Since the initial outbreak in December 2019 in Wuhan, China, the COVID-19 disease caused by the SARS-CoV-2 virus has become a global pandemic (proclaimed by the World Health Organization on March 11, 2020) [5]. Government agencies and health organizations such as the WHO, CDC, U.S National Institute of Allergy and Infectious Diseases, as well as countless scientists and Medical Doctors around the globe have scrambled to understand the alarming coronavirus and develop effective treatment methods along with a viable vaccine.
To accelerate research and triage endeavors, several enterprises have allocated resources for Artificial Intelligence (AI). For all intents and purposes, AI is an interdisciplinary science concerned with developing smart machines capable of performing complex tasks and exceeding the limitations of human intelligence [6]. Notable sectors include Machine Learning (ML), Natural Language Processing (NLP), Robotics, as well as applications that enable computers to analyze data-based models for pattern recognition, explanation, and prediction [3].
The AI functions above are playing numerous roles in each stage of the pandemic, specifically: data processing and analysis, containment, and treatment. Joel Arun Sursas, a Medical Doctor and Health Informatician with international experience, anticipates AI’s combative usefulness to surge in the following months and years. In this article, he reviews the application and progress of AI in relation to COVID-19 thus far.
Data Collection & Processing
Predictive analytics, a branch of advanced analytics used to predict unknown future events, includes several techniques ranging from statistics, modeling, data mining, machine learning, and AI for predictive data analysis [7]. The Canadian-based health surveillance startup, BlueDot, utilized AI to predict the COVID-19 outbreak at the end of 2019 [1, 8]. BlueDot is a proprietary software-as-a-service designed to track, locate, and conceptualize infectious disease spread [8]. It’s secret: big data. The AI software utilizes natural language processing and machine learning to scrutinize thousands of data sources every 15 minutes, 24 hours a day. In regards to COVID-19, BlueDot culled public health organization statements, population demographics, digital media, as well as global airline ticketing data. In addition to identifying the initial outbreak and triggering an alert, BlueDot also anticipated which cities would receive the highest volume of travelers from Wuhan, China. In retrospect, BlueDot successfully predicted 11 municipalities that were ultimately exposed to the coronavirus via international travelers [8].
AI data collection and processing are proving helpful in the research arena as well. The White House announced the COVID-19 Open Research Dataset to bridge the gap among treatment options [1]. The initiative encompasses a robust coalition in the tech sector (Semantic Scholar Project, Chan Zuckerberg Initiative, Georgetown’s University’s Center for Security and Emerging Technology, Kaggle, the National Library of Medicin, Microsoft Research) and uses natural language processing to evaluate thousands of scientific research papers [1]. It officially launched on March 16, and within four days, the program received over 594,000 impressions and 183 analyses [1].
The CORD-19 Explorer, a full-text search engine specifically designed to help people examine existing research efforts, aids the COVID-19 Open Research Dataset [9]. Understanding the virus, especially its invasive “spike proteins,” is crucial to developing therapies as well as a vaccine. The easier it is for the medical community to pinpoint relevant material, the sooner healthcare providers can save lives.
Mitigating Spread
The 2015 Zika-virus proved that AI modeling could track and predict how a disease will spread over time and space [3]. Monitoring and predicting the spread of viruses enables governments as well as public health authorities to better plan, prepare, and handle pandemic scenarios. Several institutions, including Carnegie Mellon University and the Robert Koch Institute, are building upon previous algorithms and retraining AI programs with COVID-19 datasets. If health officials can accurately forecast the virus’s movement within a targeted geographical region, they can work in conjunction with local government authorities to develop and enact policies to dampen the virus’s spread. COVID-19 task forces, as well as laypersons, can access visual representations of the tracking and forecasting data via dashboards to view the magnitude of the ongoing crisis.
AI can also be leveraged with tech in high-traffic public areas to curb the spread of the virus. Baidu, a Chinese technology company, implemented a non-contact infrared sensor system to single out feverish individuals within a crowd [4]. It was recently deployed to Beijing’s Qinghe railway station to alleviate a cumbersome screening system that relied upon human intervention.
Treatment
Image-based medical diagnosis and robotics are two of the innovative ways that hospitals are adapting treatment due to COVID-19. The former entrusts AI to analyze X-rays or Computed Tomography (CT scans) to diagnose patients at a rapid rate with high accuracy. The latter’s effectiveness is being explored in a variety of capacities. Doctors at the Providence Regional Medical Center operated a robot with a stethoscope to take an infected patient’s vitals and communicate with the patient via a video screen [1]. Similarly, robots developed by MIT’S Boston Dynamics may soon be of use in COVID clinics and inpatient wards at Brigham and Women’s Hospital and at Massachusetts General Hospital [4[. Their agility and preciseness make the AI machines good candidates to administer basic care (record vital signs, medication delivery) and significantly reduce the risk of human transmission.
Progress within the scope of this catastrophic event is measured one patient at a time. However, a consortium of supercomputers, laboratories, and medical professionals on the ground caring for the infected is building momentum in the right direction. According to Dave Turek, Vice President of Technical Computing at IBM Cognitive Systems, researchers are submitting treatment proposals to supercomputers that, in turn, analyze their viability to treat COVID-19 [2]. So, while the operational systems of healthcare facilities around the world are stretched to their limits, artificial intelligence is helping combat the virus around the clock, even when humans cannot.
About Joel Arun Sursas:
Joel Arun Sursas is a team leader and facilitator with a proven track record and a niche skill-set developed over the past seven years in his capacity as an established Medical Doctor and Health Informatician. He is most passionate about Medical Informatics, working to bridge the gap between doctors and engineers to improve patient care. His interest in the field emerged when he began working as a Project Officer for PACES — the Patient Care Enhancement System for Singapore Armed Forces (SAF).
References:
- GeekWire, How AI is helping scientists in the fight against COVID-19, from robots to predicting the future, April 8, 2020, Retrieved from https://www.geekwire.com/2020/ai-helping-scientists-fight-covid-19-robots-predicting-future/
- Spectrum News, Researchers Use Artificial Intelligence to Fight Coronavirus, April 8, 2020, Retrieved from https://spectrumlocalnews.com/nc/charlotte/news/2020/04/08/researchers-use-artificial-intelligence-to-fight-covid-19
- Towards Data Science, Artificial Intelligence against COVID-19: An Early Review, April 1, 2020, Retrieved from https://towardsdatascience.com/artificial-intelligence-against-covid-19-an-early-review-92a8360edaba
- Harvard Business Review, How Hospitals Are Using AI to Battle Covid-19, April 03, 2020, Retrieved from https://hbr.org/2020/04/how-hospitals-are-using-ai-to-battle-covid-19
- World Health Organization, WHO Director-General’s opening remarks at the media briefing on COVID-19, March 11, 2020, Retrieved from https://www.who.int/dg/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-covid-19—11-march-2020
- Builtin, Artificial Intelligence, 2020, Retrieved from https://builtin.com/artificial-intelligence
- Pat Research, What is Predictive Analytics, 2020, Retrieved from https://www.predictiveanalyticstoday.com/what-is-predictive-analytics/
- Make It, How this Canadian start-up spotted coronavirus before everyone else knew about it, March 3, 2020, Retrieved from https://www.cnbc.com/2020/03/03/bluedot-used-artificial-intelligence-to-predict-coronavirus-spread.html
- Allen Institute for AI, CORD-19 Explorer, 2020, Retrieved from https://cord-19.apps.allenai.org/