Accuracy of Preliminary Birth Defect Surveillance in Utah: A Validation Study Supporting Timely Family Referrals.
Boyce Aubree A, Pocius Stephanie S
Public health surveillance is critical for understanding disease burden, yet rigorous case confirmation often introduces significant latency. The Utah Birth Defect Network (UBDN) utilizes a multistage active surveillance process that, while highly accurate, is a long and labor-intensive process. This delay creates a barrier to timely family referrals for early intervention services. This study evaluates whether preliminary surveillance data is sufficiently accurate to initiate public health actions before final clinician confirmation. We conducted a retrospective analysis of 27 monitored birth defects for delivery years 2018-2025. Positive predictive values (PPVs) were calculated for cases identified during initial medical record review by nonclinical staff compared with the "gold standard" of final clinician-led confirmation. Analysis was performed using SAS version 9.4, with 95% confidence intervals calculated via the Wilson score method. Among 1245 unique potential cases, the overall mean PPV was 68% (range: 8%-95%). Accuracy was highest for musculoskeletal defects (95%) and chromosomal anomalies (92%), with gastroschisis (95%), limb deficiencies (93%), and Trisomy 21 (92%) showing high stability. Cardiovascular defects demonstrated lower categorical accuracy (78%), with complex conditions like single ventricle (8%) showing the greatest diagnostic discordance. However, we observed categorical stability, as 77% of reclassified cardiovascular cases remained within the same body system. Preliminary surveillance reviews demonstrate high diagnostic stability for the majority of monitored conditions. These findings validate the use of preliminary data to trigger early family-centered referrals, significantly reducing the latency between case confirmation and intervention without compromising the integrity of formal reporting.