OBJECTIVE
About one-fifth of US adolescents experienced major depressive symptoms, but few studies have examined longitudinal trends of adolescents developing depression or recovering by demographic factors. We estimated new transition probability inputs, and then used them in a simulation model to project the epidemiologic burden and trajectory of depression of diverse adolescents by sex and race or ethnicity combinations.
METHODS
Transition probabilities were first derived using parametric survival analysis of data from the National Longitudinal Study of Adolescent to Adult Health and then calibrated to cross-sectional data from the National Survey on Drug Use and Health. We developed a cohort state-transition model to simulate age-specific depression outcomes of US adolescents. A hypothetical adolescent cohort was modeled from 12-22 years with annual transitions. Model outcomes included proportions of youth experiencing depression, recovery, or depression-free cases and were reported for a US adolescent population by sex, race or ethnicity, and sex and race or ethnicity combinations.
RESULTS
At 22 years of age, approximately 16% of adolescents had depression, 12% were in recovery, and 72% had never developed depression. Depression prevalence peaked around 16-17 years-old. Adolescents of multiracial or other race or ethnicity, White, American Indian or Alaska Native, and Hispanic, Latino, or Spanish descent were more likely to experience depression than other racial or ethnic groups. Depression trajectories generated by the model matched well with historical observational studies by sex and race or ethnicity, except for individuals from American Indian or Alaska Native and multiracial or other race or ethnicity backgrounds.
CONCLUSIONS
This study validated new transition probabilities for future use in decision models evaluating adolescent depression policies or interventions. Different sets of transition parameters by demographic factors (sex and race or ethnicity combinations) were generated to support future health equity research, including distributional cost-effectiveness analysis. Further data disaggregated with respect to race, ethnicity, religion, income, geography, gender identity, sexual orientation, and disability would be helpful to project accurate estimates for historically minoritized communities.