RxSem: A rule based semantic integration method for medical informatics

Obesity and its related co-morbidities are chronic diseases with significant associated healthcare cost. Although many procedures, diets and medications have been introduced to induce weight loss, Bariatric surgery has been established as the most effective therapy for morbid obesity. However, the outcome of such surgeries is unpredictable. Bariatric Surgery Information Systems (BSIS) can be used to improve the prediction of the outcome. In this paper we will present a robust rule based semantic data integration method (RxSem) for solving the data preparation and combination problems associated with Medical Data (MD) for BSIS. RxSem enables the integration of clinical data from disparate sources by resolving canonical inconsistencies and semantic non-homogeneities. We will present ReSem using an illustrative example and share our observations regarding the difficulties and stumbling points encountered in data preparation and integration in this domain. We present solutions to some of the most pressing problems. Although we concentrate on BSIS, our observations and solutions are general and apply to a vast array of health and medical information systems.