10 August 2020 Blog Post: Quantifying the Effect of Case Reporting Delays - COVID-19 in Los Angeles County

When logging into the Los Angeles County Public Health COVID-19 website, the following disclaimer is presented in bright yellow highlighted text:
“DISCLAIMER: Public Health anticipates receiving a backlog of cases once the State electronic laboratory system issues are fixed. Data sources that track other key indicators, including hospitalizations and deaths, are not affected by this reporting issue.”
The Los Angeles Times has begun wringing their hands, stating that the “broken” tracking systems leaves us with “no idea” of what is going on (link below). This is a recapitulation of apparent helplessness that has gotten us into this nightmare of an epidemic. It occurred in March over lack of available testing kits, and again is on display 6 months later.
The difficulty with waiting for laboratory issues to be fixed (or any issue for that matter), is that we are all trying to make COVID-19 decisions in real time. So rather than waiting, I took a look at case data from Los Angeles County to quantify the duration and magnitude of this delay.
In order to do so, I looked at data snapshots of incident case rates at 4 different time points: 7/7, 7/14, 7/26 and 8/7. Epidemic curves for each of those dates are presented in Figure 1.
The four curves all overlap until mid to late June when they begin to diverge. At that point, it takes about a week for the curve to catch up to its actual value. There is a very large drop off in the 7/14 curve (Red Line) which, no doubt, represents delays from the July 4th weekend. The numbers presented by the County on 7/14 ultimately ended up being a 44% undercount (actually seems about right when considering a 3 day holiday- 3 days off divided by 7 days in a week = 43%).
Fortunately, the magnitude in undercounts due to case delays seems to have lessened since July 4th, although we are now looking at about two to three weeks to catch up. This is seen in the 7/26 data (Yellow Line) which diverges from the most recent data (Green Line) for three weeks. These undercounts are: 3.5%, 11.4% and 34.0% in each of the three successive weeks. Obviously, the further back that one goes the more likely that case data are correct and accurately tabulated.
Applying these correction factors to the most recent epidemic curve gives us a pretty good idea of where we are in the epidemic curve (Figure 2). Rather than having “no idea”, we actually have a pretty good idea. Instead of the steep drop off suggested by the most recent County data, we appear to be at a plateau of cases with, in fact, a slight uptick between last week and this week.

𝗦𝗶𝗴𝗻 𝗨𝗽 𝗳𝗼𝗿 𝗢𝘂𝗿 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿

Dig deeper into the health topics you care about most by signing up for our newsletter.

by submitting this form you indicate you have
read and agree to our Privacy Policy and Terms
of Use. Please contact us to for us for more

Recommended Posts

No comment yet, add your voice below!

Add a Comment

Your email address will not be published. Required fields are marked *