Data Monitoring Committees in Clinical Trials

SUSAN S. ELLENBERG, PHD, is Professor of Biostatistics, Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania Perelman School of Medicine, USA.THOMAS R. FLEMING, PHD, is Professor of Biostatistics and Statistics, Department of Biostatistics, School of Public Health, University of Washington, USA.DAVID L. DEMETS, PHD, is Professor Emeritus and Founding Chair of the Department of Biostatistics and Medical Informatics, University of Wisconsin–Madison, USA.

• Preface to the Second Edition xiiiPreface to the First Edition xv1 Introduction 11.1 Motivation 11.2 History of data monitoring committees in government-sponsored trials 81.3 Data monitoring committees in trials sponsored by the pharmaceutical industry 151.4 Statistical methods for interim monitoring 181.5 When are data monitoring committees needed? 211.6 Models for data monitoring committees 221.7 Where we are today 241.8 Fundamental principles of data monitoring 25References 272 Responsibilities of the Data Monitoring Committee and Motivating Illustrations 352.1 Fundamental charges 362.2 Specific tasks of the data monitoring committee 392.2.1 Initial review 402.2.1.1 Review of the study protocol 402.2.1.2 Review of procedures to ensure quality of study conduct 452.2.2 Evaluating the quality of ongoing study conduct 472.2.3 Assessing safety and efficacy data 562.2.3.1 Termination due to favorable benefit-to-risk 582.2.3.2 Termination due to unfavorable benefit-to-risk 632.2.3.3 Termination due to inability to answer trial questions 652.2.3.4 Continuation of ongoing clinical trials 682.2.3.5 Consideration of the overall picture: primary and secondary analyses 722.2.3.6 Modifying sample sizes based on ongoing assessment of event rates 762.2.4 Reviewing the final results 802.3 The data monitoring committee charter 83References 843 Composition of a Data Monitoring Committee 893.1 Introduction 893.2 Required areas of expertise 903.3 Other relevant characteristics of committee members 963.4 Committee size 983.5 Selecting the committee chair 1023.6 Responsibility for appointing committee members 1023.7 Representation of other study components on the committee 1043.8 Preparation for service on a committee 106References 1094 Independence of the Data Monitoring Committee: Avoiding Conflicts of Interest 1134.1 Introduction 1134.2 Rationale for independence 1144.3 Financial independence 1164.3.1 Commercial sponsors 1174.3.2 Government sponsors 1184.3.3 Academic investigators 1184.4 Intellectual independence 1254.5 Emotional conflicts 1314.6 Best practices to address challenges to the DMC’s independence 1324.6.1 Adequate training/experience in the DMC process 1334.6.2 Indemnification of DMC members 1354.6.3 Maintaining confidentiality of interim data 1364.6.4 Flexibility of procedures 1384.6.5 DMC meeting format 1394.6.6 Creating independent relationships and reducing conflicts of interest 1414.6.7 Adequately informative DMC reports 1424.7 Summary 143References 1445 Confidentiality Issues Relating to the Data Monitoring Committee 1475.1 Rationale 1475.2 Limits of confidentiality 1595.2.1 Interim analysis reports 1595.2.2 Access to aggregate data on efficacy and safety outcomes 1615.2.3 Providing access to interim data on a “need-to-know” basis 1645.2.4 Settings and procedures allowing broader unblinding of safety data 1665.2.5 Consequences of unblinding interim data for regulatory review in ongoing trials 1685.2.6 Some illustrations of broader unblinding 1755.2.7 The steering committee and maintaining confidentiality 1875.2.8 Indirect challenges to confidentiality 1895.3 The need for the DMC to review unblended data 1905.4 Conclusions: consensus regarding confidentiality 195References 1986 Data Monitoring Committee Meetings 2036.1 Introduction 2036.2 Specific objectives and timing of meetings 2046.2.1 Organizational meeting 2056.2.2 Early safety/trial integrity reviews 2086.2.3 Formal interim efficacy analyses 2126.2.4 End-of-trial debriefing 2136.3 Preparation of meeting reports 2146.3.1 Currentness of data in the report 2176.3.2 Inclusion of unadjudicated data 2206.4 Format for meetings 2216.4.1 The initial closed session 2236.4.2 The open session 2246.4.3 The final closed session 2276.4.4 Various formats for holding the open and closed sessions 2276.4.5 Meeting duration and venue 2296.5 DMC meeting minutes and the DMC recommendations 2306.5.1 The DMC recommendations, the open minutes, and the closed minutes 2306.5.2 The level of detail 2326.5.3 The authorship of the minutes and the sign-off by committee members 233References 2357 Data Monitoring Committee Interactions with Other Trial Components or Related Groups 2377.1 Introduction 2387.2 Study sponsors 2387.2.1 Industry sponsors 2397.2.2 Government sponsors 2417.3 Study steering committee/principal investigator 2437.4 Study investigators 2477.5 Trial statisticians and statistical centers 2477.5.1 The independent statistical center 2487.5.2 Ensuring optimal data presentations 2537.6 Institutional review boards 2537.7 Regulatory agencies 2567.8 Study participants and/or advocacy groups 2577.9 Other data monitoring committees 259References 2628 Statistical, Philosophical, and Ethical Issues in Data Monitoring 2658.1 The need for statistical approaches to monitoring accumulating data 2668.2 Overview of statistical methods 2708.2.1 Group sequential methods 2718.2.1.1 Some group sequential boundaries for establishing benefit 2738.2.1.2 Group sequential alpha spending functions 2778.2.1.3 Some group sequential boundaries when early results are unfavorable 2808.2.2 Triangular boundaries 2848.2.3 Stochastic curtailment/conditional power 2868.2.4 Bayesian monitoring 2908.2.5 The general approach to sequential stopping boundaries 2938.2.6 Software packages for sequential clinical trial designs 2948.2.7 Adaptive clinical trial designs 2948.3 Protocol specification of the monitoring plan 2998.4 Other statistical considerations in monitoring trial data 3008.4.1 Primary versus secondary endpoints 3008.4.2 Short-term versus long-term treatment effects 3028.4.3 Results in subgroups 3038.4.4 Taking external information into account 3078.4.5 Evaluating safety in the context of evidence about efficacy: role of boundaries 3098.4.6 Ensuring proper robustness when defining boundaries for establishing benefit 3118.5 Ethical considerations 3138.5.1 Early termination philosophies 3138.5.1.1 Responding to early beneficial trends 3148.5.1.2 Responding to early unfavorable trends 3188.5.1.3 Responding to unexpected safety concerns 3248.5.2 Other ethical considerations 325References 3269 Determining When a Data Monitoring Committee is Needed 3359.1 Introduction 3369.2 Typical settings for an independent data monitoring committee 3369.3 Other settings in which an independent data monitoring committee may be valuable 3399.3.1 Early trials of high-risk treatments 3399.3.2 Trials in vulnerable populations 3409.3.3 Trials with potentially large public health impact 3419.4 An alternative monitoring approach: the internal monitoring committee 3429.5 A decision model assessing need for an independent DMC or an internal monitoring committee 3469.6 Settings with little need for an independent or internal monitoring committee 3519.7 Summary 352References 35310 Regulatory Considerations for the Operation of Data Monitoring Committees 35510.1 Introduction 35610.2 Data monitoring committees in government regulations 35610.3 Regulatory guidance 35710.3.1 US Food and Drug Administration 35710.3.2 International regulatory guidance 36010.3.2.1 European Union 36010.3.2.2 International Conference on Harmonization 36110.3.2.3 The World Health Organization 36310.4 Regulatory approaches relevant to data monitoring committee operation: the US FDA 36410.5 Policies of funding agencies regarding DMC operations 36710.5.1 National Institutes of Health 36710.5.2 Other federal agencies 36910.5.3 Funding agencies outside the US 36910.6 Involvement of FDA staff in data monitoring committee deliberations 37010.7 Examples of regulatory authority interaction with data monitoring committees 372References 37911 Legal Considerations for DMCs 38311.1 DMC indemnification 38311.1.1 Motivating examples 38511.1.2 Emergence of DMCs and heightened awareness of their existence 38811.1.3 Further motivation for indemnification to protect the DMC 39011.1.4 Some specific concerns from current experiences with indemnification 39211.1.5 Potential solutions to indemnification issues 39411.1.6 Confidential disclosure agreement (CDA) 39711.1.7 Summary of indemnification, liability, and contracting issues 39911.2 Balancing legal and ethical responsibilities: a need for a mediator? 40011.2.1 A case study: the setting of Actimmune in patients with idiopathic pulmonary fibrosis 40211.2.2 IMMUNE response AIDS clinical trial 405References 406Appendix A The Data Monitoring Committee Charter 411Appendix B Performance Standards Document 431Statistics in Practice 451Index 455