Research consistently shows that subtle changes in daily routine are among the earliest indicators of decline in older adults. Hearth is designed to detect these patterns - the ones that are invisible in a phone call but obvious in the data.
Nocturia (≥2 visits per night) doubles the risk of falls and fractures, and is often the first behavioural sign of a UTI in older women - before any urinary symptoms appear. Hearth tracks bathroom motion between 10pm and 6am nightly. A rising trend triggers an alert.
Increased unpredictability in daily routines is one of the strongest passive indicators of early cognitive decline. Studies using in-home sensors detected mild cognitive impairment with ~85% accuracy by measuring deviation from established patterns. Hearth compares every day against a 14-day and 90-day baseline.
Reduced kitchen visits are a validated proxy for meal preparation decline - itself a key Instrumental Activity of Daily Living. Sensor-detected reductions predicted functional decline 6–12 months before clinical assessment. Hearth tracks kitchen session frequency, duration, and timing daily.
Later wake times, fragmented sleep, and excessive time in bed (>9 hours) are associated with incident frailty and depression. Hearth uses the bedroom presence sensor to track bedtime, wake time, and overnight restlessness - comparing each night against their personal baseline.
Reduced out-of-home activity is linked to social withdrawal, mobility decline, and incident dementia. Hearth monitors front door events to track outing frequency and duration week over week. A gradual decline is flagged in the weekly trend report before it becomes obvious.
Fewer room-to-room movements per day correlate with reduced functional capacity and predict hospitalisation. Changes in transition timing between rooms can indicate gait speed decline - a predictor of falls weeks before they occur. Hearth tracks cross-room activity and pace throughout the day.
Hearth compares every metric against a rolling 14-day baseline and a 90-day reference. Gradual decline that would be invisible in a phone call becomes clear in the data.