BioMetrics.com

Update on Independent Review

Ohad Amit and Frank Mannino

Predictive patient recruitment and supply modelling to ensure successful completion of clinical trials

Dr. Vladimir Anisimov

Abstract: Patient recruitment is an essential stage of drug development process and well recognised bottleneck of the new drug development. Drug supply stage is very costly and substantially affected by the recruitment stage. Existing techniques of recruitment planning are mainly deterministic and do not account for: (a) the uncertainties in input information; (b) variation in recruitment across centres; (c) stochastic fluctuations of recruitment over time. A large proportion of trials fail to recruit in time. Therefore, it is imperative to develop statistical modelling approaches that can account for various uncertainties, can provide more accurate prediction and avoid extra supply overages.

In the first part, an advanced statistical methodology for modelling and predicting patient recruitment is discussed. It allows: 1) Predicting recruitment at the initial (design) stage; 2) Evaluating study performance and site productivity; 3) Predicting recruitment at any interim time using data-driven approach; 4) Evaluate the number of clinical centres needed to complete the recruitment in time with a given confidence and adaptively adjust the recruitment.
In the second part, a statistical technique for predictive modelling of drug supply using a risk-based approach is discussed. The results are illustrated on case studies.

In-vitro screening for combination drug discovery

John Peterson

Predictive patient recruitment and supply modelling to ensure successful completion of clinical trials

Vladimir Anisimov

Abstract: Patient recruitment is an essential stage of drug development process and well recognised bottleneck of the new drug development. Drug supply stage is very costly and substantially affected by the recruitment stage. Existing techniques of recruitment planning are mainly deterministic and do not account for: (a) the uncertainties in input information; (b) variation in recruitment across centres; (c) stochastic fluctuations of recruitment over time. A large proportion of trials fail to recruit in time. Therefore, it is imperative to develop statistical modelling approaches that can account for various uncertainties, can provide more accurate prediction and avoid extra supply overages.

In the first part, an advanced statistical methodology for modelling and predicting patient recruitment is discussed. It allows: 1) Predicting recruitment at the initial (design) stage; 2) Evaluating study performance and site productivity; 3) Predicting recruitment at any interim time using data-driven approach; 4) Evaluate the number of clinical centres needed to complete the recruitment in time with a given confidence and adaptively adjust the recruitment.
In the second part, a statistical technique for predictive modelling of drug supply using a risk-based approach is discussed. The results are illustrated on case studies.

In-vitro screening for combination drug discovery

John J. Peterson

Recommendations for the primary analysis of continuous endpoints in longitudinal clinical trials

Peter Lane

Abstract: This talk is based on a position paper prepared by a PhRMA expert team (of which I was a member) and published last year in the Drug Information Journal. It is one strand of the PhRMA initiative to improve efficiency of late-stage clinical research. I will summarize relevant theory and current practice in the analysis of longitudinal clinical trials, concentrating on continuous response and a linear model with the goal of estimating and testing treatment differences at a given time point. Traditionally, the primary analysis of such trials handled missing data by simple imputation using the method of last, or baseline, observation carried forward (LOCF, BOCF) followed by analysis of (co)variance at the chosen time point. However, the general statistical and scientific community has moved away from these simple methods in favour of joint analysis of data from all time points based on a multivariate model (e.g., of a mixed-effects type). One such newer method, a likelihood-based mixed-effects model repeated-measures (MMRM) approach, has received considerable attention in the clinical trials literature. I will discuss specific concerns raised by regulatory agencies with regard to MMRM and review published evidence comparing LOCF and MMRM in terms of validity, bias, power, and Type I error. The main conclusion is that the mixed-model approach is more efficient and reliable as a method of primary analysis, and should be preferred to the inherently biased and statistically invalid simple imputation approaches.

Poster: Acinetobacter baumannii 2002-2007: Mathematical model insights on imipenem non susceptibility and multi-class resistance

Robertino M Mera, Linda A Miller, Heather Amrine-Madsen, Daniel F Sahm

Meta-analysis in drug safety evaluation

Peter Lane

Abstract: Meta-analysis is being carried out increasingly to investigate potential safety issues with new or existing drugs. The results can be extremely important, both to patients and to drug companies, but there are many potential problems in carrying out and interpreting the analysis. Some are due to difficulties accessing data, because the stress on reports and publications is often on efficacy and any safety issues known a priori. Others are due to the binary nature of many safety events, possibly compounded by rareness leading to difficulties in computation and approximation. I shall review these problems with reference to recent experience in GlaxoSmithKline, and to some recently published meta-analyses.

Slides: Repeated measures approach to “What if” questions for longitudinal studies with withdrawal

Handout: Repeated measures approach to “What if” questions for longitudinal studies with withdrawal

Mike Kenward and James Roger

Poster: A Bayesian approach to assessing the monotonicity of a dose response curve: a case study

Amanda Deans and James Roger