Modelo de Asignación Diagnóstica y Terapéutica (MADiT; diagnostic and therapeutic assignment model): Clinical classification and 1-year mortality among older adults admitted to an acute geriatric unit: A prospective cohort study.

Background: Older adults admitted to acute geriatric units frequently have multidimensional geriatric vulnerability, including frailty, functional and cognitive impairment, multimorbidity, and nutritional risk. The Modelo de Asignación Diagnóstica y Terapéutica (MADiT; Diagnostic and Therapeutic Assignment Model) is a pragmatic clinical-care classification framework used in routine practice to organise pre-admission multidimensional vulnerability into four ordered categories. Objectives: To examine the association between MADiT and 1-year all-cause mortality in older adults admitted to an acute geriatric unit. Design: Single-centre prospective observational cohort study. Setting and participants: Consecutive patients aged 75 years or older admitted for acute medical illness to an acute geriatric unit between January 1, 2023, and December 31, 2024. One index admission per patient was analysed. Measurements: MADiT classification was assigned at admission according to pre-admission functional, cognitive, and clinical status and categorized as green, yellow, orange, or red. The primary outcome was 1-year all-cause mortality from the index admission. Kaplan-Meier methods and Cox proportional hazards models were used. Results: The final cohort included 430 patients, with a mean age of 87.9 ± 6.8 years; 59.4% were women. One-year mortality was 33.3% and increased across MADiT categories: 17.2% in green, 25.2% in yellow, 36.9% in orange, and 43.3% in red. Compared with the green category, the orange (adjusted HR 2.63; 95% CI 1.06-6.56) and red (adjusted HR 3.40; 95% CI 1.33-8.68) categories were associated with higher 1-year mortality after adjustment for age and sex. Adding MADiT to a prespecified base model including age, sex, and the Charlson Comorbidity Index improved model fit (likelihood-ratio χ²=11.96; df=3; p = 0.0075), reduced the Akaike information criterion from 1508.06 to 1502.10, and increased the C-statistic from 0.560 to 0.605. Confidence intervals for the C-statistics and formal calibration measures were not estimated. No statistically significant associations were observed for readmission or in-hospital complications. Conclusions: In very old patients admitted to an acute geriatric unit, baseline MADiT category was associated with a graded increase in 1-year mortality. These findings support its potential use as a pragmatic descriptive framework for baseline multidimensional geriatric vulnerability in this setting, but do not establish full psychometric validity or clinical utility. External validation is warranted.