Roland Fisch

Experimental designs for small randomised clinical trials: An algorithm for choice

BackgroundSmall clinical trials are necessary when there are difficulties in recruiting enough patients for conventional frequentist statistical analyses to provide an appropriate answer. These trials are often necessary for the study of rare diseases as well as specific study populations e.g. children. It has been estimated that there are between 6,000 and 8,000 rare diseases that cover a broad range of diseases and patients. In the European Union these diseases affect up to 30 million people, with about 50% of those affected being children. Therapies for treating these rare diseases need their efficacy and safety evaluated but due to the small number of potential trial participants, a standard randomised controlled trial is often not feasible. There are a number of alternative trial designs to the usual parallel group design, each of which offers specific advantages, but they also have specific limitations. Thus the choice of the most appropriate design is not simple.MethodsPubMed was searched to identify publications about the characteristics of different trial designs that can be used in randomised, comparative small clinical trials. In addition, the contents tables from 11 journals were hand-searched. An algorithm was developed using decision nodes based on the characteristics of the identified trial designs.ResultsWe identified 75 publications that reported the characteristics of 12 randomised, comparative trial designs that can be used in for the evaluation of therapies in orphan diseases. The main characteristics and the advantages and limitations of these designs were summarised and used to develop an algorithm that may be used to help select an appropriate design for a given clinical situation. We used examples from publications of given disease-treatment-outcome situations, in which the investigators had used a particular trial design, to illustrate the use of the algorithm for the identification of possible alternative designs.ConclusionsThe algorithm that we propose could be a useful tool for the choice of an appropriate trial design in the development of orphan drugs for a given disease-treatment-outcome situation.

Bayesian Design of Proof-of-Concept Trials

The proof-of-concept (PoC) decision is a key milestone in the clinical development of an experimental treatment. A decision is taken on whether the experimental treatment is further developed (GO), whether its development is stopped (NO-GO), or whether further information is needed to make a decision. The PoC decision is typically based on a PoC clinical trial in patients comparing the experimental treatment with a control treatment. It is important that the PoC trial be designed such that a GO/NO-GO decision can be made. The present work develops a generic, Bayesian framework for defining quantitative PoC criteria, against which the PoC trial results can be assessed. It is argued that PoC criteria based solely on significance testing versus the control are not appropriate in this decision context. A dual PoC criterion is proposed that includes assessment of superiority over the control and relevance of the effect size and hence better matches clinical decision making. The approach is illustrated for 2 PoC trials in cystic fibrosis and psoriasis.

A Simplified Approach to Calibration Confidence Sets

Abstract The problem of making inference statements in calibration is treated rather heterogeneously in the literature. In this article, the classical calibration confidence sets for straight line regression are calculated in a new and easy-to-understand way by formulating a nonlinear regression model, where the value of interest is an additional parameter. Possible different types of confidence sets are discussed and a generalization to linear combinations is briefly described.

Regulatory watch: Efficiency indicators for new drugs approved by the FDA from 2003 to 2013

Regulatory watch: Efficiency indicators for new drugs approved by the FDA from 2003 to 2013

Assessing switchability for biosimilar products: modelling approaches applied to children's growth

The present paper describes two statistical modelling approaches that have been developed to demonstrate switchability from the original recombinant human growth hormone (rhGH) formulation (Genotropin(®) ) to a biosimilar product (Omnitrope(®) ) in children suffering from growth hormone deficiency. Demonstrating switchability between rhGH products is challenging because the process of growth varies with the age of the child and across children. The first modelling approach aims at predicting individual height measured at several time-points after switching to the biosimilar. The second modelling approach provides an estimate of the deviation from the overall growth rate after switching to the biosimilar, which can be regarded as an estimate of switchability. The results after applying these approaches to data from a randomized clinical trial are presented. The accuracy and precision of the predictions made using the first approach and the small deviation from switchability estimated with the second approach provide sufficient evidence to conclude that switching from Genotropin(®) to Omnitrope(®) has a very small effect on growth, which is neither statistically significant nor clinically relevant.

A nonparametric method to detect increased frequencies of adverse drug reactions over time: Method to Detect Increased Frequencies of Adverse Drug Reactions

Signal detection is routinely applied to spontaneous report safety databases in the pharmaceutical industry and by regulators. As an example, methods that search for increases in the frequencies of known adverse drug reactions for a given drug are routinely applied, and the results are reported to the health authorities on a regular basis. Such methods need to be sensitive to detect true signals even when some of the adverse drug reactions are rare. The methods need to be specific and account for multiplicity to avoid false positive signals when the list of known adverse drug reactions is long. To apply them as part of a routine process, the methods also have to cope with very diverse drugs (increasing or decreasing number of cases over time, seasonal patterns, very safe drugs versus drugs for life-threatening diseases). In this paper, we develop new nonparametric signal detection methods, directed at detecting differences between a reporting and a reference period, or trends within a reporting period. These methods are based on bootstrap and permutation distributions, and they combine statistical significance with clinical relevance. We conducted a large simulation study to understand the operating characteristics of the methods. Our simulations show that the new methods have good power and control the family-wise error rate at the specified level. Overall, in all scenarios that we explored, the method performs much better than our current standard in terms of power, and it generates considerably less false positive signals as compared to the current standard.