How COVID-19 is Changing our Relationship with Data
An increasing proportion of businesses use scientific methods to analyze data. Yet, because key decision-makers do not believe in a data-driven method, a sizable number do not turn their data into actionable information. Instead, these individuals rely on their instincts to drive decisions. If used in lieu of data,such hunch-based mindsets often lead to inefficiencies and wasted opportunities. In March of 2015, I wrote an article on the prevalence of such decision-making practices in corporate America. More broadly, I discussed how people continue to discredit the data and science that supports the claim made by climatologists: human actions are the root cause of climate change.
Five years later, we find ourselves in a world consumed by a viral pandemic and with a new angle from which to observe the same phenomenon. As involuntary participants of a natural history case study, the novel coronavirus and the insidious illnesses and fatalities caused by the resulting COVID-19 infections has given the population the opportunity to witness the power science has over a non-empirical approach. The latest survey by Quinnipiac University found that approximately 80 percent of U.S. adults back a national stay-at-home order, while less than 20 percent oppose the idea. There is an overwhelming international scientific consensus that mitigation is worth the consequential economic upheaval. Yet, even with the clear cause and effect relationship between data-driven actions and less destructive outcomes we have realized around the world, science critics still openly take issue with the data, analysis, interpretation, and most importantly, the corresponding safety actions recommended by the nation’s top healthcare experts.
As we watched the spread of infection weeks before it hit us, the coronavirus pandemic was thought to be a public health nightmare no one could have predicted. Or was it? Years before this novel strain surfaced, government strategists warned that a viral pandemic similar to the 1918 Spanish flu was inevitable. At the time I am writing this article, there are more than 3.5 million confirmed cases and over 250,000 deaths worldwide, with roughly a third of those located within the U.S. Even with those indisputable figures, both the cause of the pandemic and what would happen next was highly emotional and a matter of relentless political and social debate.
Regarding this pandemic, as whenever the stakes are high, politics and science are closely intertwined. We now have the opportunity—and quite frankly, the responsibility—to use a scientific approach to inform political decisions rather than distorting its validity to yield a different outcome. Using one’s “gut” or intuition does not have to contravene a data-driven approach. However, when the data reveals an uncomfortable truth, we must be prepared to take actions that are less than ideal. Based on the consensus of experts, we should also engage in constructive debates on how to prepare for future outbreaks using the lessons learned from this and other outbreaks.
While the origin of transmission of the novel coronavirus to humans is controversial, the first reported case came from Wuhan, China on November 17, 2019. A few weeks later, it became impossible to contain the spread of the virus and obvious to the rest of the world that the order to lockdown more than 11 million Wuhan residents came too late. The proverbial horse was already out of the barn as the viral infections quickly spread across mainland China and then, almost instantaneously, to most countries (see note 1) across the globe. The world watched the transformation of a local Chinese outbreak turn into an epidemic, and then, in just a matter of weeks, a worldwide pandemic. We obsessively watched news network reports of large and small communities alike succumbing to the stealthy, highly contagious, indiscriminate, and too-often deadly novel coronavirus.
January 21, 2020 marked the first reported case of coronavirus in the United States. However, we now know that it was more likely introduced weeks prior, circulating by asymptomatic and misdiagnosed propagators. By March, we started to witness a shrinking supply of hospital beds, ventilators, and hospital staff. We also learned to use a new acronym in casual conversation, “PPE” for “Personal Protective Equipment,” because we understood through simple predictions that we had nowhere near enough of it to protect our hospital and frontline workers. Not to mention ourselves. By the end of March, most states were on stay-at-home orders, bracing for thousands of deaths in a matter of weeks.
As local government authorities gave orders to stay home, the masses unwittingly sat through a crash course in data analysis. Concepts covered included data sufficiency, randomization in data collection and sampling, weighted results, moving averages, various biases and issues in measurement, data outliers, statistical distributions, forecasting, and margins of error. Simulations showed how the insights could be made actionable as the slope, peaks, and tails of a distribution could be altered to simulate different outcome scenarios approximately two weeks after the order had been put into effect.
Laypeople learned what it means to “flatten the curve” because they had a vested interest to do so. As the ramifications of the changing shapes of statistical distributions were contemplated, many relied on tables and graphs to help project the propagation of the coronavirus and its impact on their local communities and on their loved ones. As we have seen, fear can be a powerful motivator for learning. This contagion, with its pernicious and boundaryless global reach, has fast-tracked the learning curve for the average data savvy person and “right-brained” among us. My hope—bolstered by favorable reactions to the mitigation efforts—is that it has also converted some traditional science debunkers into believers. The unconvinced argue that the insights from the data are inaccurate, blown out of proportion, or just downright lies or scams. For most of us, however, the interpretation of an exponential slope was made crystal clear. COVID-19 cases doubling every three to four days without enough ICU beds, doctors, ventilators, test kits, and PPE was scary and easy enough to comprehend: our lives were at risk.
Not only did people want to understand the disease’s spread, recovery, and death rates, but they were also motivated to find its causes and what they could do to mitigate its impact on themselves as well as family and friends. Relevant information about how to “dodge the bullet” when the scourge hit their neck of the woods was a real-world and scalable example of how analytical insights could be made actionable. While the predictions themselves were impressive (at least to the geeks among us), action was and always will be what is most important. People recognized that changing the shape of a statistical distribution was, quite literally, a matter of life and death. Simple actions like washing hands and sneezing into an elbow gave way to more socially impactful actions like ceasing to shake hands or gathering at bars, beaches, and restaurants. Eventually, testing for COVID-19 began in earnest, and most of us were asked and then told to stay at home. Businesses deemed “non-essential” shut down within weeks of the first confirmed case in the U.S.
Because of the devastating impact this unprecedented shutdown had on the economy in the weeks that followed, many began to question whether mitigation and suppression measures “made the cure worse than the disease.” A sales executive I know questioned the reported figures: “Wouldn’t the death rate be lower if the denominator was deflated due to inadequate testing?” This is a question that any respectable analyst would consider. However, it was not expected because the sales industry is historically driven more by hunches than by data. To me, this was another sign that analytical principles were infiltrating society at a much deeper level than before the pandemic. And he is correct. We do not know precisely how many people have or have recovered from COVID-19 because the U.S. was extremely slow to begin and then ramp up testing. Not having faith in the denominator distorts the key metrics used to measure the threat: the number of confirmed cases, those who recovered, and those who died from COVID-19 infections or related complications. Even so, it is the best data available to us and not enough reason to discredit the insights we could glean from it.
Projections made by epidemiologists, virologists, and other healthcare professionals delivered answers that we so desperately sought. Some politicians not only used the same data and scientific projections to issue drastic measures in an effort to “flatten the curve,” but they also used them to educate the public on how to interpret the various metrics and charts. This was done not only to inform but also to justify and prepare the government’s future actions. Governor Andrew Cuomo of New York, for example, used data supplied by Johns Hopkins University and other sources to teach daily courses on statistics. Although intended for New York citizens, these daily briefings were eagerly watched throughout the U.S. and worldwide. His approach was simple, measured, and consistent, qualities data scientists should attempt to emulate. He warned New Yorkers (and many of us watching from elsewhere) of the horrible consequences that lay ahead. He continued to calmly interpret the information as projections solidified into new realities unfolding before our eyes. The mortality rates increased proportionally to the exponential rate of newly confirmed COVID-19 cases. Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, shared similar projections during his daily briefings with President Trump at the White House.
However, some of the evidence was not sufficient for the science skeptics, whether politically motivated or not. For example, some constituents refused to believe the veracity of the data, claiming that the U.S.’s COVID-19 death toll was artificially inflated by unrelated medical conditions. While it is always reasonable to question data and the science used for analysis and predictions, it becomes a slippery slope when the data is rejected based on non-scientific motives. Given the scale of the pandemic, it is no surprise that conspiracy theories abound. According to Dr. Fauci, “I think it falls under the category of something that’s very unfortunate – these conspiracy theories that we hear about. Any time we have a crisis of any sort there is always this popping up of conspiracy theories.” For your entertainment, you may want to see Cornell’s University’s Alliance for Science COVID: Top 10 current conspiracy theories.
Nearly every state in the union eventually issued “stay at home” orders, as did most of the world. The impact of this action clearly curbed the spread of the virus and saved millions of lives. However, economic ramifications of this caliber have not been seen since the Great Depression.
At present, we sit at a crossroads of what the data tells us and our desire to “get back to normal,” whatever “normal” looks like in the future. Many resent being confined to their homes. They want to return to work, shop, enjoy entertainment, and socialize with their friends. This is understandable. Unemployment can be debilitating, embarrassing, or even humiliating socially. Other unintended consequences of a sudden global economic shutdown are becoming obvious. It has caused and will continue to cause major supply chain disruptions, food shortages, and hunger as well as significant increases in domestic violence and deaths related to substance abuse and suicide. There are, of course, many other such “butterfly effects” related to the novel coronavirus.
Besides ignoring the impact that social distancing has had on saving lives, these same people may also be discounting the possible emergence of a new and more pernicious strain of the virus in the future that could wreak an even greater toll on the economy.
Federal guidelines recommended that states wait to begin a phased reopening until they could document 14 consecutive days of declining cases. Governors in most states simply ignored the data (or interpreted it in a way one may expect of a politician) and began to loosen restrictions in their states even as the number of confirmed cases and deaths in those states continued to rise. In the few states where restrictions remained in place, people took to the streets to protest the stay home order, some armed with assault-style weapons aggressively complaining that the orders were unconstitutional.
The economic impact to one’s livelihood due to the restrictions cannot be minimized, and it is understandable why people are frustrated with politicians who bar commerce. However, baseless repudiations of science and data should never be acceptable. Still, the blatant dismissal of facts is to be expected, even if some justifications are not. For example, as reported on NBC’s Today Show on April 29, 2020, a protestor was simultaneously parading two signs: “END THE SHUTDOWN NOW!” and “Data Over Dictators!” My question for the protester: do you have alternative data not driven by politics? Great. Please show us! As I mention in my 2015 article, “by its very nature, science is based on debunking science with better science.” Unfortunately, alternative data sources that refute the admittedly less-than-perfect coronavirus data from the World Health Organization (WHO), Johns Hopkins University and other reputable institutions either have not been shared for evaluation or, more likely, simply do not exist.
On the same episode, co-host Savannah Guthrie asked Gavin Newsom, California’s governor: “Do you feel in your gut that the worst is behind you in California?” Governor Newsom responded, “I love the way you asked that question… ‘in my gut’… I’m not sure if I’m driven by gut in this respect; I’m driven by data.” A few days later, on May 3, 2020, Governor Cuomo warned in his daily briefing (where data analysis is always a central theme), “go back to your old behavior, and the numbers are going to go up.” He added, “it is fine to use your gut as long as it is an educated gut.” He implored his constituents to rely on the insights derived from the best publicly available data, claiming, “in New York, we follow the facts, data and science to make our decisions.” Like these governors, we should follow the directives backed by science and not by dictators, especially as we face increasing uncertainties.
At the time of this writing, it is too early to know the impact of reopening too soon. According to public health officials, it is probable that such relaxations before a vaccine becomes available will result in a new and possibly larger outbreak in the future. On the other hand, since testing has been limited, we still do not know the prevalence of exposure to the novel coronavirus and our corresponding antibody response. The hope is that eventually enough people will have developed antibodies to create a desired “herd effect” that will ward off future outbreaks. As we prepare for the second and third waves of the outbreak, such scientifically grounded factors must be carefully analyzed and objectively debated.
The best answer for society may be the one that does the best job at simultaneously minimizing the human and economic impact of an impending new outbreak. In the long-term, these seemingly opposing forces may not even be at odds with one another. Regardless, we have no alternative but to rely on the best data available and scientific principles that are both inclusive and multidisciplinary. Government officials must be transparent and should not censor government institutions like the Centers for Disease Control and Prevention (CDC) (see note 2) or the National Institutes for Health (NIH). There are ways to slowly, carefully, and responsibly reopen, taking into consideration our new “normal;” it must be an orchestrated balancing act. As during the 1918 Spanish Flu pandemic, which shares eerily similar aspects to our current global scene, changes to our lives will be inconvenient and uncomfortable. But if we let it, science can help us make decisions wisely to minimize suffering. Multiple vaccine trials are underway, and agencies are hopeful that a vaccine will be available to the public by early 2021. Until then, it is encouraging to see the many creative ways people are finding to safely begin emerging back into the world, from drive-through veterinary clinics to meeting friends around the world on Zoom. After all, we must get back to living our lives. As I recently overheard a friend say, “I’m all about the science… but I just want to get my hair colored!”
1. Countries that have not reported any case of coronavirus: Kiribati, Lesotho, Marshall Islands, Micronesia, Nauru, North Korea, Palau, Samoa, Solomon Islands, Tonga, Turkmenistan, Tuvalu, Vanuatu
2. CDC barred from using ‘evidence-based,’ ‘science-based’ in reports; CDC’s 17-page draft recommendation for reopening America is rejected by President Trump
By Taymour Matin newdata.ai