of the venues selected for the intervention to verify risk and to establish intraclass correlation coefficients that could be used to calculate appropriate sample sizes for the Trial. Although ethnographic findings had indicated that each of the sites had initially identified risky populations and venues in which to conduct the Trial, the epidemiological study results indicated that the venues and populations initially selected in India and Peru did not have the characteristics required for the study. The DSMB required these two countries to refine their choice of venues and populations and conduct a second epidemiological study before beginning the Trial. The implementation of this recommendation from the DSMB greatly strengthened the Trial and enabled sample size calculations to be based on reliable data.
In conclusion, DSMB play a critical role in the conduct of phase III clinical trials sponsored or conducted by the US NIH. Because members of the DSMB have the expertise required to monitor the Trial, are not involved in the daily management of the Trial, and can review interim analyses and adverse event reports, they are in an excellent position to provide expert advice that will ensure that the goals of the Trial are achieved and that NIH funds are well invested. DSMB review and recommendations for this Trial strengthened both the study design and the procedures used for specific activities.
National Institutes of Health policy statements
The NIH has published a number of policy statements related to data and safety monitoring in the NIH guide. These statements guided the formation and operation of the DSMB for this Trial and provide useful suggestions for studies funded through other sources. The most recent of these documents are available on the Web. See Table 3 for the Web addresses of relevant documents.