Pediatrics

Sampling ~25 comparators per rare-case often approximates full-cohort performance in simulations

Using Monte Carlo simulations of 100,000 children with congenital hypothyroidism, sampled-comparator designs (SC25) often matched full-cohort performance for power and precision. SC25 deviated ≤5% from full-cohort p-values for medium effects and ≤2% for large effects across many data-generating scenarios. None of the designs achieved ≥80% power for small or medium effects, and small effects remain hard to detect. This approach can inform study planning when whole-population data access is limited, but small-effect detection remains a limitation. 1

Haemoglobin thresholds that discriminate function in 6–30-month-olds cluster near WHO anaemia cut-offs

Pooled individual participant data from 11 datasets (27,626 children, eight countries) identified Hb thresholds that best discriminated concurrent and later functional outcomes. Best discriminatory Hb ranged roughly 102–111 g/L for concurrent physical activity and 109–117 g/L for longitudinal development. The WHO anaemia cut-offs fall in the middle-to-upper range for concurrent outcomes and lower range for predicting later development. These findings add functional evidence to inform Hb cut-offs but do not by themselves mandate immediate guideline changes. 2

Symptom-based dosing shortens time to readiness for discharge in ESC-managed NOWS

In a 23-hospital cluster crossover RCT of 626 infants with NOWS, symptom-based dosing reduced time to medical readiness for discharge in the ESC cohort (9.18 vs 11.61 days; aMR 0.79). There was no difference in initiation of pharmacologic treatment or overall length of stay between dosing strategies. About 35% of infants randomized to symptom-based dosing required scheduled dosing for uncontrolled withdrawal severity. Results applied to ESC-managed infants; no significant benefit was seen in the Finnegan-managed cohort. 3