Rethinking The Numbers

Jan 26, 2022


Miller Report for the Week of January 24th, 2022; by William Miller, MD; Chief of Staff at Adventist Health Mendocino Coast Hospital

If prior respiratory, viral pandemics are any indication, we should expect this pandemic to begin to fade over the next year. The 1918 influenza pandemic caused by H1N1 went around the world in four distinct waves, mutating at least three times, then eventually fading after two years. H1N1 is still with us today and makes an occasional resurgence, but nothing like 1918. As we consider where we are with COVID, I believe it is helpful to look at the numbers being tracked with a critical eye to understand what they really mean and how useful they are.

The information presented on most databases include the following five items: (a) total number of cases to date, (b) total number of deaths to date, (c) current hospitalizations, (d) testing positivity rates and (e) current 7-day running average of new cases. Let’s examine each.

Total number of cases to date: This is really of historical interest. Since this number will only go up, never down, it doesn’t help us know where we are in the course of a pandemic. Some day we will be able to look back and say how many people became infected which will be helpful in planning for the next pandemic. However, a more helpful indicator of where we are at is the rate of growth in this number. This is usually seen best as a graph. When the rate of growth becomes less, then we will know that we are heading out of the pandemic and when the rate of growth is zero and remains zero, then we will know that the pandemic is over.

Total number of deaths to date: Like the total number of cases, this is of historical interest, but not that useful in judging our current status. There are several problems with this number and how it is used. The first problem is that it usually isn’t given in any sort of meaningful context. Today, at the end of our second year of COVID, the worldwide number of deaths is reported at about 5.6 million. However, at this point in 1918, when the worldwide population was much smaller, the number of influenza deaths due to H1N1 is estimated to have been between 30 and 45 million. In that context, 5.6 million, while terrible, is far better than 45 million.

Another major drawback to this number is what gets counted as a “COVID death”. This issue is starting to get a lot of discussion on social media. When the pandemic first hit back in January 2020, we didn’t what to expect. Thus, the definition for reporting purposes included pretty much any death of a person who was known to be COVID positive. Similarly, how long after someone recovered from COVID should a death be considered related to COVID? We have seen many deaths from cancer, heart attack, stroke and the like get counted as COVID deaths because the patient happened to test positive around the time of death. We should be able to better refine what gets counted and what doesn’t. However, this point should not be used as an argument that COVID is not serious since the majority of the 860,000+ US deaths are still likely due to COVID with some number being deaths where it was coincidental. The importance of this distinction may be a factor in how well our government and health agencies are trusted by the public.

Current hospitalizations, including ICU admissions: This is perhaps the most important value in terms of impact as it reflects whether our healthcare systems are getting overloaded. When that happens, we start seeing compromise and even deaths in persons who need hospitalization for other, non-COVID reasons. It is also the justification for potentially ordering mandates. Like with the definition of COVID deaths, this number needs further refinement. At the beginning of the pandemic, pretty much every patient admitted with COVID was admitted because of COVID. That is no longer the case. We are now seeing people admitted for reasons unrelated to COVID, but are identified as COVID positive on admission because of our screening processes. These are still considered a “COVID hospitalization” even though the person may have no symptoms related to the virus and it is irrelevant to the admission. Thus, some work is needed to better define how we count these cases.

Testing positivity rates: This was one of the factors used to determine which tier from green to purple an area was in and then used to justify lockdowns and other protective mandates. The problem is it is most accurate if we are consistently testing a large portion of the population. Instead, we have significant variables such as the availability of testing and whether people are inclined to get tested or not.

Current 7-day running average of new cases: This is the most useful in judging where we are at in a particular surge or the pandemic in general. It is reported in terms of 100,000 persons so that communities of different sizes can be compared. One limitation is that it is heavily affected by the testing rate in the community. However, when this number starts to consistently drop, then you know that the current surge is subsiding. If it stays low over a long period, then that suggests we are approaching the end of the pandemic tunnel.

After living for two years with COVID, it seems we should be able to refine some of these numbers so that we can all have greater understanding and confidence in the information we are receiving. I hope that this wasn’t too technical and as always you can access this and previous Miller Reports at