VIEWER: an extensible visual analytics framework for enhancing mental healthcare

TL;DR

The paper addresses information overload in mental health EHRs by introducing VIEWER, an open-source, extensible visual-analytics toolkit that fuses structured data with unstructured clinical notes via NLP pipelines. Retrieved data from ePJS, CRIS, and CogStack are harmonized into a time-aware model and presented through Kibana dashboards to support population health, caseload management, and patient-level decision making. The study details an interdisciplinary, participatory design process, real-world deployment in SLaM, and evaluations across physical health monitoring, medication review, and caseload management, demonstrating improved efficiency and broad user adoption. The findings highlight VIEWER’s potential to enhance data accessibility and clinical decision support at the point of care, while acknowledging NLP limitations, data privacy concerns, and the need for broader generalization and formal efficacy evaluation.

Abstract

Objective: A proof-of-concept study aimed at designing and implementing VIEWER, a versatile toolkit for visual analytics of clinical data, and systematically evaluating its effectiveness across various clinical applications while gathering feedback for iterative improvements. Materials and Methods: VIEWER is an open-source and extensible toolkit that employs natural language processing and interactive visualisation techniques to facilitate the rapid design, development, and deployment of clinical information retrieval, analysis, and visualisation at the point of care. Through an iterative and collaborative participatory design approach, VIEWER was designed and implemented in one of the UK's largest NHS mental health Trusts, where its clinical utility and effectiveness were assessed using both quantitative and qualitative methods. Results: VIEWER provides interactive, problem-focused, and comprehensive views of longitudinal patient data (n=409,870) from a combination of structured clinical data and unstructured clinical notes. Despite a relatively short adoption period and users' initial unfamiliarity, VIEWER significantly improved performance and task completion speed compared to the standard clinical information system. More than 1,000 users and partners in the hospital tested and used VIEWER, reporting high satisfaction and expressed strong interest in incorporating VIEWER into their daily practice. Conclusion: VIEWER was developed to improve data accessibility and representation across various aspects of healthcare delivery, including population health management and patient monitoring. The deployment of VIEWER highlights the benefits of collaborative refinement in optimizing health informatics solutions for enhanced patient care.

Figures (8)

• Figure 1: VIEWER system architecture. VIEWER is a Web-based application built on three pre-existing systems. 1) ePJS: the EHR system of SLaM; 2) CRIS: a case register and clinical research data repository used to develop a data model by leveraging its data linkage and NLP pipelines; 3) CogStack: an open-source data retrieval and extraction platform used to inform clinical decision-making by semantic search of free text, data analytics, visualisation and risk alerting. All data and pipelines are held securely within the Trust firewall on Azure cloud servers.
• Figure 2: VIEWER data models. The physical data model illustrates physical data stores of the source EHR data, the conceptual data model describes the conceptual structure of entity classes and the logical data model describes the semantic and logical relationships among entities.
• Figure 3: Screenshots for different dashboards in VIEWER. (a) Maps of residences for all active service users of the Trust in the "Population Health" view; (b) Patterns of anti-psychotic medication use in patients with psychosis in the "Clinical Pathways" view; (c) Case complexity in a care team's caseload, stratified by duration of service use, care coordinators and consultants in the "Caseload Management" view. and (d) Anti-psychotic prescriptions over time for a patient in the "Patient Chart" view.
• Figure 4: The number of measures completed and completion rate of each measure in annual physical health check over time for patients managed within a team.
• Figure 5: Proportions of various responses to each question in the SUS survey.