Predicting clinical improvement in youth using a national-scale multicomponent digital mental health intervention

Abstract

Introduction

Youth mental health services are characterised by high demand and modest clinical outcomes. While digital mental health interventions (DMHIs) have been shown to be clinically effective, the relationship between DMHI use and outcome is unclear. The current study sought to identify the factors affecting the relationship between DMHI use and depression and anxiety symptom improvement in sub-groups of young people.

Method

An observational cohort design included young people aged 12–25 years engaging with a DMHI (MOST) from October 2020 to October 2023. The primary outcome was improvement at 12 weeks on the Patient Health Questionnaire-4 (PHQ4). DMHIs were combinations of self-paced digital cognitive-behavioural therapy content, social network interactions, and professional support. A machine learning clustering algorithm was used to identify distinct user clusters based on baseline characteristics and multiple logistic regression models examined the relationship between DMHI usage and improvement.

Results

Two distinct user clusters emerged, differing by symptom severity, age, service setting, and concurrent external treatment. 46.7% of “Severe” users and 39.8% of “Mild-Moderate” users significantly improved. Greater use of therapy content and professional support interactions were associated with improvement for the Mild-Moderate group only (OR = 1.16, 95% CI: 1.04–1.30, p = 0.008).

Conclusion

While a greater proportion of users in the Severe group significantly improved, increased MOST use was associated with symptom improvement only for the Mild-Moderate group. These findings highlight the complexity of the relationship between DMHI use and outcome. Other unmeasured mediating or moderating factors such concurrent ‘offline’ treatment may help explain the results. Further research is required to better understand the relationship between DMHI use and clinical outcomes.