Neural tube defects (NTDs), including spina bifida, encephalocele, and anencephaly, are severe congenital conditions that are readily identifiable at birth and therefore less likely than other birth defects to be misdiagnosed or under-reported. We tested the hypothesis that, because of the relatively straightforward diagnosis of NTDs, accurate estimates of NTD prevalence can be directly estimated from reports to the Alaska Birth Defects Registry (ABDR)--a passive surveillance system receiving reports from all Alaska hospitals and other sources.
We reviewed medical records for potential NTD-affected births identified by the ABDR and Alaska birth certificates. Case verification findings were linked to Alaska birth certificate data and overall and specific positive predictive values (PPV) computed.
Alaska Birth Defects Registry-specific PPV was 44%. Positive predictive values did not differ by Alaska Native status even when data were stratified by number of NTD reports submitted, age at diagnosis, or birth year. Systematic exclusion of reporting sources with poor PPV resulted in a 28% improvement of PPV to 57%. The Alaska Native/non-Native ratio for unconfirmed NTD reports (0.7) was identical to the Alaska Native/non-Native ratio for confirmed NTD cases.
Unverified ABDR registry data is useful for identifying relative differences in NTD risk for Alaskan populations, but is unreliable for calculating NTD prevalence estimates or for evaluating preventive interventions. Systematic approaches to improving PPV may be useful for generating NTD prevalence estimates from unverified registry data when case verification is impractical.