The estimation of approximate sample size requirements necessary for clinical and epidemiological studies in vision sciences

Abstract

Purpose: The aim of this paper is to provide an overview of sample size estimations for the most frequent type of group studies that result in continuous, binary and ordered categorical outcomes.

Methods: The theory behind power and sample size calculations is explained using the basic probability concepts that underpin the most frequently used statistical significance tests.

Results: Simple formulae and tables are presented for the estimation of sample sizes necessary for efficient and effective clinical and epidemiological trials. These may be used without recourse to sophisticated and complex computer software packages. Mathematical complexity is kept to a minimum. Examples and applications from the vision sciences are specifically highlighted.

Conclusions: The paper highlights, with practical examples, the concepts and computations necessary to make sample size estimations accessible to all eye professionals involved in research, diagnostic and statutory work.