CDC’s COVID-19 Community Level vs. Transmission Indicators
August 30, 2022 11:01 AM Subscribe
I’m currently in an area where the CDC’s Transmission Indicators for COVID-19 are extremely high, but the “community level” is low. While I’ve found some definitions for the community levels, I haven’t found a comparison of the two systems that says which one I should fall back on or why. This is a bit nerve-wracking - can you please point me to a comparison between the two, why there would be such a discrepancy in values, and how I should understand it?
Best answer: Epidemiologist Jason Salemi is my favorite COVID visualizer and he has spent the last six months making nice comparison plots that show both old and new community levels data, plus visualizations of time trends by state. Really nice stuff. I hated the CDC’s new approach until I saw it in this context because I’m someone who would rather see both pieces of information (how much COVID is there, and how much is it stressing the healthcare system). Here is his Twitter and his dashboard (Warning, the latter link doesn’t render well on mobile IME).
posted by eirias at 2:49 PM on August 30, 2022 [2 favorites]
posted by eirias at 2:49 PM on August 30, 2022 [2 favorites]
Back in the olden days (2021), in California, the "Red" tier case rate was 6-10 and the "Purple" tier was > 10 cases per 100,000. Link
To me, it was amusing (read "terrifying") when CDC relaxed the rules by 20x.
posted by soylent00FF00 at 4:26 PM on August 30, 2022 [1 favorite]
To me, it was amusing (read "terrifying") when CDC relaxed the rules by 20x.
posted by soylent00FF00 at 4:26 PM on August 30, 2022 [1 favorite]
Back in the olden days (2021), in California, the "Red" tier case rate was 6-10 and the "Purple" tier was > 10 cases per 100,000. Link
To me, it was amusing (read "terrifying") when CDC relaxed the rules by 20x.
I’m pretty sure the California rates quoted are daily cases/100k, whereas the CDC’s rate thresholds are based on cases/100k/7d. That is, the fair comparison would be 70/100k/7d vs 200/100k/7d — changed by a factor of ~3, not 20.
posted by eirias at 5:51 AM on August 31, 2022 [1 favorite]
To me, it was amusing (read "terrifying") when CDC relaxed the rules by 20x.
I’m pretty sure the California rates quoted are daily cases/100k, whereas the CDC’s rate thresholds are based on cases/100k/7d. That is, the fair comparison would be 70/100k/7d vs 200/100k/7d — changed by a factor of ~3, not 20.
posted by eirias at 5:51 AM on August 31, 2022 [1 favorite]
eirias: I’m pretty sure the California rates quoted are daily cases/100k, whereas the CDC’s rate thresholds are based on cases/100k/7d. That is, the fair comparison would be 70/100k/7d vs 200/100k/7d — changed by a factor of ~3, not 20.
I'm also assuming this is an Omicron-based adjustment, because Omicron is WAY more contagious but also WAY less deadly, especially for a majority-vaccinated population. Looking at California, for example:
-In winter 2020-21, reported cases were peaking at around 40,000 per day and reported deaths were peaking at around 500 per day.
-In winter 2021-22, reported cases were peaking at around 120,000 per day (just about exactly 3x as many as before!) and reported deaths were peaking at around 200 per day.
So using this data alone, we get a mortality rate of about a percent for the delta variant and a tenth of a percent for omicron (reported cases only).
To bring this all back to OP's question, Omicron (and vaccination) is also why we have shifted to community level (how well can the hospital system handle it if someone with high risk gets a very bad covid case?) and away from the transmission risk (how likely am I to get covid?). Now, average mortality rates go out the window if you know that you personally are at a high risk of complications of covid. Omicron is NOT particularly dangerous for, e.g., me (I'm vaxxed, boosted, and have no other risk factors), but it IS dangerous for a high risk person.
posted by capricorn at 11:52 AM on August 31, 2022
I'm also assuming this is an Omicron-based adjustment, because Omicron is WAY more contagious but also WAY less deadly, especially for a majority-vaccinated population. Looking at California, for example:
-In winter 2020-21, reported cases were peaking at around 40,000 per day and reported deaths were peaking at around 500 per day.
-In winter 2021-22, reported cases were peaking at around 120,000 per day (just about exactly 3x as many as before!) and reported deaths were peaking at around 200 per day.
So using this data alone, we get a mortality rate of about a percent for the delta variant and a tenth of a percent for omicron (reported cases only).
To bring this all back to OP's question, Omicron (and vaccination) is also why we have shifted to community level (how well can the hospital system handle it if someone with high risk gets a very bad covid case?) and away from the transmission risk (how likely am I to get covid?). Now, average mortality rates go out the window if you know that you personally are at a high risk of complications of covid. Omicron is NOT particularly dangerous for, e.g., me (I'm vaxxed, boosted, and have no other risk factors), but it IS dangerous for a high risk person.
posted by capricorn at 11:52 AM on August 31, 2022
Side note - please don't use my data and calculation in place of a real mortality rate for covid calculated by actual public health professionals, I'm just using numbers on a website.
posted by capricorn at 11:53 AM on August 31, 2022
posted by capricorn at 11:53 AM on August 31, 2022
I think the other piece driving the change in thresholds was simply a tacit acknowledgement that truly low transmission is not in the cards anymore. Thresholds are not so useful if you have a ceiling effect, and the old ones were starting to look that way at least in some areas.
My own approach here, since I still think avoiding or delaying infection has value, has been to incorporate the mitigations I can tolerate into my new normal. So from that perspective, the CDC levels are mostly irrelevant to my daily life.
posted by eirias at 3:40 PM on August 31, 2022 [1 favorite]
My own approach here, since I still think avoiding or delaying infection has value, has been to incorporate the mitigations I can tolerate into my new normal. So from that perspective, the CDC levels are mostly irrelevant to my daily life.
posted by eirias at 3:40 PM on August 31, 2022 [1 favorite]
I’m pretty sure the California rates quoted are daily cases/100k, whereas the CDC’s rate thresholds are based on cases/100k/7d. That is, the fair comparison would be 70/100k/7d vs 200/100k/7d — changed by a factor of ~3, not 20.
Very good point - after doing more research, I think you are technically correct:
Old California Rules say "case rates per 100K population per day"
whereas CDC Rules say "New COVID-19 Cases Per 100,000 people in the past 7 days"
However, to be even more technically correct - the original California system was that anything over 4 cases per 100k per day (e.g. 28/100k/week) was "Red".
The CDC "Red" tier is 200 per week AND requires either (High hospital admissions OR high % of hospital beds taken by COVID patients).
So comparing "Red" to "Red" the ratio was 28 to 200, or about a 7x relaxation (all other factors being equal).
posted by soylent00FF00 at 3:23 PM on September 11, 2022
Very good point - after doing more research, I think you are technically correct:
Old California Rules say "case rates per 100K population per day"
whereas CDC Rules say "New COVID-19 Cases Per 100,000 people in the past 7 days"
However, to be even more technically correct - the original California system was that anything over 4 cases per 100k per day (e.g. 28/100k/week) was "Red".
The CDC "Red" tier is 200 per week AND requires either (High hospital admissions OR high % of hospital beds taken by COVID patients).
So comparing "Red" to "Red" the ratio was 28 to 200, or about a 7x relaxation (all other factors being equal).
posted by soylent00FF00 at 3:23 PM on September 11, 2022
This thread is closed to new comments.
New system - "Community Levels"
- Meant to reflect health system capacity
- Relies a little bit on official case count (only uses reported PCR and antigen tests - home tests not included)
- Uses 200 cases / 100k people / week to define how you get the level - if greater than that, will always be moderate
- Otherwise only uses hospital metrics (new admissions and % of beds)
- If the hospital is getting a lot of admissions and has a lot of COVID patients staying at the hospital, the community level will go up. If not, no change
Old system - "Community Transmission"
- Used just two indicators - new cases / 100k people / week and % positive PCRs per week
- Previous threshold for high transmission was 100 cases / 100k per week
- With more and more people being diagnosed at home, using % positive didn't make too much sense - I do agree with dropping this indicator
- Had four levels compared to the new system's three
- Used "scarier" colors (I do truly think the red map was a big thing they wanted to get away from)
Basically - we used to quantify one thing (how much COVID is there) and now we are quantifying something else (how much COVID are hospitals dealing with, with some adjustment for community levels). This has to do with the CDC's goals of how the pandemic should be incorporated into life. I do not find the new system useful, but I'm not sure the old system is useful in our current world either. Using official test positivity is not helpful in my experience working on test data (most people are testing at home; the people who are tested officially are often likely to be sick), and I'm not sure those case thresholds are good now that COVID is so widespread.
posted by quadrilaterals at 11:37 AM on August 30, 2022 [10 favorites]