An enhanced entity-attribute-value data model for representing high dimensional and sparse healthcare data

Healthcare data are naturally sparse, high dimensional, and require frequent schema change during storage. To overcome the above challenges, both simple and multi-data-type Entity-Attribute-Value (EAV) data storage models are often used in place of the traditional horizontal data model. However, representing data in either of the formats generally result into queries that are both inefficient and rather complex. In this paper we present an enhanced data model for representing this class of data which dramatically improves the efficiency and reduces the complexity of the queries thereof. This is achieved by finding a compromise between the conventional horizontal data model and the EAV data models. To attest the validity of the new crossbreed model, Enhanced Entity-Attribute-Value (EEAV), its performance is evaluated against that of conventional horizontal data model, simple EAV data model and multi-data-type EAV data model with regard to querying of data sets of varying sizes.