A Sample-Size Optimal Bayesian Procedure for Sequential Pharmaceutical Trials

Cressie, Noel; Biele, Jonathan

doi:10.21236/ada248512

Cited by 4 publications

(4 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, Anscombe 13 considers n pairs of patients randomised equally to two groups, a total patient horizon of N, a uniform prior on true treatment benefit, and loss function proportional to the number of patients given the inferior treatment times the size of the inferiority: he concludes it is approximately optimal to stop and give the 'best to the rest' when the standard one-sided P value is less than n/N -half the proportion of patients already randomised. Berry and Pearson 60 and others 59,120,228 have extended such theory to allow for unequal stages and so on. Backwards induction is extremely computationally demanding, but Carlin et al 91 do a retrospective analysis on a trial, 90 and claim it is computationally feasible using MCMC methods, in which forward sampling is used as an approximation to the optimal strategy.…”

Section: Monitoring Using a Formal Loss Functionmentioning

confidence: 99%

Bayesian methods in health technology assessment: a review.

Spiegelhalter¹,

Myles²,

Jones³

et al. 2000

Health Technology Assessment

289

227

View full text Add to dashboard Cite

The papers by Best and Wake®eld, Fienberg et al. and Romaniuk et al. represent an interesting and stimulating set of somewhat diverse papers with some common themes, illustrating as expected the value of (and some of the problems encountered in) applications of multilevel and random-eects models over a wide range of medical and social science applications. More speci®c common themes shared by the ®rst two papers include a consideration of aspects of underascertainment of populations and/or cancer cases, and the use of Bayesian multilevel models.Best and Wake®eld's paper on accounting for inaccuracies in population and case counts in mapping studies expounds and illustrates the use of several errors-in-variables models, and in so doing again demonstrates the gain from the use of graphical models. This allows the population (and case) count models to be deconstructed, so that information on uncertainties about the sources of data can be incorporated explicitly, although the availability of such information is inevitably limited. Hence the sensitivity of geographical inferences about risks and relative risks of disease using conventional approaches to underenumeration and other errors is critically and eectively explored. Arguably, however, more details of the implementation of the Bayesian model are required for it to be fully assessable (or reproducible); some relevant recommendations for reporting such analyses are set out and illustrated elsewhere (Spiegelhalter et al., 1999) in the form of the BayesWatch checklist.Fienberg, Johnson and Junker begin with a succinct review of multiple-recapture methods, considering important extensions to allow heterogeneity of capture probabilities. Their paper is a tour de force which successfully draws together ideas from log-linear multiple-recapture models, Rasch models for object and list heterogeneity, and Bayesian hierarchical modelling approaches in addressing heterogeneity and dependence in capture±recapture modelling. Heterogeneity and dependence eects are apparently strong in the diabetes register and World Wide Web page examples investigated. The authors note the explicit simplicity of (complex) Rasch model formulation as hierarchical Bayes models. Although this makes constraints clear, as the authors also comment, the relationship between restrictions on interactions in log-linear and constraints on moments in Bayes models need further exploration before fully meaningful comparisons can be achieved.Romaniuk, Skinner and Cooper model diary data on shampoo usage. A multilevel logistic model is formulated in terms of binary usage responses, with levels identi®ed by survey respondents and by repeated hourly measurements throughout the diary week. This leads to problems in dealing with missing initial data, which are not really resolved by use of the EM algorithm. Although an alternative event history data formulation in terms of times between shampoo usage might reduce this problem, the apparent existence of strong patterns of usage by time of day and time since last ...

show abstract

Section: Monitoring Using a Formal Loss Functionmentioning

confidence: 99%

Bayesian methods in health technology assessment: a review.

Spiegelhalter¹,

Myles²,

Jones³

et al. 2000

Health Technology Assessment

289

227

View full text Add to dashboard Cite

show abstract

“…Bayesian designs are compared with frequentist group sequential designs using decision theoretical approaches in Berry & Ho (1988) and Lewis & Berry (1994). Studies by Eales & Jennison (1992), Cressie & Biele (1994) and Barber & Jennison (2002), among others, search optimal group sequential designs under various settings using Bayesian decision theoretical approaches. The maximum sample size/block size is pre-determined for all these methods.…”

Section: Introductionmentioning

confidence: 99%

Bayesian adaptive designs for clinical trials

Cheng¹,

Shen²

2005

Biometrika

View full text Add to dashboard Cite

A Bayesian adaptive design is proposed for a comparative two-armed clinical trial using decision theoretical approaches. A loss function is specified to consider the cost for each patient, and the costs of making incorrect decisions at the end of a trial. At each interim analysis, the decision to terminate or to continue the trial is based on the expected loss function while concurrently incorporating efficacy, futility and cost. The maximum number of interim analyses is not prefixed but decided adaptively by the observed data. We derive explicit connections between the loss function and the frequentist error rates, so that the desired frequentist properties can be maintained for regulatory settings. The operating characteristics of the design are able to be evaluated on frequentist grounds. Extensive simulations are carried out to compare the proposed design with existing ones. The design is general enough to accommodate both continuous and discrete types of data. We illustrate the methods with an animal study evaluating a medical treatment for cardiac arrest.

show abstract

“…Their approach allows the trial to stop early if there is sufficient evidence to suggest that it is futile to continue. This approach has been extended by authors such as Lewis et al, 44,45 Cressie and Biele, 46 and Mü ller et al 47 More recently, Willan and Kowiger 48 considered the use of EVSI methods to determine the optimal sample size for multistage clinical trials. This work was extended by Chen and Willan 49 to an industry perspective.…”

Section: Reflection On the Current Methodological Literaturementioning

confidence: 99%

A Review of Clinical Trials With an Adaptive Design and Health Economic Analysis

et al. 2019

View full text Add to dashboard Cite

An adaptive design uses data collected as a clinical trial progresses to inform modifications to the trial. Hence, adaptive designs and health economics aim to facilitate efficient and accurate decision making. Nevertheless, it is unclear whether the methods are considered together in the design, analysis, and reporting of trials. This review aims to establish how health economic outcomes are used in the design, analysis, and reporting of adaptive designs. Methods: Registered and published trials up to August 2016 with an adaptive design and health economic analysis were identified. The use of health economics in the design, analysis, and reporting was assessed. Summary statistics are presented and recommendations formed based on the research team's experiences and a practical interpretation of the results. Results: Thirty-seven trials with an adaptive design and health economic analysis were identified. It was not clear whether the health economic analysis accounted for the adaptive design in 17/37 trials where this was thought necessary, nor whether health economic outcomes were used at the interim analysis for 18/19 of trials with results. The reporting of health economic results was suboptimal for the (17/19) trials with published results. Conclusions: Appropriate consideration is rarely given to the health economic analysis of adaptive designs. Opportunities to use health economic outcomes in the design and analysis of adaptive trials are being missed. Further work is needed to establish whether adaptive designs and health economic analyses can be used together to increase the efficiency of health technology assessments without compromising accuracy.

show abstract

A Sample-Size Optimal Bayesian Procedure for Sequential Pharmaceutical Trials

Cited by 4 publications

References 1 publication

Bayesian methods in health technology assessment: a review.

Bayesian methods in health technology assessment: a review.

Bayesian adaptive designs for clinical trials

A Review of Clinical Trials With an Adaptive Design and Health Economic Analysis

Contact Info

Product

Resources

About