Secure Secondary Use of Clinical Data with Cloud-based NLP Services. Towards a Highly Scalable Research...

Area of Science:

• Health Informatics
• Clinical Research Infrastructure
• Cloud Computing in Healthcare

Background:

• Secondary use of clinical data offers significant potential for research and quality improvement.
• Existing infrastructure often lacks flexibility, scalability, and privacy compliance for secondary data utilization.
• The cloud4health project addresses these limitations by proposing a novel architecture.

Purpose of the Study:

• To define a flexible, scalable, and privacy-compliant architecture for the secondary use of clinical data.
• To implement and evaluate a technical prototype of the proposed architecture.
• To demonstrate the system's utility through three distinct clinical use cases.

Main Methods:

• Developed a cloud-based architecture with components for data extraction from multiple hospital sites.
• Integrated text-mining services within a cloud-computing environment to process free-text clinical documentation.
• Enabled de-identification of data for anonymous or pseudonymous processing and uploading to a central study portal.

Main Results:

• A functional prototype based on the defined architecture has been successfully implemented.
• The prototype is currently undergoing evaluation across three clinical use cases.
• Clinical data from several hundred patients from university and private hospitals has been processed.

Conclusions:

• Cloud4health successfully demonstrated the integration of text-mining with existing structured data components for privacy-compliant secondary data use.
• The cloud-computing paradigm offers a flexible and adaptable service model, reducing the need for provider-side hardware investments.
• This approach facilitates wider adoption of secondary data analysis services by healthcare institutions.