Abstract
BACKGROUND: Despite the availability of several interventions, the incidence of pressure ulcers (PU) has not significantly decreased, particularly in older adults undergoing hip fracture surgery. Therefore, there is an urgent need to develop and validate a more reliable and effective predictor to enhance the prediction accuracy for PU development in this vulnerable population. METHODS: In this study, a reliable and convenient predictor of PU was developed and evaluated based on four commonly used hematological markers. The data were randomly divided into a training cohort and a validation cohort in a 7:3 ratio. The strength of the association between each predictor and the occurrence of PU was assessed using multivariate logistic regression analysis and propensity score matching (PSM). For predictors with significant associations, the receiver operating characteristic (ROC) curve and its parameters were further applied to select the best predictive model. The model was subsequently validated by a systematic evaluation of its discriminative ability, correlation, and predictive performance. Additionally, threshold analysis, subgroup analysis, and further exploration of the relationship between the GAR indicator and length of hospitalization (LOS) was conducted. RESULTS: A total of 1,279 older adults undergoing hip fracture surgery were included in this study, with 156 (12.2%) developing PU postoperatively. Multivariate logistic regression and PSM analyses revealed a nonlinear positive correlation between GAR and postoperative PU (OR = 1.84, 95% CI: 1.44-2.35). The area under the ROC curve (AUC) for GAR was 0.72, indicating moderate predictive ability. Furthermore, each 0.1-unit increase in preoperative GAR was associated with an approximately 0.17-day increase in the length of hospitalization. CONCLUSION: Preoperative GAR levels are a moderate predictor of the risk of postoperative PU and LOS in older adults with hip fractures.