Predicting therapy outcome in a digital mental health intervention for depression and anxiety: A machine learning approach.

OBJECTIVE: Predicting the outcomes of individual participants for treatment interventions appears central to making mental healthcare more tailored and effective. However, little work has been done to investigate the performance of machine learning-based predictions within digital mental health interventions. Therefore, this study evaluates the performance of machine learning in predicting treatment response in a digital mental health intervention designed for treating depression and anxiety. METHODS: Several algorithms were trained based on the data of 970 participants to predict a significant reduction in depression and anxiety symptoms using clinical and sociodemographic variables. As a random forest classifier performed best over cross-validation, it was used to predict the outcomes of 279 new participants. RESULTS: The random forest achieved an accuracy of 0.71 for the test set (base rate: 0.67, area under curve (AUC): 0.60, p = 0.001, balanced accuracy: 0.60). Additionally, predicted non-responders showed less average reduction of their Patient Health Questionnaire-9 (PHQ-9) (-2.7, p = 0.004) and General Anxiety Disorder Screener-7 values (-3.7, p < 0.001) compared to responders. Besides pre-treatment Patient Health Questionnaire-9 and General Anxiety Disorder Screener-7 values, the self-reported motivation, type of referral into the programme (self vs. healthcare provider) as well as Work Productivity and Activity Impairment Questionnaire items contributed most to the predictions. CONCLUSIONS: This study provides evidence that social-demographic and clinical variables can be used for machine learning to predict therapy outcomes within the context of a therapist-supported digital mental health intervention. Despite the overall moderate performance, this appears promising as these predictions can potentially improve the outcomes of non-responders by monitoring their progress or by offering alternative or additional treatment.

Significant reduction in depressive symptoms among patients with moderately-severe to severe depressive symptoms after participation in a therapist-supported, evidence-based mobile health program delivered via a smartphone app.

Depression is a debilitating disorder associated with poor health outcomes, including increased comorbidity and early mortality. Despite the advent of new digital health interventions, few have been tested among patients with more severe forms of depression. As such, in an intent-to-treat study we examined whether 218 patients with at least moderately severe depressive symptoms (PHQ-9 ≥ 15) experienced significant reductions in depressive symptoms after participation in a therapist-supported, evidence-based mobile health (mHealth) program, Meru Health Program (MHP). Patients with moderately severe and severe depressive symptoms at pre-program assessment experienced significant decreases in depressive symptoms at end-of treatment (mean [standard deviation] PHQ-9 reduction = 8.30 [5.03], Hedges' g = 1.64, 95% CI [1.44, 1.85]). Also, 34% of patients with at least moderately severe depressive symptoms at baseline and 29.9% of patients with severe depressive symptoms (PHQ-9 ≥ 20) at baseline responded to the intervention at end-of-treatment, defined as experiencing ≥50% reduction in PHQ-9 score and a post-program PHQ-9 score lower than 10. Limitations include use lack of a control group and no clinical diagnostic information. Future randomized trials are warranted to test the MHP as a scalable solution for patients with more severe depressive symptoms.

Feasibility and Efficacy of the Addition of Heart Rate Variability Biofeedback to a Remote Digital Health Intervention for Depression.

A rise in the prevalence of depression underscores the need for accessible and effective interventions. The objectives of this study were to determine if the addition of a treatment component showing promise in treating depression, heart rate variability-biofeedback (HRV-B), to our original smartphone-based, 8-week digital intervention was feasible and whether patients in the HRV-B ("enhanced") intervention were more likely to experience clinically significant improvements in depressive symptoms than patients in our original ("standard") intervention. We used a quasi-experimental, non-equivalent (matched) groups design to compare changes in symptoms of depression in the enhanced group (n = 48) to historical outcome data from the standard group (n = 48). Patients in the enhanced group completed a total average of 3.86 h of HRV-B practice across 25.8 sessions, and were more likely to report a clinically significant improvement in depressive symptom score post-intervention than participants in the standard group, even after adjusting for differences in demographics and engagement between groups (adjusted OR 3.44, 95% CI [1.28-9.26], P = .015). Our findings suggest that adding HRV-B to an app-based, smartphone-delivered, remote intervention for depression is feasible and may enhance treatment outcomes.

Long-Term Outcomes of a Therapist-Supported, Smartphone-Based Intervention for Elevated Symptoms of Depression and Anxiety: Quasiexperimental, Pre-Postintervention Study.

BACKGROUND: Depression is one of the most common mental health disorders and severely impacts one's physical, psychological, and social functioning. To address access barriers to care, we developed Ascend-a smartphone-delivered, therapist-supported, 8-week intervention based on several evidence-based psychological treatments for depression and anxiety. A previous feasibility study with 102 adults with elevated depression reported that Ascend is associated with a postintervention reduction in depression symptoms. OBJECTIVE: We aimed to examine whether Ascend is associated with a reduction in symptoms of anxiety, and importantly, whether reductions in symptoms of depression and anxiety are maintained up to 12-months postintervention. METHODS: We assessed whether the previously reported, end-of-treatment improvements seen in the 102 adults with elevated symptoms of depression extended up to 12 months posttreatment for depression symptoms (measured by the Patient Health Questionnaire-9 [PHQ-9]) and up to 6 months posttreatment for anxiety symptoms (added to the intervention later and measured using the Generalized Anxiety Disorder-7 [GAD-7] scale). We used linear mixed effects models with Tukey contrasts to compare time points and reported intention-to-treat statistics with a sensitivity analysis. RESULTS: The intervention was associated with reductions in symptoms of depression that were maintained 12 months after the program (6.67-point reduction in PHQ-9 score, 95% CI 5.59-7.75; P<.001; Hedges g=1.14, 95% CI 0.78-1.49). A total of 60% of the participants with PHQ-9 scores above the cutoff for major depression at baseline (PHQ≥10) reported clinically significant improvement at the 12-month follow-up (at least 50% reduction in PHQ-9 score and postprogram score <10). Participants also reported reductions in symptoms of anxiety that were maintained for at least 6 months after the program (4.26-point reduction in GAD-7 score, 95% CI 3.14-5.38; P<.001; Hedges g=0.91, 95% CI 0.54-1.28). CONCLUSIONS: There is limited evidence on whether outcomes associated with smartphone-based interventions for common mental health problems are maintained posttreatment. Participants who enrolled in Ascend experienced clinically significant reductions in symptoms of depression and anxiety that were maintained for up to 1 year and 6 months after the intervention, respectively. Future randomized trials are warranted to test Ascend as a scalable solution to the treatment of depression and anxiety.