Clinical trials dsmb

Cutlip and Dr. Tcheng contributed equally to this article. Ellenberg, T. Fleming, D. Fleming, and D. Chartering a Data Safety Monitoring Board. July 1, Download a DSMB template charter to use in your own clinical trials here.

Photography: Getty Images Board responsibilities. These components include: Delineation of the roles and responsibilities of the DSMB Membership requirements inclusive of qualifications related to independence and conflict of interest Meeting schedule and format Provisions for both open and closed DSMB meetings Safety and efficacy data reporting and the analysis plan Delineation of predefined criteria for early termination Communication channels between the DSMB and study sponsor Confidentiality requirements.

At the start of each new member's term, the individual should sign a confidentiality statement promising not to disclose any proprietary and nonproprietary data or deliberations of the DSMB. The Chairperson of the DSMB should be selected among the voting members and have previous experience in monitoring clinical trials. The Chairperson should be a good facilitator, communicator, and consensus builder.

The Executive Secretary may not vote or be present during closed or executive sessions of the DSMB, and must maintain all information reviewed, discussed and recorded during DSMB meetings with strict confidentiality. Non-voting, non-member ex officio attendees may participate in DSMB meetings to provide information and answer questions as needed, e. Ex officio attendees must be limited in number and are not permitted to attend the closed or executive sessions of DSMB meetings.

The clinical trial sponsor or principal investigator may draft this charter and present it to the DSMB for agreement, or the DSMB may draft the charter with subsequent concurrence by the clinical trial sponsor or principal investigator.

The charter should at a minimum specify: the meeting schedule and format, the format for presentation of data, the individuals who may attend all or part of the DSMB meetings, the individuals who will have access to interim data, the method and timing for providing interim reports to the DSMB, procedures for assessing conflict of interest of potential DSMB members, and other issues relevant to committee operations.

The frequency and format of DSMB meetings depends on the nature and risk of the studies to be monitored. Meetings may be face-to-face or by teleconference. The Board shall have the option for expedited meetings to review unexpected Serious Adverse Events [3] or other urgent issues that may arise during the course of the trial. Unscheduled meetings may be recommended and initiated by the DSMB Chairperson, the clinical trial sponsor, or the principal investigator.

Clinical trial documentation and data should be available to the Board members at least two weeks prior to the meeting. The purpose of the open session is to provide relevant information to the Board about general aspects of the trial.

The open session may focus on: the background of the study, the protocol, status of the study, problems with accrual and follow-up, baseline demographic data, compliance issues, frequency of adverse events [4] , documentation of endpoints, data quality issues, flow of forms, data based protocol modification issues, external monitoring of coordinating center operations in multicenter trials, and any other study-related issues that can be discussed without reference to interim comparative results.

The principal investigator, co-investigators, and statisticians may attend the open session and present information during the meeting. The issue should be addressed in the DSMB charter. The NIAAA program official involved with the study should be informed of upcoming board meetings at least two weeks in advance, and receive the appropriate meeting materials at the same time as the Board members. During the closed session, the DSMB reviews and votes on all issues. This session is usually attended by the DSMB members only.

The principal investigator or clinical trial sponsor may attend the closed session at the request of the DSMB. During the closed session, the discussions should focus on: treatment safety, efficacy data, whether the primary study question has been answered, the interim results by treatment arm usually masked , determination of when study data may be released, review of requests for access to the results of the interim analysis, and results of Board actions and recommendations made in the previous meeting.

During these sessions, the Board may discuss any unmasked analysis of a blinded clinical trial and other sensitive issues related to the clinical trial. The upper boundary implies that any treatment that is highly likely to incur an increased risk over standard treatment is of concern and may not warrant further development. It should be noted that the monitoring guidelines discussed above for the SPRT and the TS methods are for the general cases.

When monitoring safety events, the lower boundaries are usually not relevant and can be ignored. Trials will be stopped when it is considered unsafe to continue. However, whether to stop a trial could depend on many factors, such as other safety observations and efficacy in the context of risk-benefit trade-off.

In practice, monitoring guidelines may be considered after a minimum number of subjects have been treated to ensure reliable estimates can be obtained [ 18 ]. We recommend close collaboration with clinical experts in soliciting realistic input. Computer simulation of potential outcomes given various scenarios can also help make a decision. For the randomized, controlled clinical trials, we consider similar explicit statistical boundaries for safety monitoring.

We adopt a Bayesian Beta-Binomial model in setting up the monitoring. The posterior distributions can be expressed as:. In situations when additional information regarding exposure in the form of person-time is available, Bayesian methods with the Gamma-Poisson model may be applied instead of the Beta-Binomial model.

We apply the methods discussed in Section 3. Subjects will be randomized in a ratio to the investigational drug arm or the control arm. Based on previous knowledge about the disease and the biological mechanism of action of the investigational drug, the team feels that a serious AESI is a potential risk and requires heightened surveillance.

The trial has an independent DSMB. We first consider a situation where the sponsor monitors the event of the AESI from the combined treatment arms. By closely monitoring the overall event rate in the combined arms, the sponsor may be able to decide whether an ad hoc DSMB meeting is warranted. If the pre-determined monitoring boundary is crossed, the team may call the DSMB into an ad hoc session to avoid any delays in waiting for the next scheduled DSMB meeting.

The boundary values for the first 10 events are shown in Table 1. For example, if three events of the AESI are observed in 11 or fewer subjects, the boundary is crossed and the team may consider convening an ad hoc DSMB meeting. The boundary values for SPRT can be obtained from the online calculation tool [ 19 ]. In addition to the SPRT, Bayesian methods can be used if some knowledge is available to form the prior distributions.

Note that the distribution of Beta 3, 57 represents information from prior experiences equivalent to three events from 60 subjects.

Similarly, Beta 3, 11 represents prior information equivalent to three events out of 14 subjects. Our choices for the prior distributions assume more knowledge about the control group from historical data, but less knowledge about the experimental group given limited experience. The derivation of the Bayesian safety boundary follows Section 3. The boundary values based on the Bayesian method for the first 10 events are shown in Table 1. They are also plotted in Figure 1. It can be seen that the SPRT method has boundary values that are easier to cross in declaring a safety concern than the Bayesian method, except when only two events are observed.

We now consider setting up the monitoring boundary for the DSMB who has access to the actual treatment assignments. The Bayesian boundaries when monitoring the two arms cannot be easily plotted, but the DSMB can use a posterior probability table as a reference. In this example, there are eight subjects in the treatment arm and 11 subjects in the control arm at the analysis time.

Table 2 shows the posterior probability table that summarizes the values of the criterion function under all possible scenarios. In this table, all the bolded numbers highlight the scenarios in which the stopping boundary is crossed. We have discussed several statistical methods that can be applied to monitor ongoing clinical trials in either a blinded or an unblinded fashion.

In addition, Bayesian methods allow sponsors to take advantage of information originating from multiple sources both internal and external to the trial. This is an important advantage, as safety signals identified in clinical trials alone may be limited. The globalization of clinical trials has posed additional challenges. A great deal of coordination is required of sponsors to ensure timely communication of new safety findings among all stakeholders in all regions.

Efforts in building a standard safety data warehouse across all trials in a development program will lay a solid foundation for integrated safety analyses. Innovative statistical methods can be applied to increase the efficiency in reviewing a large volume of safety data, to identify safety trends and to establish prospective monitoring guidelines, as described in this article.

National Center for Biotechnology Information , U. Journal List Pharmaceutics v. Published online Jan Amy Xia. Author information Article notes Copyright and License information Disclaimer. Amgen Inc. This article has been cited by other articles in PMC.

Abstract Monitoring patient safety during clinical trials is a critical component throughout the drug development life-cycle. Introduction Clinical trials provide the evidentiary basis for regulatory approvals of safe and effective medicines. Common Practice in Safety Monitoring 2. Stakeholders in Safety Monitoring 2. Sponsor Clinical trial sponsors, usually pharmaceutical companies, are responsible for developing the clinical trial protocol.

Subjects Subjects are patients or healthy volunteers who agree to participate in a clinical trial and have signed the ICF. Investigators Investigators are qualified individuals who are trained and experienced to provide medical care to subjects enrolled in the trial.

Medical Community and Patients Clinical trials generate data that contribute to the body of knowledge about the treatment and the disease that benefit the broader medical community and, ultimately, the patients. Communicating Safety Information among Stakeholders Timely communication among the various stakeholders is critical to ensure subject safety in clinical trials.

Statistical Methods in Safety Monitoring 3. Methods for Single Arm Trials With the continued focus on safety monitoring and increased amount of work on case processing and reporting, quantitative approaches to safety evaluations will become increasingly important.

Methods for Randomized, Controlled Trials For the randomized, controlled clinical trials, we consider similar explicit statistical boundaries for safety monitoring.

Open in a separate window. Figure 1. Conclusions Monitoring patient safety during clinical trials is a critical component throughout the drug development life-cycle. Conflict of Interest The authors declare no conflict of interest.