Development And Validation Of A Nomogram Model Based On Ultrasonic Images For Predicting The Sacrococcygeal Pressure Injury In Adult Critically Ill Patients

Objective:To develop a nomogram model based on ultrasound images for predicting sacral pressure injuries in adult critically ill patients and assess its predictive performance and clinical.Methods:A total of 361 adult patients admitted to the ICU of a general hospital in Xinjiang from September 2021 to December 2022 were enrolled in this study.The first 60 percent of patients admitted to ICU were selected and assigned to the model development group and the remaining 40 percent of patients were divided into the model validation group.Demographic information,treatment parameters,laboratory test results were collected.Sacral ultrasonic image(subcutaneous tissue thickness,discontinuous superficial fistula,unclear subcutaneous layer structure,subcutaneous hypoechoic injury and left ventricle ejection fraction(LVEF)were measured by a bedside ultrasound equipment.Logistic regression analysis was conducted to determine the sacrococcygeal pressure injuries-related factors and ultrasonic images to develop a nomogram model.The Braden scale and the nomogram model were conducted respectively to evaluate the risk of sacrococcygeal pressure injury.The performance of the model was assessed and validated by the area under the receiver operating characteristic curve(AUC),calibration curve,concordance index,Hosmer-Lemeshow test,decision curve and clinical impact curve.Results:1)Pressure injury occurred in 59 of 361 patients(16.3%).42 cases developed pressure injury in the model development group and 17 cases in the model validation group.2)In the model development group,there were significant differences between the patients with and without pressure injury in APACHEII,VIS,white blood cell,serum lactate level and oxygenation index,unclear subcutaneous layer structure,and LVEF(P<0.05).3)The factors associated with the occurrence of pressure injuries occurred in sacral region were APACHEII(OR=1.201),VIS(OR=1.145),serum lactate(OR=1.733),unclear layered structure(OR=22.274),and LVEF(OR=0.899).4)Predictive performance of the nomogram model:the AUCs were 0.894 in the model development group and 0.875 in the model validation group.The sensitivity and specificity were 0.899 and 0.711 in the model development group and 0.857 and 0.883 in the model validation group.The concordance index in the model development group was 0.963 in the model validation group was 0.920.Hosmer-Lemeshow test:χ~2=13.176,P=0.109.A good consistency between the observed probability and predicted probability presented in both the model development group and validation group according to the calibration curve.The nomogram model had a satisfactory clinical benefit when the risk threshold was between 0.05 to 0.95.5)Predictive performance of the Braden scale:the optimal value was 9,the AUC was 0.643and the sensitivity and specificity were 0.714 and 0.548 respectively.5)Clinical application value:the nomogram model can obtain clinical benefits when the high risk threshold ranged from 0.05 to 0.95.The Braden scale had clinical benefits when the high risk threshold ranged from 0.22 to 0.47.The clinical impact curve showed that the number of patients at high risk of pressure injuries predicted by the nomogram was closer to the number of cases actually developed pressure injuries.Conclusion:The prediction model based on ultrasound image is superior to Braden scale in predicting sacrococcygeal pressure injury.