How to make sense of COVID-19 health indicators


There are so many numbers from state health offices and the federal government. Making sense of them can be a confusing and overwhelming experience.

So we’ve created a guide on what each number reported by the Tennessee Department of Public Health means. We have also included the limits of each metric. We do this because no metric or raw number can fully capture such a complicated situation.

Number of cases

  • This is the raw number of lab-confirmed cases during the entire pandemic. It is a measure of the extent of a disease, but it is incomplete for several reasons.
    • Confirmed cases are almost always underestimated from the true number of cases, especially when testing is limited and targeted to the sickest.
    • With COVID-19, this is especially true, given that it can take up to 14 days to show symptoms. The number of raw cases will always lag behind the actual cases.
    • The increase in confirmed cases does not necessarily mean that more people are infected, it just means that new cases are detected.
    • The increase in confirmed cases is lower than the actual number of cases due to delays in finding medical care, supplies and testing capacity.
    • Daily confirmed cases are “noisy”, meaning they won’t be consistent day to day, but monitoring them is useful for finding the top of the “curve”.
    • The raw number of cases says nothing about who is most vulnerable to infection. To do this, you need to break down the workload by age, race/ethnicity, or other demographic information.
    • The number of cases without a test number is an incomplete measurement. 20 positives out of 100 tests means something very different from 20 positives out of 100,000.

Number of deaths


  • This is the raw number of deaths among confirmed cases. This can give you an idea of ​​the lethality of a disease but, again, it doesn’t tell the whole story.
    • Death totals only count people with confirmed cases of COVID-19, which are undercounted. Some people could have died untested, at home without seeking medical care. Some may have died of COVID-19, but their deaths have been attributed to other causes.
    • Death totals could include people with a confirmed case of COVID-19 who died of other causes.
    • Crude death totals don’t tell you which segments of the population are most vulnerable to the disease. To do this, you need to disaggregate the percentages of deaths from COVID-19 by demographic factors such as age, race/ethnicity, and gender.



  • This can give you a rough estimate of how many people are seriously ill from an illness. This is important to consider. When epidemiologists and public health officials talk about “flattening the curve,” they’re talking about spreading serious illnesses out over time so hospitals aren’t overwhelmed. In Tennessee, the state’s hospitalization rate is about 10%. The TNDPH notes that hospitalization information is collected at the time of diagnosis, which may exclude people who were hospitalized later.

Recovered or Recoveries

  • It is another way of estimating the severity and lethality of a disease. Not all states collect this number in the same way. Tennessee defines recovered people as people confirmed to be asymptomatic by their local health department after completing a period of mandatory isolation or people who are asymptomatic 21 days after testing positive.
    • It is important not to extrapolate too much from recoveries. It’s unclear whether people develop lasting immunity or suffer long-term complications from COVID-19.

Laboratory tests

  • This is a total of tests performed by or reported to the state health department. This is an important number because it gives you an idea of ​​where confirmed cases are coming from. A low number of laboratory tests may mean that the total number of cases is underestimated. The test numbers don’t tell the whole story either.
    • If lab tests are not random, they are subject to sampling bias, which can impact case counts.
    • Testing only symptomatic and at-risk people will inflate the rate of positive tests, the death rate, and miss mild cases.
    • Organizing driving tests during working hours, in more affluent neighborhoods or far from public transport will take the tests away from poorer populations.
    • Tests performed only in areas with hospital coverage will miss rural populations.

You may have noticed that all raw metrics have limitations. This is one of the reasons epidemiologists try not to rely on raw daily numbers. They plot things like the growth of new cases over time as percentages to try to see if the rate of new cases is going down.

  • One metric you might see is something called a “test positivity rate.” This is used in cases where there are not many tests available to get an idea of ​​the number of missed cases. This is the percentage of tests that come back positive. Test positivity rates of around 20% are considered high. The United States has a test positivity rate of 20%. New York has one at around 41%. Tennessee has a positivity rate of about 7%.
  • That may sound good, but testing and COVID-19 are unevenly distributed. Davidson County has a rate of about 18%, which means there are likely many cases that are going undetected.
  • Knox County has a rate of around 5%, which is encouraging, but the county has an admitted sampling bias in favor of whiter, wealthier people. That’s why they’re doing a test drive in East Knoxville this weekend. It’s unclear how precisely sampling bias influences the positivity rate, but bad sampling means bad data all around.


Comments are closed.