School of Mathematics

Drugs and the Statistics group

Congratulations are due to two members of the Statistics research group who had adjacent papers in a recent issue of the Journal of the Royal Statistical Society, Series C, Applied Statistics.

JRSS C is a journal with long-standing reputation in attracting high-quality papers from a wide range of application areas, in which specific problems lead to interesting statistical challenges and creative solutions to these challenges.

The first was by Colin Aitken and Amy Wilson (now at the University of Durham) and co-authors from Mass Spec Analytical in Filton near Bristol:

The evaluation of evidence for auto-correlated data in relation to traces of cocaine on banknotes.

There has been much work in recent years in the development of statistical models for the evaluation of forensic scientific evidence for multivariate hierarchical random effects continuous models with independent data. The methods described in this article extend the univariate model to autocorrelated data. Application of the methods is illustrated with data concerning the quantities of cocaine on seizures of banknotes. Calculations are made for the strength of the support provided by the evidence for the proposition that the banknotes are associated with a person who is associated with a criminal activity involving cocaine in contrast to the proposition that the banknotes are associated with a person who is not associated with a criminal activity involving cocaine.”

The second was by Ioannis Papastathopoulos with a co-author from Lancaster University:

Stochastic ordering under conditional modelling of extreme values: drug-induced liver injury.

The clinical trial statistician concerned with efficacy is concerned with characterising the expected response patients have to an experimental drug. When it comes to safety, what is important is the characterization of the unexpected responses, i.e., the extreme values. For safety data, unlike efficacy, the questions are typically not well-defined, some of them will not be known until the data has been studied in some detail, and the data are usually messy. For these reasons, statisticians have tended to shy away from the analysis and particularly modelling of safety data, leaving interpretation to clinicians. In this article, new statistical methods that identify signals of drug toxicity are developed and applied to safety data from a phase 3 clinical trial of a drug that has been linked to toxicity. The methods allow for the testing of the hypothesis of ordered dependence between doses in laboratory variables and for estimating the probability of joint extreme elevations. This work was supported by AstraZeneca and is used to aid their decision making in accepting/rejecting experimental drugs“