Eigen RadarScience
Archive
Analysis

Environmental measures improve estimates of adolescent depression risk

Environmental measures outperformed genetic scores alone in estimating later depression in a United States child-development cohort. Combining both improved discrimination further in the peer-reviewed observational study. Family depression, exposure to violence, sleep and dieting featured in the environmental score, which still needs testing in independent populations before its findings can be generalised.

Science··Morning
In a plain bedroom lit by morning daylight, a child sleeps facing away while a parent sits beside the bed with their back to the camera.

Two years of childhood follow-up

A peer-reviewed observational analysis followed depression-related outcomes in 9,797 children from the Adolescent Brain Cognitive Development study, a United States child-development cohort. They were 9–11 years old at the starting assessment and 11–13 at follow-up. Environmental measures gave stronger discrimination of later depression than a genetic score used alone.[1], [2]

Researchers drew on 31 earlier reviews and meta-analyses to identify 38 possible environmental risk factors. Of these, 23 could be matched to the cohort data. The final statistical model retained 12 variables, including parental depression, witnessing violence, sleep and dieting. Cross-validation helped select the variables during development.[1], [2]

The dataset was split, with 80% used for training and 20% for internal validation. Longer sleep and parental supervision were associated with lower symptom scores. The dieting measure concerned dieting behaviour rather than nutritional quality. These associations describe information carried by the selected variables; the design did not test an intervention changing them.[1], [2]

Combining the scores improved discrimination

The environmental score explained 12.3% of variation in later depressive symptoms. For distinguishing depression from parent assessments, its discrimination measure was 0.69, compared with 0.64 for the genetic score. Combining both raised the value to 0.71. The corresponding measures for young people’s own reports were 0.67 for environmental information and 0.70 for the combination.[1], [2]

The discrimination measure, known as the area under the receiver-operating-characteristic curve, describes how well the model separates groups. Its values are not percentages of children correctly diagnosed. Parent and youth reports were assessed separately, and genetic comparisons did not include every child in the environmental analysis. The combined results therefore remain comparisons within the study’s validation data.[1], [2]

The cohort leaves gaps in generalisation

Higher-income and more highly educated families were overrepresented, and some genetic-ancestry groups had small samples. Independent populations are needed to assess transfer across places and cultures. The children had not yet reached the peak ages of depression onset, leaving later recurrence and persistence outside this two-year outcome. Community connectedness and academic pressure were among the influences not fully included in the model. The score was not tested as a treatment or as a standalone basis for care.[1]

References

  1. News sourceMolecular PsychiatryStudy examines adolescent depression risk using environmental and genetic measures↩1↩2↩3↩4↩5↩6
  2. News sourceTıbbiye BülteniEnvironmental factors lead adolescent depression risk assessment↩1↩2↩3↩4↩5
Space and explorationEnergy transition and climate

More analysis from this desk

View all Brief analyses from the Science desk →