Leveraging Artificial Intelligence for Pandemic Preparedness and Response

Ava Alsubai, MA, BA
Department of Medicine
Poster Overview

The coronavirus (COVID-19) is having a big impact on people’s health and on society. Preparing for future diseases that spread around the globe is important. One field that might help with this is Artificial Intelligence (AI). We looked at published studies and new studies that haven’t been published yet to learn how AI could be applied to help with pandemics. We found a few areas where AI could play a useful role. The first is to predict how the disease will spread and results from government policies. It can also help with tracking new diseases. Another use for AI is to help people follow public health recommendations such as wearing a mask and to detect flu-like symptoms in real time. We also found two ways it could be used in the hospital or clinic: helping doctors to diagnose COVID-19 quickly and predict who will have severe disease.

The examples of uses came from studies on COVID-19, H1N1 flu, Severe Acute Respiratory Syndrome (SARS), and hypothetical pandemics.

Scientific Abstract

Background: Coronavirus disease 2019 (COVID-19) has had significant impacts on society and highlighted the importance of preparedness for future pandemics. Artificial intelligence (AI) could be used to inform clinical and policy decision-making. The objective of this scoping review was to identify key uses for AI in pandemic preparedness and response.

Methods: Relevant studies were located by searching five peer-reviewed databases (PubMed, Embase, Web of Science, IEEE Xplore, and ACM Guide to Computing Literature) and two preprint servers (medRxiv and bioRxiv). A structured Google search was conducted to identify relevant grey literature. AI applications were synthesized and categorized by use case.

Results: Two hundred and four references met the inclusion criteria and reported on the use of AI for COVID-19 (78%), 2009 H1N1 influenza (10%), Severe Acute Respiratory Syndrome (SARS) (7%), and hypothetical pandemics (5%). We identified six key use cases: (1) forecasting; (2) surveillance; (3) monitoring adherence to public health recommendations; (4) real-time detection of influenza-like illness; (5) triage and diagnosis; and (6) prognosis of illness.

Discussion: AI-based solutions have been developed in response to COVID-19, but few have been optimized for practical application. These findings can support policymakers and other stakeholders in prioritizing operation of AI for future pandemics.

Clinical Implications
The results of this study can help hospital administrators and clinicians make informed decisions on how to effectively employ AI for use in the clinical setting. Most specifically, they focus on triage, diagnostics, and prognosis.
Research Areas
Ania Syrowatka, PhD; Masha Kuznetsova, MPH; Ava Alsubai, MA; Adam L. Beckman, BS; Kelly Jean Thomas Craig, PhD; Jianying Hu, PhD; Gretchen Purcell Jackson, MD, PhD; Kyu Rhee, MD, MPP; David W. Bates, MD, MSc.
Principal Investigator
David Bates, M.D., MSc

Explore Other Posters

One reply on “Ava Alsubai, MA, BA”

On your poster, it said you data (Lung images) are based on EMR … Did your work (data mining part) is based exclusively on the EMR at BWH or can your work extended to the data from EMR at other facilities & other EMR vendors?
I know this is a bit too early, but is your finding going to be published or has been submitted to be published? If it is, what is the name of the paper and the journal the paper will be in?

Comments are closed.