The COVID Denominator
by Jason S. Brinkley, PhD, MS, MA
COVID-19 is here and with it comes concern and anxiety about the effects on America. Central in peoples’ mind are questions of risk. How likely am I to get it? How likely am I to die from it? There are a lot of measures related to virus epidemiology, but the most fundamental are simple rates that most everyday folks use to assess personal risk. The crude or naïve ratio of how many bad outcomes divided by how many people are affected. It’s that fraction that can often drive both social policy and personal decision making. Like all ratios, it involves a numerator and a denominator, and while most public health surveillance focuses on the numerator, it is the denominator that may yield more valuable insight.
Public health systems worldwide are designed to focus on incidence counts for monitoring outbreaks, and COVID-19 is no different. Currently, WHO provides routine reports on the current status of the pandemic with detailed country by country outlines of the number of current cases, cumulative cases, and total deaths. This information is critical to public health monitoring in terms of determining overall impact on a specific region. Number of current cases tells us the stress level to the medical system, and total deaths is self-explanatory as an important metric. But in the assessment of risk, those are only the numerator in important ratios, and the WHO reports give little information as to the total population at risk. In fact, the question of at risk is actually somewhat fluid depending on point of view. Total deaths is a fixed value, but should we compare that to the total number of individuals hospitalized? The total number tested? The total number exposed? The total population? Each of these answers gives a different denominator and a different perspective on risk.
Determining the appropriate denominator for risk estimates is actually quite hard and not something any health system is set up to do well. Mortality estimates for COVID-19 vary dramatically. This article from the Guardian sums things up quite well. (The whole article is worth a read):
“It is probably about or a bit less than 1%. Much higher figures have been flying about, but the chief medical officer, Chris Whitty, is one of those who believes it will prove to be 1% or lower. The World Health Organization’s director general, Dr Tedros Adhanom Ghebreyesus, talked of 3.4%, but his figure was calculated by dividing the number of deaths by the number of officially confirmed cases. We know there are many more mild cases that do not get to hospital and are not being counted, which would bring the mortality rate significantly down.”
Actual range of estimates seems to vary from 0.6% to 3.4%; and while that is useful on one hand (to help people understand that very few COVID-19 infected patients actually die), it is problematic on the other. Both are very small numbers, but 3.4% is more than 5 times higher than 0.6%, and with numbers that small, that degree of imprecision has the potential to play havoc on public health planning when taken to a macro level. Are we talking about 1,000 deaths or 5,000 deaths? Are we talking 1,000,000 or 5,000,000? What starts out as small difference on small fractions has a big impact when we start talking about population impact, which is where knowing the denominator is key.
So what can we do to form better denominators? For COVID-19 there may not be much we can do. Large-scale testing may help, but there has been little published on the sensitivity of COVID-19 tests, and the false positive rate may be higher than expected. In general, there is a place where each and every American can help. The US Census is the most important data collection effort in the United States as it helps establish the denominator for all of our monitoring systems, not just health. The decennial Census is mandated by law and requires years of extensive planning to get right. In fact, COVID-19 has the potential to threaten Census data efforts, and just this week the Census department released an update on its planning in the face of the virus.
The Census will continue, but for 2020 there is an added feature. In addition to mailing in responses, the Census will be available on the web. Web access for the Census has the potential for many errors, but those trade-offs may pale in comparison to the ease of access for families who are facing self-isolation due to COVID-19. Completing your Census early online (or by mail) helps the US to form better denominators not only for nationwide risk estimates but also for local risk estimates. COVID-19 hits just as our current regional population data is the most dated and that hinder public health planning efforts. Post 2020 Census, we will have more accurate counts for estimating population impact. More directly, Census employs a veritable army of Census takers who are tasked with field data collection that may include door-to-door interviewing. As we work to limit exposure and prevent the spread of this and other viruses, every person who completes the Census early on is someone that the field staff doesn’t have to visit.
April 1, 2020 is Census Day, but you can get started now.
Jason S. Brinkley, PhD, MS, MA is a Senior Researcher and Biostatistician at Abt Associates Inc. where he works on a wide variety of data for health services, policy, and disparities research. He maintains a research affiliation with the North Carolina Agromedicine Institute and serves on the executive committee for the NC Chapter of the American Statistical Association and the Southeast SAS Users Group. Follow him on Twitter. [Full Bio]
Previous posts in this series:
- Halloween by the Numbers 2019
- We R-Naught Ready for an Epidemic
- How Do Machines Learn? Part 3: They Recover
- How Do Machines Learn? Part 2: They Fight
- How Do Machines Learn? Part 1: They Learn
- Taste Testing Generic Drugs
- Halloween by the Numbers
- What Kills Kids?
- The Golden Age of Health Research Funding
- Does Living on a Prayer Work?
- The Opioid Data Crisis
- Income Lost from Snow Days*
- What the #$@&*! Is Blockchain?
- Opportunistic Research Opportunities
- Text Mining UFO Data: Little Green Aliens or Santa’s Elves?
- Should You Know Your Doctor’s Home Address?
- The Population Bullet
- The Unknown Unknowns of Missing Data
- Communicating Science–More Than Just Good Words?
- Counting Alabamas
- The Third World in Your Own Backyard
- The Unrealistic Gold Standard
- Does MACRA Signal the Beginning of the End for Medicare Claims Data?
- Think You Aren’t Extraordinary? Odds Are You’re Wrong
- Mapping by Words
- Are We Asking Too Much From Surveys?
- Making Better Comparisons
- What Kills Us?