Three billion people, over a third of the world’s population, are currently in lockdown because of the coronavirus. This is the gravest public health danger since the Spanish flu virus in 1918, which killed 50 million people. Fortunately, Artificial Intelligence and Machine Learning are catalyzing the convergence of technologies, including genetics, imaging, diagnostics, sensors and cloud computing, helping to usher in a new era of data-driven and digitized healthcare.
Out of times of crisis comes innovation. COVID-19 is forcing the US Healthcare sector to rapidly evolve to combat the pandemic. Let’s look at a few ways in which Machine Learning is at the forefront of the efforts to diagnose, treat, and prevent the coronavirus.
Companies are working hard to find a test that provides accurate results quickly, and can also be produced at speed, but we aren’t there yet. Currently, testing for the coronavirus is based on looking for its genetic sequence in a specialized lab.
Gene sequencing is being democratized thanks to machine learning and cloud computing. For instance, Illumina can sequence genes 5X faster than the previous generation of genetic sequencing during the SARS outbreak. The cost of sequencing continues to decline, from about $750K during SARS to just about $1K today. It is becoming easier than ever to organize an ever-larger database of genomes for researchers to study and to track how the virus mutates. Understanding the virus’s genetic makeup gives health systems a way to understand how the virus is transmitted and facilitate the development of accurate testing.
One of the ways in which data analytics can help defeat COVID is by tracking the locations of those who test positive and using algorithms to recognize unusual patterns. As data analytics become an increasingly standardized module of healthcare, it will be even more important that health systems be able to integrate data from multiple sources into a unified platform. Machine learning software is playing a key role in enabling hospitals to undertake scenario-planning so that they can forecast where future coronavirus-related demand is likely to occur and thereby manage hospital capacity and resource constraints (personnel and supplies) better.
As we emerge from this pandemic, hospitals and health systems will probably seek to build greater automation into their infrastructure. As more monitoring devices and techniques are established, medics will need a standardized, unified view with which to triage patients. Data analytics will be embedded into these platforms in a way that helps doctors to make more informed decisions through things like automatic risk scores and predictive health warnings. In addition to reducing costs, systems like this will improve the quality of patient care and save lives.
The Telemedicine adoption curve had advanced by light-years amid the coronavirus. While most people associate Telemedicine with remote clinical consultations through video conferencing, such consultations can also be led by Artificial Intelligence. Patients can be diagnosed and counseled without having to leave their homes. With the help of AI, telemedicine can be scaled up much faster than traditional methods, thereby reducing the burden on the traditional healthcare system and freeing up resources to deal with the most acute cases.
In China, Ping An Good Doctor, has reported that daily new registrations jumped 10X amid COVID, while daily online consultations rallied 9X. Tokyo-based LINE Healthcare Corporation saw its online consultations increase 40X month-over-month. Teladoc, a US telehealth company, reported that it conducted 100,000 virtual appointments during the week of March 8th: “The demand for virtual care visits has accelerated as several health plans have waived consumer cost-sharing and public health officials at all levels of [the US] government have encouraged the public to take advantage of virtual care services. These actions have driven many people to use telemedicine for the first time, with more than half of all the Teladoc Health visits this month being from first time users.”