Sleep disturbance has become a considerable factor affecting the quality of life for middle-aged and elderly people; however, there are still many obstacles to screening sleep disturbance for those people. Given the growing awareness of the association between gastrointestinal function and sleep disturbance, our study aims to predict the risk of sleep disturbance using gastrointestinal electrophysiological signals.
The Pittsburgh Sleep Quality Index and gastrointestinal electrophysiological signals of 914 participants in western China were used to establish the model. Demographic characteristics and routine blood test were collected as covariates. Participants were randomly assigned into two sets with a 7:3 ratio for training and validation. In the training set, the least absolute shrinkage and selection operator (LASSO) regression and stepwise logistic regression were used, respectively for variables selection and optimization. To assess the model performance, receiver operator characteristic (ROC) curve, calibration curve and decision curve analysis (DCA) were utilized. Then, validation was performed.
Thirteen predictors were chosen from 46 variables by LASSO regression. Then, age, gender, percentage of normal slow wave and electrical spreading rate on the pre-meal gastric channel, dominant power ratio on the post-meal gastric channel, coupling percent and dominant frequency on the post-meal intestinal channel were the seven predictors reserved by logistic regression. The area under ROC curve was 0.65 in the training set and 0.63 in the validation set, both exhibited moderate predictive ability. Furthermore, by overlapping the DCA results of two data-sets, there might be clinical net benefit if 0.35 was used as reference threshold for high risk of sleep disturbance.
The model performs a worthy predictive potency for sleep disturbance, which not only provides clinical evidence for the association of gastrointestinal function with sleep disturbance, but also can be considered as an auxiliary assessment for screening sleep disturbance.