An intelligent community-based system for healthcare prioritisation

Healthcare rationing is unavoidable in systems constrained by limited resources. While decisions about who should be treated are ethically complex, they must reflect not only efficiency concerns but also socially accepted values. This study aims to develop a multi-criteria decision-support system - Vital Priority System, that prioritise patients using a Random Forest algorithm trained on multiple rationing criteria endorsed by Portuguese civil society. Based on a Portuguese online survey data, the model incorporates nine dimensions: clinical need, life expectancy gain, quality of life improvement, age, waiting time, parental status, lifestyle responsibility, and social role. Our results show that clinical need, expected treatment effectiveness, waiting time and age were the most influential, followed by parental status. Lifestyle and social role factors were least weighted. The proposed system enables the classification of patients as 'priority' or 'non-priority', providing healthcare professionals with a transparent, consistent, and ethically grounded tool to support decision-making. This study advances the literature by operationalising, for the first time in the Portuguese context, public preferences in a replicable AI-based framework for fairer patient prioritisation.