The predictive value of three risk score scales for postoperative pulmonary complications in critically ill patients: A retrospective cohort study

Abstract

Abstract Background ICU patients are critically ill and complex, with a high risk of PPCs. An effective risk prediction scale can help healthcare professionals quickly identify patients at high risk of PPCs and guide them to take targeted preventive measures as early as possible. Objectives To compare the predictive value of the risk assessment tools LAS VEGAS, Assessment of Respiratory Risk in Catalan Surgical Patients (ARISCAT) and Chinese Brief Predictive Risk Index (CHI-BPRI) of Postoperative Pulmonary Complications (PPCs) regarding the risk of critically ill patients. Methods A retrospective cohort study was conducted with 495 patients transferred to the intensive care unit (ICU) after surgery between January 2020 and December 2020 who were scored using each of the three scales, and the predictive value of the scales was compared by the area under the receiver operating characteristic curve (AUC). Results (1) LAS VEGAS and ARISCAT could be used to predict PPCs; (2) the scores in the group of patients with confirmed PPCs (29 (24,35), 58 (44,72)) were greater than those in the group of patients without PPCs (25.5 (21,30), 52 (38,68)) (P<0.05). (3) In the receiver operating characteristic curve (ROC) analysis, the AUC values were 0.625 (0.567, 0.683) for the LAS VEGAS score, 0.598 (0.537, 0.658) for the ARISCAT score, and 0.547 (0.483, 0.610) for the CHI-BPRI score, with AUC differences of 0.0269, 0.0781, and 0.0513, respectively, which were not statistically significant (P > 0.05). Conclusion The three scales for assessing the occurrence of PPCs in critically ill Chinese patients showed average performance and poor predictive accuracy. In the next study, the characteristics of critically ill Chinese patients can be combined to better explore the development of a predictive model regarding risk factors for PPCs.