Use of a Novel Toilet Seat to Passively Collect Digital Biomarkers in Assisted Living Settings

使用新型马桶座圈在辅助生活环境中被动收集数字生物标记

• 批准号: 10325655
• 负责人: Paul Cristman
• 金额: $ 29.97万
• 依托单位: TOI LABS, INC.
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-09-01 至 2024-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10325655
• 关键词: Adverse event; Affect; Algorithms; Antibiotic Resistance; Antibiotics; Assisted Living Facilities; Behavior; Biological Markers; Blood; Cause of Death; Characteristics; Chronic Kidney Failure; Clinical; Clinical Management; Clinical Research; Clostridium difficile; Colon Carcinoma; Color; Computational algorithm; Constipation; Coprophagia; Data; Data Set; Defecation; Dehydration; Diarrhea; Disease; Documentation; Early Diagnosis; Early treatment; Elderly; Event; Fecal Impaction; Feces; Female; Frequencies; Future; Gastrointestinal Hemorrhage; Gastrointestinal tract structure; Goals; Health; Health Personnel; Health Professional; Health care facility; Healthcare; Hospitalization; Image; Independent Living; Individual; Infection; Inflammatory Bowel Diseases; Lead; Left; Legal patent; Link; Liquid substance; Long-Term Care; Machine Learning; Manuals; Medical; Medical Records; Methods; Monitor; Morbidity - disease rate; Names; Onset of illness; Operative Surgical Procedures; Patient Self-Report; Patient-Focused Outcomes; Patients; Pattern; Phase; Pilot Projects; Population; Prevalence; Primary Health Care; Records; Reference Standards; Reference Values; Reporting; Source; Speed; System; Technology; Testing; Time; Urinary system; Urinary tract infection; Urination; Urine; Vulnerable Populations; base; caregiving; clinical decision-making; community living; cost; detection method; digital; healthcare-associated infections; hospital readmission; infection rate; instrument; machine learning algorithm; male; memory care; novel; patient health information; residence; resistant strain; screening; stem; tool

ABSTRACT In senior living facilities, some of the most common health related problems stem from the GI tract and urinary system including Clostridium difficile Infections (CDI), Urinary Tract Infections (UTI), Constipation and Colon Cancer. The best way to lower costs and treat these conditions effectively is early diagnosis and treatment. Current clinical management for these conditions, as well as many others, include monitoring of specific excreta characteristics including urine color, urination frequency, urine duration, stool color, stool frequency and stool consistency. These logs are the best tool doctors currently have to screen for conditions such as UTIs, infectious diarrhea, dehydration, chronic kidney disease, GI bleeding, GU surgery, inflammatory bowel disease, and constipation, that can lead to hospitalizations and readmissions. However, these logs are on average 61% inaccurate at reporting adverse episodes such as diarrhea. Toi Labs has developed the patented TrueLoo technology to take pictures of excreta and, using machine learning algorithms, classify the toileting event using Digital Biomarkers (DBMs). The ability to create an excreta log to accurately deliver detailed information to doctors and healthcare providers can revolutionize healthcare by notifying when further screening (urine or fecal) is necessary. This novel, low-cost approach of machine learning and image identification technology that requires no change in behavior of the user will enable currently undetectable links between medical records and specific excreta patterns. In the future, the machine learning algorithm may be able to determine links between these excreta logs and the onset of specific diseases. In this study we will be collecting manual excreta records and automated TrueLoo digital excreta records in memory care and assisted living facilities, and compare them against each other and against patients’ deidentified medical records. We will determine correlative data between excreta logs and adverse events to establish conditional threshold for each type of adverse event and, using ML, try and establish individual conditional thresholds for reporting to caregiving staff. We will compare the manual and digital logs to assess the difference between speed of adverse episode identification when using TrueLoo as compared to manual logs.

摘要 在老年生活设施，一些最常见的健康相关的问题源于胃肠道和泌尿系统 系统包括艰难梭菌感染（CDI）、尿路感染（UTI）、结肠和直肠感染 癌降低成本和有效治疗这些疾病的最佳方法是早期诊断和治疗。 目前对这些疾病的临床管理，以及许多其他疾病，包括监测特定的排泄物 特征包括尿液颜色、排尿频率、排尿持续时间、粪便颜色、粪便频率和粪便 一致性这些日志是最好的工具，医生目前有屏幕的条件，如尿路感染，传染性 腹泻、脱水、慢性肾病、胃肠道出血、GU手术、炎症性肠病，以及 便秘，这可能导致住院和再入院。然而，这些日志平均为61%， 在报告腹泻等不良事件时不准确。Toi Labs开发了专利TrueLoo 技术来拍摄排泄物的照片，并使用机器学习算法， 数字生物标志物（DBMs）。能够创建排泄物日志，以准确地向 医生和医疗保健提供者可以通过通知进一步筛查（尿液或粪便）时， 是必要的.这种新颖、低成本的机器学习和图像识别技术方法， 不需要改变用户的行为，将使医疗记录之间目前无法检测到的联系， 特殊的排泄物模式在未来，机器学习算法可能能够确定 这些排泄物日志和特定疾病的发作。在这项研究中，我们将收集人工排泄物记录 和自动化TrueLoo数字排泄物记录在记忆护理和辅助生活设施，并比较他们 以及病人的匿名医疗记录我们将确定相关数据之间 排泄日志和不良事件，以建立每种类型不良事件的条件阈值，并使用ML， 尝试建立个人的有条件的阈值，以便向新员工报告。我们将比较手册 和数字日志，以评估使用TrueLoo时不良事件识别速度之间的差异 与手动日志相比。