Social selection in cohort studies and later representation of childhood psychiatric diagnoses: The Danish National Birth Cohort

Aim: This study aimed to estimate the relative representation of childhood psychiatric diagnoses and use of psychotropic medication in the Danish National Birth Cohort (DNBC) compared to the general population. Methods: The general population was identified as all childbirths in Denmark during 1998-2002 (N=344,160). Linking the DNBC (N=91,442) and the general population to the Danish national health registries, all children were followed until they received an ICD-10 psychiatric diagnosis, had a prescription of psychotropic medication or to the end of follow-up in 2013. The prevalence ratios (PRs) with corresponding 95% confidence intervals (CI) were estimated for each psychiatric diagnosis and by sex. Age at first diagnosis presented as means were compared using the one-sample t-test. Results: In the DNBC, the selected childhood psychiatric diagnoses were underrepresented by 3% (PR=0.97, 95% CI 0.94-0.99), ranging from a 20% underrepresentation for schizophrenia (PR=0.80, 95% CI 0.59-1.09) to a 6% over-representation for anxiety disorder or obsessive-compulsive disorder (PR=1.06, 95% CI 0.97-1.17). The majority of the specific diagnoses were modestly underrepresented in the DNBC compared to the general population, while use of psychotropic medication had similar representation. Girls were generally more underrepresented than boys. Depression was on average diagnosed 0.4 years earlier in the DNBC than in the general population (p=0.023). Conclusions: These findings suggest that the social selection may influence the prevalence of diagnosed childhood psychiatric disorders in the DNBC.