DIAGNOSTIC NETWORK FOR SECONDARY HYPERTENSION: INTEGRATING KNOWLEDGE, EXPERIENCE AND PREDICTIVE MODELS

Objective: Secondary hypertension (SH) accounts for up to 30% of hypertension cases yet remains underdiagnosed and almost not tackled in hypertension guidelines due to its complex and diverse etiologies, reliance on diverse clinical experience and avoidance of costly unnecessary testing. To help improve this status quo, the Diagnostic Network for Secondary Hypertension (DNeSH) combines (1) an interactive digital network of all the physiological causal pathways that can be involved in hypertension (Knowledge) with (2) a dynamic database of real-world cases (Experience) manually diagnosed by European clinicians. Then, (3) a Bayesian predictive model uses information to predict the probability of each differential diagnosis. Meanwhile, the physiological network and the statistical analysis of the database of clinical cases themselves are directly interactive, with a user-friendly interface for clinicians.Design and method: DNeSH is anchored on three pillars: - Foundational Network: A comprehensive and interactive map of all physiological pathways and causal mechanisms of SH, providing clinicians with an organized framework for exploration. - Dynamic Database: A crowdsourced repository of real-world cases manually contributed by European clinicians, offering an evolving statistical foundation for SH insights and trends. - Bayesian Predictive Model: Leveraging the foundational network and patient-specific data, the model predicts probabilities for differential diagnoses of SH. This transparent system empowers clinicians with intuitive, explainable AI to enhance decision-making. These components are unified through a user-friendly website interface, facilitating accessibility and real-time collaboration across borders. Results: The ambitious original plan was feasible and a fully working website was built (dnesh.me) and tested with medical students and interns. Initial qualitative feedback from test users indicates high usability, trustworthiness, and promise for clinical adoption. Conclusions: DNeSH epitomizes the European Society of Hypertension’s values by making the best knowledge and expertise accessible in real time. This system fosters a paradigm shift in SH management, enabling dynamic, evidence-based, and personalized medicine. By integrating advanced AI with clinician-friendly transparency, DNeSH transcends traditional expert systems, presenting a scalable model for diagnosing SH and other complex conditions. This ’European digital brain’ sets a new standard for clinical decision support, education, and research collaboration.