We reviewed Atlanta area hospital records to determine the following regarding neonatal circumcision: incidence in July 1985; incidence after publicized serious complications of circumcision in August 1985; medical record documentation; and the complication rate. After stratified sampling from hospital birth logs, we abstracted information from medical charts and calculated weighted estimates and P values. The circumcision incidence was 89.3% in July 1985, 87.5% in September 1985, and 84.3% in September 1986. Circumcision was recorded on the medical record face sheet for 84.3% of circumcised boys. The complication rate was 3.1%; no serious complications were recorded. We conclude the following: circumcision incidence was high during the study period; publicity regarding adverse outcomes may have decreased the subsequent incidence of the procedure; hospital discharge data, which rely on medical record face sheet information, underestimate the true incidence of neonatal circumcision; and neonatal circumcision is usually safe, but serious complications may occur.