I created this page for my own use as Illinois has weekly trends that make it hard to tell the overall trend. By doing a 7 day moving average, I think it is easier to see where we are.
Update 6/24/20 – I’ve added a new post on tracking COVID cases in Illinois, as tracking deaths will not show us trends as quickly as tracking cases. This was a good graph to tell when we were peaking and starting to show a downward trend, as case counts were tough to interpret based on the changes in testing rates at that time. Now that we clearly peaked and were trending down, I’m concerned that just tracking deaths will not notify us early enough that things are getting worse. I know the state is using the measure of percentage of positive tests, which I do think is a reasonable indication of how widespread the disease is in any particular area when testing numbers were somewhat limited. We now do have widely available testing, so I think case counts are now the most reasonable measure to try to identify cases trending upward again.
Update 6/2/20 – there was a moderate uptick in the 7 day moving average and I anticipate it will come right back down tomorrow. I believe this is an artifact of Memorial day reporting many of the weekend cases on Wednesday last week instead of the usual spike on Tuesdays (like today). If the moving average doesn’t go down much tomorrow, that would be worrisome.
5/24/2020 – you can see it seems we just started to come down from our peak but had a slight increase the last day or two. Hopefully the trend will continue downward next week as things move into Stage 3 in Illinois. A return to increasing would mean we are reversing course too quickly. Based on looking at a 7 day moving average, it looks to me like we hit our peak on May 13, 2020 with a moving average of 116.9 deaths/day.
A few comments. While morbid, tracking deaths seems to be the best way we currently have to tell the prevalence of the infection. Cases are a poorer estimate of our illness prevalence for several reasons. First, we do not test everyone, but we are much more likely to identify COVID-19 in people who have died. Second, the test is still not perfectly sensitive. The Abbot test is thought to detect only 85% of those actually positive for the disease. Other tests seem to detect upwards of 95% of cases, but the detection seems best when it is done soon after symptoms start. If the test is done more than five days after symptoms start, the detection rate may drop dramatically. We have had a major increase in documented cases over the last few weeks, but that makes sense just from the large expansion of testing. Illinois is using the same logic I had used in my post where I started tracking the percentage of positive test as one indicator of how well we are doing with managing the disease. At what was possible the worst time in Illinois, there were around 25% of all tests coming back positive, and that was when we were only testing those who were at high risk (travel, exposure, severe symptoms). We weren’t testing most people who were sick at that time. If only 25% of the high risk people were positive, it is likely we had a fairly high percentage of people in the community who were negative.
While our percent positive has gone down, it is partially due to the large increase in testing. I think that is another reason the measure is a good one to use. More testing, even if most are negative, is a good thing. The sooner we catch positive cases, the sooner we can isolate those people and track any contacts they had.