A Microservice-Based Architecture for Clinical Decision Support System for Addressing Non-Specific Musculoskeletal Disorders: The SupportPrim Project

Abstract

There is a dearth of applied interdisciplinary research on the implementation of a clinical decision support system (CDSS) that employs case-based reasoning (CBR) based on patient data from a subjective questionnaire for non-specific musculoskeletal disorders (MSDs). Furthermore, there are numerous practical obstacles to creating a CDSS system for such a complex and challenging domain, which can impede progress due to several cold-start issues. This applied research introduces a microservice-based backend architecture developed for the interdisciplinary SupportPrim project that mitigates critical real-world challenges while providing a scalable, resilient, and viable CDSS. The implemented CDSS is deployed for randomized controlled trial (RCT) in Norwegian primary care using subjective patient-reported non-specific MSD patient data. The microservice architecture provides scalability, reliability, and a user-centric approach that aligns with the dynamic nature of clinical workflows. The research is based on the applied research methodology while leveraging a design thinking approach to iteratively develop a scalable backend framework that can accommodate the idiosyncrasies of subjective patient data and facilitate the implementation of CBR systems in the CDSS. The results of a laboratory deployment of SupportPrim patient data in the Norwegian primary healthcare setting support the technical competency of the proposed architecture in finding patients with similar non-specific MSDs. Medical researchers have demonstrated the competence of the proposed architecture in the SupportPrim project through iterative experiments carried out by domain experts and the RCT within the Norwegian healthcare system, and clinicians have reported good acceptability and usability of the CDSS. This research might serve as a guide for further investigation and development of an adaptable and scalable patient-centered CDSS employing a CBR system.