Combining Two Developmental Screening Tests to Improve Predictive Accuracy.

Area of Science:

• Pediatric developmental screening
• Clinical decision support

Background:

• Single developmental screening tests often lack sufficient predictive value for clinical decisions, with less than 50% of referred children actually exhibiting delays.
• Combining screening tools can potentially enhance accuracy and reduce false positives/negatives in developmental assessments.

Purpose of the Study:

• To evaluate the predictive value (PV) of combining the Ages and Stages Questionnaires (ASQ) and Parents Evaluation of Developmental Status (PEDS) screening tests.
• To assess how different likelihood ratio (LR) thresholds impact the accuracy of developmental delay referrals.

Main Methods:

• Utilized data from 328 children (ages 12-60 months) from primary care, including ASQ, PEDS, and diagnostic assessments.
• Calculated likelihood ratios (LRs) for various ASQ/PEDS combinations and analyzed predictive value using different referral thresholds.

Main Results:

• Setting a referral threshold at LR 2.0 for combined scores reduced false positives from 66 to 28 and improved positive predictive value (PPV) from 30% to 47%.
• Sensitivity decreased from 82% to 74%, while specificity improved from 72% to 90%.
• Increasing LR referral thresholds improved PV but created an inconclusive group requiring further assessment.

Conclusions:

• Combining ASQ and PEDS significantly enhances the predictive accuracy for developmental delay referrals.
• Higher referral thresholds improve predictive value but necessitate strategies for managing an emerging inconclusive group.
• Further research is needed to identify information aiding decision-making for children in the inconclusive category.