Construction and validation of a risk prediction model for device-related pressure injuries in critically ill patients
- VernacularTitle:急危重症患者医疗器械相关压力性损伤风险预测模型的构建与验证
- Author:
Wei LI
1
;
Ying LIU
;
Weiping TIAN
;
Qiyan LIU
Author Information
- Publication Type:Research Article
- Keywords: critically ill patients; device-related pressure injuries; nomogram model; hemoglobin; diabetes
- From: Journal of Clinical Medicine in Practice 2024;28(24):15-19
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore the influencing factors of device-related pressure injuries (DRPI) in critically ill patients, construct a risk prediction model, and validate its effectiveness. Methods A retrospective analysis was conducted on the clinical data of 136 critically ill patients. Based on the occurrence of DRPIs, the patients were divided into occurrence group (32 patients) and non-occurrence group (104 patients). Univariate analysis and binary Logistic regression analysis were used to investigate the influencing factors of DRPIs, and a binary Logistic regression model was constructed. The Hosmer-Lemeshow test was used to assess the goodness of fit of the model, and the area (AUC) under the receiver operating characteristic (ROC) curve was used to evaluate the predictive performance of the model.Additionally, the Bootstrap method was employed for internal validation of the model. Results Univariate analysis revealed statistically significant differences between the two groups in terms of age, diabetes, ICU stay duration, duration of non-invasive ventilation mask use, Acute Physiology and Chronic Health Evaluation Ⅱ (APACHE Ⅱ) score, Braden scale score, use of vasoactive drugs, prone position ventilation, hemoglobin levels, and lactate levels (
P < 0.05). Binary Logistic regression analysis indicated that age, diabetes, ICU stay duration, APACHE Ⅱ score, duration of non-invasive ventilation mask use, use of vasoactive drugs, prone position ventilation, lactate levels, Braden scale score, and hemoglobin levels were independent influencing factors for DRPIs in critically ill patients (P < 0.05). A binary Logistic regression model for predicting the risk of DRPIs in critically ill patients was constructed based on these independent factors. The overall prediction accuracy of the model was 99.26%. The Hosmer-Lemeshow goodness-of-fit test showed that the model had good fitness (χ 2 < 0.001,P >0.999). Internal validation using the Bootstrap method and ROC curve analysis showed that the AUC of the model for predicting DRPIs in critically ill patientswas 0.952, with a sensitivity of 87.5% and a specificity of 93.3%. Conclusion The risk prediction model for DRPIs in critically ill patients constructed in this study demonstrates good stability and predictive performance. It can assist clinical healthcare professionals in screening high-risk patients and formulating personalized intervention plans.
