Predictive Models on Pain and Severity in FM Patients

Overview

The primary goal of this research project is to develop different prediction models in fibromyalgia disease through the application of machine learning techniques and to assess the explainability of the results. As specific objectives the research project intends: to predicting Fibromyalgia severity of patients based on clinical variables; to assess the relevance of social-psycho-demographic variables on the fibromyalgia severity of the patients; to predict the pain suffered by the patients as well as the impact of the fibromyalgia on patient's life; to categorize fibromyalgia group of patients depending on their levels of Fibromyalgia severity.

Fibromyalgia (FM) is a condition characterized by chronic musculoskeletal pain whose pathophysiology is still unclear. Furthermore, this pathology is frequently associated with sleep disturbances, pronounced fatigue, morning stiffness, poor quality of life, cognitive disturbances (mainly memory problems) and psychological problems (depression, anxiety and stress). FM is associated with greater negative affect, which implies a general state of anguish composed of aversive emotions such as sadness, fear, anger and guilt. Patients with FM commonly suffer from high rates of anxiety, depression, pain catastrophizing, and stress levels, which are associated with a worsening of symptoms, including own cognitive. Machine learning (ML) and data mining had been successfully applied, over the past few decades, to build computer-aided diagnosis (CAD) systems for diagnosing complex health issues with good accuracy and efficiency by recognizing potentially useful, original, and comprehensible patterns in health data. Thus, machine learning provides useful tools for multivariate data analysis allowing predictions based on the established models and hence offering a suitable advantage for risk assessment of many diseases including heart failure. Machine learning offers advantages not only for clinical prediction but also for feature ranking improving the interpretation of the outputs by clinical professionals. Explainable ML models, also known as interpretable ML models, allow healthcare experts to make reasonable and data-driven decisions to provide personalized treatment that can ultimately lead to high quality of service in healthcare. These models fall into eXplainable Artificial Intelligence (XAI) field, defined as suite of ML techniques that 1) produce more explainable models while maintaining a high level of learning performance, and 2) enable humans to understand, appropriately trust, and effectively manage the emerging generation of artificially intelligent partners.