# Artificial Intelligence-based Clinical Decision Support for COVID-19 -- Where Art Thou?

**Authors:** Mathias Unberath, Kimia Ghobadi, Scott Levin, Jeremiah Hinson, Gregory D Hager  
**arXiv:** [2006.03434](https://arxiv.org/abs/2006.03434)  
**Date:** 2020-06-05  
**Summary page:** https://www.summarizepaper.com/en/paper/2006.03434v1/

## Key points

- COVID-19 pandemic has brought unprecedented challenges in healthcare
- AI-based clinical decision support systems have shown maturity in other areas of healthcare, but their application for COVID-19 has been limited
- Opportunities and requirements for AI-based clinical decision support systems are identified to address the rapidly emergent challenges posed by COVID-19
- Need for AI readiness to overcome challenges such as data heterogeneity and bias in training datasets
- Importance of collaboration between clinicians and data scientists to ensure that AI-based tools are clinically relevant and effective in improving patient outcomes
- Suggest a framework that combines machine learning algorithms with domain knowledge from medical experts to create an AI system tailored to fit the specific needs of each hospital or clinic.

## Summary

The COVID-19 pandemic has brought about unprecedented challenges in healthcare, including new clinical questions, workflows, and the need for accelerated distributed healthcare. While artificial intelligence (AI)-based clinical decision support systems have shown maturity in other areas of healthcare, their application for COVID-19 has been limited thus far. In a perspective piece authored by Mathias Unberath, Kimia Ghobadi, Scott Levin, Jeremiah Hinson and Gregory D Hager, opportunities and requirements for AI-based clinical decision support systems are identified to address the rapidly emergent challenges posed by COVID-19. The authors highlight the need for AI readiness to overcome challenges such as data heterogeneity and bias in training datasets. They also emphasize the importance of collaboration between clinicians and data scientists to ensure that AI-based tools are clinically relevant and effective in improving patient outcomes. To this end, they suggest a framework that combines machine learning algorithms with domain knowledge from medical experts to create an AI system tailored to fit the specific needs of each hospital or clinic. Overall, this perspective piece provides insights into how AI can be leveraged to enhance clinical decision-making during the ongoing pandemic.