Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

Health

In reply to the discussion: Covid data links, revised 12/12/22 [View all]

progree

(10,907 posts)
2. Testing Rates and Test Positivity Rates
Mon Jan 4, 2021, 09:14 AM
Jan 2021

Last edited Wed Mar 31, 2021, 03:42 AM - Edit history (3)

Yes, some states test more than others, and differ in positivity rates too. Thus ranking states by their reported cases (per capita or not) is unfair. Such a list like the above that lists states according to their daily new cases per capita is just a starting point. It is what it is. If there is a website that tries to adjust a list like the above for these factors, please let me know.

About testing rates - Early in the pandemic, there was a great shortage of testing facilities. In some places like Minnesota, they have become abundant, and so testing rates is as much or more a "demand side" issue as a "supply side" issue.

After soaring new cases and testing rates, our governor (Minnesota) completely shut down bars and restaurants for indoor dining in mid-November. Within a few days, daily case levels dropped and then so did testing rates. Testing facilities keep getting added, but with fewer people getting infected, the testing rate also dropped. Most people don't get tested unless they feel some need to do so -- such as feeling symptoms or having been in close contact with a Covid-positive person.

(The declining number of cases in Minnesota was corroborated with rapidly dropping test positivity rates and soon afterwards by dropping hospitalization rates, and after 4 or so weeks, dropping death rates).

So I don't make an enormous hoo hah about a state's low testing rate (compared to other states or the U.S. average) if the test positivity rate is low as well. But a low testing rate combined with a high positivity rate is a big red flag. An even bigger red flag is if the foregoing is true, and additionally, new cases are going up.

Positivity rates and testing rates are compared here:

. . https://coronavirus.jhu.edu/testing/tracker/overview ::

If the page doesn't show all 3 columns ( New Cases, Tests per 100k people, and Percent Positive ), then try lowering your zoom level or widen the window. All 3 of the above statistics are 7 day moving averages, good.

Missing: New Cases PER CAPITA, GRRR (it does have the raw count of new cases). Also missing: the U.S. overall test positivity rate and the testing rate. Duh. Would be nice to have these, especially for comparing specific states to it. Another shortcoming: It also doesn't number the states when it gives a list of states sorted by some statistic, sigh. Also, there are no trend mini-charts, but it does indicate if a state is trending up or down in some statistic over the past week, and separately, over the previous week


. . https://covidactnow.org/?s=1301312 ::

This also has test positivity and also daily new cases per capita, both 7dma, so it's more useful for state ranking in daily new cases than NYTimes, since it has everything I want in such a listing except that it doesn't have testing rates. It also has Population, Infection Rate, ICU Capacity Used (%), and Tracers Hired (%). It is also good that the states are numbered when states are sorted by some metric or another.

In the second half of the page, there are graphs of Cases, Deaths, Hospitalizations, and ICU Hospitalizations all in 7 day moving average. And one can select per 100K and compare 2 or more states.

(BUT the NY Times has those mini-charts of each state's daily case trends all on one page)

Missing: Testing rates. Also missing the U.S. overall test positivity rate. Duh.


. . https://coronavirus.jhu.edu/testing/individual-states ::

U.S. Test Positivity Rate (just a graph) - If one chooses "US" in the pull down dark blue box near the top left, it shows a daily graph of the test positivity rate, 7 day moving average. (It also shows the number of daily tests in millions, and the number of daily positive tests in millions -- the latter is and hopefully remains a small fraction of a million). Unfortunately, for the test positivity rate, one must either read the graph, or hover your mouse over it -- in which case it will pop up the testing rate, as well as the number of tests and the number of positive tests.

Problem is positioning your mouse exactly on the day you want, usually the last day. For that, it comes down to manual dexterity Or using Mouse Keys, which is explained in the OP, and which I prefer. You can also zoom in with your browser's zoom feature.

(no reason they can't just show the last day's 7dma numbers to the right of the graph, sigh).

By the way, it is apparent from the graph that the daily tests fluctuate a lot from day to day. This is one reason I hate Worldometers and other sites that just present the most recent day or one day results. I very strongly prefer those that present the 7 day moving average as that eliminates the day-of-the-week effect, and more generally averages out the day-to-day volatility.

Note also that when the mouse is on the graph, a set of icons appears above the graph for things like zoom, pan, select, and many more. I find them very difficult to use. One that may be useful is "download as .png"

Lastly, note that one can show individual states by selecting them from the pull down dark blue box near the top left.

Missing: U.S. testing rate. But at least it has the U.S. positivity rate, though one has to read it from the graph or hover the mouse just right. ALSO, very annoyingly, on the graph, in the pop up, the date is missing, except for the very last (right-most) data point.


. . https://covidtracking.com/data ::

This has U.S. daily tests, daily cases, currently hospitalized, and daily deaths, all for the latest day and 7dma as well. And all as graphs. The U.S. overall daily tests, 7dma is above the first graph "7-Day Avg". See "The U.S. overall positivity rate and testing rate" section below for how to calculate the testing rate per thousand, per 100k, and per 1 million.

. . . . . Scroll down to see results for states (note the "Jump to a state" purple box with links to the right of Alabama, so you don't have to scroll and scroll)


. . https://coronavirus.jhu.edu/testing/tracker/map/new-tests-per-100-k-people ::
A color-coded map of Daily Testing rates per 100k people, 7dma (hover mouse over a state on the map to see the number for that state).

Missing: U.S. overall, duh



The U.S. overall positivity rate and testing rate :

. . U.S. overall test positivity rate (the percent of tests that are positive), 7dma:

OurWorldInData.org -- https://ourworldindata.org/coronavirus -- choose the "Share of positive tests" radio button above the graph. Has views: Chart, Table, Map.

https://coronavirus.jhu.edu/testing/individual-states where one has to read it from the graph or hover the mouse just right.

https://covidtracking.com/data
(Some math required, see details in the "U.S. overall testing rate" section below.


. . U.S. overall testing rate:

OurWorldInData.org -- https://ourworldindata.org/coronavirus -- choose the "Tests" radio button above the graph, then 7-day rolling average and "Per 1,000 people" checkbox. Has views: Chart, Table, Map. Also can Download. Yes, the Tests per 1000 people 7dma in the chart match the table (the table show the latest day's info (7dma if that's what was selected)

https://covidtracking.com/data
This has U.S. daily tests, daily cases, currently hospitalized, and daily deaths, all for the latest day and 7dma as well. And all as graphs. The U.S. overall daily tests, 7dma is above the first graph "7-Day Avg".

As an example, the 7 day average on 1/2/21 was 1,535,081. Divide by the U.S. population of 332.0 million to get 1.535 M / 332.0 M = 0.0046 tests per person = 4.6 tests per thousand = 460 per 100k = 4,600 per million.

(source of U.S. population : https://www.worldometers.info/coronavirus/ )

Latest Discussions»Issue Forums»Health»Covid data links, revised...»Reply #2