A parsimonious internally validated model for predicting in-hospital mortality using predictors available within the first 24 h in acute myocardial infarction: evidence from the Yazd Cardiovascular Disease Registry

Abstract

Abstract Background Accurate prediction of in-hospital mortality in patients with acute myocardial infarction (AMI) is clinically critical. However, many existing risk models are complex, rely on variables not routinely available early during hospitalization, or have limited applicability in resource-constrained settings. Therefore, there remains a need for a clinically interpretable, parsimonious, and data-driven model with strong discrimination and calibration. Methods This retrospective registry-based study used data from the Yazd Cardiovascular Disease Registry and included hospitalized AMI patients between 2016 and 2019. The primary outcome was in-hospital mortality. A clinically informed parsimonious logistic regression model was developed using five routinely available predictors: age, fasting blood sugar, ejection fraction, sex, and Killip class. Model performance was assessed using AUC, Brier score, calibration plots, calibration intercept, and calibration slope. Internal validation was performed using bootstrap resampling with 200 repetitions. Missing data were handled using multiple imputation by chained equations (MICE) as the primary analysis, with complete-case analysis performed as a sensitivity analysis. Results A total of 1,396 AMI patients were included, of whom 74 (5.3%) died during hospitalization. In the primary multiple imputation analysis, the mean AUC across five imputed datasets was 0.893 (SD 0.007). In sensitivity complete-case analysis of the parsimonious model, 614 patients and 24 deaths were available; the model showed good discrimination with an apparent AUC of 0.863 and an optimism-corrected AUC of 0.844 after bootstrap validation. The Brier score was 0.029. Optimism-corrected calibration slope and intercept were 0.934 and −0.134, respectively, indicating mild overfitting and slight overestimation of risk. Conclusions The proposed five-variable model demonstrated good discrimination and acceptable calibration after internal validation. Given the limited number of events, substantial missingness in key predictors, and absence of external validation, the model should be considered a preliminary early in-hospital risk stratification tool requiring further validation before routine clinical implementation.