Findings from this study support our hypothesis that SDOH, specifically public insurance (as a marker of poverty), minority race, and non-English primary language, independently increase the risk of low BSID-III language scores in preterm infants at 18–24 months corrected age. Infants with public insurance had a significantly lower mean BSID-III language composite score, and were more likely to have a score more than one SD below the mean compared to those with private insurance. Mean receptive and expressive communication subscales were also significantly lower than those of children with private insurance.
According to Rhode Island Department of Health, from 2012 to 2018 the percentage of individuals in the state who had public insurance increased from 31–42.2%.(21) Data from Rhode Island Kids Count 2023 Factbook indicates 45% of infants born in RI were on public insurance.(22) Among the infants participating in this preterm study, 57% had public insurance. With the increasing percentage of families with public insurance, more infants are at risk of developmental delays. Green et al.(23) previously reported that public insurance was associated with decreased BSID-III total language composite score and receptive language subscale test scores between 8 to 20 months. Other studies have shown that race/ethnicity and primary language have effects on language outcomes with POC children scoring lower on BSID-III language compared to White race children.(7, 24) Family health literacy can also amplify the differences in health outcomes by affecting healthcare use, relationship with provider, and health behavior, especially if there is a language barrier or cultural dissimilarities.(25) In addition, decreased access to resources has been linked to cognitive impairment and adverse effects on learning.(15) Comfort et al.(13) and Jang et al.(17) describe the effects that type of health insurance can have on the accessibility and quality of services offered. Studies have also shown that public insurance is a risk factor associated with poverty, and with less optimal perinatal and post-discharge outcomes.(14, 16, 17) As expected, public insurance was also associated with lower cognitive scores in bivariate analysis and independently in multivariable analysis, demonstrating that type of health insurance is a valid marker for SDOH.
Language development in infants starts through daily interactions and experiences with caregivers, even prior to speaking their first words.(26–28) The exposure to early sensory experiences and language postnatally has a considerable influence on the developing brain, since the first few years of life represent a critical time for neural circuitry expansion related to language development, hearing, vision, and socialization, especially for preterm neonates.(29–31) Maternal report of adverse mental health was more prevalent in the public insurance group, but did not achieve significance in our regression modeling when controlled for other factors. There continues to be growing prevalence rates of adverse maternal mental health disorders world-wide with reported antenatal depression and/or anxiety ranging from 8–30%.(32, 33) In the maternal cohort of this study, rates were significantly higher for those with public versus private insurance (57% and 34%) respectively. Multiple adverse outcomes have been linked to maternal depression in the perinatal period including social isolation, marital discord, restricted fetal growth, increased stress reactivity in infants, and delays in infant motor or intellectual development.(32) Maternal mental health can impact the relationship between mother and infant, affecting interactions and provision of stimuli needed for language development.(24, 31) Although multiple additional social risk factors were associated in bivariate analyses with public insurance, they did not achieve independent effects on BSID-III language scores in adjusted analyses. This may be attributed to our specific patient population, inadequate sample size, and collinearity since many of our mothers had multiple co-morbidities.
Premature infants are also at higher risk of language impairments associated with prolonged hospitalization in a NICU in conjunction with increased rates of neonatal medical morbidities. Studies have also shown an association between low birth weight, low gestational age, and medical conditions associated with prematurity and linguistic development.(7, 8, 23, 34) Although our findings did not show a relationship between public insurance and birth weight or length of stay, it did confirm a significant association between public insurance and necrotizing enterocolitis. Immune mediated responses caused by infection, such as necrotizing enterocolitis, have been thought to impact the central nervous system, and a history of sepsis has also been associated with reduced BSID-III language scores.(23) Multivariable logistic regression analysis did show a protective effect of exclusive or partial breast milk feeding at discharge. Breast milk is well known to have protective effects, which include decreased risk of necrotizing enterocolitis, and multiple studies have reported the positive relationship between breast milk and development.(35–38) Belfort et al.(38) found that higher maternal milk intake during the NICU hospitalization and after discharge was associated with better academic achievement and IQ, with stronger beneficial associations of maternal milk feeding with neurodevelopment for neonates with lower gestational ages.
Secondary outcomes evaluated in our study included assessing the relationship between BSID-III cognitive and motor scores and socioeconomic factors. Infants with public insurance were more likely to have lower mean cognitive and motor composite scores compared to infants with private insurance, consistent with other studies of very preterm infants. (23) Gross and fine motor subscores were also lower for the public Insurance group, although the fine and gross motor subscore cutpoints of < 7 indicating a more significant delay did not differ between the groups. Predictive value of Bayley motor scores for later development remains uncertain.(23, 39, 40)
Strengths of this study include data collection of social determinants of health in the NICU, and neurodevelopmental assessment at 2 years of age. Limitations include that this is a single tertiary care center retrospective analysis, which may limit the generalizability of the study’s findings and there is no standardized Spanish version of the BSID-III. Furthermore, the maternal data regarding some of the psychosocial variables could be affected by maternal reporting. A small subset of children did not complete all components of Bayley testing either due to lack of cooperation or fatigue.