Predicting Health Risks Using Activity and Sleep Data

A new study shows that sleep patterns can predict disease risk more accurately than traditional markers like age and body-mass index We often think of sleep studies as tools for diagnosing sleep apnea or other sleep disorders. But what if those same tests could reveal far more—like your risk of diabetes, heart disease, or even frailty? A new study suggests just that. By analyzing the fine details of how we sleep, researchers have uncovered links between sleep patterns and a wide range of bodily functions—from metabolism to heart health. The study, published in Nature Medicine, comes from the Human Phenotype Project—a large Israeli cohort that combines wearable sleep monitors, blood tests, body scans, microbiome analysis, and more. Researchers examined 448 distinct sleep traits across over 16,000 nights of home monitoring in more than 6,300 adults. It’s the most comprehensive sleep phenotyping study to date, showing that sleep patterns—especially disruptions in breathing, oxygenation, and sleep stages—can predict health outcomes, often more accurately than traditional risk factors like BMI or age. Key findings: - More visceral fat meant worse sleep quality, especially higher levels of sleep-disordered breathing (measured by the apnea–hypopnea index, or AHI). This effect was seen even after accounting for age, sex, and BMI. - Sleep data predicted insulin resistance and cholesterol levels more accurately than traditional risk markers like age, BMI, or even visceral fat—especially in women. - Lower oxygen levels during sleep, less deep sleep, and more breathing disruptions at night increased the risk of future heart, metabolic, and hormone-related diseases. - People with less physical activity or more screen time tended to have worse sleep, even more so than those with higher BMI or body fat. - Gut microbiome composition and dietary patterns were strong predictors of daytime sleepiness, particularly in women—surpassing age or BMI as predictors. - Lower pulse rate variability (PRV) during sleep—a marker of autonomic nervous system function—was linked to higher frailty, poorer bone #health, and elevated blood lipids, offering new potential tools for cardiovascular risk screening. The results underscore a broader role for sleep testing beyond diagnosing sleep apnea. Multi-night sleep data—capturing oxygen drops, breathing patterns, and autonomic signals—can reveal early signs of metabolic, cardiovascular, and even bone health issues, often outperforming traditional risk markers like BMI or age. With machine learning models, the researchers showed that sleep metrics alone explained over 15% of the variance in crucial biomarkers like blood glucose and triglycerides. This suggests that home-based sleep monitoring could evolve into a powerful, non-invasive screening tool for early detection of systemic health risks—turning sleep from a symptom to watch into a signal to act on.   Study: https://lnkd.in/dHFZxAV3 #sleep #research

Your smartwatch knows more than your medical record. Most wearable AI models today are built from noisy second-by-second sensor streams: heart rate, steps, accelerometer blips. But a new study just flipped that script. Researchers trained a foundation model on 2.5 billion hours of behavioral data from over 162,000 people. Not raw pulses. Actual human behavior. Sleep patterns. Walking speed. Exercise consistency. Recovery trends. The result? Better predictions. Across 57 health tasks. • Sleep quality • Pregnancy detection • Cardiovascular fitness • Hypertension status • Even medication use Behavior beats biosignal. Why does this matter? Because behaviors operate on human timescales. Days and weeks, not milliseconds. And they reflect your life, not just your pulse. Even better: when behavioral data was combined with raw sensor signals like PPG or ECG, accuracy jumped even higher. Clinical-grade insight from consumer-grade devices. This model was trained on data from the Apple Heart and Movement Study - one of the largest wearable health datasets ever assembled. This is not just an Apple Watch story. It signals a broader shift in healthcare that may soon be: → Grounded in behavior → Powered by AI → Driven by consumers And it reframes the big question. Not can AI read your heart rate? But can AI learn your habits well enough to warn you before your doctor does? “The future of medicine won’t be written in code or blood. It’ll be written in how you move, rest, walk, sleep, and show up every day.”

This paper explores the use of wearable activity tracker data and individual characteristics to improve mortality prediction using XAI techniques and ML models. 1️⃣ Data from the UK Biobank, including wrist-worn accelerometer readings and demographic, lifestyle, and health factors, was analyzed. 2️⃣ Random Forest models were the most accurate, achieving an AUC of 0.78, followed by Gradient Boosted Machines. 3️⃣ Key predictors include physical activity levels, age, and body fat percentage, with nonlinear relationships observed between these variables and mortality. 4️⃣ XAI methods like SHAP and LIME provided transparency by explaining variable importance and individual predictions. 5️⃣ Practical implications include applications in healthcare planning, personalized medicine, and insurance modeling. 6️⃣ Ethical considerations such as privacy and data use were acknowledged as barriers to implementation. ✍🏻 Byron Graham, Mark Farrell. Mortality prediction using data from wearable activity trackers and individual characteristics: An explainable artificial intelligence approach. Expert Systems With Applications. 2025. DOI: 10.1016/j.eswa.2024.126195

Sleep patterns and risk of chronic disease as measured by long-term monitoring with commercial wearable devices in the All of Us Research Program Link: https://lnkd.in/ekkUW96E Poor sleep health is associated with increased all-cause mortality and incidence of many chronic conditions. Previous studies have relied on cross-sectional and self-reported survey data or polysomnograms, which have limitations with respect to data granularity, sample size and longitudinal information. Here, using objectively measured, longitudinal sleep data from commercial wearable devices linked to electronic health record data from the All of Us Research Program, we show that sleep patterns, including sleep stages, duration and regularity, are associated with chronic disease incidence. We conducted the largest study to date that analyzes sleep patterns objectively and longitudinally across many years using direct measures of sleep from commercial wearable devices linked to EHR data. We observed clinically and statistically significant relationships between sleep quantity, quality and regularity and the onset of important chronic diseases even after accounting for daily activity (that is, step counts). Our discovery analyses demonstrated several expected findings, such as the association between increased restless sleep duration and increased odds of incident sleep disorders or insomnia. Similarly, participants who had fewer sleep onset times between 8:00 p.m. and 2:00 a.m. had higher odds of incident sleep disorders, including insomnia, hypersomnia and circadian rhythm disorders. In addition, many of our findings are supported by previous studies using large-scale population surveys, such as our observations that decreased daily sleep duration and increased sleep irregularity are associated with obesity and sleep apnea4,7. Daily average sleep duration had nonlinear, J-shaped associations with hypertension, major depressive disorder and generalized anxiety disorder, a well-documented phenomenon reported in previous studies1,2,3,4,5,6,31,32. Our findings, along with those from previous epidemiologic studies, support the notion that 7 h of objectively measured sleep may be the middle of the healthy range for adults rather than the floor33, highlighting that the range of healthy sleep depends on the tool used to measure sleep. #sleep #sleephealth #sleepmedicine #sleep2025 #sleeptrends #hme #healthcare #health #cpap #osa #sleepapnea #sleepresearch #phenotype