Feasibility of recruiting adolescents into a prospective cohort study of the effects of social isolation during COVID-19.

Background: Social connection and loneliness in adolescence are increasingly understood as critical influences on adult mental and physical health. The unique impact of the social isolation imposed by the COVID-19 lockdown on emerging adults is therefore expected to be especially profound. We sought to investigate the feasibility of using ecological momentary assessment (EMA) and wearable accelerometers to characterize the effects of social isolation and/or loneliness experienced by adolescents during the COVID-19 pandemic.

Methods: We recruited 19 participants aged 13-18 from an Adolescent Medicine practice in Atlanta, GA. Participants completed surveys at baseline and throughout a 2-week study period using EMA regarding their degree of social isolation, loneliness, family functioning, school climate, social media use, and COVID-19 experiences surrounding their physical, mental, and social domains. Six participants agreed to wear an activity tracker and heart rate measurement device for 14 days to monitor their emotional state and physical health. Participant feedback was collected via open-ended exit interviews. Feasibility of recruitment/retention, adherence, and outcome measures were investigated. Implementation was also assessed by evaluating the barriers and facilitators to study delivery. Associations between the social isolation and loneliness variables and all other variables were performed with univariate linear regression analysis with significance set at p < 0.05. The progression criteria were a recruitment rate of > 30% and a retention rate of > 80%.

Results: Progression criteria were met for recruitment (76%) of participants, but not retention (38%). Adherence to EMA survey completion was highly variable with only 54% completing ≥ 1 survey a day, and accelerometry use was not feasible. Social isolation was significantly correlated with lower school climate, higher COVID-19 experiences, higher depression scores, and lower sleep quality. Loneliness also showed a significant correlation with all these factors except COVID-19 experiences.

Conclusions: EMA and wearable accelerometer use was not feasible in this longitudinal study of adolescents during the COVID-19 pandemic. Future research should further investigate barriers to conducting long-term research with adolescents and the potential effects of the pandemic on subject recruitment and retention.