Health-PRIOR: An intelligent recommender system architecture for Case prioritization in healthcare

Abstract

Smart city environments offer a data flow and constant monitoring, aiming for the well-being of the population. When applied to healthcare, they improve the quality of people’s lives, making it possible, for example, the prediction of diseases and treatment monitoring. The case prioritization in medical centres is of great importance, which favours patients’ health and physicians in their daily job. Recommender systems are an alternative to automatically integrate the data generated in these environments with predictive models and recommend actions, content or services that can benefit patients in their context. The main goal of this research work is to assist patients and doctors in the early diagnosis of diseases or in the detection of worsening in postoperative cases through constant monitoring. To achieve this goal, this research work proposes an architecture for recommender systems applied to healthcare, that can prioritize emergency cases. The architecture brings an ensemble approach for prediction, which adopts multiple Machine Learning algorithms. The methodology for carrying out the work follows three steps: (I) conducting a systematic mapping, seeking to identify the gaps present in the state of the art in the context of health recommender systems; (II) construction and development of the architecture; (III) evaluation through case studies, aiming to evaluate specific parts of the proposal, its completeness and adherence in medical contexts. The adoption of the architecture is justified since the data produced by smart devices are accurate and reliable for predictive and decision-making contexts. The results obtained in each case study showed the feasibility of the proposal, where, for accurate data sets with a low amount of noise or missing values, the predictions are promising and adherent to the application context.