Intelligent Patient Management in Viral Diseases: An Integrated Regression Model and Multi-Criteria Decision-Making...

Area of Science:

• Public Health
• Biomedical Informatics
• Health Services Research

Background:

• Viral diseases pose significant global health risks, especially during outbreaks, necessitating efficient resource allocation.
• Convalescent Plasma (CP) transfusion shows promise but lacks integrated frameworks for patient prioritization and donor selection.
• Current allocation methods are fragmented, hindering rapid and fair distribution of therapeutic resources.

Purpose of the Study:

• To propose a converged smart framework for data-driven convalescent plasma allocation.
• To integrate multicriteria decision-making (MCDM) and regression-based validation within a telemedicine environment.
• To enable transparent and equitable distribution of CP during public health emergencies.

Main Methods:

• Utilized Analytic Hierarchy Process (AHP) to weight five key clinical biomarkers.
• Employed Order Preference by Similarity to Ideal Solution (TOPSIS) and Višekriterijumsko Kompromisno Rangiranje (VIKOR) with Group Decision-Making (GDM) for patient and donor prioritization.
• Implemented regression-based model selection for robust prioritization and validated using an independent dataset.

Main Results:

• The external GDM AHP-VIKOR model achieved high predictive performance (R² = 0.971, MSE = 0.0010).
• Internal consistency and correlation analysis confirmed the reliability and stability of biomarker rankings.
• The framework demonstrated accuracy, interpretability, and timeliness in patient-donor matching.

Conclusions:

• The developed framework offers an accurate, interpretable, and timely solution for convalescent plasma allocation.
• This study represents a foundational step towards AI-enabled transfusion systems in telemedicine.
• The framework supports transparent and equitable CP distribution, crucial for future outbreak management.