Article Text
Abstract
Background Various patterns of early childhood family adversity characterised by poverty and/or other stressors are major risk factors for middle childhood mental health. Efforts to reduce mental health inequalities would benefit from understanding whether childcare moderates effects of different adversity patterns. This study examined hypotheses regarding centre-based (eg nursery) and non-centre (eg grandparent, childminder) childcare in: (1) buffering effects of adversity mainly characterised by poverty; (2) potentiating effects of adversity characterised by multiple stressors.
Methods We used the Growing Up in Scotland first birth cohort (born in 2004/5, n=5217), selecting families with a singleton birth, where mothers provided information on adversity and childcare at child age 10, 22 and 34 months, and where there was parent-reported information on children’s externalizing and internalizing problems at 46, 58, 70, 94, 122 and/or 152 months (n=3419). Using Mplus, latent class analysis of 19 indicators identified four adversity subtypes: Low (66%), Health-related (20%), Poverty/discord (9%), Multiple (5%). Growth mixture modelling identified five childcare patterns: Low (28%), Moderate Non-Centre (30%), High Non-Centre (16%), High Centre (12%), High Combined (13%). Latent growth models of problem trajectories (approximate ages 4 to 12 years) on adversity, childcare and interaction terms controlled for child gender, low birth weight, mothers’ age, ethnicity, education, smoking in pregnancy, family type and number of children at baseline.
Results Compared to the Low adversity subtype, children from higher-risk subtypes had higher 8.25-year intercepts of externalizing problems (coefficients with 95% confidence intervals: Health-related 1.46, 1.15–1.78; Poverty/discord 1.34, 0.82–1.87; Multiple 2.56, 1.86–3.26) and internalizing problems (Health-related 1.20, 0.93–1.47; Poverty/discord 1.16, 0.70–1.61; Multiple 2.60, 2.02–3.17); and steeper linear growth in internalizing problems (Health-related 0.08, 0.04–0.12; Poverty/discord 0.08, 0.01–0.16; Multiple 0.24, 0.15–0.34). Adding childcare and interactions with adversity yielded negative interaction terms for High Centre x Poverty/discord on the externalizing intercept (-1.83, -3.18 to -0.47); and for High Combined x Poverty/discord on the externalizing intercept (-1.70, -3.23 to -0.15), internalizing intercept (-1.52, -2.44 to -0.60), and internalizing linear slope (-0.23, -0.44 to -0.02). Corresponding interaction terms for Health-related adversity were generally small and negative, but for Multiple adversity were larger and all positive: for both these adversity subtypes, confidence intervals spanned zero. Sensitivity analyses incorporating teacher- and child-reported 122/152-month outcomes confirmed main findings.
Conclusion Centre-based care, especially when combined with individual non-parental care, buffered children’s mental health from family adversity characterised by poverty and interparental discord. Limitations include low statistical power for small subgroups and lack of childcare quality information. Findings indicate childcare may be protective against some, but not all, patterns of family adversity.