Context: Affective symptoms are common, but many individuals are unable to access formal mental health treatment. Digital mental health interventions (DMHIs) are a promising solution to this treatment gap, yet real-world engagement is low, which remains a barrier to impact of DMHIs. In order to more effectively engage diverse users, it is critical to understand how individual differences and demographics are related to engagement. The goal of this study was to understand individual predictors of engagement in a publicly available text messaging intervention, Small Steps SMS.

Methods: Users who took an online self-screening through Mental Health America were invited to use Small Steps SMS, an 8-week automated text messaging intervention grounded in evidenced-based psychological strategies. At sign-up, users completed the Kessler Psychological Distress Scale (K10) and a demographics survey. A series of GLS regression models were estimated separately for predictors including K10, age, gender, sexual orientation, race, and ethnicity, and the engagement outcome was operationalized as response day rate (i.e., sending at least one reply on a day when a response was solicited from the program).

Results: 5,052 users registered for Small Steps. The mean response day rate was 27% (SD=30%) across users. Greater age predicted higher response day rate (b=.005,p<.0001), while greater K10 was negatively associated with response day rate (b=-.005,p<.0001). Men had lower response day rates compared to women (b=-.03,p=.01) and non-binary users (b=-.06,p=.02). Heterosexual users had greater response day rates than LBGTQ-identifying users (b=.02,p=.018). Multiracial users had lower response day rates compared to White (b=-0.03,p=.03) and Asian (b=-.05,p=.05) users. Non-Hispanic users had greater response day rates than Hispanic users (b=.05,p=.0001).

Conclusions: Demographic factors and psychological distress significantly predicted engagement with Small Steps SMS. Findings highlight the need to better understand real-world engagement in younger users and those with more symptoms. Findings also suggest that there may be a need to design content and delivery approaches to meet the unique needs of specific populations that are underrepresented in clinical trials in psychology/psychiatry (i.e., men, and LGBTQ+, Multiracial, and Hispanic users). Results have implications for future iterations of Small Steps and for designing scalable DMHIs that achieve real-world impact.