Background: Coronavirus disease 2019 (COVID-19) caused an unprecedented healthcare crisis and warranted a need to use artificial intelligence (AI) and machine learning (ML) for enhancing caller screening and triage within pre-hospital Emergency Medical Services (EMS) specifically tailored to COVID-19 cases. This study aimed to analyze existing AI and ML models and assess their accuracy and precision.
Methods: A comprehensive assessment of artificial intelligence (AI) applications used to improve EMS responses in the context of COVID-19 instances was done. The dataset produced by Mexican government was used. This dataset was assessed over different models encompassing logistic regression, random forest, gradient boosting, neural networks, k-nearest neighbors (KNN), Naive Bayes, and clustering (K-means).
Results: Multiple models performance evaluation was done employing metrics such as accuracy, precision, recall, and F1-score to comprehensively assess the strengths and limitations of these models.
Conclusion: The study's findings underline the complexities inherent in caller screening and triage for COVID-19 cases, showcasing diverse strengths and limitations within the deployed machine learning models. The discourse underscores the necessity for a multifaceted approach to effectively manage the intricate challenges associated with caller classification and triage, offering invaluable insights for future research endeavors and guiding the enhancement of emergency healthcare systems.
Key words: COVID-19, EMS, machine learning, caller screening, healthcare management.
|
