COVID-19 Fuels Innovation and Advancements in… | PRA Health Sciences

Some of these tools were in place before the pandemic began. We will use many long after the pandemic ends. Read on for an overview of some key innovations from the pandemic.

Artificial Intelligence

Artificial intelligence tools are poised to help researchers sift through the ever-growing medical research collection focused on COVID-19. As of November 2020, the National Institute of Health’s COVID-19 portfolio has a record of 88,561 publications on the disease. For researchers seeking to parse this information into practicable insights, combing through the research may seem insurmountable.

One way researchers parse this data is through freely available web code on websites like GitHub. Artificial intelligence is another way. Advances in natural language processing technology allow researchers to quickly scan large amounts of data and identify relevant papers or even specific research findings.

Although artificial intelligence’s utility is still unproven in many ways, these tools allow researchers on the front line of other critical developments — for example, a COVID-19 vaccine — to access potentially game-changing data with little effort.

Search engines are yet another tool in the artificial intelligence arsenal. Google’s contribution, the COVID-19 Research Explorer, allows users to ask clinical questions and receive a list of relevant research in return, with highlighted key passages for easy perusal. COVIDScholar, an offering from Lawrence Berkeley, uses artificial intelligence to power a simple search function. The Allen Institute for AI developed SPIKE-CORD, which prioritizes information extraction using Boolean queries. Users can download that information to a spreadsheet for ease of access.

The applications for artificial intelligence are near-endless, even beyond the pandemic. The healthcare industry has already embraced artificial intelligence in many disciplines, including robotic surgery, diagnostics, and the development of precision medicine. As they refine these tools, artificial intelligence can play a role in healthcare for many years to come.

Machine Learning and App-Based Diagnostics

The identification of asymptomatic carriers has been one of the biggest challenges in the COVID-19 pandemic. Massachusetts Institute of Technology (MIT) has developed an artificial intelligence-powered model that, using forced-cough recordings, can distinguish asymptomatic COVID-19 carriers from healthy individuals.

They fed a researcher-trained machine learning model tens of thousands of samples of both coughs and spoken words. When presented with new cough recordings, the model accurately identified 98.5% of coughs from confirmed cases of COVID-19 and 100% coughs from asymptomatic people. More than 200,000 forced-cough audio samples have been collected, representing the largest research-based cough dataset.

The model’s team is currently developing a user-friendly app and seeking US Food and Drug Administration (FDA) approval. If approved, this technology represents a large-scale, free, convenient, and noninvasive screening tool that can be incorporated into the current contact tracing process.

This is not the first time that researchers have evaluated how trained algorithms can diagnose via cough: Several groups have already tested these models to diagnose respiratory conditions like pneumonia and asthma. Researchers at MIT were training neural networks to determine if forced-cough recordings could detect Alzheimer’s disease.

Remote Heart Monitoring

Remote heart care has a moment in the spotlight during COVID-19. Although remote technology for heart monitoring is nothing new — see the Apple Watch EKG functionality, launched in late 2018 — remote technology for heart monitoring from health technology companies and device manufacturers has exploded since the pandemic began.

Companies like AliveCor, iRhythm, and Medtronic have all identified the opportunity to keep at-risk patients safe while providing uninterrupted care; also while assessing patients’ response with COVID-19 to experimental medications that may impact heart health.

Wearables like Fitbit and the Apple Watch are extremely popular among consumers, but these devices are not yet sensitive enough to catch a wide range of cardiac issues. Conversely, devices from AliveCor and iRhythm have been used across clinical settings to conduct remote EKGs. Other healthcare technology companies are examining how they can deploy remote monitoring for blood oxygen levels, another key marker in COVID-19. In the longer term, we can use such tracking to identify other respiratory conditions like sleep apnea.

Post-Acute Care Technologies

The distribution of a COVID-19 vaccine does not quite represent the end of the pandemic. Officials might lift stay-at-home orders, but researchers still will have their work cut out for them, particularly in the field of post-acute COVID-19 care.

Estimates suggest that approximately 10% of those affected experience prolonged symptoms after COVID-19. Also known as “long haulers,” this population of patients requires long-term care — beyond three weeks — for this multisystemic disease. Researchers are still unclear why some patients experience this prolonged recovery, but they do know that a whole-patient perspective can often improve care.

Part of this whole-patient picture includes self-management of symptoms. To facilitate this, patients can use technology like pulse oximetry to self-monitor oxygen saturation over a prescribed period — three to five days, for example — to evaluate and reassure patients experiencing shortness of breath. The creation of “virtual COVID wards” can allow physicians to monitor a group of patients for respiratory, cardiac, and neurologic conditions over time and identify symptomatic improvement or deterioration before it happens.

Technology in the Future

Will these technologies continue to be embraced by the healthcare industry once the COVID-19 pandemic reaches its end? Many of the technologies currently being used existed in some form before the pandemic, and for many healthcare organizations, embracing technology was a top priority—the global outbreak of COVID-19 simply fast-tracked these needs. Healthcare organizations worldwide will need to balance both short- and long-term goals as the industry approaches its so-called new normal.

Hutson M. Artifical-intelligence tools aim to tame the coronavirus literature. Nature. https://www.nature.com/articles/d41586-020-01733-7. Published June 9, 2020. Accessed November 30, 2020.

National Institutes of Health Office of Portfolio Analysis. COVID-19 Portfolio. https://icite.od.nih.gov/covid19/search/#search:searchId=5fc548331288440672b58ee1. Accessed November 30, 2020.

Chu J. Artificial intelligence model detects asymptomatic Covid-19 infections through cellphone-recorded coughs. MIT News. https://news.mit.edu/2020/covid-19-cough-cellphone-detection-1029. Published October 29, 2020. Accessed November 30, 2020.

Brodwin E. Fueled by the Covid-19 pandemic, remote heart monitoring could become tech’s next big target. STAT News.https://www.statnews.com/2020/05/13/remote-heart-monitoring-coronavirus/. Published May 13, 2020. Accessed November 30, 2020.

Greenhalgh T, Knight M, A’Court C, Buston M, Husain L. Management of post-acute covid-19 in primary care. BMJ. https://www.bmj.com/content/370/bmj.m3026. Published online August 11, 2020. Accessed November 30, 2020.

Torjesen I. Covid-19: Patients to use pulse oximetry at home to spot deterioration. BMJ. https://www.bmj.com/content/371/bmj.m4151. Published October 27, 2020. Accessed November 30, 2020.

Landi H. How COVID-19 shifted healthcare executives’ technology priorities and what to expect in 2021. Fierce Healthcare. https://www.fiercehealthcare.com/tech/these-are-technology-innovations-health-systems-fast-tracked-during-covid-19-pandemic. Published October 21, 2020. Accessed November 30, 2020.