Feature Article

Minding your Bs and Ps: The Purpose of the Data Safety Monitoring Board and its Role in the Overall Data Safety Monitoring Plan
January 2009 Issue

Mary-Tara Roth, RN, MSN, MPH
GCRC Research Subject Advocate (RSA)

Author has nothing to disclose with regards to commercial support.

Educational Objectives:

At the end of this activity, participants should be able to:

• Identify the components of a DSMP;
• Describe the difference between a DSMP and a DSMB;
• Describe the purpose of a DSMB and how it fulfills this purpose;
• Identify two important documents related to the DSMB necessary for IRB protocol review.

PRINT | CLOSE WINDOW

• Introduction
• The Monitoring Plan (DSMP), and How DSMBs Fit In
• Some History
• What Does a DSMB Do?
• The DSMB Charter
• Your Study’s DSMB and the IRB
• Summary
• References
• Quiz

Introduction

The main obligation in the conduct of clinical research is that the safety, rights, and well-being of human research subjects are protected. This obligation is shared among all clinical research staff and by the institutions at which the research is conducted, but is the ultimate responsibility of the Principal Investigator at a research site and the overall Sponsor of the research study (sometimes this is the same individual). One major component to fulfilling this obligation includes defining and implementing a data safety monitoring plan (DSMP). (See CR Times Nov. 2006.) In some studies, this monitoring plan will call for a multidisciplinary group of approximately 3 to 8 clinical, biostatistical and clinical trials experts who are independent from the investigator and from the sponsor; they are charged with monitoring the accumulating data at periodic intervals and making recommendations regarding study conduct, including whether the study should be changed or stopped early.

The goal of this article is to provide an introduction about these monitoring groups. As a note, these groups go by several names. The NIH uses the term Data Safety Monitoring Board (DSMB). The FDA uses the term Data Monitoring Committee (DMC) because this is the common term internationally and is used in the ICH Good Clinical Practice Guidelines. It is not the name of the group which determines whether a study has an adequate monitoring plan, rather it is the defined relationship of the group to the sponsor and its purpose. If the group is independent from the sponsor and charged with reviewing the accumulating study data at periodic time points, then whatever the specific name chosen, this group is a DSMB for the purposes of any guidance, regulation or requirements regarding DSMBs. For simplicity, this article will use the term DSMB.

Back to the Monitoring Plan (DSMP), and How DSMBs Fit In

Sometimes, probably due to the similarity of the two acronyms, people confuse DSMPs with DSMBs. Briefly, at a minimum, EVERY study must have a DSMP (a monitoring plan).

Key elements of a DSMP for any study include:

• Description of risks involved in the study (and how will these risks be minimized);
• A plan for safety review (what is monitored, who monitors it and how frequently);
• Adverse Event definitions, grading (severity), and attribution (relatedness);
• A reporting plan for AEs and Unanticipated Problems: what gets reported, to whom, and in what timeframe;
• Description of the criteria for ending the research study (i.e., “stopping rules”).

Though EVERY study will have a DSMP, only a subset of studies will have a DSMB (an independent monitoring board) as one of several components of their safety monitoring plans. Though the NIH recommends that an “external [to the sponsor] monitoring body” be considered for any study (NIH 1998), studies that are more likely to need an external monitoring body (DSMB) will typically be higher risk, based on factors such as:

• the specific study interventions/procedures,
• population(s) being studied,
• the size and complexity of the study (i.e., a blinded multi-center study will be more complex in its operations and reporting than a single-center unblinded study),
• how much is known about the study interventions (conversely, how much is not known),
• number of subjectss,
• length of the trial (more subjects and a longer trial equals more exposure to the intervention and greater potential for AEs),
and
• the severity of endpoints (i.e., major morbidity or mortality).

Some History

The idea of having a separate committee monitor accumulating data for on-going clinical trials was first proposed in the mid-1960’s when the National Heart Institute commissioned an advisory committee to determine the best way to manage large trials, spurred in part by the complexity of managing the Coronary Drug Project (CDP) trial. The committee’s report pointed to the need for the monitoring of accumulating data by experts “who are not data contributing” to ensure the safety of study participants, and in 1968 this recommendation was implemented with the institution of a Safety Monitoring Committee for the CDP (DeMets, Furberg, & Friedman, 2006). More NIH guidance related to DSMPs and DSMBs followed in 1979, 1998, 1999, and 2000. The summary of these is that all clinical trials should have a provision for timely and effective monitoring by those with appropriate expertise and that this monitoring should be commensurate with the nature, risks, size, and complexity of the study. In regards to DSMBs, these guidance documents advise that DSMBs are required in Phase III clinical trials and may be appropriate in some early phase trials such as multicenter, blinded, or with particularly high-risk interventions or vulnerable populations.

Though the idea of DSMBs is not new, for a long time there was little formal guidance or regulation specifically related to their structure and operation, so there has been much variation in how DSMBs have functioned through the years (Gordon, Sugarman & Kass, 1998). However, each NIH institute and center has its own policies regarding data safety monitoring, and many do have very detailed recommendations regarding DSMBs. In addition, in recent years, several important detailed guidelines have been introduced regarding management and function of DSMBs specifically. In 2005, the World Health Organization (WHO) released its guideline document on monitoring committees with Operational Guidelines for the Establishment and Functioning of Data and Safety Monitoring Boards (WHO, 2005). Then in March, 2006, the FDA released its recommendations regarding monitoring committees titled: Guidance for Clinical Trial Sponsors: Establishment and Operation of Clinical Trial Data Monitoring Committees (FDA, 2006).

What Does a DSMB Do?

The DSMB helps to ensure that research participants are protected from harm. Harm to participants involved in research could stem from several factors. For example, the research intervention may be found to have more, or more dangerous, side effects than was previously known, so that the potential risks of the study come to outweigh the expected benefits. The intervention may be shown to be ineffective, so that continuing the study could be harmful, especially if there are potential side effects and/or there are other treatment options that the subject could consider. The intervention may be found to be effective before the study’s scheduled end, so that it would be unethical to continue subjects on placebo or standard treatment within the context of the study until its scheduled end. The study design (or the assumptions on which major aspects of the design are based) may be found to be inappropriate or inadequate to answer the primary study question(s) so that continuing the research would also be unethical (WHO, 2005). A second purpose of a DSMB is to ensure the integrity of the conduct of the study and in doing so, protecting the study subjects by ensuring that the data and results of the study are valid (Freemantle & Stocken, 2004; Hampton, 2000).

The DSMB fulfills its purpose in several ways. First, before the study begins, a DSMB will typically review the protocol and study design to ensure that the study is ethical, the research questions are relevant, and the plan to conduct and monitor the study are acceptable. Often, recommendations can be made at this early stage that improve the study moving forward (Ellenberg, Fleming, & DeMets, 2003). After the study begins enrolling subjects, DSMB looks at the data that have accumulated up to certain specified time points. The DSMB may assess both data accumulated from the trial and also any external data available on the intervention from other trials or from clinical practice, always focusing on the risk-benefit ratio to the study subjects (Broklehurst, Elbourne & Alfirevic, 2000; Meinert, 1998). In addition to data on the main study outcomes (safety and efficacy measures), most DSMBs also look at administrative/logistical study data such as the trial progress, recruitment, adherence to the protocol, data quality and timeliness of reporting, participant compliance to the study intervention, withdrawals, dropouts, loss to follow-up, etc. (DeMets & Califf, 2002; Wittes, 2004). Although not all of these administrative data and observations are linked directly to subject safety in the same way that AEs or measures of efficacy are, if these aspects have a detrimental effect on the study’s ability to answer the research question(s), then all of the subjects may be put at unnecessary risk without the benefit of advancing knowledge.

Unlike other monitoring specified in the DSMP (such as by a sponsor pharmacovigilance group, a sponsor monitor, the site PI/staff, the IRB, etc.), the DSMB has a privileged vantage point: it views comparative data (i.e., safety and efficacy data by intervention group) and typically this data is unblinded (Fleming, Ellenberg & DeMets, 2002). This enables those involved in the day-to-day conduct and management of the trial (sponsor, site PIs and staff, etc.) to remain blind and not be influenced by data trends associated with a specific arm of the trial. This knowledge among those conducting the trial could potentially affect how they interact with the research participants or follow the study procedures, thereby biasing the results of the study, and rendering the results questionable. The privileged vantage point of the DSMB enables it to assess whether the trial continues to satisfy ethical criteria required for the study to continue. Most often, the DSMB acts in an advisory capacity (making recommendations) rather than in an executive capacity (making decisions) (DAMOCLES Study Group, 2005). After meeting, the DSMB will issue its recommendations. In general, the DSMB will communicate its recommendations to the Study Sponsor or Steering Committee who then notifies the site PIs, who then notify their respective IRBs. The style of the DSMB's report is usually brief and to-the-point, noting when the DSMB met, on which study it met, its recommendations, and when the board is meeting again. Per Whitehead (1999) recommendations usually fall into the following four categories:

1. The study should continue without modification.
2. The study should continue with the following modifications...
3. The study should be stopped due to the following safety concerns...
4. The study should be stopped as a result of the formal stopping rule for efficacy...

The DSMB Charter

The DSMB is a formal entity, and as such follows a set of written policies that describes the roles, rules, and functioning of the DSMB. This is often referred to as the DSMB Charter. A typical Charter will include, but not be limited to details about the purpose of the DSMB, responsibilities of the members, the operation and format of DSMB meetings, monitoring guidelines, reporting to and from the DSMB, what data it will monitor and how the data is provided to it, and responsibilities of DSMB administrators. The Charter is often written by the sponsor, with input from other study leadership such as the steering Committee and Investigators. A DSMB may develop its own Charter, or, at the very least, the DSMB should review, amend if necessary, and approve the sponsor-written Charter as one of its first orders of business. Here's a link to several examples of DSMB Charter templates.

Your Study’s DSMB and the IRB

In regards to DSMBs and IRB submissions, there are a few important points to keep in mind. First, when you are submitting your protocol for review, you are required to develop a DSMP (detailed in Section H of INSPIR). If your DSMP includes a DSMB, you should attach the DSMB Charter to Section S of INSPIR. The Charter may be included in the study protocol, or it may be a separate document. Inclusion of the Charter enables the IRB to understand how the DSMB functions, including frequency of meetings, and so that it can make a determination on whether this is acceptable as a part of the overall DSMP.

Second, per NIH Guidance issued in June 1999, investigators of studies who have a DSMB as a part of their monitoring plan are required to submit DSMB reports to the IRB “as soon as they are received.” If you are the site investigator of a study that has a DSMB, you should understand how frequently the DSMB is scheduled to meet so you can comply with this requirement. It does happen that DSMBs do not meet when they are supposed to, or that a DSMB issues recommendations for study changes that are not communicated promptly by the site investigators to the IRB. Since the IRB approves a particular study on the basis of a specific monitoring plan, if that plan is not followed, this can have serious consequences for participant safety. Likewise, if the IRB is not notified of DSMB recommendations in a timely manner, this can also have serious consequences for participant safety.

Conclusion

Understanding the purpose and function of DSMBs, how a DSMB relates to the overall DSMP and reporting requirements for the IRB is crucial for studies that utilize DSMBs.

References

Click here and close window if you are a BUMC researcher
and would like to take the quiz now.

Close Window