Please activate JavaScript!
Please install Adobe Flash Player, click here for download


ISCB 2014 Vienna, Austria • Abstracts - Oral Presentations 47Tuesday, 26th August 2014 • 11:00-12:30 Monday25thAugustTuesday26thAugustThursday28thAugustAuthorIndexPostersWednesday27thAugustSunday24thAugust the trial. Early stopping for both futility and efficacy will be discussed, for trials of varying numbers of drugs and stages. Moreover, controlling the Family Wise Error Rate of such designs will also be explored. Using results from the four-treatment four-period TOMADO trial for sleep apnoea de- vices, the performance of the group sequential cross-over designs will be examined. C24.3 Optimal sequential clinical trials in small populations S Nikolakopoulos1 , KCB Roes1 , I van der Tweel1 1 UMC Utrecht, Utrecht, The Netherlands A sequential clinical trial will on average require fewer patients to reach a conclusion, possibly resulting in superior treatments becoming avail- able faster. The vast majority of the developed and most commonly used sequential methods rely on asymptotic distributions of the test statistics. Recent guidelines and published research suggest the use of sequential methods, among other options, as an alternative for the design of clinical trials in orphan diseases. In this work we thus investigate the behavior of sequential tests in small to very small sample sizes. We explore their operational characteristics and point out - usually overlooked - simple corrections that can preserve type 1 errors accurately. In addition we look into the case where the sample size has a given maximum due to, e.g., a very rare disease. When exploring such a case in a fixed-sample size, design and sample size considerations might become irrelevant and, for a given type I error level, there is one design with a respective power function. In the sequential case however, different design ingredients - e.g. timing and number of interim analysis , efficacy and futility boundaries - can lead to designs with different op- erational characteristics and the choice will matter in terms of power and efficiency. We suggest an optimization rule taking into consideration a maximum sample size and prior belief of the treatment effect.We illustrate our meth- od with a real example using a clinical trial for a rare disease. C24.4 Group sequential designs for verifying whether effective drug concentrations are similar in adults and children L Hampson1 , T Jaki1 , R Fisch2 1 Lancaster University, Lancaster, United Kingdom, 2 Novartis Pharma AG, Basel, Switzerland New medicines for children should be subject to rigorous testing while avoiding unnecessary experimentation in children. In particular, paedi- atric dosing recommendations should be informed by existing relevant data. If the effective concentration of a drug can be assumed to be similar in adults and children, an appropriate paediatric dosing rule may be found by ‘bridging’, that is, conducting pharmacokinetic studies in children to find doses that produce concentrations therapeutic in adults. However, this strategy may result in children receiving an ineffective or hazardous dose if, in fact, effective concentrations differ between adults and children. When there is uncertainty about the equality of effective concentrations, some pharmacokinetic-pharmacodynamic (PK-PD) data may be needed in children to verify whether differences between adults and children are small. In this presentation, we develop adaptive procedures that can be used to verify this assumption efficiently. Asymmetric inner wedge group sequential tests are constructed which permit early stopping to accept or reject an assumption of similar effective drug concentrations in adults and children. Asymmetry arises because the consequences of under- and over- dosing may differ. Using exact calculations we compare the efficiency of error spending inner-wedge tests with optimal designs which minimise the expected sample size needed to reach a conclusion. If there is time, we will show how stopping rules based on predictive tail area probabili- ties can be derived for testing whether observed paediatric PK-PD data are consistent with an assumption of similar effective concentrations in adults and children. C24.5 Group-sequential strategies when considering multiple outcomes as co-primary in clinical trials T Hamasaki1 , K Asakura1 , SR Evans2 1 Osaka University Graduate School of Medicine, Suita, Japan, 2 Harvard School of Public Health, Boston, United States   Many recent clinical trials, especially in pharmaceutical drug develop- ment, have utilized more than one endpoint as co-primary, thus evaluat- ing the intervention’s multidimensional effects. In this paper, we consider group sequential strategies in clinical trials with multiple co-primary endpoints when appropriate planning for the varying number of analyses and information space, is conducted, where the trial is designed to evaluate if the intervention is superior to the control on ALL endpoints. Then we investigate operating characteristics of group se- quential strategies in terms of overall power, Type I error rate and sample sizes. Based on the investigations, selecting different numbers of analyses for each endpoint with equal/unequal information space among the end- points may not decrease the overall power and will not inflate the over- all Type I error rate. Strategic selection regarding the testing procedure and the number of analyses with equal/unequal information space may reduce the average sample number. Early interim evaluations should be carefully evaluated as they can provide higher power, but larger average sample numbers. These results are useful when constructing efficient group-sequential strategies in clinical trials with multiple co-primary endpoints.

Pages Overview