Validation and comparison to the standard of care of the TrueLoo smart device to record toileting sessions in older adults: A retrospective validation and acceptance study (Preprint)

Abstract

Due to the relationship between independent living and activities of daily living, care teams spend significant time managing senior living residents’ toileting problems. Recently, the TrueLoo (TL) was developed as a connected toilet seat to automatically log and monitor toileting sessions.

The purpose of this retrospective study was to demonstrate the validity of the TL to (a) accurately record and identify toileting sessions with regard to stool and urine events, (b) compare results to the person-reported standard of care methods, and (c) establish metrics of user acceptability and ease of use in a senior living facility population.

We utilized two phases: (1) initial development of the TL algorithms to accurately identify urine and stool events; (2) evaluation of the algorithms against person-reported standard of care methods, commonly employed in senior living facilities. Phase 2 analyzed data over a 3-day period from 52 devices. Participants ranged from 63 and 101 years old (average age = 84 ± 9.35 years). Acceptability and ease of use data were also collected.

Regarding development of the TL algorithm for urine assessment, sensitivity and specificity of 96% and 85% were observed when evaluating a gold-standard labeled dataset, respectively (F1 score = .95). For stool, sensitivity and specificity of 90% and 79% were observed, respectively (F1 score = .85). Regarding the evaluation of the TL algorithm in senior-living settings, classification performance statistics for urine assessment revealed a sensitivity and specificity of 84% and 94%, respectively (F1 score = .90). For stool, a sensitivity and specificity of 92% and 98% were observed, respectively (F1 score = .91). Throughout the study, 46 person-reported instances of urine were documented, compared to 630 recorded by the TL. For stool events, 116 person-reported events were reported, compared to 153 events reported by the TL. This indicates person-reported events were captured 7% of the time for urine and 76% of the time for stool. Overall, 45.1% of participants said the new toilet seat was better than their previous seat, 84.5% participants reported using TL was easy, requiring no effort, and 97.6% said they believed the monitoring system had the potential to help aging adults. Over 98% of participants reported they would find alerts related to their health valuable and would be willing to share this information with their doctor. When asked about sharing information with caregivers, 65.7% reported they would rather the TL send information and alerts to their caregiver, as opposed to the participant having to personally communicate those details.

The TL demonstrated the feasibility to accurately record toileting sessions compared to standard of care methods, while successfully establishing metrics of user acceptability and ease of use in senior living populations. While additional validation studies are warranted, the data presented in this paper support the utilization of the TL in senior living settings as a model of event monitoring during toileting.