Passive non‐wearable technology for real‐time detection of COVID‐19 infection in people living with dementia

• Published in: Alzheimer's & dementia Vol. 19; no. S11
• Main Authors: Kolanko, Magdalena A., Soreq, Eyal, Lai, Helen, Purnell, Matthew, Walsh, Chloe, Whitethread, Nicole, Sharp, David J
• Format: Journal Article
• Language: English
• Published: 01.12.2023
• ISSN: 1552-5260; 1552-5279
• DOI: 10.1002/alz.082219

Background People with dementia (PLWD) are disproportionately affected by COVID‐19, as they are at a greater risk of becoming infected (1), and have an increased risk of disease‐related morbidity and mortality (2,3). The clinical presentation of COVID‐19 in PLWD is often asymptomatic or atypical, reducing early recognition of symptoms and timely treatment (4). Continuous monitoring of physiological parameters such as heart rate and respiratory rate using digital health technologies holds promise for early disease detection. Recent studies demonstrated that consumer wearable devices such as smartwatches can identify COVID‐19 by capturing subtle intra‐individual changes in physiological parameters (5‐11). However, the use of wearable devices in PLWD is significantly limited due to the lack of acceptance and poor compliance (11‐14). Here we present the first non‐contact monitoring technology using an under‐mattress pressure‐sensor mat to detect and predict COVID‐19 infection in PLWD. Method We used contactless Withings Sleep Analyser (WSA) to continuously monitor the night‐time physiology of 120 PLWD enrolled in the CR&T MINDER cohort study. A total of 38 positive COVID‐19 cases were reported since late 2021. Based on the minute‐to‐minute time series extracted from the WSA, average nocturnal heart and respiratory rates were calculated. We examined whether physiological deviations from a person’s baseline were detected in retrospect around the time of the reported illness. Result Out of the 38 individuals who tested positive for COVID‐19 during the study (Fig.1), 25 had sleep data available around the time of their positive test. Of these 25 cases, 24 (96%) had abnormally high respiratory rate (Fig.2a), and 22 also had abnormally high heart rates (Fig.2b) within the 2‐week period preceding the COVID‐19 diagnosis. Our anomaly detection algorithm, which flags deviations in heart and respiratory rates above a certain threshold, correctly identified all 24 cases of abnormal physiology associated with COVID‐19 infection. Conclusion Passive monitoring of night‐time physiology using non‐invasive and non‐wearable technology allows early identification of changes in an individual’s baseline physiology, which can be used for real‐time detection of respiratory infections, including COVID‐19 in vulnerable groups of patients. Such technology could help healthcare providers to proactively identify and treat infections early, potentially improving patient outcomes.