Bayesian Methods / Small Population Trials

A qualitative assessment of Bayesian elicitation via SHELF (Quality TREATs)

Qualitative Assessment of the SHELF protocol

Team: Giles Partington, Suzie Cro, Alexina Mason, Victoria Cornelius

As part of the TREAT trial (a trial in serious childhood asthma) a Bayesian prior was elicited through the SHELF procedure (a method of gaining consensus between expert opinions). This procedure has not been frequently used in late stage clinical trials; therefore, during this session qualitative interviews were carried out to evaluate the robustness of this technique.

The findings from this review will be added once published.

Eliciting Informed Priors for Small Population RCTs

Assessing current use of Bayesian methods in rare disease and small population randomised controlled trials (RCTs).

Team: Giles Partington, Suzie Cro, Alexina Mason, Victoria Cornelius

Large populations are often infeasible in the case of rare disease trials, making reaching sensible power calculations difficult. Alternatives such as observational studies are used instead, one such option is using a Bayesian approach. Bayesian methods require prior distributions for all unknown parameters. There are several ways of eliciting prior distributions. We were interested in identifying how often Bayesian methods are used, which elicitation methods they use, and how frequentist rare disease RCTs handle the lack of population.

The findings from this review will be added once published.