Predicting neurodevelopmental delay among two-year-old infants in Northern Uganda.
David Mukunya, Grace Ndeezi, Gertrude Namazzi et al.
✦ AI-curated · Sources linked
In 30 seconds
This study developed and evaluated a risk-prediction model for neurodevelopmental delay among two-year-old children in Northern Uganda. Children were followed from a cluster-randomized trial cohort, and development was assessed using the Malawi Developmental Assessment Tool. Among 1,421 children assessed, 70 (4.9%; 95% CI: 3.6-6.7%) had delays. Six readily obtainable factors were retained in the model, which had an area under the curve (AUC) of 0.75 and reasonable calibration. The findings suggest that maternal, birth, and early infant characteristics may help identify children at increased risk.
Key findings
- Neurodevelopmental delay was identified in 70 of 1,421 children (4.9%; 95% CI: 3.6-6.7%).
- The final model retained moderate wasting, prolonged neonatal hospitalization, maternal education, place of birth, preterm birth, and absence of skin-to-skin contact.
- The model's discriminative performance was AUC 0.75, with a calibration slope of 1.37 and calibration belt 0.536-0.727.
Why it matters
Neurodevelopmental assessment and follow-up resources may be limited in low-resource settings. A model based on routinely obtainable maternal, birth, and early infant characteristics could help identify children who may warrant closer developmental follow-up. Its reported discrimination is promising, but the findings concern risk prediction and do not establish that using the model improves developmental outcomes.
What to keep in mind
The model was developed and assessed in children from a single regional cohort, and 21.2% of the children enrolled at the trial's final visit were not followed up.