Intelligent Patient Management in Viral Diseases: An Integrated Regression Model and Multi-Criteria Decision-Making Approach to Convalescent Plasma Transfusion

<b>Background/Objectives:</b> Viral diseases remain a major threat to global public health, particularly during outbreaks when limited therapeutic resources must be rapidly and fairly distributed to large populations. Although Convalescent Plasma (CP) transfusion has shown clinical promise, existing allocation frameworks treat patient prioritization, donor selection, and validation as separate processes. This study proposes a credible, converged smart framework integrating multicriteria decision-making (MCDM) and regression-based validation within a telemedicine environment to enable transparent, data-driven CP allocation. <b>Methods:</b> The proposed framework consists of three stages: (i) Analytic Hierarchy Process (AHP) for weighting five clinically relevant biomarkers, (ii) dual prioritization of patients and donors using Order Preference by Similarity to Ideal Solution (TOPSIS) and Višekriterijumsko Kompromisno Rangiranje (VIKOR) with Group Decision-Making (GDM), and (iii) regression-based model selection to identify the most robust prioritization model. An external dataset of 80 patients and 80 donors was used for independent validation. <b>Results:</b> The external GDM AHP-VIKOR prediction model demonstrated strong predictive performance and internal consistency, with R² = 0.971, MSE = 0.0010, RMSE = 0.032, and MAE = 0.025. Correlation analysis confirmed biomarker behavior consistency and stability in ranking, thereby reinforcing the reliability of the prioritization outcomes. <b>Conclusions:</b> The proposed patient-donor matching framework is accurate, interpretable, and timely. This work presents an initial step toward realizing safe AI-enabled transfusion systems within telemedicine, supporting transparent and equitable CP allocation in future outbreak settings.