1256 Daily Stress and Objective Sleep Outcomes in a Real-World Sample of over 83,000 U.S. Adults

Abstract Introduction While daily stress is believed to affect sleep, associations with objective sleep architecture in large real-world samples are not well established. Using longitudinal data from nearly 2 million nights captured by a noninvasive under-mattress monitoring device, we tested whether reports of daily stress were associated with differences in sleep duration, sleep stages, and sleep disturbance indices. Methods From 11/15/2023-11/14/2025, overnight sleep across 83,288 U.S. adults (1,982,889 nights) was recorded using the Sleeptracker-AI Monitor (Fullpower Technologies Inc., California, USA), which continuously captures sleep stages, respiration, and movement via piezo-electric sensors. Individuals responded each day to the question: “Was your day stressful?” with three options: stress-free, normal, or stressful. Outcomes included total sleep time (TST), wake after sleep onset (WASO), REM sleep percentage, deep sleep percentage, arousal index (AI), and breathing anomaly index (BAI). We used mixed linear models with participant as a random effect to estimate the percentage changes in sleep parameters associated with self-reported daily stress. Results The sample was 47% female and had a mean age of 51.6±14.5 years. Relative to “stress-free” days, “normal” days were associated (p 1×10 -6 for all) with 0.9% lower TST (standard error SE=0.04%), 1.4% lower WASO (SE=0.13%), 1.0% higher REM sleep percentage (SE=0.06%), and 1.7% higher deep sleep percentage (SE=0.10%); and “stressful days” were associated with 2.6% lower TST (SE=0.05%), 0.9% higher WASO (SE=0.18%), and 3.5% higher deep sleep percentage (SE=0.14%). Lower BAI and AI were also observed on “normal” (BAI: -0.3%, SE=0.11%; AI: -0.7%, SE=0.06%), and “stressful” days (BAI: -1.3%, SE=0.16%; AI: -1.7%, SE=0.09%), relative to “stress-free” days (p 0.001). Conclusion In this large real-world sample, subjective daily stress ratings were associated with significant differences in objective sleep measurements. Higher daily stress was linked to shorter TST, higher WASO, higher REM and deep sleep percentages, and lower breathing anomaly index and arousal index. Future work should explore how daily stress interacts with individual sleep patterns over time. Continued use of large-scale at-home monitoring will be central to understanding these relationships in real-world settings. Support (if any) Fullpower Technologies