Preschool care early in life and mental health in adolescence in Sweden: a cohort study

OBJECTIVE: Despite the high rates of early enrolment in preschool and of poor mental health in adolescence in Sweden, knowledge regarding their association in Sweden is lacking. We investigated whether age at starting preschool and weekly hours spent at preschool in different ages are associated with mental ill-health in Swedish adolescents. DESIGN: A cohort study based on data from KUPOL (Swedish acronym for 'Knowledge about Adolescents Mental Health and Learning'). SETTING AND PARTICIPANTS: We used data from KUPOL, a longitudinal study conducted during 2013-2018, involving Swedish adolescents born 2000-2001. We included in the current analyses adolescents with available information on the exposures and the outcomes of interest (n=2261). OUTCOME MEASURES: Study participants and their parents completed questionnaires concerning the child's age (in months) at start of preschool, the average weekly hours in preschool in different ages, the adolescent's mental health, lifestyle, school-related, psychosocial and parental sociodemographic factors. We analysed the association between preschool-related factors and mental health using logistic regression. RESULTS: Children enrolled in preschool at the age of 12-15 months had increased odds of high overall and externalising problems score on the Strengths and Difficulties Questionnaire relative to those enrolled at 20 months or later. The corresponding ORs (95% CIs) were 1.39 (95% CI 1.02 to 1.90) and 1.52 (95% CI 1.08 to 2.16), while the corresponding population attributable fractions were 8% and 9%, respectively. There were no associations between age at start of preschool and internalising problems, nor between the average weekly hours at preschool and mental health. CONCLUSION: We found weak and inconsistent evidence for a link between early preschool attendance and mental health in adolescence; population attributable fractions suggest limited public health implications for the studied associations. The results should be interpreted in light of the methodological constraints of observational studies, the multitude of our comparisons and the sample selection.