Use of capture-recapture analyses in fetal alcohol syndrome surveillance in Alaska.

Abstract

Capture-recapture methods were used to estimate the prevalence of fetal alcohol syndrome among Alaska Natives born during the period 1982-1989. Potential cases were identified through an Indian Health Service (IHS) patient case file, a pediatric practice case file, and Medicaid claims from private physicians. A total of 74 Alaska Native children aged 3-10 years were identified with a notation of fetal alcohol syndrome by a physician in a medical record. Because not all of these cases had supporting documentation regarding the syndrome, they were classified as possible cases. Of these possible cases, 50 met all five criteria for chart verification of the syndrome: physician notation of fetal alcohol syndrome, growth deficiency, facial features of the syndrome, central nervous system impairment, and a maternal history of alcohol abuse. These data provided observed prevalence rates of chart-verified fetal alcohol syndrome of 3.1 per 1,000 live births for children born 1982-1985 (age 7-10 years), and 2.0 per 1,000 live births for children born 1986-1989 (age 3-6 years). Capture-recapture analyses were conducted using cases identified by IHS and private physicians. These analyses estimated a prevalence of the syndrome of 3.8 per 1,000 live births for children born 1982-1985, and 3.1 per 1,000 live births for children born 1986-1989. Based on the capture-recapture predicted number of cases, the IHS case file ascertained a greater percentage of cases among the older cohort (75%) than among the younger cohort (56%). These data illustrate the use of capture-recapture analyses in identifying the extent to which observed trends in rates may reflect differences in cases ascertainment over time (or by birth cohort). The application of capture-recapture in fetal alcohol syndrome surveillance, however, requires careful attention to the underlying assumptions of capture-recapture methods.