Modelling of clinical risk groups (CRGs) classification using FAM

• Main Authors: Mohd. Asi, Salina, Saad, Puteh
• Format: Conference or Workshop Item
• Language: English
• Published: 2006
• Subjects: QA76 Computer software
• Online Access: http://eprints.utm.my/id/eprint/10683/1/PutehSaad2006_ModellingofClinicalRiskGroupsClassification.pdf; http://eprints.utm.my/id/eprint/10683/; http://eprints.utm.my/10683/1/PutehSaad2006_ModellingofClinicalRiskGroupsClassification.pdf

Clinical Risk Groups (CRGs) are a clinical model in which each individual is assigned to a single mutually exclusive risk group which relates the historical clinical and demographic characteristics of individuals to the amount and type of resources that individual will consume in the future [1]. CRGs based risk adjustment system is a potential risks adjustment to be used in the capitation-based payment system, a budgetary system for healthcare resource and care management [I. 2. 3]. The purpose of CRGs is to provide a conceptual and operational means through diagnosis and procedure code information routinely available from claims and encounter records. Basically, CRGs classifies patient population by presents of chronic health condition, type of chronic health condition, severity of chronic health condition and presence of significant acute health condition. Fuzzy ARTMAP (FAM) is an incremental supervised learning of recognition neural networks in response to input and target pattern [4, 5]. FAM is a fast learning algorithm and used less epoch training [4]. Based on its performance in doing the classification, FAM is theoretically suitable to do the CRGs classification. This paper views CRGs clinical logic and the data elements focus on identification of CRGs features using FAM. Previous studies (in USA and Canada) used claimed base data such Medicare, Medicaid and private insurance provider data for few years back. Some of the material use in this paper is based on research proposal titlcd, "Development Of Clinical Risk Groups Based Intelligent System For Future Prediction Of Health Care Utilization And Resources" by UKM CRGs researchcrs and KUKUM AI Embedded researchers.