Designing a risk assessment system for cardiovascular diseases and non-alcoholic fatty liver disease in metabolic syndrome patients

Abstract

Background and objective: Metabolic syndrome is a collection of important risk factors for cardiovascular disease, including obesity, insulin resistance, diabetes, high blood pressure, and dyslipidemia. Individuals with metabolic syndrome are at increased risk for cardiovascular diseases, type 2 diabetes, non-alcoholic fatty liver disease, and ultimately premature death. Smart systems such as fuzzy expert systems can assist in early diagnosis and prevention of diseases. Methodology: The present descriptive-developmental study was conducted in two main stages. In the literature review stage, variables and rules related to each of the three categories of non-alcoholic fatty liver disease, cardiovascular disease, and metabolic syndrome were extracted from articles and guidelines such as AHA/ACC and confirmed by clinical consultants and specialists in gastroenterology, liver, and heart. In the next stage, patient information from Sheikh Al-Reis Clinic was collected based on the variables to develop the system, and descriptive rules extracted in the literature review stage were defined in the system as if-then rules. Finally, the system was written and designed using the Python programming language. The system evaluation was conducted in two stages: technical evaluation by a health informatics specialist and three specialists in liver and gastroenterology using the QUIS questionnaire, and for content evaluation, the system results were compared with physician diagnoses. Findings: The informational elements present in the system design include demographic information, laboratory test results, and blood pressure data. Thirty-nine rules for non-alcoholic fatty liver disease and forty-one rules for cardiovascular diseases were extracted from articles and guidelines. Based on entry and exit criteria, thirty-eight patients were examined. Ultimately, the system correctly identified healthy cases from patients with an accuracy close to 90%. Conclusion: Medical decision support systems, including tools for timely diagnosis and prediction of risk situations, can assist physicians. This system, serving as an interpretable decision support tool coded in Python, can be applied to both physicians and patients. With an accuracy exceeding 90%, this tool has successfully identified the risk of cardiovascular diseases and non-alcoholic fatty liver disease in patients at risk. Keywords: Non-alcoholic fatty liver, Cardiovascular disease, Fuzzy logic system