Abstract

The digitalization of healthcare has created abundant and rich health-related data. To exploit the wealth of information in these healthcare data, modern applications often need to support rich queries that access heterogeneous data from diverse sources. This raises a number of data management challenges on data placement, data integration, and data querying. In this paper, we demonstrate how to address these challenges using an example healthcare application, which helps physicians match drugs against patient conditions. Three datasets are collected and placed into three disparate stores: a relational database, a text search engine, and a graph database. Domain specific data integration methods are applied to link the different pieces of data together. And finally, a simple polystore architecture is developed to support rich queries across the different datasets stored in disparate stores.