Identifying falls risk using wearables data in older adults: an observational cohort study

Background Falls are a major cause of morbidity in older adults. Low-cost wearable devices track potential falls risk factors, but adoption in older adults remains uncertain. Methods We conducted a 6-month prospective observational study in community-dwelling adults who self-reported a recent fall or were deemed at increased risk. Participants were given a wrist-worn wearable device (Fitbit, Garmin or Polar), synced with a smartphone application (Smplicare app) to collect additional information by questionnaires, including self-reported falls. We analysed adherence wearing the devices, and studied step count and sleep data in relation to falls. Results Of 284 people (74.2 +/- 9.0 years, 68% women) in the study, 266 (94%) provided at least 7 days of data, with 196 (76%) engaged on at least half of study days. Engagement did not differ by self-reported technology confidence. There were 81 (30%) people who reported a fall during follow-up, but only 5 (6%) resulted in hospital attendance. Each additional hour of average sleep was associated with a 24% reduction in falls risk (HR 0.76, 95% CI 0.63 to 0.92), but in multivariable models only carer support (aHR 3.47, 95% CI 1.46 to 8.26) and incontinence (aHR 2.26, 95% CI 1.34 to 3.82) remained independently associated with falls. No changes in step or sleep patterns were noted after falls, but there was high individual heterogeneity. Conclusion Wearable adoption, risk factor identification and digital self-reporting of falls is feasible in older adults using low-cost commercial technology. The importance of simple wearable measures like sleep for fall risk were outweighed by markers of frailty. Future research should understand how these granular wearable data could add to proactive falls risk assessment.