Stable computational methods for additive binomial models with application to adjusted risk differences

Mark W. Donoghoe*, Ian C. Marschner

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)

Abstract

Risk difference is an important measure of effect size in biostatistics, for both randomised and observational studies. The natural way to adjust risk differences for potential confounders is to use an additive binomial model, which is a binomial generalised linear model with an identity link function. However, implementations of the additive binomial model in commonly used statistical packages can fail to converge to the maximum likelihood estimate (MLE), necessitating the use of approximate methods involving misspecified or inflexible models. A novel computational method is proposed, which retains the additive binomial model but uses the multinomial-Poisson transformation to convert the problem into an equivalent additive Poisson fit. The method allows reliable computation of the MLE, as well as allowing for semi-parametric monotonic regression functions. The performance of the method is examined in simulations and it is used to analyse two datasets from clinical trials in acute myocardial infarction. Source code for implementing the method in R is provided as supplementary material (see Appendix A).

Original languageEnglish
Pages (from-to)184-196
Number of pages13
JournalComputational Statistics and Data Analysis
Volume80
DOIs
Publication statusPublished - Dec 2014

Keywords

  • Additive binomial model
  • Multinomial-Poisson transformation
  • Risk difference
  • Semi-parametric regression

Fingerprint Dive into the research topics of 'Stable computational methods for additive binomial models with application to adjusted risk differences'. Together they form a unique fingerprint.

Cite this