Sociodemographic Predictors of Clinical Effectiveness, Therapeutic Program Engagement, and Device Usability for an In-Home Virtual Reality Program for Chronic Low Back Pain: Secondary Analysis of a Ra

Depression and anxiety are the most common mental health conditions. They are among the leading contributors to the global burden of disease, affecting several hundred million people worldwide. Reducing this burden will require advances in understanding their underlying causes, developing more effective treatments, and more accurately predicting individual patient outcomes. Current classification frameworks for depression and anxiety largely ignore their heterogeneity, grouping individuals experiencing symptoms into discrete diagnostic categories. For example, a diagnosis of Major Depressive Disorder (MDD) represents more than 200 different possible combinations of symptoms. Assessment of depression and anxiety is most commonly based on elicitation of snapshots of symptoms through clinical interviews or self-report scales. These evaluations are subjective, susceptible to inaccurate recall and other biases, and inadequately capture core elements of these conditions (e.g., those relating to physiology or behavior) or their dynamic nature. The ubiquity of smartphones and wearable devices provides an opportunity to transform the classification and assessment of depression and anxiety, through the continuous collection of multimodal data from their sensors to measure features of physiology, behavior, and emotional states that are associated with these conditions. Such digital phenotyping offers the potential to objectively characterize consenting individuals, on an immense scale. However, despite this potential, the utility of digital phenotyping as a tool for mental health research and clinical care has not yet been convincingly demonstrated.