Predictive Value of the BSID-II and the Bayley-III for Early School Age Cognitive Function in Very Preterm Infants

OBJECTIVE: To compare the predictive validity of the Bayley Scales of Infant Development, Second Edition (BSID-II) and the Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III) for cognitive function at early school age in very preterm infants. METHODS: Seventy-seven former preterm infants (born <32 weeks gestation and ≤2000 g) completed both the BSID-II and the Bayley-III at 2 years corrected age. Children enrolled at hospitals that perform follow-up beyond 2 years had cognitive assessments with the Wechsler Preschool and Primary Scale of Intelligence Fourth Edition (WPPSI-IV). Associations between Bayley and WPPSI scores were assessed using correlation coefficients, linear regression, and Bland-Altman plots. RESULTS: Thirty-one of 45 eligible children were tested with the WPPSI-IV at 47 ± 11 months. Average BSID-II Mental Development Index (MDI) was 86 ± 19, Bayley-III Cognitive composite score was 101 ± 12 and WPPSI Full Scale IQ (FSIQ) was 96 ± 12. Correlation between MDI and FSIQ was 0.54 (P < .001); correlation between Bayley-III cognitive composite score and FSIQ was 0.31 (P = .03). Bayley-III language composite had a modestly stronger correlation with FSIQ than cognitive composite (correlation coefficient 0.39; P = .005). Linear regression models also demonstrated that BSID-II was more closely correlated with FSIQ than Bayley-III. This bias was consistent across the full range of scores. CONCLUSION: The BSID-II underestimated FSIQ and the Bayley-III overestimated FSIQ. Children at risk for impairment might be missed with the Bayley-III. As the Bayley-4 is introduced, clinicians and researchers should be cautious about interpretation of scores until performance of this new measure is fully understood.