Regional bias in birth defect prevalence rates for Arkansas: influence of incomplete ascertainment along surveillance system borders.

Abstract

As part of the continuing evaluation of the Arkansas Reproductive Health Monitoring System (ARHMS), we assessed the effects on birth defect prevalence rates introduced by incomplete case ascertainment along surveillance boundaries. Using data from ARHMS and Arkansas Vital Statistics for 1993-1998, we determined birth defect prevalence rates (per 10,000 live births), stratified by race, among three geographic comparison groups of counties. These included: (1) the Northeast Group, near the state border at Memphis, Tennessee; (2) the Central Group, surrounding Little Rock, Arkansas; and (3) the Southwest Group, near Texarkana, Texas. These counties have similar socioeconomic measures and proximity to health care facilities, but are differentiated by limitations imposed by ARHMS' surveillance borders. Maternal age-standardized rates from the control groups were used to impute expected rates, for the Northeast Group and statewide, which were compared with reported rates. We found that there were 620 fewer reported birth defect cases than expected for the Northeast Group. The Northeast Group's prevalence rates were approximately half of the control groups' rates (310.6 vs. 529.8, respectively, for Whites, and 240.8 vs. 550.1, respectively, for African-Americans). Incorporating the missed cases into statewide prevalence calculations could increase prevalence rates from 502.6 to 523.2 for Whites and from 527.4 to 590.7 for African-Americans. This study identified significant regional differences in reported birth defect rates in Arkansas. Case ascertainment might be incomplete in other surveillance systems lacking the means to share data with neighboring systems. Regional inaccuracy can hinder evaluation of localized birth defect trends or targeted prevention efforts.