Work

Improving Collaborative Management in Chronic Diseases through AI-Based Decision Support Tools

Year: 2026
Type: dissertation
Abstract: Accurate and timely treatment decision-making is a core practice in medicine, but the process is complex as clinicians and patients navigate uncertainty, information overload, and cognitive biases. Collaborative decision-making (CDM) helps address these challenges by integrating clinical expertise with patient values, ultimately improving patient satisfaction, adherence to care, and health outcomes. Patient-facing decision support tools can strengthen CDM by operationalizing how evidence, clinical judgment, and patient preferences are integrated. In chronic disease, decision-making focuses on long-term management and requires ongoing collaboration. Advances in artificial intelligence (AI), from machine learning recommender systems to large language models (LLMs), offer opportunities to personalize information, reduce cognitive burden, and dynamically align decisions with patient needs. This thesis examines how AI-based decision aids enable collaboration in pediatric chronic disease management through their development and evaluation in real-world contexts. Using a multimethod, three-paper dissertation design, the research is guided by a blended conceptual framework integrating the Knowledge to Action model, User-Centered Design, and the Ottawa Decision Support Framework. Study 1 is a scoping review of 45 studies on health recommender systems (HRSs) and their application in chronic disease management. Most systems aimed to promote behavior change but rarely assessed real-world use or utilized relevant frameworks, indicating a need for evaluations focused on collaboration rather than algorithms alone. Study 2 is a qualitative descriptive evaluation of an HRS integrated into the Power over Pain youth chronic pain portal. Interviews with ten participants showed that, beyond tailoring recommendations, the system supported self-reflection, social identification, and validation of lived experience, which can help sustain engagement in long-term care. Study 3 is a multimethod feasibility study of an LLM-based decision aid for juvenile idiopathic arthritis. Thirty-four patients evaluated its usefulness, and three rheumatologists and an LLM-Judge assessed the quality of its outputs. The tool was acceptable to clinicians, improved patients’ preparedness and confidence for dialogue, and produced recommendations comparable in quality to clinician assessments. These studies illustrate how AI-based decision support strengthens collaboration in pediatric chronic disease care by personalizing information, integrating values, and supporting decision-making across the care continuum, ultimately contributing to adaptive and patient-centered models of decision support.
Language: English
Field: Medicine
Open Access status: green
OpenAlex ID: W7204018996
Cites: 107
Cited by:
Related to: 10