Harnessing Artificial Intelligence to Fight COVID-19

Harnessing Artificial Intelligence to Fight COVID-19

By Alba Kuqi, MD, CCS, CDIP, CICA, CRCR, CCDS, CSMC

As COVID-19 continues to spread across the globe, researchers have turned to artificial intelligence (AI) to address the myriad challenges of mounting an effective pandemic response. AI has been deployed to identify effective drug treatments, reveal the structure of the virus, detect emerging outbreaks, and identify signs of infection in medical images.1

It’s impossible for humans to absorb and apply all of the literature on COVID-19, so a branch of AI—natural language processing (NLP)—is being applied to this vast data set to extract useful information about the virus. We can apply NLP to this literature data to better understand the protein structure, develop vaccinations more quickly, understand treatment options and targets, predict adverse effects, determine the dosage, and more.

How NLP Works

Natural language processors also can break apart narrative text by words or phrases and encode the words and phrases for later processing. NLP also enables narrative data to be converted into structured data elements and stored in a database. The programs then assign computer codes to the individual packets, thereby codifying the text for storage, analysis, and, later, retrieval.

One of the latest algorithms for text processing is an open-source NLP technique from Google called Bidirectional Encoder Representations from Transformers (BERT). This algorithm overcomes the limitations of prior NLP algorithms by looking at words and sentences from both directions, left to right and right to left, so that it can understand the language’s full context. Sentences are mapped to vectors or points in multi-dimensional space.

By giving more context, the vector falls closer in proximity to other vectors that convey the same or similar meaning. According to Google, “unlike these previous models, BERT is the first deeply bidirectional, unsupervised language representation, pre-trained using only a plain text corpus.”2

Other forms of textual data or social media data where people share and talk about the coronavirus are platforms such as Facebook, Google searches, Twitter feeds, and so on. The unstructured nature of textual data is what makes advanced NLP a great prospective tool to deal with them.

Identifying Potential Treatments and Vaccines

A partnership between three companies—Excientia, Calibr, and Diamond Light Source—used AI technology to discover that the virus binds to a protein on the surface of cells called Angiotensin-converting enzyme (ACE) 2.

According to Drug Target Review, “Three prioritized targets include; the 3CL protease, the NSP12-NSP7-NSP8 RNA polymerase complex (both of which are vital components for viral replication) and the virus’s Spike (S) protein, which interacts with the human cell receptor ACE2 to gain entry to human cells.”3

This information can be used by researchers first to identify opportunities in the existing set of known drugs and then to work on a new molecule to protect the human population from this invasive disease.

AI and Patient Care

The use of AI for healthcare must be balanced by the appropriate level of human clinical expertise, and it helps humans improve effectiveness and efficiency. Beyond screening, AI is being used to monitor COVID-19 symptoms, provide decision support for CT scans, and automate hospital operations.4

In the midst of the global COVID-19 pandemic, a Beijing-based robotics company sent 14 robots that could take vital signs and deliver food and medications to Wuhan, China, to help with patient care and limit the possibility of virus spread. This helped limit physician exposure to the virus and ease the workload to healthcare workers experiencing exhaustion. Doctors and nurses monitored patients’ vital signs remotely using connected thermometers and bracelet-like devices on one AI platform called HARIX (Human Augmented Robot Intelligence with eXtreme Reality). The government of South Korea released an app allowing users to self-report symptoms and alert them if they left a “quarantine zone” to curb the impact of “super-spreaders” who would otherwise go on to infect large populations.

In this epidemic, consumers and healthcare professionals can get real-time data on the spread of this virus through the media and other data sources. John Zarocostas wrote in The Lancet that “it is important that the public health community help the media to understand better what they should be looking for because the media sometimes gets ahead of the evidence.”5

Tracking Outbreaks, Fighting Misinformation

Some observers say there is also an “infodemic” or rampant spread of misinformation going on, which some are countering by offering up accurate sources of data. Johns Hopkins University created an online coronavirus dashboard to help visualize and track the reported cases on a daily timescale. They also made the complete set of data downloadable as a Google sheet. The map shows new cases, confirmed deaths, and recoveries. They collect data from the World Health Organization, the Centers for Disease Control and Prevention, local Chinese websites, and other reports that help provide more regional data than other national organizations can.

Computer-assisted coding uses NLP to review the documentation in the electronic health record (EHR) and assign a code number. Using NLP techniques, we can extract keywords from more than 100,000 articles and publications about COVID-19. Within healthcare, artificial intelligence or machine learning is more frequently called expert systems, and the process that produces knowledge from such systems is called knowledge management. Knowledge production technology is most commonly found in clinical decision support (CDS) components of EHRs, although it is increasingly being applied in other components of EHRs and other forms of health IT. As a result, NLP can serve as a tool to bridge the gap between the unfathomable amount of data generated daily on COVID-19 and the limited capacity of the human mind.

On December 30, 2019, BlueDot, a Toronto-based startup that uses a platform built around artificial intelligence, machine learning, and big data to track and predict the outbreak and spread of infectious diseases, alerted its private sector and government clients about a cluster of “unusual pneumonia” cases happening around a market in Wuhan, China.6

But AI is not a perfect predictor of outbreaks. In 2008, Google launched a flu-detection service that analyzed peoples’ search queries and tracked supermarket purchases, web browsing patterns, and private messages. Developers ultimately decided it wasn’t successfully predicting flu because it consistently overestimated the pervasiveness of the disease. In the end, AI cannot replace experienced epidemiologists, but it can serve as a tool in helping them.

Microsoft has been collaborating with adaptive biotechnologies to map the immune system and using machine learning to hopefully bring an antibody blood test to market for COVID-19. Microsoft is also using the supercomputing capacity it previously used to train deep neural networks to run molecular simulations to try to identify potential therapies for COVID-19. While Microsoft, physicians, and computer scientists are collaborating on several potential therapies, the timeline for when they can be tested safely on humans is unknown.

Notes
  1. Best, Jo. “AI and the coronavirus fight: how artificial intelligence is taking on COVID-19.” ZDNet. April 9, 2020. https://www.zdnet.com/article/ai-and-the-coronavirus-fight-how-artificial-intelligence-is-taking-on-covid-19/.
  2. Devlin, Jacob and Ming-Wei Chang. “Open Sourcing BERT: State-of-the-Art Pre-training for Natural Language Processing.” November 2, 2018. Google AI Blog. https://ai.googleblog.com/2018/11/open-sourcing-bert-state-of-art-pre.html.
  3. Rees, Victoria. “AI technology to screen existing drugs for use against COVID-19.” April 2, 2020. Drug Target Review. https://www.drugtargetreview.com/news/59188/ai-technology-to-screen-existing-drugs-for-use-against-covid-19/.
  4. Wittbold, Kelley A., Colleen Carroll, Marco Iansiti, Haipeng Mark Zhang, and Adam B. Landman. “How Hospitals Are Using AI to Battle Covid-19.” April 3, 2020. Harvard Business Review. https://hbr.org/2020/04/how-hospitals-are-using-ai-to-battle-covid-19.
  5. Zarocostas, John. “How to fight an infodemic.” February 29, 2019. The Lancet. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30461-X/fulltext.
  6. Bowles, Jerry. “How Canadian AI start-up Blue Dot spotted Coronavirus before anyone else had a clue.” March 10, 2020. Diginomica. https://diginomica.com/how-canadian-ai-start-bluedot-spotted-coronavirus-anyone-else-had-clue.

 

Alba Kuqi, MD, CCS, CDIP, CCDS, CRCR, CICA, is the CDI supervisor at Prime Healthcare. She is an ACDIS Leadership Council member, PHIMA member, and AHIMA Foundation Research Network Volunteer.

Leave a comment

Send a Comment

Your email address will not be published. Required fields are marked *