Street Experiences, Affect, and Coping: Harnessing Computational Models for the Development of a Cellphone-Based Intervention to Prevent Drug Use among Youth Experiencing Homelessness

摘要 经历无家可归(YH)的青少年面临许多重大风险，这些风险对他们的长期健康不利- 生存和生存，包括从事消极的应对行为，如酗酒和吸毒。：叶诗文 报告终生饮酒，77%使用大麻，48%报告使用甲基苯丙胺。吸毒和酗酒 YEH的患病率比他们的同龄人高2到3倍，导致精神健康状况较差 以及艾滋病毒对这一人群的影响。叶氏吸食毒品的近期危险因素包括街头受害： 39%和94%的人分别遭受过性侵害或身体伤害；精神健康状况不佳的人：80%和74%的人 分别为创伤后应激障碍和抑郁症的诊断标准；以及危险的社交网络：62%的Yeh报告吸毒 与网络成员一起使用。实质性的另一个风险因素是影响监管。普通人口研究 年轻人发现，那些有更大负面影响的人有更高的物质使用率。情感调节的作用 很少被检测为YEH；因此，调解机制尚不清楚。有可能最近的风险是 因素增加了负面影响，导致叶中使用更多的物质。理解这些联系是 发展应对预测模型的关键。该项目使用社会生态框架来检查 近邻风险因素、影响调节和应对行为之间的联系。具体地说，我们将确定 影响积极应对行为与消极应对行为的因素叶氏制定数据驱动的干预措施：1) 与这一弱势群体相关，并提供个性化的支持，以及2)在外部可以持续 研究背景。我们的团队率先使用手机进行生态瞬时评估(EMA) 在公共卫生研究(可汗)和叶(泰勒和奥尔森)。叶诗文的手机拥有量增加 随着城市WiFi热点覆盖率的增长，共同创造了一个公共卫生机会 项目目标：1)确定影响调节是否在YEH的近期风险之间起中介作用 通过分析调查数据和最近风险因素的EMA数据得出的因素和应对行为影响 通过观察性队列中基于APP的响应性EMA(REMA)收集的规则和应对行为 60天内N=300 YEH的研究；2)开发数据驱动的基于APP的即时个人支持干预 (JIT-PSI)提供响应性消息，以缓解消极应对，有利于积极应对。这个目标是 通过建立基于机器学习的应对行为预测模型实现--《目标1》S的分析 调查和REMA数据；组织N=30个YEH的焦点小组和N=15个机构工作人员的会议，以 设计针对一系列应对行为的干预信息内容；3)测试可行性和 JIT-PSI的可接受性，通过60天内N=300YEH的观察性队列研究实现，使用 调查，以及基于手机的REMA和JIT干预交付。利用越来越多的公众 免费WiFi热点和不断缩小的手机硬件成本，这一项目将产生重大的公共健康影响 通过开发可持续的JIT行为干预措施，预防叶诗文中的药物使用。

ABSTRACT Youth experiencing homelessness (YEH) face many significant risks that are detrimental to their long-term well- being and survival, including engaging in negative coping behaviors such as alcohol and drug use. 89% of YEH report lifetime alcohol use, 77% marijuana use, and 48% report methamphetamine use. Drug and alcohol use prevalence rates are 2 to 3 times higher among YEH than in their housed peers, leading to poorer mental health and HIV outcomes for this population. Proximal risk factors for substance use in YEH include street victimization: 39% and 94% have been sexually or physically victimized, respectively; poor mental health: 80% and 74% meet diagnostic criteria for PTSD and depression, respectively; and risky social networks: 62% of YEH report drug use with a network member. Another risk factor for substance is affect regulation. Studies of general population youth find that those with greater negative affect have higher rates of substance use. The role of affect regulation has rarely been examined for YEH; thus, mediating mechanisms are unclear. It is possible that proximal risk factors increase negative affect, leading to greater substance use among YEH. Understanding these linkages is key to developing predictive models of coping. This project uses a socio-ecological framework to examine linkages between Proximal risk factors, Affect regulation, and Coping behaviors. Specifically, we will identify factors that influence positive vs. negative coping behavior for YEH to develop a data-driven intervention that: 1) is relevant to this vulnerable population, and provides individualized support, and 2) can be sustainable outside the research setting. Our team has pioneered the use of cellphones for ecological momentary assessment (EMA) in public health research (Khan), and with YEH (Tyler and Olson). Increased cellphone ownership among YEH alongside growth in urban WiFi hotspot coverage together create a public health opportunity leveraged in this project's aims: 1) determine whether affect regulation mediates the relationship between YEHs’ proximal risk factors and coping behaviors, achieved by analyzing survey data and EMA data on Proximal risk factors, Affect regulation, and Coping behaviors collected via app-based responsive EMA (rEMA) in an observational cohort study of N=300 YEH over 60 days; 2) develop a data-driven app-based just-in-time personal support intervention (JIT-PSI) that delivers responsive messaging to mitigate negative coping in favor of positive coping. This aim is achieved by developing predictive models of coping behavior based on machine learning analyses of Aim 1’s survey and rEMA data; and conducting focus groups with N=30 YEH and meetings with N=15 agency staff to design intervention messaging content responsive to a range of coping behaviors; 3) test the feasibility and acceptability of the JIT-PSI, achieved by an observational cohort study of N=300 YEH over 60 days, using surveys, alongside cellphone based rEMA, and JIT intervention delivery. Leveraging growing number of public free WiFi hotspots and shrinking cellphone hardware costs, this project will yield significant public health impacts by developing a sustainable, JIT behavioral intervention for the prevention of drug use among YEH.