ABSTRACT
This chapter considers tests for association, effect measures, and confidence intervals for paired binomial probabilities. Paired binomial probabilities arise in study designs such as matched and cross-over clinical trials, longitudinal studies, and matched case-control studies. The data consist of two samples of dichotomous events: Event A and Event B. Each observation of Event A is matched with one observation of Event B. Stem cell transplantation (SCT) is a recognized treatment option for patients with haematological malignancies. SCT is, however, associated with pulmonary complications. In a prospective longitudinal study, Bentur et al. measured the airway hyper-responsiveness status of 21 children before and after SCT. A confidence interval for the number needed to treat is obtained by first calculating a confidence interval for the associated difference between probabilities. The calculation of a confidence interval for the number needed to treat is based on the confidence limits for the associated difference between probabilities.
