COVID-19 Diagnostics SB

The fight against COVID-19 needs an urgent lesson in data management from the finance and retail sectors (Simon Brightman,

The way to get back to a “new normal”, our national economies going, workers back to work and our kids back to school is with a massive user-contributed data strategy and simple, iterative execution plan

Author’s note: This article was written on April 6th, a few days before the now widely shared announcements by Google Apple about their collaboration (see “Apple and Google partner on COVID-19 contact tracing technology”, April 10th, 2020 on


The global fight against COVID-19 has claimed a tremendous toll in lives as well as economic well being. Sadly, as of this writing, an end is not yet in sight. The thesis of this article is that a massive, distributed individual-level data gathering system is needed in order to proactively detect symptoms, diagnose the virus as well as develop immunizations and therapeutics with greater success. In addition, such an early warning system would allow local, regional and national governments to far better deploy critical resources, ultimately saving lives. The solution lies in the individual-level collection via distributed sensory technology that uses existing and new hardware and software including phones and wearables, governed by a robust infrastructure to ensure the legal, compliant and proper use of this sensitive data to achieve important goals in the current fight against COVID-19 as well as future fights which we will undoubtedly face. Such a system can be launched in weeks and needs urgent collaboration from government, industry, medical and other key stakeholders.


I have never had the opportunity to work in the key sector of healthcare. As a data and analytics product professional and an academic, I’ve focused primarily on data-driven decision support products for financial services, retail, travel, loyalty and enterprise marketing. These are dynamic, challenging markets without question, but as we find ourselves deep in the throes of COVID-19, they appear less urgent than the fight which the healthcare system, health care professionals and supporting staff are battling.

The recent epidemic has given me an opportunity to reflect on many of the similarities between financial (and similar markets) and healthcare industries from a data management perspective. What can I and other data and analytics professionals, deeply embedded in what are deemed “non-essential” markets, bring forward to the Healthcare system to assist them in their current battle as well as possible future epidemics? And…how can we use these lessons to get our economies rolling again and life back to “normal”?

The Approach

The answer lies, in part, in enhanced data management, specifically diagnostic analytical capabilities. In the financial services and retail industries, for example, we have a strong appreciation for massive data collection from consumer banking or purchase transactional activity.

Every one of us generates hundreds or thousands of transactions monthly through the use of our credit and debit cards in stores and online. These massive sources of data provide banks, lenders, retailers and others with the opportunity to use this data, when permitted, to better understand client profiles, past and current behaviors. They can then predict their future actions, known as propensity models, allowing firms to optimize acquisition and product pricing to detect fraud and help ensure loyalty.

What would be the equivalent massive transactional data asset Healthcare professionals could use to better understand their patients/clients?

Unfortunately, there isn’t one and the lack of such a resource is, in my opinion, the very basis of the challenges the world is now facing as it attempts to play catchup in a fight against a deadly global virus which it is yet impossible to inoculate against, extremely hard to detect at its early stages, and successfully treat.

A Proposal

However, data alone is insufficient to achieve better outcomes. What we need is enhanced end to end data management practices that define the goals one is trying to achieve, the metrics we are seeking to evaluate, and as a result, the types of data assets required. For example, in the loyalty space, a goal may be to detect the loyalty of existing customers in order to mitigate abandonment. As a result, there is a need to measure loyalty through analytical metrics, achieved in part through the acquisition of customer transactional data from credit card companies (as most purchases in North America are via cards). Without such an approach, companies are relegated to handling abandoned clients when they cancel accounts or simply fail to renew, a much harder but common feat.

Within the Healthcare domain, the goal is the ability to detect COVID-19 symptoms at an early stage before the patient becomes infectious and or ill, thereby reducing the chances of spreading the infection to others, decreasing the burden on the healthcare systems, and mitigating personal and financial losses to individuals and families. Achieving this goal demands an early detection mechanism able to collect data and establish health profiles on individuals on a continuous basis, measure these over time, and detect trends as well as anomalies.

The underlying data may be in the form of basic health measures such as heart rate, oxygen levels, temperature, perspiration, breathing rate, sleeping patterns, degree and speed of mobility. As an analytics professional, I anticipate that data scientists in the Healthcare industry would be most interested in such data in order to develop analytical models that aim to correlate patterns in these data attributes to outcomes, in this case, patients with and without COVID-19. Those that exhibit clear symptoms such as severe shortness of breath, fever or cough are already symptomatic, and testing them is important from the perspective of administering care and identification of contacts, yet it does not contribute to curbing the spread of the virus (similar to the bank manager who contacts clients after they switch their mortgage to competitors). The focus, however, must also be on those who are asymptomatic and may be carrying the virus, and consequently equally contagious to others.

This is not to say that any one of these attributes or composite metrics would certainly work, but rather to illustrate that we must begin with a widespread, goal-based collection of foundational health data in order to develop and validate models that achieve the goals we set forth. Where though, would such data originate?

One proposal may be to focus on widespread screening using COVID-19 diagnostic kits. This would, in essence, cut straight to the outcome — simply test everyone. While desirable, this is not feasible for several reasons. The first challenge is that the tests are taken and sent to specific laboratories with varying processing times. The kits need to be manufactured, are expensive, require trained staff to administer, and a period of time to provide a result. As a result, they are currently administered only to those displaying specific symptoms and/or in high-risk situations, which hardly achieves the goal of early detection.

Case in point, Iceland, which has a very small population, has tested 5% of their 350,000 citizens as reported by CNN on April 1st, 2020 (this is considered a tremendously high percentage that only such a small nation could test with existing technology) and found that 50% of those that tested positive were asymptomatic. The CDC announced on April 4th (published in Politico) that “CDC begins blood tests to find undetected coronavirus cases” in an effort to learn more about the true proliferation of the disease.

In addition, unless tests are administered continuously over a period of time, a person tested on Monday, as an example, could be infected on Tuesday, and while the original test will be negative, it is not reflective of the patient’s current situation, resulting in a risk to themselves and others. There is an urgent need for a more affordable, widespread, self-administered and ongoing method to collect this data.

Furthermore, the current reality whereby global economies, markets and normal patterns of daily life are nearly 100% shut down is not sustainable. Eventually, there will be a need to “release” populations back into society and economy, and while we hope this occurs with a far lower risk of illness, such a solution is currently not foreseen, nor is a mechanism to address resurgences and future occurrences. Even if life were to resume to a new normal during the course of 2020, what would happen if a mutation of this virus or a new offender rears its head in 2021?

The type of early warning analytical diagnostic solution described is intended to be deployed and used on a continuous basis as a seamless part of everyday life, to help detect threats and risks at an early stage and prevent mass casualties and market crippling events such as the ones we are currently experiencing. See the following NYT from April 6, 2020 “Lockdown Can’t Last Forever. Here’s How to Lift It”, which reinforced the need for a massive, distributed data-gathering infrastructure.


The answer lies in the use of mobile and wearable or other smart devices (current and future hardware and software products), essentially turning every individual on the planet with a cell phone and/or a wearable device into a data gathering node as part of a massive, distributed data gathering and healthcare monitoring system. Several recent articles have been written in relation to this topic, including “This company claims its smart thermometer could help detect coronavirus hot spots faster than the CDC”, by Mollie Bloudoff-Indelicato on, April 2nd, 2020, “Could wearables like Apple Watch, Fitbit fitness trackers help detect coronavirus?” by Mike Feibus on April 3rd, 2020, in USA Today, as well as “Estimote launches wearables for workplace-level contact tracing for COVID-19”, by Darrell Etherington on April 2, 2020, published in TechCrunch.

What this would allow for is the daily or continuous measurements of the aforementioned attributes (and others), with little to no effort from the “client”. Geo-location data would clearly augment this data by adding in the precise location and travel patterns to these attributes. All of this information would need to be very carefully gathered from these individuals in a highly secured manner, analyzed by trusted parties and used only for very specific purposes. This is not to overlook the inherent risks and challenges, primarily with respect to data governance, security, and privacy, as well as costs of hardware and maintenance. All these challenges can be addressed, and while not insignificant, they should appear quite manageable in the face of the current global crisis we are experiencing.

Governance could be via distributed or federalized approaches, mandatory or optional. The benefits from this program grow exponentially with the rate of participation given the network effect and the reality whereby most individuals do not remain isolated during the course of normal days. A single household would benefit greatly from such remote diagnostics, and a neighborhood or entire city, state or region even more. As a non-invasive and easy to adopt data collection mechanism underlying what could be critical diagnostics and treatments, a forced-enrollment or participation model may be suitable, for example, as an integral component of healthcare systems and eligibility for care, or for privileges such as a drivers’ or business licensing.

While it is proposed that certain attributes be made available to a central, government-mandated analytical center for early detection of critical illnesses such as COVID-19, additional innovative service models would evolve from the availability of this data. Using an Open Banking model, service providers can provide early detection as well as ongoing monitoring of other health ailments, possibly for a fee. For example, the early onset of diabetes, heart disease or cancers could be monitored, complementing existing medical practices that collect and analyze data far less frequently and almost never proactively, but rather only after certain advanced symptoms occur. New software, hardware innovations will arise to enable as well as to make use of this data in a responsible way, extending the benefits across the economy.


While such a solution may not appear to solve the immediate crisis at hand, this is only partially accurate. The gathering of essential data could provide information essential to the development of new tools for diagnosis, therapeutics for treatment and help curb the spread of the virus itself through early detection and monitoring tools. The setup of such a system on local and national bases may appear daunting but could be deployed in a matter of weeks to an initial large segment of the population, with ongoing rollouts to follow.

Companies such as Apple, Google, and Microsoft, as well as Amazon, are obvious choices to help lead this initiative, although there are hundreds of smaller firms that could assume the leadership of such an initiative which may start locally and rapidly expand. This would require a government mandate and collaboration with trusted scientific, medical, technology and industry partners. While imperfect, models for such a capability do exist in the financial services sector and should be used as references to assist in the urgent implementation of such a capability in the Healthcare domain.

About the Author

Simon Brightman is a recently appointed Adjunct Professor at the University of Ottawa, Telfer School of Management. His research interests include global strategy, data analytics and Artificial Intelligence-driven decision making in government, international organizations, as well as private corporations.

Simon has lectured on technology commercialization, innovation, and product development for over 12 years in private, public and academic organizations, has served in various executive roles leading data analytics product technology firms in the USA and Canada, including Head of Data Strategy & Open Banking at TransUnion (Credit Reporting Agency), Vice President at Panvista Analytics (a US wearable geo-location analytics firm for tradeshows and conferences), Head of Agile & Product at (leading global loyalty e-Commerce provider to Airlines & Hotels), Senior Manager at KPMG (program management), presently a Senior Partner with Global Data Insights (data analytics advisory & investments across Fintech and other xTech domains).

Simon holds a BSc. in Computer Software & Business Management, an MBA as well as a Masters in International Relations from Cambridge University. Connect with Simon on Linked in.

Leave a Reply