This paper describes the Antiretroviral Pregnancy Registry's (APR) monitoring and analysis plan. APR is overseen by a committee of experts in obstetrics, pediatrics, teratology, infectious diseases, epidemiology and biostatistics from academia, government and the pharmaceutical industry. APR uses a prospective exposure-registration cohort design. Clinicians voluntarily register pregnant women with prenatal exposures to any antiretroviral therapy and provide fetal/neonatal outcomes. A birth defect is any birth outcome > or = 20 weeks gestation with a structural or chromosomal abnormality as determined by a geneticist. The prevalence is calculated by dividing the number of defects by the total number of live births and is compared to the prevalence in the CDC's population-based surveillance system. Additionally, first trimester exposures, in which organogenesis occurs, are compared with second/third trimester exposures. Statistical inference is based on exact methods for binomial proportions. Overall, a cohort of 200 exposed newborns is required to detect a doubling of risk, with 80% power and a Type I error rate of 5%. APR uses the Rule of Three: immediate review occurs once three specific defects are reported for a specific exposure. The likelihood of finding three specific defects in a cohort of < or = 600 by chance alone is less than 5% for all but the most common defects. To enhance the assurance of prompt, responsible, and appropriate action in the event of a potential signal, APR employs the strategy of 'threshold'. The threshold for action is determined by the extent of certainty about the cases, driven by statistical considerations and tempered by the specifics of the cases.
Copyright 2004 John Wiley & Sons, Ltd.