Multicriteria decision support system for triage and ethical allocation of scarce resources to COVID-19 patients

• Published in: Multimedia tools and applications Vol. 83; no. 9; pp. 27463 - 27480
• Main Authors: Chandra, Tej Bahadur, Singh, Bikesh Kumar
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.03.2024; Springer Nature B.V
• Subjects: Computer aided decision processes; Computer Communication Networks; Computer Science; Coronaviruses; Correlation coefficients; COVID-19; Data Structures and Information Theory; Decision support systems; Diagnosis; Ethics; Health care; Health services; Multimedia Information Systems; Multiple criterion; Patients; Signs and symptoms; Special Purpose and Application-Based Systems; Subject specialists; Track 2: Medical Applications of Multimedia; Viral diseases; COVID-19; Prioritization; Decision Support System; Ethics; Scarce resources; Multicriteria
• ISSN: 1573-7721; 1380-7501; 1573-7721
• DOI: 10.1007/s11042-023-16617-x

Mitigating the rapid surge of Coronavirus disease (COVID-19) is one of the challenging tasks for the healthcare industry. While offering adequate healthcare services to the best of their ability, scarce medical resources like medicines, ICU beds, ventilators, test kits, personal protective equipment (PPE), domain experts, etc., forks an additional ethics dispute. To help with difficult triage decisions, developing appropriate triage protocols and rationing resources is of vital importance. In this paper, we proposed a multicriteria decision support system (MDSS) that performs weighted aggregation of different associated symptoms, clinical and radiological findings. The model assists physicians to priorities patients based on disease severity. In this study, 20 commonly used symptomatological, clinical and radiological variables were considered in addition to computer-aided diagnosis (CAD) system’s decision. Subsequently, the robustness of the proposed method is evaluated using a private dataset and compared with the results of subjective evaluation by domain experts. The obtained experimental results with positive correlation coefficient r  = 0.9554 (between MDSS rank and ground-truth rank) and r  = 0.8622 (between MDSS rank and computer-aided diagnosis (CAD) based rank) at 95% confidence interval confirm the strong agreement between proposed method and domain expert. The proposed system could be useful in low resource settings, specifically in pandemic situations and could also be updated to prioritize resources in completely new scenarios.