Simon Fong, Dana Wang, Jinan Fiaidhi, Sabah Mohammed, Libo Chen* and Li Ling Pages 1 - 11 ( 11 )
Background: Clinical pathways are part of the diabetes therapy management which involves medical interventions versus the patient’s conditions. For instance, in insulin-dependent diabetes mellitus, insulin dosage can be injected in the right quantity, and intake timings are recorded for each individual patient, so that a suitable level of blood glucose is used for his/her body.
Objective: Despite the availability of general health guidelines, the interactions between the medicine, the lifestyles and the patient’s conditions are complex; the dosages and the pattern of therapy may differ from one patient to another.
Method: In this article, a hybrid data stream mining approach with a fusion of fuzzy unordered rule induction is proposed to computationally derive real-time decision rules as clinical pathways for the regulation of the prescription for IDDM on daily basis records, as well as the blood glucose level of the patient patterns. The decision rules are generated timely from the fresh health patterns that are continuously updated from the patient’s instead of a historical data archive of a population accumulated over years. Results: The rules are adaptive and reflect more accurately of the patient’s conditions, as the glucose levels fluctuate differently under medical effects of both long and short terms, such as changes in lifestyles, type of medications, or other external factors. The most suitable data stream algorithms for this task are evaluated by a computer simulation that is presented in this paper.
Conclusion: The patient-specific rules can be applied to a personalized diabetic advisor, which are customized to the individual lifestyle and health requirements.
Insulin-dependent diabetes mellitus; Diabetes therapy; Decision rules; Data stream mining; Fuzzy Unordered Rule Induction; Clinical pathways
Department of Computer and Information Science University of Macau, Macau SAR, Department of Computer and Information Science University of Macau, Macau SAR, Department of Computer Science Lakehead University, Thunder Bay, Department of Computer Science Lakehead University, Thunder Bay, Department of Endocrinology Guangdong Medical College Affiliated Shenzhen Nanshan Hospital Shenzhen 518052, Department of Endocrinology, Guangdong Medical College Affiliated Shenzhen Nanshan Hospital, No.89 Taoyuan Road, Nanshan District, Shenzhen 518052