Predictive model of mild neurocognitive disorder due to Alzheimer's disease in Cuban adults

Objective: Develop a predictive model for mild neurocognitive disorder (NCD) based on the risk factors for Alzheimer's disease (AD).The diagnosis of NCD associated with the presence of risk factors for AD is still incipient and requires greater diagnostic precision.Methods: A cross-sectional, analytical, observational case-control study was carried out in 100 subjects classified as mild NCD due to possible AD according to Diagnostic and Statistical Manual of Mental Disorders-5 criteria and twice controls with normal cognitive functions, aged 50-64 years and both sexes, selected by simple random sampling, during the period from January 1 to December 31, 2024.The variables that constituted prognostic factors were determined by multivariate binary logistic regression.Results: Low educational attainment, low skill level, physical inactivity, smoking, diabetes mellitus, hypertension, obesity, weight loss and COVID-19 were included in a model that explained 82.5% of the dependent variable (Nagelkerke's R 2 ); with 83% sensitivity and 95% specificity; calibration was good (Hosmer-Lemeshow test = 0.940) with an area under the receiver operating curve (AUROC) of 0.976 (95% confidence interval [CI]: 0.961-0.990).A risk score was calculated from the model, defining two categories: low risk (< 33) and high risk ( 33), with an AUROC of 0.918 (95% CI: 0.877-0.958).Conclusions: The predictive model included nine variables that were easy to determine and interpret; therefore, it can constitute a useful tool in decision-making aimed at early and probably more effective interventions.