Objective To screen the influencing factors associated with adverse clinical outcomes in hospitalized elderly patients with Alzheimer's disease (AD) using LASSO-Logistic regression analysis and to construct a nomogram prediction model.

Methods A retrospective selection of 214 hospitalized elderly patients with AD who visited the Department of Geriatric Medicine in the hospital from February 2021 to March 2023 was conducted, and clinical data of all patients were collected. Patients were divided into adverse events group (n=53) and non-adverse events group (n=161) based on the occurrence of adverse clinical outcomes. After variable screening using LASSO regression, multivariate Logistic regression analysis was performed to identify independent factors influencing adverse clinical outcomes in hospitalized elderly patients with AD. A Nomogram model for predicting adverse clinical outcomes in these patients was established based on the results of multivariate analysis. The predictive performance, calibration, and clinical utility of the Nomogram model were evaluated using the concordance index, calibration curve, and decision curve analysis (DCA). The diagnostic performance of the Nomogram model for adverse clinical outcomes in hospitalized elderly patients with AD was assessed using the receiver operating characteristic (ROC) curve and the area under the curve (AUC).

Results LASSO-Logistic regression analysis revealed that the Mini-Mental State Examination (MMSE) score was an independent protective factor against adverse clinical outcomes in hospitalized elderly patients with AD (P < 0.05), while the Charlson Comorbidity Index (CCI score), creatinine, urea, and fasting blood glucose (FBG) levels were all independent risk factors for adverse clinical outcomes in these patients (P < 0.05). The Nomogram model constructed based on the influencing factors screened by LASSO-Logistic regression analysis showed a concordance index of 0.994 (95%CI, 0.958 to 1.000) for predicting adverse clinical outcomes in hospitalized elderly patients with AD. The Hosmer-Lemeshow test results indicated χ²=1.909, P=0.984, suggesting good model fit. The DCA result demonstrated that the model had favorable threshold probabilities and net clinical benefits.

Conclusion The Nomogram model for predicting clinical outcomes in elderly inpatients with AD constructed based on LASSO-Logistic regression analysis exhibits high predictive value, and can be used to forecast the occurrence of adverse clinical outcomes in these patients.