JMIR Publications

ABSTRACT

Background:

Chronic obstructive pulmonary disease (COPD) remains a major global health challenge. In primary care, inconsistent recording of symptom scales and lung function hinders timely risk stratification and proactive management. There is an urgent need for objective, automated tools that leverage routinely collected clinical data to identify patients with poor disease control and support equitable healthcare resource allocation.

Objective:

To validate the predictive performance of Seleida— a previously developed, bijective, deterministic model for real-time COPD control assessment and automated phenotyping—using real-world electronic health record (EHR) data, and to evaluate its applicability within routine clinical informatics workflows.

Methods:

Seleida applies deterministic analytics to two predefined, routinely collected variables: annual use of rescue inhalers [short-acting β-agonist (SABA)/ short-acting muscarinic antagonist (SAMA)] and antibiotic prescriptions for respiratory exacerbations. We implemented methods to verify the model’s bijectivity, compare two phenotyping approaches (126- and 21-combination systems), and demonstrate its feasibility as an automated screening tool for integration into healthcare informatics systems supporting COPD management.

Results:

In a real-world primary care cohort, Seleida demonstrated perfect concordance between its dual phenotyping systems (κ = 1.00, p < 0.001) and substantial agreement with real-world clinical phenotypes (κ = 0.70, p < 0.001). Its validated forward risk estimation and reverse phenotypic inference enable real-time control assessment and automated patient stratification. These results highlight Seleida’s potential as an interoperable screening tool for integration into clinical informatics workflows, supporting proactive and data-driven COPD management.

Conclusions:

This study confirms the clinical validity of Seleida as an innovative, automated, and scalable decision-support tool for COPD control assessment and phenotyping. Its ability to integrate seamlessly into routine healthcare informatics systems addresses a key gap in the early identification of patients at risk of poor disease control.

Conclusions:

Seleida offers a practical solution to enhance precision-guided and equitable COPD management. External validation and cost-effectiveness studies are ongoing to confirm its broader applicability and sustainability.