September 21 update: COVID-19 fatality rate continues to drop in Florida

ANALYSIS

BY JENNIFER CABRERA

This is an update of the information in our August 12August 18 , and August 26 articles on this topic. The graph above shows that even with deaths reported as of September 21 added to the data, the death rate for people who tested positive for COVID in June, July, or August was much lower than it was in April.

There was a great deal of concern about rising numbers of COVID-19 cases in Florida in June and July (and again in September with in-person classes starting), but the real question is whether these cases have translated into correspondingly large numbers of deaths; as we’ll show below, this is unlikely because people who get sick now are far less likely to suffer adverse consequences than they were in April or May.

We used Florida’s case line data (September 21 update), which does not exaggerate the number of cases because it includes one line per person. The case line data also shows whether or not that person was hospitalized or passed away (but not the date of either). Using this data allows a comparison of cases to deaths without worrying about a specific lag time for the disease process (we only look at cases through August 31 to allow for reporting lags).

The graphs below show the percentage of daily cases that later died, regardless of death date (i.e., the graph shows the percentage of people who tested positive on each date who later died). The overall rate has dropped dramatically since the start of May. The 65+ rate has been dropping since mid-April and started dropping even faster after May 27. Note that these graphs stop with people whose case dates were August 31, allowing 20 days for the cases to resolve. As you can see from the featured graph above, deaths continue to be reported for previous months (“new” deaths from July are announced almost every day), so every time we do this, the previous curves move up a bit.

This table shows that the overall average COVID-19 death rate for Florida residents has dropped by 68% from April to June, 72% from April to July, and 78% from April to August; the death rate for those over 65 has dropped by 36% from April to June, 47% from April to July, and 65% from April to August.

Note that these are case fatality rates; various studies have shown that a large number of cases are never detected, leading to an infection fatality rate that is on the order of 1/10 the case fatality rate.

All65+<65
Apr6.63%21.15%1.33%
May5.35%19.82%0.92%
June2.12%13.43%0.54%
July1.85%11.30%0.40%
Aug1.48%7.66%0.26%

The following graphs show how the case fatality rate has changed since the beginning of April for all cases, for people over 65, and for people under 65.

Below is a graph of the case fatality rate by age in Florida, followed by the same graph broken out by month to demonstrate the decreasing fatality rate.

This is where we can see the extremely low risk to college students. The overall case fatality rate for the 15-24 age group is .03% (3 of 10,000 cases). For the 18-23 age group, there have been 21 deaths out of 78,064 cases. It was difficult for young people to even get tested before June and young people are more likely to have mild illness, so the proportion of missed cases in this age group is at least as high as the number missed overall. Using the 1/10 factor leads to an estimate of about 3 deaths in 100,000 college-aged students who become infected with COVID. While every one of these 21 deaths was a tragedy, many of them were reported to have pre-existing major health issues. The fatality rate for healthy young people is close to zero.

If we remove residents of long-term care facilities, the case fatality rate for Florida residents 65 and over is much lower. Unfortunately, the Florida Department of Health has not updated their age-stratified long-term care facility death data since July 24.