Semantic Technologies for Data Analysis in Health Care

A fruitful application of Semantic Technologies in the field of healthcare data analysis has emerged from the collaboration between Oxford and Kaiser Permanente a US healthcare provider (HMO). US HMOs have to annually deliver measurement results on their quality of care to US authorities. One of these sets of measurements is defined in a specification called HEDIS which is infamous amongst data analysts for its complexity. Traditional solutions with either SAS-programs or SQL-queries lead to involved solutions whose maintenance and validation is difficult and binds considerable amount of resources. In this paper we present the project in which we have applied Semantic Technologies to compute the most difficult part of the HEDIS measures. We show that we arrive at a clean, structured and legible encoding of HEDIS in the rule language of the RDF-triple store RDFox. We use RDFox’s reasoning capabilities and SPARQL queries to compute and extract the results. The results of a whole Kaiser Permanente regional branch could be computed in competitive time by RDFox on readily available commodity hardware. Further development and deployment of the project results are envisaged in Kaiser Permanente.