Search PubMed⌕ Search

Biomedical subjects

David L DeMets

Publications and source records attributed to David L DeMets.

At least 19 recordsLinked to original sources
PubMed2006

Training of the next generation of biostatisticians: a call to action in the U.S.

Two workshops (2001, 2003) were held by the National Institutes of Health (NIH) to examine the need to train more biostatisticians in the U.S. to meet the increasing opportunities in the biomedical research enterprise. The supply of new PhD graduates in biostatistics in the U.S. has been relatively steady for the past two decades while the demand has increased dramatically. These workshops concluded that a renewed effort must be made in the U.S., led in part by the NIH, to add to and expand the existing training programs to increase the supply. This article summarizes those two workshops and their recommendations. Some progress has been made through a new biostatistics training program with emphasis in bioinformatics sponsored by the National Institute of General Medical Sciences (NIGMS).

Biological Science Disciplines↗
PubMed2006

Conditional and unconditional confidence intervals following a group sequential test.

After a group sequential test, the naive confidence interval (CI) is usually biased in the sense that it does not cover the true parameter at the correct nominal level. Furthermore, when the stopping time is taken into account, the actual conditional confidence coverage probability can be much less accurate. In this article, we study the conditional coverage probability and other related properties of the naive CI and different versions of exact CI's. It is demonstrated that only correcting the overall confidence level does not necessarily improve the confidence level at any given stopping stage. Conditional inference can be applied to construct an exact conditional CI but it is not without serious undesirable properties. We propose a two-step restricted conditional confidence interval (RCCI) which considerably improves the conditional confidence level while minimizing the undesirable properties. Numerical comparisons are made between the proposed method and existing methods. The results show that the RCCI not only improves the conditional coverage probability considerably from the exact CI's but also is free of the major undesirable properties displayed by the pure conditional CI. Differences between the conditional and unconditional CI's and their respective strengths are also discussed.

Clinical Trials as Topic↗
PubMed2006

An Institutional Review Board dilemma: responsible for safety monitoring but not in control.

Clinical trials have become a major research tool to evaluate new medical interventions. Most trials require some level of data monitoring for quality control and many trials require special monitoring for participant safety. For national multicenter trials, independent data monitoring committees have become the standard for monitoring for evidence of participant benefit or harm in trials with irreversible outcomes such as death or serious morbidity. The Institutional Review Board (IRB) is held responsible for monitoring local trials. Often local institutions do not have an infrastructure in place to meet this responsibility, and therefore local IRBs cannot fulfill this obligation. In addition, IRBs are currently inundated with individual safety reports from local and multi-institutional trials which may appear to provide some level of safety monitoring, but in fact gives a false sense of security. We propose the establishment of institutional data monitoring committees and appropriate informatics infrastructure to monitor local trials.

Clinical Trials Data Monitoring Committees↗
PubMed2004

The independent statistician for data monitoring committees.

Clinical trials are an essential part of the clinical research process. Recently, independent Data Monitoring Committees (DMCs) have been widely implemented to provide scientific and ethical oversight of pivotal clinical trials having irreversible outcomes such as death, stroke, disease recurrence or a serious adverse event. To carry out their responsibility, the DMC reviews interim analyses of accumulating data. We address the motivation for having the preparation and presentation of these interim analyses be conducted by an independent statistician who is not a member of the DMC and who is not the trial's lead or steering committee statistician. These views are based on having served as members of many DMCs as well as having been the independent statistician for several trials.

Bias↗
PubMed2004

Increasing the sample size when the unblinded interim result is promising.

Increasing the sample size based on unblinded interim result may inflate the type I error rate and appropriate statistical adjustments may be needed to control the type I error rate at the nominal level. We briefly review the existing approaches which allow early stopping due to futility, or change the test statistic by using different weights, or adjust the critical value for final test, or enforce rules for sample size recalculation. The implication of early stopping due to futility and a simple modification to the weighted Z-statistic approach are discussed. In this paper, we show that increasing the sample size when the unblinded interim result is promising will not inflate the type I error rate and therefore no statistical adjustment is necessary. The unblinded interim result is considered promising if the conditional power is greater than 50 per cent or equivalently, the sample size increment needed to achieve a desired power does not exceed an upper bound. The actual sample size increment may be determined by important factors such as budget, size of the eligible patient population and competition in the market. The 50 per cent-conditional-power approach is extended to a group sequential trial with one interim analysis where a decision may be made at the interim analysis to stop the trial early due to a convincing treatment benefit, or to increase the sample size if the interim result is not as good as expected. The type I error rate will not be inflated if the sample size may be increased only when the conditional power is greater than 50 per cent. If there are two or more interim analyses in a group sequential trial, our simulation study shows that the type I error rate is also well controlled.

Anti-HIV Agents↗
PubMed2004

Conditional bias of point estimates following a group sequential test.

Repeated significance testing in a sequential experiment not only increases the overall type I error rate of the false positive conclusion but also causes biases in estimating the unknown parameter. In general, the test statistics in a sequential trial can be properly approximated by a Brownian motion with a drift parameter at interim looks. The unadjusted maximum likelihood estimator can be potentially very biased due to the possible early stopping rule at any interim. In this paper, we investigate the conditional and marginal biases with focus on the conditional one upon the stopping time in estimating the Brownian motion drift parameter. It is found that the conditional bias may be very serious for existing point estimation methods, even if the unconditional bias is satisfactory. New conditional estimators are thus proposed, which can significantly reduce the conditional bias from unconditional estimators. The results of Monte-Carlo studies show that the proposed estimators can provide a much smaller conditional bias and MSE than the naive MLE and a Whitebead's bias reduced estimator.

Effect Modifier, Epidemiologic↗
PubMed2004

Design and analysis of group sequential clinical trials with multiple primary endpoints.

In many phase III clinical trials, it is desirable to separately assess the treatment effect on two or more primary endpoints. Consider the MERIT-HF study, where two endpoints of primary interest were time to death and the earliest of time to first hospitalization or death (The International Steering Committee on Behalf of the MERIT-HF Study Group, 1997, American Journal of Cardiology 80[9B], 54J-58J). It is possible that treatment has no effect on death but a beneficial effect on first hospitalization time, or it has a detrimental effect on death but no effect on hospitalization. A good clinical trial design should permit early stopping as soon as the treatment effect on both endpoints becomes clear. Previous work in this area has not resolved how to stop the study early when one or more endpoints have no treatment effect or how to assess and control the many possible error rates for concluding wrong hypotheses. In this article, we develop a general methodology for group sequential clinical trials with multiple primary endpoints. This method uses a global alpha-spending function to control the overall type I error and a multiple decision rule to control error rates for concluding wrong alternative hypotheses. The method is demonstrated with two simulated examples based on the MERIT-HF study.

Adrenergic beta-Antagonists↗
PubMed2004

Liability issues for data monitoring committee members.

In randomized clinical trials, a data monitoring committee (DMC) is often appointed to review interim data to determine whether there is early convincing evidence of intervention benefit, lack of benefit or harm to study participants. Because DMCs bear serious responsibility for participant safety, their members may be legally liable for their actions. Despite more than three decades of experiences with DMCs, the issues of liability and indemnification have yet to receive appropriate attention from either government or industry sponsors. In industry-sponsored trials, DMC members are usually asked to sign an agreement delineating their responsibilities and operating procedures. While these agreements may include language on indemnification, such language sometimes protects only the sponsor rather than the DMC members. In government-sponsored trials, there has been even less structure, since typically there are no signed agreements regarding DMC activities. This paper discusses these issues and suggests sample language for indemnification agreements to protect DMC members. This type of language should be included in DMC charters and in all consulting agreements signed by DMC members.

Clinical Trials Data Monitoring Committees↗
PubMed2003

Her-2/neu overexpression and response to oophorectomy plus tamoxifen adjuvant therapy in estrogen receptor-positive premenopausal women with operable breast cancer.

PURPOSE: Studies evaluating the relationship of HER-2/neu breast tumor status and response to adjuvant endocrine therapy have reached conflicting conclusions about resistance of HER-2/neu-positive tumors to this treatment. We studied 282 patients participating in a randomized controlled trial of adjuvant oophorectomy and tamoxifen or observation who had estrogen receptor-positive tumors and whose tumors were evaluated for HER-2/neu overexpression by immunohistochemistry. PATIENTS AND METHODS: Univariate and multivariate Cox proportional hazards regression models and Kaplan-Meier disease-free and overall survival estimate methods were used. RESULTS: HER-2/neu overexpression was a negative prognostic factor for overall survival. In univariate analyses, in HER-2/neu-positive patients, the hazard ratio (HR) for disease-free survival (DFS) with adjuvant endocrine therapy was 0.37 (95% confidence interval [CI], 0.26 to 0.89); for HER-2/neu-negative patients, the corresponding HR for DFS was 0.48 (95% CI, 0.31 to 0.71). The overall survival (OS) data were HR=0.26 (95% CI, 0.07 to 0.92) and HR=0.68 (95% CI, 0.32 to 1.42) for HER-2/neu-positive and HER-2/neu-negative patients, respectively. In multivariate models, the P values for tests of interaction of HER-2/neu status and response to adjuvant endocrine therapy were 0.18 and 0.07 for DFS and OS, respectively. Kaplan-Meier DFS and OS curves and 3-year DFS estimates were consistent in showing greater benefit to the HER-2/neu-positive subgroup given adjuvant treatment. CONCLUSION: HER-2/neu overexpression does not adversely and may favorably influence response to adjuvant oophorectomy and tamoxifen treatment in patients with estrogen receptor-positive tumors.

Adult↗
PubMed2003

Monitoring mortality at interim analyses while testing a composite endpoint at the final analysis.

Mortality is often used as the clinical endpoint in clinical trials for acute diseases and takes precedence over any other outcome. A composite outcome such as death plus disease occurrence (or recurrence) or death plus hospitalization may also be considered, sometimes even as the primary outcome due to practical sample size issues. That is, a composite endpoint should have a higher event rate and thus a smaller sample size than for mortality alone to reach the same power. Two different scenarios are considered: in Scenario 1, the composite outcome is the primary endpoint and the mortality outcome is secondary; in Scenario 2, the mortality outcome is the primary endpoint and the composite outcome is secondary. In either scenario, the trial will be stopped if the simple mortality outcome shows an adverse effect or a significant benefit at an interim analysis, while the composite outcome will be tested at the final analysis if the mortality outcomes fails to show significance. These scenarios are typical in many trials sponsored by industry for regulatory approval. We refer to them as a switching the primary endpoint process. Two switching-endpoint procedures are proposed to calculate the efficacy boundary for the composite test statistic at the final analysis. The Bonferroni method is used in Method 1. In Method 2, the calculation is based upon the joint distribution of the test statistics for the simple mortality and the composite outcomes. A completed clinical trial, prospective randomized amlodipine survival evaluation (PRAISE-1), is used to illustrate the two switching-endpoint procedures. A simulation study shows that the two switching-endpoint procedures allow a trial to be stopped early due to a clinically relevant benefit in the mortality while preserving the overall alpha level.

Clinical Trials as Topic↗
PubMed2003

Toward protecting the safety of participants in clinical trials.

It is a widely held belief that the current system of oversight of clinical research, particularly the means of assessing risks and minimizing harms to participants in clinical trials, could be improved. In particular, the system is inefficient with overemphasis on the monitoring ability of some groups such as research ethics review boards and investigators, underemphasis on others such as data monitoring committees (DMCs) and sponsors, confusion about responsibilities for safety and imperfect communication between these different groups. Research ethics review boards are not able to perform safety monitoring by review of individual adverse events and are often burdened by duplicative reviews of large multicenter studies. There are no standards for DMCs to ensure they can reliably identify safety issues. Sponsors may be overreliant on data audits and slow to disseminate safety data in a coherent summary. Investigators, their staffs and clinical sites may not fully appreciate all the nuances of good clinical practice or may be inattentive to the daily conduct of studies. Regulators, particularly those in the United States, have failed to completely harmonize their policies with each other or with international regulatory agencies. We recommend well-designed monitoring plans for all studies that are appropriate to their scope and risk, more centralized review of large multisite studies and closer local scrutiny of single-institution studies. In addition, sponsors should pay greater attention to monitoring adverse events and keeping up-to-date databases or investigator's brochures emphasizing safety issues. A minimal standard of education or expertise in good clinical practice should be established for investigators, their staffs and research ethics review board members. DMC composition and functions should be standardized and regulations should be harmonized nationally and internationally. Finally, there should be a concerted effort to study the efficacy of various components of the system.

Clinical Trials Data Monitoring Committees↗
PubMed2002

Effect of carvedilol on the morbidity of patients with severe chronic heart failure: results of the carvedilol prospective randomized cumulative survival (COPERNICUS) study.

BACKGROUND: Beta-blocking agents improve functional status and reduce morbidity in mild-to-moderate heart failure, but it is not known whether they produce such benefits in severe heart failure. METHODS AND RESULTS: We randomly assigned 2289 patients with symptoms of heart failure at rest or on minimal exertion and with an ejection fraction <25% (but not volume-overloaded) to double-blind treatment with either placebo (n=1133) or carvedilol (n=1156) for an average of 10.4 months. Carvedilol reduced the combined risk of death or hospitalization for a cardiovascular reason by 27% (P=0.00002) and the combined risk of death or hospitalization for heart failure by 31% (P=0.000004). Patients in the carvedilol group also spent 27% fewer days in the hospital for any reason (P=0.0005) and 40% fewer days in the hospital for heart failure (P<0.0001). These differences were as a result of both a decrease in the number of hospitalizations and a shorter duration of each admission. More patients felt improved and fewer patients felt worse in the carvedilol group than in the placebo group after 6 months of maintenance therapy (P=0.0009). Carvedilol-treated patients were also less likely than placebo-treated patients to experience a serious adverse event (P=0.002), especially worsening heart failure, sudden death, cardiogenic shock, or ventricular tachycardia. CONCLUSION: In euvolemic patients with symptoms at rest or on minimal exertion, the addition of carvedilol to conventional therapy ameliorates the severity of heart failure and reduces the risk of clinical deterioration, hospitalization, and other serious adverse clinical events.

Adrenergic beta-Antagonists↗
PubMed2002

Clinical trials in the new millennium.

Since the introduction of the randomized clinical trial (RCT) over 50 years ago, this method has become the corner stone for evaluating new pharmacologic or biologic agents with many disease areas benefiting. There are numerous examples demonstrating beneficial interventions as well as others either not beneficial or harmful. Statistical methodology for clinical trials has grown rapidly. Advances in information technology for data collection have allowed trials to be conducted around the world. Academia, industry and government have worked in partnership to conduct RCTs. Despite these many RCT achievements, the most interesting and challenging era of clinical trials lies ahead of us. With the human genome now sequenced, we face a new set of challenges to transform vast amounts of data into useful information. The post-genomics era will better identify the disease mechanism and thus help design better treatments, and be more selective in screening patients, yielding more efficient clinical trials. For some areas of medicine, such as medical devices, standards of acceptance and regulatory approval are changing. Other areas, such as medical procedures and alternative medicines, are generally not well evaluated and could benefit greatly from a wider use of RCT methodology. As the ICH guidelines facilitate and encourage international clinical trials, the scientific and ethical dimension of conducting trials in Third World countries are raised. For example, Western society's best standard of care may not be available or affordable to these countries as the control Western investigators should not exploit patients in the Third World. Those and many other challenges face us in the decade ahead. It is truly an exciting time with new opportunities for the RCT to contribute to medicine and health care or prevention.

Data Interpretation, Statistical↗
PubMed2002

Monitoring clinical trials: issues and controversies regarding confidentiality.

During phase III clinical trials in life-threatening disease settings, it is important to ensure that the Data Monitoring Committee (DMC) has exclusive access to the interim efficacy and safety data generated by the data analysis centre, in order to minimize the risk of widespread prejudgement of unreliable trial results based on limited data. This prejudgement could adversely impact rates of patient accrual, continued adherence to trial regimens and ability to obtain unbiased and complete assessment of trial outcome measures. This also could result in publications of early results that might be very inconsistent with final study data on the benefit-to-risk profile of the study interventions. Circumstances arise only rarely in which unblinding of interim data beyond the DMC would enhance the ability of the trial to provide reliable results. However, to address the ethical imperative to protect the interests of study participants, the DMC itself should have access to unblinded efficacy and safety results.

Acquired Immunodeficiency Syndrome↗