Development and internal validation of a nomogram for predicting 90- day mortality in intensive care unit patients with pressure injury: a retrospective cohort study based on the MIMIC-IV database

Abstract Background Pressure injury (PI) is a frequent and serious complication in critically ill patients and is independently associated with increased mortality; however, dedicated tools for individualized short-term prognosis prediction in intensive care unit (ICU) patients who already have PI remain scarce. Methods We retrospectively enrolled 2880 adult patients with PI on their first ICU admission from the MIMIC-IV database and randomly assigned them to a training cohort (n = 2016, 70%) and a validation cohort (n = 864, 30%). Least absolute shrinkage and selection operator (LASSO) Cox regression followed by backward elimination was used for variable selection, and a multivariable Cox model was constructed and visualized as a nomogram. Discrimination was assessed by the Harrell concordance index (C-index) and the area under the receiver operating characteristic curve (AUC); calibration and clinical utility were evaluated by calibration plots and decision curve analysis (DCA). Internal validation used 1000 bootstrap resamples. Results Overall 90-day mortality was 41.3%. Six independent predictors were retained: age, body mass index (BMI), Acute Physiology and Chronic Health Evaluation (APACHE) III score, Charlson comorbidity index, red cell distribution width (RDW), and arterial partial pressure of oxygen (PaO2). The nomogram achieved a C-index of 0.699 in the training cohort and 0.690 in the validation cohort (AUC 0.738 and 0.731, respectively). After 1000 bootstrap resamples, the optimism-corrected C-index was 0.694 (95% CI 0.684–0.711), and the calibration slope and calibration-in-the-large (intercept) were 0.959 and 0.001, respectively, indicating excellent calibration. Decision curve analysis demonstrated net clinical benefit across a wide range of threshold probabilities. Compared with the generic APACHE III, SOFA, and SAPS II scores used alone, the nomogram showed superior discrimination (category-free NRI 0.588 and IDI 0.121 vs APACHE III alone in the training cohort). The associations of all six predictors remained stable at 28-, 90-, and 365-day endpoints (all P < 0.001). Conclusions We developed and internally validated a six-variable nomogram that reliably predicts 90-day mortality in ICU patients with PI, with good discrimination, calibration, and clinical utility, which may support individualized risk stratification and early intervention in this high-risk population. An accompanying free online calculator is provided to facilitate bedside use.