Data Mining approaches for Bio-Demographic and Morbidity analysis

Abstract

This work addressed, in a structured manner, the challenge of mining a healthcare database, demonstrating that, even with a simple approach, it is possible to extract relevant clinical information about the population under study. Thus, the main objective was to characterise a population from a biodemographic perspective and morbidity profile (TOP 10 most prevalent pathologies), by municipality of residence. The research question was: What are the biodemographic characteristics and morbidity profile of an adult population that sought care from the Lusíadas Health Group? The target population of this study were adult users (≥ 18 years old) who sought care from the Lusíadas Health Group (Braga, Porto, Lisbon, Almada, Amadora, Sacavém, Vila Nova de Gaia, Albufeira, and Faro). A cross-sectional observational study was conducted based on a database of electronic medical records. The data collection enabled the analysis and characterisation of the population that attended at least one scheduled or unscheduled consultation between 1 August 2022 and 1 August 2023. Focusing on the extraction of relevant clinical information, the studied group's most recorded diagnosis was Lumbago, followed by Depression and Lower Respiratory Tract Infections, completing the Top 3. Focusing on the main pathology of this group, Lumbago, the areas of the country where individuals in this studied group reside were presented. These areas were delimited by the first four digits of the postal code, without the risk of individual identification. Additionally, through K-means clustering, it is shown that it is possible to obtain relevant information from this database to develop an action plan for multidisciplinary teams, with areas of operation limited to a specific area, without compromising ethical or data protection requirements. In a national context where value-based healthcare is gaining more prominence, it is hoped that this study can serve as a starting point for the use of data by healthcare professionals to develop well-justified strategies and health plans appropriate to the communities in which the target populations are located. The analysis aims to contribute to the continuous improvement of healthcare, highlighting the importance of applying data mining techniques to complex databases to obtain significant insights.