Wisniewski Hannah, Henson Philip, Torous John
Divison of Digital Psychiatry, Department of Psychiatry, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States.
Front Psychiatry. 2019 Sep 23;10:652. doi: 10.3389/fpsyt.2019.00652. eCollection 2019.
The use of smartphone apps for research and clinical care in mental health has become increasingly popular, especially within youth mental health. In particular, digital phenotyping, the monitoring of data streams from a smartphone to identify proxies for functional outcomes like steps, sleep, and sociability, is of interest due to the ability to monitor these multiple relevant indications of clinically symptomatic behavior. However, scientific progress in this field has been slow due to high heterogeneity among smartphone apps and lack of reproducibility. In this paper, we discuss how our division utilized a smartphone app to retrospectively identify clinically relevant behaviors in individuals with psychosis by measuring survey scores (symptom report), games (cognition scores), and step count (exercise levels). Further, we present specific cases of individuals and how the relevance of these data streams varied between them. We found that there was high variability between participants and that each individual's relevant behavior patterns relied heavily on unique data streams. This suggests that digital phenotyping has high potential to augment clinical care, as it could provide an efficient and individualized mechanism of identifying relevant clinical implications even if population-level models are not yet possible.
在精神卫生领域，使用智能手机应用程序进行研究和临床护理越来越普遍，尤其是在青少年心理健康方面。特别是数字表型分析，即通过监测智能手机的数据流来识别诸如步数、睡眠和社交能力等功能结果的替代指标，由于能够监测这些临床上有症状行为的多个相关指标而备受关注。然而，由于智能手机应用程序之间存在高度异质性且缺乏可重复性，该领域的科学进展一直缓慢。在本文中，我们讨论了我们部门如何利用智能手机应用程序，通过测量调查问卷分数（症状报告）、游戏（认知分数）和步数（运动水平），回顾性地识别精神病患者的临床相关行为。此外，我们展示了个体的具体案例，以及这些数据流在他们之间的相关性如何变化。我们发现参与者之间存在很大差异，而且每个人的相关行为模式严重依赖于独特的数据流。这表明数字表型分析有很大潜力增强临床护理，因为即使尚未建立人群水平的模型，它也可以提供一种高效且个性化的机制来识别相关的临床意义。
