Maternal expressions of warmth and negativity and adolescent mental health: using longitudinal monozygotic twin-difference analyses to approach causal inference.

Background: Emotions that mothers express about their offspring are associated with offspring mental health during childhood, but little research has explored whether this extends into adolescence. We investigated associations between maternal warmth and negativity towards twin offspring at age 10, and subsequent mental health outcomes in early and late adolescence.

Methods: The Environmental Risk (E-Risk) Longitudinal Twin Study is a population-based cohort of 2,232 same-sex twins born in 1994-1995 across England and Wales. Maternal warmth and negativity were assessed from Five-Minute Speech Samples obtained when twins were aged 10. Continuous mental health outcomes were assessed in interviews with twins at ages 12 (depression, anxiety, attention-deficit hyperactivity disorder and conduct disorder) and 18 (general psychopathology, internalising, externalising and thought disorder). Linear regressions were conducted on 1906 participants with available data and adjusted for sex, family socioeconomic status, and age-5 emotional and behavioural problems. We then conducted a monozygotic twin-difference analysis to control for unmeasured shared environmental and genetic factors.

Results: Lower maternal warmth and higher maternal negativity were associated with worse mental health outcomes at ages 12 and 18. For example, when comparing differences in mothers' expressed emotions and mental health outcomes within monozygotic twin pairs, higher negativity remained associated with externalising symptoms (b = 1.77, 95% CI = 0.68 to 2.86, β = .14) and poorer general psychopathology (b = 1.82, 95% CI = 0.63 to 3.01, β = .13), and lower warmth with externalising symptoms (b = -1.96, 95% CI = -3.54 to -0.37, β = -.11). These associations remained after adjusting for twin-differences in age-5 emotional and behavioural problems and birth weight. Null findings were more frequently observed for maternal warmth and internalising outcomes.

Conclusions: Using a genetically sensitive design to approach causal inference, we found evidence for associations between maternal warmth/negativity and adolescent mental health outcomes. Maternal expressed emotion ratings might provide an early opportunity to identify families who would benefit from interventions and mental health disorder prevention programmes.