Prediction of risk factors of sleep disturbance in patients undergoing total hip arthroplasty

The purpose of this study was to assess sleep quality in patients undergoing total hip arthroplasty (THA) from preoperatively to 12 weeks postoperatively and to establish a risk predictor for postoperative sleep disturbance to enable early care and intervention. A self-designed data collection form was used. Patients were assessed preoperatively and at 5 postoperative time points using visual analog scale (VAS) for pain, sleep quality and neuropsychological status with the following assessment tools: the Chinese versions of the Pittsburgh Sleep Quality Index (CPSQI), the Epworth Sleepiness Scale (CESS), the Zung Self-Rating Anxiety Scale (ZSAS) and the Epidemiological Studies Depression Scale (CESD). Univariate and multivariate logistic regression analysis was used for the identification of risk factors for postoperative sleep disturbance. The receiver operating characteristic (ROC) curve was plotted to evaluate the regression model. Of the 290 eligible patients, 193 (133 women) were included in the study. There was a 60.6% prevalence of preoperative sleep disturbance. The CPSQI score increased significantly at 2 weeks postoperatively compared to preoperative baseline, but appeared to decrease at 4 weeks postoperatively. Multivariate logistic regression analysis showed that pain (VAS score: OR = 1.202 [95% CI = 1.002–1.446, P < 0.05]), daytime sleepiness (CESS score: OR = 1.134 [95% CI = 1.015–1.267, P < 0.05]) and anxiety (ZSAS score: OR = 1.396 [95% CI = 1.184–1.645, P < 0.001]) were risk factors associated with postoperative sleep disturbance at 2 weeks. The ROC curve showed that the AUC was 0.762, the sensitivity was 83.19% and the specificity was 64.86%. Postoperative sleep disturbance is highly prevalent in the first 2 weeks after THA. The risk prediction model constructed according to the above factors has good discriminant ability for the risk prediction of sleep disturbance after THA. The use of this risk prediction model can improve the recognition of patients and medical providers and has good ability to guide clinical nursing observation and early screening of sleep disturbance after THA.