Accrual monitoring in cardiovascular trials

Prior Elicitation for Use in Clinical Trial Design and Analysis: A Literature Review

Bayesian inference is increasingly popular in clinical trial design and analysis. The subjective knowledge derived from an expert elicitation procedure may be useful to define a prior probability distribution when no or limited data is available. This work aims to investigate the state-of-the-art Bayesian prior elicitation methods with a focus on clinical trial research. A literature search on the Current Index to Statistics (CIS), PubMed, and Web of Science (WOS) databases, considering “prior elicitation” as a search string, was run on 1 November 2020. Summary statistics and trend of publications over time were reported. Finally, a Latent Dirichlet Allocation (LDA) model was developed to recognise latent topics in the pertinent papers retrieved. A total of 460 documents pertinent to the Bayesian prior elicitation were identified. Of these, 213 (45.4%) were published in the “Probability and Statistics” area. A total of 42 articles pertain to clinical trial and the majority of them (81%) reports parametric techniques as elicitation method. The last decade has seen an increased interest in prior elicitation and the gap between theory and application getting narrower and narrower. Given the promising flexibility of non-parametric approaches to the experts’ elicitation, more efforts are needed to ensure their diffusion also in applied settings.

Modelling the patient accrual with truncated Gaussian mixture distribution for the accurate estimation of sample size in survival trials

In survival trials with fixed trial length, the patient accrual rate has a significant impact on the sample size estimation or equivalently, on the power of trials. A larger sample size is required for the staggered patient entry. During enrollment, the patient accrual rate changes with the recruitment publicity effect, disease incidence and many other factors and fluctuations of the accrual rate occur frequently. However, the existing accrual models are either over-simplified for the constant rate assumption or complicated in calculation for the subdivision of the accrual period. A more flexible accrual model is required to represent the fluctuant patient accrual rate for accurate sample size estimation. In this paper, inspired by the flexibility of the Gaussian mixture distribution in approximating continuous densities, we propose the truncated Gaussian mixture distribution accrual model to represent different variations of accrual rate by different parameter configurations. The sample size calculation formula and the parameter setting of the proposed accrual model are discussed further.

Statistical models to predict recruitment in clinical trials were rarely used by statisticians in UK and European networks

OBJECTIVE

Identify the current practice for recruitment prediction and monitoring within clinical trials.

STUDY DESIGN AND SETTING

Chief investigators (CIs) were surveyed to identify data sources and adjustments made to support recruitment prediction. Statisticians were surveyed to determine methods and adjustments used when predicting and monitoring recruitment. Participants were identified from the National Institute for Health Research recently funded studies, the UK Clinical Research Collaboration registered Clinical Trial Units network or by the European Clinical Research Infrastructure Network.

RESULTS

A total of 51 CIs (UK = 32, ECRIN = 19) and 104 statisticians (UK = 51, ECRIN = 53) were contacted. Response rates varied (CIs UK = 53% ECRIN = 32%; statisticians UK = 98% ECRIN = 36%). Multiple data sources are used to support recruitment rates, most commonly audit data from multiple sites. Variation in individual site recruitment rates are frequently incorporated, but staggered site openings were featured more commonly among UK respondents. Simple prediction methods are preferred to rarely used statistical models. Lack of familiarity with statistical methods are barriers to their use with evidence needed to justify the time required to support their implementation.

CONCLUSION

Simplistic methods will continue as the mainstay of prediction; however, generation of evidence supporting the benefits of complex statistical models should promote their implementations. Multiple data sources to support recruitment prediction are being used, and further work on the quality of these data is needed. Pressure to be optimistic about recruitment rates for the trial to be attractive to funders was felt by a sizable minority.