Meta-Analysis of Rare Adverse Event Data - 1=Healthcare technologies 2=Biological Informatics

Many adverse events such as pneumonitis with EGFR and immune-oncology treatments, and liver toxicity with targeted agents and immune-oncology, occur at very low rates. Meta-analyses are used to increase the number of observations, however low event rate can still be an issue, e.g. the meta-analysis of cardiovascular events of rosiglitazone by Lane [1] and Cai et al [2] using Bayesian and Poisson random effects models. The objectives are: In collaboration with AstraZeneca (Cambridge) through data provision, co-supervision and placements at the company, 1) To increase our ability to understand rare adverse events in pharmacology through application of systems modelling, data analytics and statistical methodologies (e.g. zero-inflated Poisson or zero-inflated negative binomial) that can be applied to better detect the rate at which such events occur. 2) To propose best practice in analysing rare adverse events in clinical studies.
Expected results: A combined modelling and statistical approach to better determine the incidence of rare events compared to background levels that can be used for go/no-go guidance at early and late stages of drug development and to differentiate specific drugs from other drugs of the same class.
Refs: [1] Lane PW Meta-analysis of incidence of rare events Stat Methods Med Res 2013 22: 117, [2] Cai et al. Meta-Analysis For Rare Events Stat Med. 2010 September 10; 29(20): 2078-2089, [3] Zero-Inflated Boosted Ensembles for Rare Event Counts