Preface to the First Edition
The randomized clinical trial has been recognized as the gold standard for evaluation of medical interventions for only half a century (Doll, 1998). Over the past several decades, the increasingly central position of randomized clinical trials in medical research has led to continual advances in the development of methodology for the design, conduct, and analysis of these studies. An enormous body of literature relating to clinical trials methodology is now available, a professional society focusing on clinical trials has been established (Roth, 1980; www.sctweb.org), and a large number of statisticians, clinicians, and epidemiologists consider clinical trials as their primary area of research and/or application.
One area of clinical trials that has received relatively little attention but that can be critical to the ethics, efficiency, integrity, and credibility of clinical trials and the conclusions of such trials is the process of interim monitoring of the accumulating data. To an increasing extent, interim monitoring is becoming the province of formally established committees. While a great deal has been written about statistical methods for interim data monitoring, the practical aspects of who should serve on data monitoring committees (s) or otherwise be involved in the monitoring process, what data should be monitored and how frequently, and what are the necessary and appropriate lines of communication have received limited discussion. Since DMCs are given major responsibilities for ensuring the continuing safety of trial participants, relevance of the trial question, appropriateness of the treatment protocol, and integrity of the accumulating data, it is important to understand the ways in which these committees meet such responsibilities.
A word about terminology. Committees to monitor accumulating data from clinical trials go by a variety of names. The two most frequent of these are probably "data and safety monitoring board" and "data monitoring committee," but there are many other variations (Ellenberg, 2001). We have arbitrarily selected "data monitoring committee," in part because of its simplicity and in part because this is the term used by international regulatory authorities (http://www.ifpma.org/ich1.html).
From time to time, papers describing the experience of particular DMCs, as well as papers addressing general approaches for operating and serving on such committees, have been published; a number of these are referenced in Chapter 1.
These papers have provided some valuable insights into the monitoring process. In 1992 an international workshop was held at the National Institutes of Health to discuss different approaches to data monitoring that had been or were being used in a variety of settings, and the proceedings were published as a special issue of the journal Statistics in Medicine (Ellenberg et al., 1993). At this workshop, individuals with substantial practical experience in interim data monitoring reported on their preferred operating models, and there was substantial discussion of the advantages and disadvantages of the different approaches presented. Up to now, those workshop proceedings plus the aforementioned papers have constituted the primary references for those interested in learning about the various operating models in use for DMCs, as well as the diversity of issues these committees may consider.
The use of DMCs has continued to grow, especially with respect to trials sponsored by pharmaceutical companies. The demand for individuals to serve on these committees is high; it is increasingly difficult to ensure that any DMC will include at least some members with prior experience on other DMCs. As individuals with extensive experience coordinating and/or serving on such committees, the authors of this book are frequently asked for advice concerning their operation (from trial organizers/sponsors) and the scope of responsibilities of committee members (from new members of such committees). The increasing interest in these issues led us to believe that a comprehensive reference on the practice of interim data monitoring and the structure and operation of DMCs was needed; that was our primary motivation for writing this book.
The book is intended for those involved with or otherwise interested in the clinical trials process. We expect this group will include statisticians, physicians and nurses, trial administrators and coordinators, regulatory affairs professionals, bioethicists, and patient advocates. The issues are relevant to trials sponsored by government funding agencies as well as by pharmaceutical and medical device companies, although approaches taken may differ in different contexts.
We also believe this book should be of interest to those involved in the evaluation and reporting of trial results - for example, medical journal editors and science journalists for lay publications - as the process of trial monitoring has important implications for the interpretation of results. We have attempted to keep the material non-technical, so as to make it accessible to as large a part of the clinical trials community as possible.
Every chapter in the book addresses an issue that has been debated among those with DMC experience in different settings. Our intent is to describe the issues clearly as well as to describe the arguments that have been made for and against different approaches that might be taken. We will identify areas where there appears to be a general consensus, and occasionally recommend a particular approach even when there is no widespread consensus on that issue. For the most part, however, our goal is to clarify the types of decisions that must be made in implementing DMCs and not to provide a prescription for their operation. There is no "one size fits all" for DMCs; different models may be needed for different situations.
We begin with some introductory background and some historical notes on the use of DMCs in different contexts. Next, we address the scope of responsibilities that may be assigned to a DMC. Some committees are charged with reviewing outcome data only (or even safety data only); others are asked to review the initial protocol, monitor the conduct of the study by assessing accrual, eligibility, compliance with protocol, losses to follow-up, and other issues that are ultimately relevant to the value and credibility of a trial. The specific responsibilities delegated to a committee monitoring a particular trial will influence other operational aspects, such as committee composition.
In Chapter 3 we consider the committee membership: what types of expertise should be represented on all committees, other relevant factors in selecting committee members, optimal committee size, methods of selecting committees (and committee chairs). An important issue regarding committee membership that we discuss in some detail is conflict of interest.
Chapter 4 continues the consideration of conflicts of interest in the broader context of the independence of the committee. We discuss what is meant by an "independent" committee, and the potential consequences for the trial and its credibility when the committee's independence is called into question. We also discuss the various types of trials for which independence of the DMC may be most critical.
Chapter 5 deals with one of the most controversial issues relating to the interim monitoring of clinical trial data: the extent to which any interim data, and unblinded interim data in particular, should be released to individuals or groups other than the committee itself. It has been argued that there may be a "need to know" for some groups such as the sponsor or the regulatory authority; it has also been argued there is a "right to know" for participating investigators, study subjects, and the general public. Others believe that limiting access to interim results is essential to the successful completion of clinical trials. This chapter focuses on such debates, and their potential implications for trial integrity.
In Chapter 6 we deal with the logistical issues - how often a committee should meet, how long the meetings need to be, how they are conducted, the content of the report the committee is to consider, the preparation and content of meeting minutes, and a number of other issues. Many groups who regularly sponsor and/or coordinate clinical trials have developed their own approaches to these issues, but these approaches can be quite different, even for similar types of clinical trials. Some might consider these types of issues part of the "minutiae" of clinical trials; our experience, however, is that the quality and reliability of the monitoring process may depend very heavily on just these types of issues.
Chapter 7 addresses the very important but little discussed topic of how the DMC interacts with other trial components. There are many constituencies involved in any given trial, including the sponsor(s), the investigators, the statistical coordinating center, the study steering committee, the institutional review board(s), and of course the patients. There is also a variety of modes of interaction, both formal (e.g. submitting reports) and informal (e.g. attending meetings of other components where unstructured discussion may take place).
Chapter 8 provides an overview of the various statistical approaches for interim monitoring of clinical trial data, and some discussion of why some approaches...
