Sample size determination for a Bayesian cost-effectiveness model with structural zero costs

Abstract

The idea that medical treatment costs and outcomes might be connected is not new. Likewise, as long as researchers have been designing clinical trials and public opinion polls, there has been interest in the sample size necessary to obtain a desired level of precision and certainty before collecting the data. However, researchers continue to adapt cost-effectiveness models for scenarios of ever-increasing complexity, and equally adaptable sample size determination schemes are required. Of interest for the current study are those instances wherein a non-trivial proportion of patients incur zero costs associated with their treatment. We propose a sample size determination scheme for a cost-effectiveness model fit to such a scenario. Furthermore, we display our method’s usefulness on multiple parameter configurations derived from applications presented in already published research.

Keywords:

• Sample size determination
• Cost-effectiveness models
• Bayesian methods

Disclosure statement

Both authors declare that there is no conflict of interest associated with this work. Furthermore, this research was not supported by any grant funding.