Researchers have developed a predictive tool that can recognise adolescents who are at high or low risk of depression in young adulthood.
Published in the Journal of the American Academy of Child and Adolescent Psychiatry, the study also evaluated the performance of the tool in samples of adolescents from New Zealand and the UK, demonstrating differences in its ability to predict depression across these countries and highlighting the need to consider local variations when developing predictive tools.
"In our study we tried to go beyond more traditional ways of identifying youths at high risk of depression and learn from other medical specialties that combine multiple variables to generate composite risk scores, such as the Framingham cardiovascular risk score," said study lead author Christian Kieling from Universidade Federal do Rio Grande do Sul in Brazil.
"This is relatively new in the field of mental health," Kieling added.
Predictive tools have already been developed for psychosis and many physical health conditions, but little has been done for depression, despite its prevalence.
Current methods for assessing the risk of depression are based on family history and subthreshold symptoms which do not reach the criteria for depression.
Based on existing research and clinical expertise the researchers identified 11 sociodemographic variables that could be combined into a single score to recognise those adolescents at risk of developing depression.
In addition to a strong association with depression the variables were chosen because they were also easy to collect and simple to evaluate.
Using data on these 11 variables from 2,192 Brazilian 15-year-old adolescents and mental health assessments of the same adolescents at 18 years old, the researchers developed the tool to assess the risk of development of a major depressive disorder.
The study compared the ability of this tool to predict depression in a sample of 1,144 British 12 year olds (no data was available at age of 15) from the E-Risk Study and 739 New Zealand 15 year olds from the Dunedin study.
The predictive ability of the score was not as strong in the UK and New Zealand samples.
According to researchers, this was to some extent expected as not all the information used from the Brazilian sample was available in the other datasets and different diagnostic measures were used to assess mental health at 18 years old.
'The existence of these discrepancies does not discount the value of our tool but provides important insight into adapting the score according to where it will be used," Kieling said.
"Adaptation is necessary for most predictive tools; for instance, tools used to assess the risk of cardiovascular disease developed in the US require adjustments when used in other countries," Kieling added.