Digital Phenotyping of Anxiety and Anxiety-Related Alcohol Comorbidity and Treatment

项目总结

Project Summary Anxiety and anxiety-related disorders (AARD) are notable risk factors for alcohol use problems (AUP) and their co-occurrence (AARD-AUP) results in tremendous health costs and societal burden. The self-medication model states that those with AARD use alcohol to numb or cope with arousing and distressing negative emotional states; however, longitudinal evidence for self-medication is mixed and psychological treatments integrate cognitive-behavioral therapy (CBT) skills for substance use disorders with CBT interventions for AARD have demonstrated modest success. The heterogeneity in these findings is likely influenced by dynamic and person-specific factors. We hypothesize that results from person-specific analyses can improve our understanding of this heterogeneity and can be harnessed to personalize interventions for AUP. Personalized interventions have the potential to increase symptom reduction and rate of change, as well as therapy retention and alliance. Modeling person-specific patterns of psychopathology has been challenging due to historical limitations with data collection and statistical analyses. However, increasing access to digital phenotyping – an innovative method for collecting single-subject intensive ecological momentary assessment (EMA) integrated with passive sensing smartphone data streams (e.g., GPS data, accelerometer, Bluetooth/call/SMS data) – offers a compelling solution for the development and evaluation of truly personalized psychological interventions. Such data can be used to generate statistical models of a single individual’s pattern of behavior over time. Previous data-driven (i.e., feedback-based) interventions have demonstrated success; however, it is unclear whether using person-specific models to adapt an active intervention reduces symptoms or improves therapeutic efficiency (i.e., rate of change), therapeutic retention, or therapeutic alliance. This K99R00 award will provide training in passive sensing data collection methods and conduct of clinical trials, as well as continued training in alcohol research. In the K99 phase, I will build on previous expertise constructing person-specific models and identifying person-specific factors testing self-medication models of AARD-AUP, developing a method for personalizing a CBT skills-based intervention. In the R00 phase, the feasibility of this intervention will be tested in a pilot three-group parallel randomized clinical trial (RCT), which will test the feasibility of a personalized CBT skills-based interventions compared to two control conditions testing the relative efficacy of personalization and personalization method, respectively. In the personalized condition, the therapist will review identified influential factors from the person-specific model, treat these factors using CBT skills, and track them using the mindLAMP digital phenotyping platforms – testing a data- driven method for personalizing and adapting the content of each session. Findings will inform future R01 applications testing personalized interventions for problematic alcohol use.