A wrist biosensor-based mHealth suite to support alcohol intervention in young people living with HIV

点击翻译按钮获取中文摘要

ABSTRACT 1 Despite the overall decline in new HIV infections in U.S., adolescents and emerging adults aged 13 to 29 2 remain disproportionately affected by HIV who account for 37% of all new HIV diagnoses. Alcohol use is more 3 common among young people living with HIV (YPLWH) than other age groups, creating a significant barrier for 4 mitigating transmission, achieving viral suppression, and reducing comorbidities. Despite the high prevalence 5 of alcohol use and its detrimental consequences in this population, developmentally-tailored intervention 6 targeting alcohol use in YPLWH is rare. Self-management emerged as an important strategy in the context of 7 medical adherence and chronic disease management, with self-monitoring and personalized feedback being 8 two key elements to improve health outcomes including alcohol use. Recent development of wrist alcohol 9 sensors provides a promising tool for self-monitoring and feedback based on objective data. However, two 10 critical challenges must be addressed to leverage biosensors in alcohol intervention for YPLWH. First, without 11 existing analysis software, accurate detection of alcohol use (e.g., start time, consumption level) based wrist 12 sensor data remains a significant barrier. Some literature also suggests that HIV may impact blood alcohol 13 concentration, so a detection algorithm specific to this population is needed. Second, engagement in mobile 14 health (mHealth), including wearables in youth, is suboptimal. To address these challenges and to achieve the 15 overall P01 goal to improve self-management of alcohol use and HIV care among YPLWH, we propose to 16 develop the first wrist biosensor-based mHealth suite for alcohol use monitoring with tailored detection 17 algorithem and optimized engagement strategies for YPLWH. We will conduct a 30-day micro-randomized trial 18 (MRT)--a clinical trial design for developing and optimizing mHealth interventions, to innovatively and efficiently 19 address the challenges. YPLWH will be asked to wear the Skyn biosensor for 30 days and report alcohol use 20 via smartphone-based ecological momentary assessment (EMA). They will also be micro-randomized twice 21 daily to test two empirically-based engagement strategies – reciprocity and personalized feedback. Data from 22 the 30-day MRT will be used to (a) develop and validate machine learning algorithm for alcohol detection 23 based on Skyn data; and (b) test which engagement strategy works better, for whom and under what 24 conditions in facilitating proximal engagement with the sensor. We will also collect alcohol biomarker (i.e., 25 PEth) at 1-month follow up, which will allow us to (c) explore the potential impact of sensor engagement (and 26 thus self-monitoring) on distal alcohol use. By trial end, we will have the first integrated, optimized, and 27 developmentally appropriate biosensor-based mHealth suite for alcohol self-monitoring that can be used as the 28 foundation for future alcohol interventions in YPLWH. Findings from this project may also inform the 29 development of new translational behavioral research model to guide adolescent and YPLWH research 30 agenda by leveraging powerful sensing technology to improve self-management in an engaging manner.