A risk prediction model integrating smart wearable device sleep data with clinical assessments achieved an AUC of 0.92 for identifying preoperative sleep disorders in older adult surgical patients, with Hospital Anxiety and Depression Scale score, number of awakenings, REM sleep duration, and light sleep duration as independent predictors.
Key Findings
Results
Hospital Anxiety and Depression Scale (HADS) score was an independent risk factor for preoperative sleep disturbances in older adults.
Odds ratio 3.21 (95% CI 1.54-6.69; P=.002) from multivariable logistic regression
Higher HADS scores were associated with increased risk of preoperative sleep disorder
Study population consisted of 242 older surgical patients at the Second Affiliated Hospital of Zunyi Medical University
Patients were monitored using smart rings on the night before surgery
Results
Number of awakenings was an independent predictor of preoperative sleep disorders in older surgical patients.
Odds ratio 3.33 (95% CI 1.82-6.12; P<.001), the strongest predictor among the wearable-derived variables
Data on awakenings were collected objectively via smart ring wearable devices on the night before surgery
This was identified through univariable and multivariable logistic regression analysis
Results
Longer duration of REM sleep was associated with reduced risk of preoperative sleep disturbances.
Odds ratio 0.96 (95% CI 0.93-0.99; P=.04), indicating a protective effect per unit increase in REM sleep duration
REM sleep duration was measured objectively using smart ring wearable devices
This was identified as an independent predictor via multivariable logistic regression
Results
Longer duration of light sleep was independently associated with reduced risk of preoperative sleep disorders.
Odds ratio 0.98 (95% CI 0.96-0.99; P=.01), indicating a protective association per unit increase in light sleep duration
Light sleep duration was derived from smart ring monitoring data
Identified alongside REM sleep duration, HADS score, and number of awakenings as the four independent predictors in the final model
Results
The risk prediction model demonstrated high discriminative ability for identifying preoperative sleep disorders in older surgical patients.
Area under the receiver operating characteristic curve (AUC) of 0.92
Model was internally validated using 1000 bootstrap samples
The calibration curve indicated good model calibration
Decision curve analysis showed the model improved maximum net benefit across risk thresholds ranging from 0.2 to 0.8, indicating high clinical utility
Methods
The study employed smart ring wearable devices to objectively collect sleep data from older adult surgical patients on the night before surgery.
A prospective cohort design was used with 242 older surgical patients
Data collected included sociodemographic factors, cognition, and psychological status alongside wearable sleep metrics
Patients were classified into sleep disorder and non-sleep disorder groups based on preoperative sleep assessments
Both univariable and multivariable logistic regression were used to identify independent predictors
What This Means
This research suggests that a combination of anxiety and depression scores and objective sleep measurements collected by a smart ring worn the night before surgery can accurately predict which older adults are at risk for sleep problems before their operation. The study followed 242 older patients scheduled for surgery at a hospital in China, monitoring their sleep using wearable smart rings and also assessing their mental health and other clinical factors. Four key factors were found to independently predict preoperative sleep disorders: higher anxiety and depression scores, more nighttime awakenings, shorter REM (deep dreaming) sleep, and shorter light sleep duration.
The prediction model built from these four factors performed very well, achieving an accuracy score (AUC) of 0.92, where 1.0 would be perfect. The model was also shown to be well-calibrated and clinically useful across a wide range of risk thresholds, meaning it could help doctors make practical decisions about which patients need extra attention before surgery.
This research suggests that routinely using wearable devices to monitor sleep combined with mental health screening could help medical teams identify older surgical patients who may need targeted interventions — such as anxiety management or sleep support — before their operation. Since poor sleep before surgery is linked to worse outcomes afterward, early identification of at-risk patients could potentially improve recovery and reduce healthcare costs for this vulnerable population.
Li J, Yang B, Gao P, Feng D, Shao X, Cai X, et al.. (2026). Predictive Modeling of Preoperative Sleep Disorder Risk in Older Adults by Using Data From Wearable Monitoring Devices: Prospective Cohort Study.. JMIR formative research. https://doi.org/10.2196/79008