Tech-based evaluation of healthcare quality during the COVID-19 pandemic

The coronavirus disease 2019 pandemic profoundly disrupted global health systems, exacerbating preexisting inequalities in access to and quality of care. Disparities in service accessibility, continuity, coordination, and comprehensiveness highlight the need for a multidimensional, data-driven approach to healthcare evaluation. We conducted a retrospective analysis using 52,490 responses from the 2021 Global Burden of Disease COVID-19 Health Service Disruption Survey. We developed a 4-dimensional evaluation framework aligned with World Health Organization principles for people-centered care. Following variable selection, standardized preprocessing, and manual labeling, we trained and assessed 6 machine learning (ML) algorithms. We further optimized the 2 best-performing models - support vector machine (SVM) and random forest (RF) - using interpolation-based data augmentation, cross-validated hyperparameter tuning, and particle swarm optimization. SVM showed the highest baseline performance (accuracy 0.93), whereas RF achieved competitive macro-F1 scores. Interpolation-based data augmentation improved the generalizability of the SVM model, increasing its accuracy to 0.96. The RF model benefited most from particle swarm optimization-based hyperparameter optimization. Model predictions revealed persistent inequalities in healthcare service quality: respondents with lower income, lower educational attainment, and rural residence were disproportionately classified as receiving lower-quality services, whereas urban residence and higher income were associated with higher-quality care. Our scalable ML framework successfully assessed healthcare service quality during the pandemic, with optimized SVM and RF models demonstrating robust performance on high-dimensional, imbalanced survey data. Post-pandemic recovery strategies must prioritize socioeconomically disadvantaged populations. Furthermore, integrating ML tools can enhance real-time quality monitoring within health systems.