GUIDE-AI

Today, chronic lung, heart, and kidney diseases constitute a large proportion of chronic illness in Europe’s population, with nearly 85% of deaths in the EU due to chronic diseases. Guideline-directed medical therapy is evidence-based and the cornerstone of pharmacological therapy for patients with chronic illness. Yet, a large number of patients are not receiving all medications as recommended given the length and complexity of guidelines. The ambition of GUIDE-AI is to leverage generative AI to increase guideline adherence of chronically ill patients in Europe. GUIDE-AI will be developed as a blueprint on three indications: Chronic obstructive pulmonary disease, chronic heart failure, and chronic kidney disease. This setup can be expanded to other chronic diseases where clinical practice guidelines exist or be updated with the new guideline once published. GUIDE-AI has three objectives: OBJ-A[people-centricity]: GUIDE-AI integrates the perspectives of patients, physicians, healthcare administrators, and payers, addressing both inpatient and outpatient care and covering the full treatment continuum—from diagnosis and therapy initiation to long-term monitoring and treatment adjustments. OBJ-B [Innovate & Integrate]: GUIDE-AI computer scientists and clinicians will develop the technical set-up of the guideline navigators including benchmarking and red teaming on retrospective data. OBJ-C [evaluate in real world]: GUIDE-AI will evaluate the guideline navigators in a real-world evaluation study in a 2-arms randomized design. The primary outcomes are (1) the percentage of guideline-conform therapies and (2) the number of patients whose medication changed after the LLM-intervention. Taken together, GUIDE-AI is a public-private partnership uniquely positioned to bring together private partners with leading academic institutions to advance beyond existing solutions to increase guideline-directed medical therapy by employing generative AI, making the latest evidence actionable. The impact will be demonstrated across multiple diseases, sectors, countries and languages.