Prevalence estimates of childhood and adolescent mental health disorders appear to vary between 20 to 30% worldwide. It is therefore unsurprising that studies have yielded inconsistent findings in regards to the trends of prevalence of mental health disorders. Some reasons for the discrepancy in findings include use of survey data and its associated attrition and selection bias.

First, to determine and compare the prevalence of mental health disorders derived from a survey and a population cohort. Second, to evaluate trends of mental health prevalence over time.

As population data (i.e., linked health records) may be used to overcome the issues presented by survey data, we compared the prevalence estimated from a prospective survey cohort (the Western Australian Pregnancy Cohort (Raine study) to another estimate from a prospective population cohort (linked population data; data from the Hospital Morbidity Records and Mental Health Registration).

As expected, the Raine cohort yielded a larger estimate of prevalence when compared to the linked population data. However each cohort also revealed opposite trends of prevalence, where the Raine cohort showed the prevalence of mental health disorders to decrease as children age.

We therefore recommend that estimates of prevalence be interpreted with the type of cohort in mind, as estimates from survey cohorts will provide different information to that from population cohorts.