During the coronavirus epidemic, it’s easy to become overwhelmed by data — confirmed cases, deaths per million, case-fatality rates, etc. They are imprecise numbers that are often misquoted or quoted out of context. This, in turn, generates political pronouncements, scientific observations and media punditry that is frequently wrong, but rarely uncertain. Obtaining accurate information can be like drinking from a fire hose. So much information, what to believe? Here are eight tips on how to consider the numbers related to the coronavirus pandemic:
All models are wrong. Some are useful. Models are mathematical descriptions of the real world used for calculations and predictions. They depend on assumptions and the numbers entered. Every model is an imperfect tool, some more than others. This is evident from the wide variance in predictions of COVID-19 deaths, from several million to the current figure nearing 60,000, which we will surpass shortly. Politicians use models to formulate policy, but models are always incomplete and wise policymakers must understand this. No single model should ever be accepted as the final word.
Long-range projections are typically less accurate than short-range projections. It’s easier to predict tomorrow’s weather than it is to predict next week’s. Conditions are always changing, and forecasters must account for uncertainty, make assumptions and anticipate unknowable future events. In the coronavirus pandemic different countries use different models and develop different strategies. Great Britain changed its policy midstream when scientists revised their models.
Numbers are a representation of reality, not reality itself. The point of numbers is to help understand the reality of what’s happening. A philosopher once said, “the map is not the territory,” meaning there is a difference between a description of something and the thing itself. A good example of using numbers but ignoring reality was how some people describe COVID-19 as not much worse than the flu, based on the number of deaths from both. The comparison is inapt; flu deaths are a roughly defined estimate per season derived from multiple projections. In contrast, COVID-19 deaths have actually been observed over a period of weeks, and most deaths are clearly attributable to the virus (though not all). Only rarely has seasonal flu forced the creation of temporary field hospitals and morgues. The coronavirus is clearly not the routine seasonal flu.
Numbers require context. Adjust numbers to the size of the population. California illustrates how case numbers and deaths should be normalized to the size of the population. California is fifth in the number of total cases in the United States, but 30th in cases per capita. The state remains a mystery as to why it has a dramatically lower caseload and death total than expected, normalized for its size.
There is no perfect number. No number is exact; every number is subject to the limitations of measurement. The number of cases of COVID-19 is far from precise — it is an obvious undercount because not everyone has been tested. But it is a useful number for establishing trends. Likewise, some death figures are overcounts, others are undercounts and no one can say which predominates or by how much. But the measured number is a reliable, uniform outcome and helps us understand these trends, which is our ultimate goal. Rather than automatically discarding imperfect numbers as some advocate, try to understand their imperfections when looking at them.
Trends are more important than single values. The coronavirus has proved unpredictable — what was true yesterday may not be true today. It is tempting to pick out a single number and draw conclusions, but a single value may prove to be an outlier. To observe trends, information is more accurate when it is collected over time. A single value is a snapshot, many values over time are a movie.
Beware of cherry-picking. We are often treated to articles with headlines such as “What the U.S. can learn from some other country” or “What one state got right.” No two regions are completely comparable — every region has a different population, climate, policy and a different situation. It’s tempting to select some place to illustrate a point that does not hold up under scrutiny. Canada is widely praised for its approach to the pandemic, but the Canadian profile of cases and deaths is similar to that of South Carolina, which has been excoriated for its approach. Similar numbers, different interpretation.
Known knowns. Then-Secretary of Defense Donald Rumsfeld explained the uncertainty of complex situations when he said of the Iraq War, “As we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns — the ones we don’t know we don’t know.”
Among the known unknowns in the uncertainty of the coronavirus epidemic are how the virus is transmitted and especially how it is transmitted by asymptomatic carriers; whether recovery means immunity to the virus; whether viral transmission will be deterred by warm weather; whether antivirals currently being tested will be effective; whether a timely vaccine can be developed; and how many people in the United States actually carry the virus. The unknown unknowns? There are many, and we don’t even know what they are.
There are pitfalls everywhere in determining what’s happening. When reading about this pandemic, don’t be discouraged. No one, anywhere, has all the answers.
———
ABOUT THE WRITER
Cory Franklin is a Wilmette, Ill., physician and author of the book “The Doctor Will See You Now.”
———
©2020 Chicago Tribune
Visit the Chicago Tribune at www.chicagotribune.com
Distributed by Tribune Content Agency, LLC.