Just What the Doctor Ordered – Towards Design Principles for NLP-Based Systems in Healthcare

Abstract

Patient data is mainly transmitted in the form of unstructured free texts in medical documentation. Natural language processing (NLP)-based systems can help to structure and extract information from these free texts to support the work of healthcare professionals. However, the healthcare sector must meet certain information quality requirements to comply with regulations and provide optimal patient care. Therefore, we argue that a design guideline is needed to tailor NLP-based systems to the unique requirements of clinical processes and to catalyze the practical application of such systems. In this paper, we report the results of a design science research study, focusing on the requirements of NLP-based systems used by healthcare professionals. In doing so, we shed light on the needs of practitioners when working with sophisticated NLP-based systems that extract and analyze text-based information from medical documentation. By providing evaluated, testable propositions and detailed design principles, we support the practical endeavor of such systems.