Researching adolescent mental health during a pandemic: A registered report of a natural experiment within two pragmatic school-based trials

  • Boehnke, J. (Speaker)
  • Rosie Mansfield (Contributor)
  • Joao Santos (Contributor)
  • Jessica Deighton (Contributor)
  • Daniel Hayes (Contributor)
  • Tjasa Velikonja (Contributor)
  • Praveetha Patalay (Contributor)

Activity: Talk or presentation typesOral presentation


Aims: Research on the impact of the pandemic on adolescent mental health is limited by several methodological factors, such as the lack of pre-pandemic scores for comparison, control for developmental change, and wider geographic and social representation. Empirical evidence which can potentially causally attribute changes to the pandemic remains therefore limited. The presentation will focus on how a natural experiment within two ongoing school-based cluster randomised pragmatic trials was used to overcome such challenges to investigate whether and to which degree differences in adolescents' mental health could be attributable to the pandemic.

Methods: The multi-partner structure of the trials separated relevant aspects of the data collection from the research team, therefore the analyses could be prospectively registered. The necessity of two recruitment cohorts led to one cohort of schools being assessed in Sept-Oct 2018 (baseline) and Jan-Mar 2020 (1-year follow-up); and the other cohort in Sept-Oct 2019 and Feb-Apr 2021, respectively. Participants in the first cohort (90 schools, N=6,419) acted as unexposed controls, and participants in the second cohort (88 schools, N=5,031) were exposed to the COVID-19 pandemic between assessments. The outcomes were depressive symptoms (Short Mood and Feelings Questionnaire, primary outcome), externalising difficulties (Me and My Feelings questionnaire), and life satisfaction (Huebner Life Satisfaction Scale). Analyses were conducted with random intercept regression models with exposure to the COVID-19 pandemic as focal predictor while controlling for baseline scores and individual- and school-level covariates.

Results: The primary outcome analysis showed higher levels of depressive symptoms (adjusted d=.11). For the secondary outcomes, life satisfaction scores were lower (adjusted d=.12), and no differences were detected for externalising difficulties (adjusted d=.01). The analyses further indicated potential differential effects based on socio-economic background and gender. The results were robust in sensitivity analyses exploring the impact of missing data, drop-out, and comparability of participants in both cohorts.

Conclusion: A multi-site and interdisciplinary approach to trial planning laid the foundation for a responsive research structure that in addition to work on high-quality evidence for school-based universal interventions provided evidence that the COVID-19 pandemic increased adolescent depressive symptoms and potentially decreased life satisfaction.
Period21 Oct 2022
Event titleInternational Society for Quality of Life Research (ISOQOL) 29th Annual Conference: Redefining boundaries – breaking new ground in patient-centered outcomes research
Event typeConference
LocationPrague, Czech RepublicShow on map