JAMA Netw Open. 2026 Sep 1;9(9):e2631747. doi: 10.1001/jamanetworkopen.2026.31747.
ABSTRACT
IMPORTANCE: Gestational age at birth plays a critical part in childhood neurodevelopment, especially for individuals born preterm (<37 weeks’ gestation).
OBJECTIVE: To investigate the association between gestational age and educational outcomes at primary and high school age, controlling for unmeasured confounding by family-level factors.
DESIGN, SETTING, AND PARTICIPANTS: This population-based cohort study conducted in Australia included all individuals born at 23 to 42 weeks’ gestation in New South Wales (NSW) between 2001 and 2011 and in Victoria between 2005 and 2012 who had school outcomes assessed by standardized assessments in grade 3 (age 8-9 years) and/or grade 9 (age 14-15 years). Analyses were conducted between June 2024 and February 2026.
EXPOSURE: Gestational age at birth.
MAIN OUTCOMES AND MEASURES: Standardized numeracy and reading assessment z-scores were calculated in grade 3 and grade 9 and compared across gestational age groups using multivariable general linear mixed models. A clinically meaningful difference was considered an adjusted mean difference (AMD) in z-score of 0.2 or greater.
RESULTS: The study included 1 095 816 children in grade 3 (772 589 NSW-born [50.7% male] and 323 227 Victoria-born [50.5% male]) and 292 027 NSW-born adolescents in grade 9 (50.7% male). Among children assessed at grade 3, 6.4% in NSW and 7.0% in Victoria were born before 37 weeks’ gestation and 54.1% in NSW and 41.1% in Victoria were born at 39 to 40 weeks’ gestation; 6.3% of NSW adolescents assessed at grade 9 were born before 37 weeks’ gestation and 53.8% at 39 to 40 weeks’ gestation. Compared with children born at 39 to 40 weeks, children born at 38 weeks had lower numeracy (AMD, -0.02; 95% CI, -0.03 to -0.02) and reading (AMD, -0.02; 95% CI, -0.02 to -0.01) z-scores in grade 3. z-Scores decreased in a stepwise manner for each gestational age group, with the greatest difference for those born at 23 to 27 weeks vs 39 to 40 weeks (numeracy: AMD, -0.54 [95% CI, -0.58 to -0.49]; reading: AMD, -0.29 [95% CI, -0.34 to -0.25]). However, among siblings at grade 3, these differences were attenuated, and clinically meaningful differences compared with birth at 39 to 40 weeks were only seen for numeracy in children born at 28 to 31 weeks (AMD, -0.23; 95% CI, -0.27 to -0.19) and 23 to 27 weeks (AMD, -0.45; 95% CI, -0.53 to -0.38) and for reading, only for children born at 23 to 27 weeks (AMD, -0.28; 95% CI, -0.36 to -0.21). Results were similar in grade 9, with clinically meaningful differences for numeracy (AMD, -0.53; 95% CI, -0.62 to -0.45) and reading (AMD, -0.39; 95% CI, -0.48 to -0.30) only seen for adolescents born at 23 to 27 vs 39 to 40 weeks’ gestation in the full-cohort analysis.
CONCLUSIONS AND RELEVANCE: In this cohort study of over 1 million Australian children, the independent association of gestational age at birth with educational performance was most pronounced for children born at less than 28 weeks’ gestation; for children born from 28 to 38 weeks’ gestation, differences were not clinically meaningful and were likely primarily attributable to unmeasured confounding by shared familial factors. These findings support prioritizing prevention of the earliest preterm births and suggest that children born before 28 weeks’ gestation may benefit from early identification of learning difficulties and targeted educational support.
PMID:42678696 | DOI:10.1001/jamanetworkopen.2026.31747
