Characterizing the Complexity of Advanced Cancer Pain in the Home Context by Leveraging Smart Health Technology

项目摘要 癌症疼痛是复杂的，普遍存在，并对患者、家庭照顾者和 医疗保健系统。管理不当的癌症疼痛可能会给患者应对带来特别的问题 患有晚期转移性疾病。大多数症状管理发生在家庭环境中，而家庭 照顾者通常在帮助管理癌症疼痛方面发挥关键作用，但往往发现这项任务令人望而生畏 压力很大。使癌症疼痛管理复杂化的是阿片类药物，这是一种主要的药物类别 用于控制严重的癌症疼痛，受到更严格的审查，因为公众对 国家“阿片类药物流行”现在，比以往任何时候都更迫切的是，我们必须了解患者和家属如何 护理人员尝试在家中管理癌症疼痛，以便我们可以为他们提供个性化的支持，以便有效地 并安全地减轻疼痛。移动和无线技术(智能健康)可以帮助缓解症状 在家庭环境中进行管理，但必须仔细设计，以考虑到患者和 应对晚期疾病的家庭照顾者。我们假设，个人和患者家属照顾者 DUADS，将显示先进的癌症疼痛体验的独特的“数字指纹”(或表型)--如果 更好的理解可以用来告知和提供个性化、及时的干预措施。这样做的目的是 这项研究建立在初步试点工作的基础上，目的是部署一种不引人注目的智能医疗系统，即 癌症的行为和环境感知和干预(BESI-C)，以监测和描述-以及 最终预测和帮助管理-家庭环境中晚期癌症疼痛的体验。BESI-C是 由可穿戴(智能手表)和环境传感器组成，收集生理、行为和 可集成的个人、对偶和家庭级别的上下文数据，以提供全面的 一幅与健康有关的现象的图片。BESI-C系统的一个独特功能是患者和 护理人员使用自定义从他们自己的角度记录和描述癌症疼痛事件 应用程序在他们各自的智能手表上。具体地说，这项观察性研究将分析收集的数据 通过BESI-C从门诊肿瘤学姑息治疗诊所招募的患者-家庭照顾者夫妇和 一项家庭临终关怀计划，以发展家庭晚期癌症疼痛的全面“数字表型” 布景。这些数字表型将表征疼痛的频率、强度和对生活质量的影响 事件；监测药理学和非药理学战略的使用和自我报告 有效性；将环境、背景、行为和生理传感器数据与报告的疼痛相关联 事件；并评估患者和照顾者数据的一致性。这项研究还将探索偏好 用于与患者、家庭护理人员和医疗保健提供者交流收集的数据，方法是 共享数据可视化。此外，我们还将探索哪些传感数据最具预测性 突破疼痛事件，构建简约的疼痛预测算法。

Project Summary Cancer pain is complex, prevalent and has serious consequences for patients, family caregivers and healthcare systems. Inadequately managed cancer pain can be particularly problematic for patients coping with advanced, metastatic disease. Most symptom management occurs in the home setting, and family caregivers often play a key role in helping to manage cancer pain, but often find this task daunting and stressful. Complicating cancer pain management is the reality that opioids, a main-stay class of medications used to control serious cancer pain, are subject to increased scrutiny given well-publicized concerns about the national `opioid epidemic.' Now more than ever, it is imperative we understand how patients and family caregivers attempt to manage cancer pain at home so we can offer them personalized support to effectively and safely alleviate pain. Mobile and wireless technology (`smart health') can help support symptom management in the home setting, but must be carefully designed to account for the realities of patients and family caregivers coping with advanced disease. We hypothesize that individuals, and patient-family caregiver dyads, will display a unique `digital fingerprint' (or phenotype) of the advanced cancer pain experience – that if better understood can be utilized to inform and deliver personalized, timely interventions. The purpose of this study, which builds upon preliminary pilot work, is to deploy an unobtrusive smart health system, the Behavioral and Environmental Sensing and Intervention for Cancer (BESI-C), to monitor and describe – and ultimately to predict and help manage – the experience of advanced cancer pain in the home setting. BESI-C is comprised of wearable (smart watch) and environmental sensors that collect physiological, behavioral, and contextual data at the individual, dyad and home level that can be integrated to provide a comprehensive picture of a health related phenomenon. A unique feature of the BESI-C system is the ability of patients and caregivers to record and characterize cancer pain events from their own perspective using a custom application on their respective smart watch. Specifically, this observational research will analyze data collected via BESI-C from patient-family caregiver dyads recruited from an outpatient oncology palliative care clinic and a home hospice program, to develop comprehensive `digital phenotypes' of advanced cancer pain in the home setting. These digital phenotypes will characterize the frequency, intensity and impact on quality of life of pain events; monitor the use of pharmacological and non-pharmacological strategies and self-reported effectiveness; correlate environmental, contextual, behavioral and physiological sensor data with reported pain events; and evaluate concordance of patient and caregiver data. This research will also explore preferences for communicating collected data with patients, family caregivers and healthcare providers by creating and sharing data visualizations. Additionally, we will explore which sensing data are most predictive of breakthrough pain events to build parsimonious pain prediction algorithms.