Towards a Sustainable DWH Approach for Evidence-Based Healthcare
By Nevena Stolba .
This work has been finished in September 2007.
The healthcare industry is one of the world’s largest, fastest developing and most information-rich industries. Rapid growth of information technologies has brought immense opportunities for patient data sharing, development and dissemination of evidence-based medical knowledge and analysis across distributed, heterogeneous healthcare data sources.
In contrast to other industries, where data warehouses have been successfully applied, healthcare is an area in which the information technology had only been able to permeate the administrative and logistic aspects of information processing. The growing need for integrated healthcare has led this industry to open towards adoption of extensive clinical decision support systems.
Evidence-based medicine (EBM) offers a collection of proven best practise guidelines for recommending drugs and medical treatments. This thesis states that the way data warehouse (DWH) technology can facilitate EBM is twofold: (1) by supporting the rule development process, and (2) by providing the EBM-enriched knowledge base to support the decision-making process of the care givers.
With respect to (1), we explain that data warehousing and data mining support the creation of the evidence-based rules by providing a platform and tools for knowledge discovery and pattern recognition. Large amounts of data can be analysed to confirm known or discover unknown trends and correlations in data.
Regarding (2), we argue that the care giving process can benefit significantly from the application of DWH technology at the point of care. Given an integrated knowledge base, built upon broad variety of patient-related information sources and incorporating evidence-based rules, our approach offers a unique decision support for the practitioners in their every-day work.
In order to guarantee the confidentiality of patient data in today´s increasingly information-based, multi-site health delivery environment, this thesis recommends a federated DWH approach instead of collecting data from remote sources into a centralized system. We endorse the application of de-personalisation, pseudonymization and role-based access mechanism for protection of sensitive healthcare data.
This dissertation is intended to provide a roadmap for achieving sustainable healthcare decision support system based on federated data warehouses, facilitating evidence-based medicine that safeguards patient´s personal privacy. It postulates four rules to follow when building a modern medical decision support system and we hope that its advisory nature will prove to be helpful in designing future healthcare projects.
Download the paper