JMIR Publications

ABSTRACT

Background:

Sleep disorders frequently affect elderly surgical patients, significantly contributing to postoperative complications, delayed recovery, and increased healthcare expenditures. Age-related physiological decline combined with the stress of surgery further disrupts sleep architecture in this vulnerable population. However, existing predictive models for sleep disorders do not sufficiently account for the specific risk factors unique to elderly surgical patients. This gap limits the ability to accurately identify high-risk individuals and implement timely interventions, thereby impeding improvements in perioperative care and outcomes.

Objective:

To develop and validate a risk prediction model for preoperative sleep disorders in elderly surgical patients utilizing smart wearable device monitoring data and clinical assessments, aiming to facilitate early identification of influencing factors and provide a scientific basis for personalized care planning.

Methods:

This prospective study was conducted at the Second Affiliated Hospital of Zunyi Medical University to address the need for accurate risk prediction of preoperative sleep disorders in elderly surgical patients. We developed and internally validated a risk prediction model tailored to elderly surgical patients by integrating objective sleep data from smart wearable devices with comprehensive clinical and psychosocial assessments. A prospective cohort of 242 elderly patients was monitored using smart rings on the night preceding surgery, with simultaneous collection of sociodemographic, cognitive, and psychological data. Patients were categorized into sleep disorder and non-sleep disorder groups based on preoperative sleep assessments. Univariate and multivariate logistic regression analyses identified independent predictors of preoperative sleep disorders, which were then used to construct the risk prediction model. Internal validation was performed via 1000-bootstrap resampling. The model’s predictive performance was assessed through discrimination (ROC curve), calibration, and decision curve analyses.

Results:

Multifactorial logistic regression analysis showed that HADS score (OR=3.21, 95% CI=1.54-6.69), number of wakefulness (OR=3.33, 95% CI=1.82-6.12), duration of rapid eye movement sleep (OR=0.96, 95% CI=0.93-0.99), duration of light sleep (OR=0.98 , 95% CI=0.96 to 0.99) were independent risk factors for preoperative sleep disturbances in elderly patients (P<0.05).The ROC curve showed an AUC of 0.92, and the calibration curve indicated that the model had good calibration. Decision curve analysis showed that the model improved the maximum net benefit between the risk thresholds of 0.2 and 0.8, which has high clinical utility.

Conclusions:

The risk prediction model developed using smart ring-derived data effectively identifies elderly surgical patients at elevated risk of preoperative sleep disturbances, thereby facilitating timely and individualized interventions. This advancement provides a robust scientific foundation for delivering personalized perioperative care, with the potential to improve postoperative outcomes and alleviate the healthcare burden in this vulnerable population.