Konsolidierte Datenmodellierung von Versorgungsdaten mit dem Entity-Attribute-Value-Modell und Data Vault

Abstract

Distributed and heterogeneous data must be integrated for health services research in a way, which is open to new requirements and easily expandable for new data sources. For data integration in the health services research domain, mainly data warehouses have been used that model data either as a dimensional or an entity-attribute-value (EAV) model. However, these data models are either not flexible enough or lack data management capabilities, which makes longitudinal data analyses more difficult. We have extended the EAV approach with data vault modelling and hereby modelled the data structures of the hospital quality reports by the Gemeinsamer Bundesauschuss (G-BA) and integrated data from the years 2011 to 2015 accordingly. This makes it possible to historicise metadata of features, in particular those of quality indicators, and establishes a high degree of extensibility towards new heterogeneous data sources. The proposed approach allows a free selection of the abstraction level for the entities to be modelled, so that a completely generic EAV model with historicised metadata can be created.

Keywords

• health services research
• data warehouse
• data collection
• common data elements
• information storage and retrieval