Arkansas has a low hospitalization rate for Covid-19 compared to what we would expect from published studies. Estimates from the CDC project that 20.7-31.4% of Covid-19 cases will be hospitalized. Deaths are estimated at 1.8-3.4% of cases.

Using those estimates for the deaths in Arkansas, we get:

  • 30 (deaths) / .018 (low estimate for deaths) = 1,666 cases

or

  • 30 (deaths) / .034 high estimate for deaths) = 883 cases

Since the death number is cumulative, we use the cumulative number of known cases in Arkansas: 1,475. We are in line with this estimate – albeit closer to the low end on deaths than the high end.

If we do the same thing with hospitalizations, we get:

  • 74 (hospitalizations) / .207 (low estimate for hospitalizations) = 357 cases

or

  • 74 (hospitalizations) / .314 (high estimate for hospitalizations) = 235 cases

Since the hospitalizations are “point in time” and not cumulative, we will compare this to our active cases – not cumulative. We had 989 active cases as of 4/13 (the same time we also had 30 deaths). You can see that our number of cases is way outside the expected number based on hospitalization estimates. In fact, just 7.5 percent of our currently ill cases are hospitalized.

What’s going on? Some theories:

Arkansans have super immune systems and are less likely to get very ill than other places in the U.S. or the world. While this would be fun to believe, Arkansas is actually positioned to have a worse time than many places. Our state has the fourth-highest death rate from heart disease in the country, and the second highest death rate from diabetes. Both of these are key co-morbidities for Covid-19. And we can see from the exercises above that Arkansas does fall in line with statistics for death rates, so we are getting severely ill enough to die at the expected rates.

The hospitalization estimates are way off for everyone, not just Arkansas. Maybe? Sadly, states are reporting (or just not reporting) hospitalization differently. Here’s a great article on the problem. But until we get more guidance, the CDC estimates are the best we have. They do seem to be holding true in New York and in many places around the world.

It’s so early in the outbreak that people haven’t had time to get sick enough to need hospitalization. If this were the case, deaths should also be lagging behind.

There is a systemic problem with the reporting of Covid-19 hospitalizations. We can see from the article referenced above that there is definitely a national issue with reporting. I’ve reached out to the Arkansas Department of Health with the following questions:

  • What guidance has the Arkansas Department of Health given to hospitals on reporting Covid-19 patients?
  • At what point does a hospital consider a patient a Covid-19 patient for state reporting purposes?
  • Does a patient require a positive Covid-19 test before they are reported?
  • Are presumed positive Covid-19 patients reported?
  • How frequently are hospitals reporting Covid-19 hospitalizations?
  • How are Covid-19 hospitalizations reported – online or through some other means?

Arkansas is out-testing all of the models and thus catching more cases than predicted. This would seem unlikely, as Arkansas has a much lower testing rate by population than many of its peers.

The hospitalization rate is an incredibly important metric to identifying the breadth of the outbreak. While a low hospitalization rate is fantastic and reassuring to see, it’s also important to ensure the integrity of the number – especially when it’s a significant outlier.