A carregar...
Publicação
The influence of socioeconomic factors on disease risk and medical costs in Malta using Machine Learning on health insurance data
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.38 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
This study investigates how socioeconomic factors, such as age, gender, relationship to the poli-cyholder, and coverage type, influence disease risk and medical costs in Malta. Using anonymized health insurance data from MAPFRE Malta, we apply supervised and unsupervised Machine Learning techniques to improve client segmentation and cost prediction. Our models outperform traditional actuarial methods (GLMs), especially in identifying high-risk profiles and forecasting elevated costs. These findings support fairer premium pricing, personalized product design, and strategic decision-making in health insurance. The study is motivated by the increasing prevalence of chronic diseases, such as diabetes and cardiovascular conditions, which place pressure on healthcare systems. In Malta, the growing demand for private health insurance highlights the need for insurers to better understand risk pat-terns and service utilization. Socioeconomic variables such as age, gender, relationship to the policyholder, and coverage type were found to be key predictors. Traditional actuarial models, such as Generalized Linear Models (GLMs), often fail to capture complex interactions between these factors. The dataset includes anonymized records of insured individuals with at least three years of med-ical history. Model performance was evaluated using RMSE, MAE, R², AUC-ROC, F1-score, and Silhouette Score. Results show that ML models outperform GLMs, especially in identifying high-risk clients and forecasting elevated medical costs. Feature importance analysis highlights premium paid, age, and insurance duration as the most influential variables. The findings support the integration of ML into insurance decision-making, enabling more equi-table pricing, better resource allocation, and the development of preventive care strategies. This approach aligns with MAPFRE Malta’s innovation goals and contributes to academic research in health economics and actuarial science.
Descrição
Tese de Mestrado, Engenharia Biomédica e Biofísica, 2026, Universidade de Lisboa, Faculdade de Ciências
Palavras-chave
Machine Learning Health insurance Socioeconomic factors Risk prediction MAPFRE Malta