|
|
Therapist Directory: Find a Psychologist, Find a Therapist, Find a Marriage Counselor
PSYCHOLOGY TOPICS
Selected topics in psychology
and mental health.
|
|
|
|
|
THE THERAPIST PSYCHOLOGIST BOOK STORE
 | |
 View Larger |
Data Monitoring in Clinical Trials: A Case Studies Approach ( Springer )
Release Date: 2005-09-30
Average Customer Rating:
List Price: $54.95
Price: $47.97
Eligible for FREE Super Saver Shipping on orders over $25.
Availability: Usually ships in 24 hours
|  Add to Cart |
|
|
Product Description
Randomized clinical trials are the gold standard for establishing many clinical practice guidelines and are central to evidence based medicine. Obtaining the best evidence through clinical trials must be done within the boundaries of rigorous science and ethical principles. One fundamental principle is that trials should not continue longer than necessary to reach their objectives. Therefore, trials must be monitored for recruitment progress, quality of data, adherence to patient care or prevention standards, and early evidence of benefit or harm. Frequently, a group of external experts, independent from the investigators and trial sponsor, is charged with this monitoring responsibility, especially for safety and early benefit. This group is referred to by various names, such as a data monitoring committee or a data and safety monitoring board. This book, through a series of case studies presented by many distinguished clinical trial experts, illustrates the complexity of this monitoring process. The editors provide an overview of the process and a summary of a multitude of the lessons learned from the cases presented. This book should be useful to anyone serving on a data and safety monitoring board, or planning to do so, for colleagues in academia, industry and governmental agencies, and for teaching students in biostatistics, epidemiology, clinical trials and medical ethics. No other text has as extensive a collection of cases which provide insight into the many issues, often conflicting, that must be examined before recommendations to continue or discontinue a trial can be made. While depth in statistical methods is not required, some familiarity with statistical design and analysis issues in clinical trials is helpful. The cases cover trials which were terminated early for convincing evidence of benefit, or for harmful effects. Cases with complex issues are also included. This series of cases should provide broad background information for potential monitoring committee members and better prepare them for the challenges that may exist in the trials for which they are responsible. The three editors have contributed two overview chapters as well as several case studies to go along with cases contributed by a distinguished group of colleagues experienced in the design, monitoring and analysis of clinical trials. Dr. David DeMets is currently Professor and Chair, Department of Biostatistics and Medical Informatics at the University of Wisconsin-Madison. He is past president of the Eastern North American Region (ENAR) of the International Biometric Society, a past member of the Board of Directors of the American Statistical Association and an elected Fellow. He recently received the Robert Gordon Lectureship Award, given by the National Institutes of Health, for significant contributions to the field of clinical trials. Dr. Curt Furberg, is currently Professor (and former Chair) of the Department of Public Health Sciences at Wake Forest University. Previously, he was Head of the Clinical Trials Branch and Associate Director of the Clinical Applications and Prevention Program at the National Heart, Lung, and Blood Institute. Dr. Lawrence Friedman is a former Director of the Division of Epidemiology and Clinical Applications and a former Assistant Director for Ethics and Clinical Research at the National Heart, Lung, and Blood Institute. All three are past presidents of the Society of Clinical Trials. The editors have collaborated previously as coauthors of a text: the Fundamentals of Clinical Trials.
|
|
This excellent book provides overall guidance, as well as subtle tips, on designing clinical trials. ( tom5brody2 )
DATA MONITORING IN CLINICAL TRIALS, by DeMets, Furberg, and Friedman, is 374 pages long. There are three chapters totaling 49 pages, followed by 29 case histories, each about ten pages in length. The first three chapters provide narratives to the case histories. Thus, the reader knows when to jump ahead in the book to one or more of the case histories, depending on his or her interest. Although the book is "about" Data Monitoring Committees (DMCs), the book provides guidance to most aspects of managing a clinical trial. This book discloses many subtle facts on the logistics of clinical trials, including how various groups (DMC, Institutional Review Board, study subjects, sponsor) are supposed to communicate with each other. In view of the excellent guidance provided by this book, I would recommend DATA MONITORING IN CLINICAL TRIALS to all managers and medical writers involved in clinical trials. The book walks the reader through many interesting decision-making trees.
My criticisms are that the book fails to provide any examples of a DMC Charter or a DSMB Charter, fails to provide an example of a DMC's recommendation or report, and that chapter 3 is on the sketchy side.
In chapter 1, we learn that the main purposes of a DMC is to ensure that subjects are not being harmed by the study drug or by the standard of care, and to ensure that the Clinical Study Protocol (CSP) is being followed (page 3). We learn that, for small single-center studies, an Institutional Review Board (IRB) is sufficient to do this job, but that for multi-center studies, both an IRB and DMC are needed. We learn that DMC had its origin in the "Greenberg Report" from 1967. The Greenberg Report was innovative in that it provided for an independent board (independent of investigators) that received study data, and where this data was kept secret from the investigators.
MEMBERSHIP OF DMC. Regarding members of the DMC, we learn that "the more distant and independent the better, but complete independence from the sponsor should not come at the expense of needed expertise" (page 6). We learn that DMC members must disclose conflicts of interest, on an ongoing basis during the clinical trial. We learn that DMC meetings have an open session component, dealing with administrative issues, baseline data, and adherence to the CSP, and a closed session, where unblended data is discussed. Administrative aspects include, e.g., are the subjects completing their forms; are subjects being enrolled on a timely basis; are the subjects adhering to the drug regimen?
We learn the definition of clinical equipoise: The state of not knowing which of the study arms is preferred, and gives the best results, as far as efficacy and safety are concerned.
We learn these things. The DMC needs to be alert of other trials using the same drug (pages 10 & 32, Cases 2, 8,15, 24, 27). Regarding other trials, the authors caution that study design may differ, for example, the relative time between the myocardial infarction and time of first administering the drug (the 2 studies were judged not likely relevant to each other) (page 32). The DMC may, if needed, recommend modification or termination of the present clinical study depending efficacy and safety results from the other trial (page 10, case 24).
RECOMMENDATIONS FROM A DMC. We learn that the DMC can recommend extending a trial, where the rate of outcome (primary variables) is too low, and that this strategy can be built into the CSP (page 11, Cases 8 & 27). We learn that the DMC can recommend stopping for (1) Overwhelming evidence for efficacy, (2) Serious harm, (3) Hopelessness that the study drug will work (Case 3), and (4) Extremely poor compliance by subjects with the CSP, and (5) Evidence from another trial that the drug works (or fails to work) (page 10). We that if changes are to be made to the CSP, they should be made sooner, not later (page 16, Case 27). We learn that where changes are made in the CSP, the FDA can become suspicious that the sponsor had a knowledge of the results (efficacy data and safety data) (page 43). In other words, the problem is that the sponsor can change the nature of the primary endpoint so that it is consonant with actual data that is coming out of the study population (pages 42-43). We learn of examples where the DMC was unable to provide a recommendation, and where an ad hoc committee or policy board was formed (pages 17, 29, 32 & Cases 12, 15, 18). We learn of the fact-pattern where the DMC recommended change in the CSP, where one change was a more stringent exclusion criterion, and the other related to drug administration (a slower drug infusion rate) (page 18 & Case 23). We learn this tidbit about scheduling DMC meetings--where materialization of certain adverse events (AEs) required the scheduling of an emergency teleconference by the DMC (page 19 & Case 16).
LAGGING DATA. We learn an interesting detail of the common problem of lagging reports from study subjects, where a "sweep" was implemented, where every subject was contacted at the same time (page 19 & Cases 9 & 22). We learn that under-reporting can result in an accumulation of data showing lack of efficacy (or poor safety) and that correcting the non-reported data can result in the disappearance of these disappointing trends (pages 20 & 24 & Case 22). We learn that the CSP should include a schedule for interim analyses (page 20 & Case 28). We learn of pre-determined stopping rules, where the rules were linked to the amount of data collected, and where stopping for benefit required data to be more convincing, than stopping for harm (page 21 & Case 19).
UNEXPECTED ADVERSE EVENTS. We learn that where an unexpected AE presents during a trial, e.g., cataracts or pulmonary embolism, this may require all study subjects to be informed of the risk, and to fill out a Re-Consent Form (pages 21 & 26 & Cases 7 & 17). We learn that a consent form is a contract between the investigator and subject (page 22). We learn of the hierarchy in obtaining data: data on mortality must be very current, data on serious adverse events (SAEs) must also be very current, data on primary outcome other than mortality can be a few months old, data on lab results and data on concurrent medications can also be old and not current (page 24). We learn that, at any given time, the DMC can choose to unblind the data or to maintain blinding (page 25 & Case 13). We learn that the sponsor can improve the efficiency of the DMC, by providing the DMC with a mock report at the DMC's first meeting (page 25).
EARLY DATA MISLEADING. We learn that early data from study subjects can be misleading, and that early trends can disappear later on (pages 25 & 31 & Cases 5, 11, 12, 17). Where the DMC can't decide whether to recommend stopping the trial or to recommend continuing, we learn that a third avenue is available--deciding to hold more frequent DMC meetings to evaluate incoming safety data (page 26). We learn the value of knowing the mechanism of a disease--where several symptoms are known to be part of the same disease, each of these symptoms can be listed as an outcome in the CSP, and a "composite outcome" can be set forth as an endpoint (page 26). A composite outcome is especially justified where each symptom responds in the same way to the study drug. For example, Case 10 discloses the composite outcome of death, infarction, and stroke (page 27 & Cases 7, 11, 27).
DEFINE SUBGROUPS. We learn of the advantages of defining subgroups (strata) in the CSP. For example, if an unacceptable AEs occur in the study, but where these are clustered in one of the subgroups, the DMC can recommend terminating this subgroup (rather than terminating the entire study (pages 28-29 & Case 12). We learn that the CSP must be written in a manner that adequately defines the subgroups. We also learn that what at first might be reasonable subgroups(ischemic and non-ischemic), turn out to be irrelevant to response of subjects to the drug (page 29). We learn that the group of subjects defined by a subgroup is not the same as a group of subjects presenting with a specific outcome for a surrogate marker, though these might seem to be the same, at first glance, to a layperson. We learn that it is preferable for a surrogate (or biomarker) to capture the full effect of the study drug (page 30). We learn of an example of surrogates (arrhythmia suppression; lipoprotein levels; cardiac function, microaneurisms) that were misleading, and that was not predictive of mortality or of coronary events (pages 30-31 & Cases 5, 13, 14, 17). We learn about follow-up procedures that occur long after the trial has ended or has been terminated early (pages 33-34, Cases 1, 12, 17). In Case 12, the trial itself showed no efficacy of the drug, while a follow-up after 9 years showed good effect. We also learn that studies can be terminated because the money ran out, or because the company changed hands (pages 34-36).
GOVERNMENT GUIDANCE ON ROLE OF DMC. Chapter 3 informs us on guidance on the nature of DMCs from the U.S.government ("Guidance for Clinical Trial Sponsors on the Establishment and Operation of Clinical Trial Data Monitoring Committees") and from International Conference on Harmonization.
INTERIM DATA. Chapter 3 provides interesting guidance on what to do with interim data. We learn about use of interim data (comprising data on surrogate endpoints such as viral load) and whether it should be used as a basis for stopping a trial for efficacy--we learn that it is more acceptable to stop a trial on the basis of interim data if the trial is nearly complete, and less acceptable to stop a trial still in its early stages (because that particular trial will not yet allow any valid conclusion) (page 42).
In chapter 3, we find a case history disclosing interactions between FDA, DMC, and sponsor, where the issue was agreement between interim data and data from a separate study on the same drug. The purpose of this commentary is to show how a DMCs decision can influence FDA's decision. The outside study showed that the drug worked. But the interim data showed that the drug did not work, and where the DMC recommended continuing the trial (until some difference might have materialized). After taking in all the information, including the info from the outside study showing that the drug worked, FDA required that the study be completed (page 47). Another case history disclose the situation where one sponsor was running several trials on the same drug, where each of these had its own DMC. In view of emerging safety issues from one of these trials, the FDA recommended that the manufacturer tell all of these DMCs to share safety data with each other (page 47).
Overall, chapter 3 is somewhat incomplete. This chapter would benefit by several more examples of interactions between DMCs and FDA. Apparently, the authors did not have much information to share on this particular topic.
RETINOPATHY CASE HISTORY. Case 1 concerns diabetic retinopathy, where treatment was photocoagulation (one eye treated, one eye not treated). Early results showed dramatic efficacy. The DMC recommended stopping the trial (and treating both eyes of each subject with photocoagulation), even though the trial was at an early stage, and only a few patients had been on therapy, and even though there was no opportunity yet to observe late complications. This posed a dilemma to the sponsor. The final decision was to continue the trial, but where both eyes were treated only in high risk subjects (high risk for blindness). This decision took the form of a compromise(pages 55-63).
PROPANOLOL CASE HISTORY. Case 2 (pages 64-72) concern the beta-blocker, propanolol, for treatment of myocardial infarction. Inclusion criterion was having a myocardial infarction 5-21 days before randomization. Case 2 provides useful nuts & bolts information for running any clinical study. We learn that dosing for subjects was increased for any given subject, where serum drug levels were used as a basis for stepping up the dose (where blinding was maintained by changing dose formulation for PLACEBO subjects). The DMC reviewed the efficacy data at various times (May 1979, April 1981, Oct. 1981), but it was not until Oct. 1981 that the DMC recommended stopping the trial for efficacy. What turned the tide was that the "monitoring boundary" had been crossed. (A full understanding of this particular case history requires an understanding of the O'Brien-Fleming procedure of monitoring boundaries, alpha values, the logrank statistic, Z value, and conditional power.) This case history also discloses the interesting fact-pattern where the DMC educated itself on results of a similar drug (the beta-blocker, timolol) for the same indication. We also learn that the consent form contained a clause on stopping for efficacy and for harm. This case history discloses the list of points in favor of stopping and points in favor of continuing the trial (the trial was stopped for efficacy).
This book clearly spells out many useful and subtle things useful to a trial manager or medical writer. Every case history provides a few unique take-home lessons. FIVE STARS.
|
|
|
|
|
|
Addiction
