Preventing Medication Mismanagement in People Living with Dementia through Automated Medication Dispensing with Facial Recognition and Video Observation

通过面部识别和视频观察自动配药，防止痴呆症患者用药管理不善

• 批准号: 10461514
• 负责人: Charles Gellman
• 金额: $ 44.96万
• 依托单位: HIDO TECHNOLOGIES, INC.
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-15 至 2024-07-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10461514
• 关键词: Accountability; Activities of Daily Living; Address; Adherence; Alzheimer' s Disease; Artificial Intelligence; Automation; Caregiver Burden; Caregivers; Caring; Cellular Phone; Clinic; Complex; Consumption; Dementia; Dementia caregivers; Development; Devices; Disease; Dose; Elderly; Emergency department visit; Ensure; Face; Family; Feedback; Goals; Health; Health Care Costs; Home; Hospitalization; Incidence; Individual; Ingestion; Intuition; Label; Lead; Legal patent; Manuals; Medical; Medication Management; Modeling; Monitor; Nursing Homes; Outcome; Patient Care; Patient Rights; Patients; Performance; Persons; Pharmaceutical Preparations; Phase; Pilot Projects; Population; Protocols documentation; Provider; Reporting; Rights; Risk Factors; Route; Sales; Schedule; Secure; Serious Adverse Event; Small Business Innovation Research Grant; Sorting - Cell Movement; Stream; System; Target Populations; Technology; Testing; Time; Update; Validation; adverse drug reaction; age related; aging in place; base; cognitive function; comorbidity; compliance behavior; cost; dashboard; design; dosage; executive function; health care service utilization; hospital readmission; improved; innovation; instrumental activity of daily living; machine vision; medication administration; medication nonadherence; new technology; novel; prevent; prototype; sensor; smartphone Application; socioeconomic disparity; software as a service; success; tool; trend; usability

Globally, over 47 million individuals are living with dementia, with new incidence of 7.7 million annually. Medication mismanagement is one of the most common and concerning risk factors in people with dementia (PwD), as it leads to undertreatment of disease, emergency department visits, hospital admissions/readmissions, and serious adverse events. In the U.S. an estimated 3 million older adults are admitted to nursing homes due to drug-related adherence problems with annual cost exceeding $14 billion. The challenge is complex medication management requires moderate executive functioning. However, as cognitive function declines, PwD can no longer perform such Instrumental Activities of Daily Living (IADLs) safely, effectively, and independently. While the goal is to keep older adults at home as long as possible, caregivers are not available 24/7 & costs of external care are often prohibitive. The HiDO platform will solve these market challenges by automating medication administration for PwD to eliminate mismanagement, decrease caregiver burden, reduce healthcare utilization and facilitate the ability for PwD to age in place. While still premarket, HiDO is being designed and validated as an automated, AI driven medication dispensing and direct observation platform to optimize adherence. The innovative device integrates medication dispensing, dose administration time, medication synchronization, and a pair of front-facing video cameras to validate the right medications, the right route, right time, right dosage to the right patient (5R’s). The cameras record every dose using facial recognition & provide real-time medication consumption recordings for medical review if needed by monitoring the time a patient interacts with the device. Through cloud connectivity, providers & caregivers have access to video observation logs, dose administration time, adherence trends, & longitudinal adherence through the platform’s dashboard. Patients & caregivers can easily setup complex medication protocols in minutes using a smartphone app. The device then alerts patients and dispenses up to 7 different types of meds simultaneously, with up to 40 doses each. The fully commercialized HiDO platform will integrate the full feature suite above. However, to demonstrate feasibility, Phase I will target an in-clinic usability study, platform enhancements & novel AI to confirm ingestion, and remote pilot study to document independent usability & adherence in PwD. An existing prototype HiDO platform, which already integrates facial recognition AI, will be leveraged as a base technology to increase likelihood of project success. First, using the existing prototype we will complete an in-clinic usability study to validate use cases and product features in the target population. The existing platform will then be enhanced with machine vision AI to confirm medication ingestion, as well as updates to address challenges found in early usability. Once the enhanced platform has been technically verified, it will be deployed in a remote field usability study with PwD and caregivers.

在全球范围内，超过4700万人患有痴呆症，每年新发770万人。 药物管理不当是痴呆症患者最常见和最令人担忧的危险因素之一 (PwD)，因为它导致疾病治疗不足，急诊室就诊，住院/再住院， 和严重不良事件。在美国，估计有300万老年人被送进养老院， 每年花费超过140亿美元的与药物相关的依从性问题。挑战在于复杂的药物治疗 管理需要适度的执行功能。然而，随着认知功能的下降，PwD不能 安全、有效和独立地进行这些工具性日常生活活动（IADL）。而 目标是尽可能长时间地让老年人呆在家里，照顾者不是24/7和外部费用 护理往往是禁止的。 HiDO平台将通过为PwD自动化药物管理来解决这些市场挑战， 消除管理不善，减轻护理人员的负担，减少医疗保健利用率，并提高 PwD在原地老化。虽然仍处于上市前阶段，但HiDO正在被设计和验证为自动化的人工智能驱动的 药物分配和直接观察平台，以优化依从性。该创新设备集成了 药物分配、给药时间、药物同步和一对前置视频 摄像头来验证正确的药物，正确的路线，正确的时间，正确的剂量给正确的病人（5 R）。的 摄像头使用面部识别记录每次剂量，并提供实时药物消耗记录， 如果需要，通过监测患者与器械交互的时间进行医疗审查。通过云连接， 提供者和护理人员可以访问视频观察日志、剂量给药时间、依从性趋势， 通过平台的仪表板纵向附着。患者和护理人员可以轻松设置复杂的 使用智能手机应用程序在几分钟内完成药物治疗方案。然后，该设备会提醒患者并分配多达7 同时使用不同类型的药物，每种最多40剂。 完全商业化的HiDO平台将集成上述全部功能套件。然而，为了证明 可行性，第一阶段将针对临床可用性研究，平台增强和新型AI以确认摄入， 和远程试点研究，以记录PwD中的独立可用性和依从性。现有的HiDO原型 平台已经集成了面部识别AI，将被用作基础技术，以增加 项目成功的可能性。首先，使用现有原型，我们将完成一项临床可用性研究， 在目标人群中验证用例和产品功能。现有的平台将得到加强 使用机器视觉AI来确认药物摄入，以及更新以解决早期发现的挑战。 可用性一旦增强的平台得到技术验证，它将被部署在一个偏远的外地可用性 与PwD和护理人员一起学习。