Beat Future Infectious Diseases by Crowdsourcing Strategic Data
By Rich Murr, CIO, Epicor Software
To blunt the impact of COVID-19, a vast number of medical, industrial, and financial resources are being deployed. And while these efforts are important, we should also look ahead so that we are better able to prevent future pandemics by crowdsourcing strategic, available data that would offer a bigger picture.
As someone who previously served in the United States Marines and who now serves as a CIO for a software company, I can emphatically say that regardless of your line of work, one of the best first lines of defense is information – specifically, quality information that can be speedily obtained and assessed. In the case of fighting infectious diseases, this information is critical for our epidemiologists. They need information so that we can combat potential threats before they become widespread.
It’s clear that relying solely on closed societies, public health institutions, or the media to deliver quality information is insufficient. Fortunately, the imperfect information they do offer can be challenged, enriched, and ultimately made more accurate and actionable by the creative use of information technologies.
So, what would proactive, efficient, quality information look like in action when fighting infectious disease?
Picture passengers deplaning a flight. As they walk off the plane, inexpensive IoT (Internet of things) devices are utilized to unobtrusively take their temperatures as they stroll past a checkpoint. This temperature information is then delivered instantaneously to a cloud data warehouse and combined with the temperature data of millions of other passengers from thousands of airports. When a spike in temperatures correlates to a common origin, alerts would be generated, investigations launched, and if deemed necessary, preventative measures taken. All this would occur without infringing on civil liberties. Anonymized data is all that’s required (e.g. flight departure and destination times and locations, and the passengers’ temperatures).
Also consider ways we could responsibly use the data available on social media, which is ubiquitous. Even in countries where governments seek to suppress social media and free speech, the data on these platforms is typically unfiltered. Many companies already use commercially available SaaS (Software-as-a-Service) solutions to scan social media for comments and keywords that reference their companies or products. These same solutions can easily be configured to search for keywords that suggest a potential infectious outbreak in a location. If we combine and correlate this information with passengers’ temperature data, the dataset becomes more valuable.
To further enrich the data sourced from passengers’ temperatures and social media keywords, we can look to gather information from supply chain replenishment operations. Most large retailers and pharmacies instantly know where and when a product is purchased, to include over the counter and prescription products used to treat infectious illnesses. This data can easily be anonymized and fed into the data warehouse.
Innovative healthcare companies like Kinsa are already leveraging crowdsourced data and artificial intelligence in practical yet impressive ways, publishing a US Health Weather Map that’s informed by temperature data gleaned from Kinsa’s patient thermometers. It’s easy to envision how Kinsa’s weather map could be enriched with passenger temperature data taken at airports, telemetry gathered from social media, supply chain operations, and any number of other sources.
Even contact tracing, typically a manual exercise that becomes exponentially more challenging as an infection spreads, can be made easier with technology. Companies are developing mobile apps that can allow individuals that have tested positive for an infection to publish their statuses, and for other individuals to be alerted when they come into contact with an infected individual. And these applications can be tailored for time and location. Quickly passing by an infected individual on a hiking trail may not require an alert, while sitting near them in a movie theater likely would. While this capability understandably raises serious privacy concerns, like any technology, it’s possible to anonymize the relevant data, and to periodically audit such applications and the data they send and receive to ensure privacy concerns are honored.
Similar opportunities to glean useful data will undoubtedly also emerge as the behaviors of communities and individuals exposed to COVID-19 are studied and understood. Was there a change in transportation utilization? Was there a change in foods purchased? Did they modify their entertainment choices and consumption methods? Was there a change in how they communicated with each other? Were home thermostats set higher or lower than the average? So many of these activities, behaviors, and actions are enabled through information technology today, which means they can be measured and provide an incredible, readily available source of early-warning indicators.
Even better, none of this data would need to be delivered by under-resourced and overwhelmed health departments. None of it requires healthcare providers to pivot away from patients so they can fill out additional forms. Its collection and delivery are automated, and thus less prone to adulteration, error, and delay. It’s in quantities that are statistically significant. And it can be enabled through existing robust and affordable cloud and other information technologies.
With many Fortune 500 and technology companies having long ago mastered the complexities of business intelligence (i.e., the collection and delivery of actionable information), the leadership and technical experience and expertise needed to deliver and sustain this capability for the epidemiological community exists. It’s a project with an extraordinarily high return on investment – lives saved and economic damage avoided. So let’s get started.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.