Reporting Total Tests to Positive Tests ratio/ Death rate increase
March 28, 2020 12:34 PM   Subscribe

I have two questions about the Corona pandemic.

1. Wouldn't it be quite informative to know the total number of tests conducted in relation to the positive ones? As there is widespread concern about undertesting, this would show the extent of it? Why isn't this done?

2. In the projections for the casualties from the Pandemic, are the likely deaths from non-Corona causes included that would occur because the hospitals are full? For example in the one study where the worst case scenario was 2.2 million deaths?
posted by SweetLiesOfBokonon to Health & Fitness (12 answers total)
Best answer: I can only guess, but I'm willing to guess.

1. Since testing is still limited an erratic, any result would be hard to evaluate.
2. Deaths as reported are only confirmed cases. Otherwise some deaths from ordinary flu would be included. Besides, it can be hard to identify. A person goes to the hospital and dies. Would he have died if the hospital wasn't crowded by covid-19 patients? Maybe, maybe not.
posted by SemiSalt at 12:49 PM on March 28, 2020

Best answer: The UK reports on the total number of tests administered:
As of 9am on 28 March 2020, a total of 120,776 people have been tested, of which 103,687 were confirmed negative and 17,089 were confirmed positive.
posted by EndsOfInvention at 1:12 PM on March 28, 2020

Best answer: 1. Many jurisdictions are reporting the total number of tests, Quebec for example. I suspect that some counties/states in the US and other countries too are reporting these stats, even if not all of them are. However, as SemiSalt noted there is still the issue that different areas are prioritizing testing differently — different rules about who can get tested. So that could cause two areas with a similar infection rate and similar testing rate to nonetheless have different positive/negative case ratios. Quebec is claiming this is part of the reason they have a higher number of positive tests than Ontario, though I’m not informed enough to know whether that claim has merit.
posted by mekily at 1:15 PM on March 28, 2020 [2 favorites]

I no longer believe that the deaths as reported are accurate. Some countries are only reporting hospital deaths and some are only testing patients in the hospital, just for example. I’m going to start reading the weekly report on excess deaths that is available for the uk... I think that will give me a better idea.
posted by pairofshades at 1:24 PM on March 28, 2020 [1 favorite]

It's not just the number of tests you have to consider, but the rules for administering the tests. It's very different if you only test people in hospital who have been admitted with symptoms, for example, versus doing widespread testing of the population. There's a quantity difference, of course, but also the positive versus negative rates are going to be different.

Questions like this will provide decades of work for epidemiologists.
posted by Winnie the Proust at 1:34 PM on March 28, 2020 [2 favorites]

Best answer: Governor Cuomo’s daily briefings have included county-specific figures for total tests and total positive tests in New York. I was surprised that only 5.6% of them have been positive out of several thousand tests in my county, considering that tests have only been available to sick people. In NYC, it is 45% at last report.
posted by jkent at 2:22 PM on March 28, 2020

Best answer: The COVID Tracking Project (run by a journalist from The Atlantic, and other contributors) collects statistics on the number of tests performed in the US, broken down by state/territory, with links to official sources and annotations about differences in reporting across jurisdictions.
posted by mbrubeck at 2:52 PM on March 28, 2020

Best answer: If you are looking at this, be aware that the number of tests reported may constitute swabs taken, not completed testing for results. I came to that realization on a day when the Vermont testing statistics showed greater increase in cases than tests administered on a certain date.
posted by meinvt at 3:06 PM on March 28, 2020

Best answer: They are reporting this in my state. Number of positive, number of negative , deaths, recoveries and last week pending tests results. I’m sure that last one won’t keep up because they are beginning to use more than one lab and they won’t be able to keep up with the changing number.

Here last I knew the number of positive vs negative was around 10%
posted by domino at 5:55 PM on March 28, 2020

Best answer: Positive vs. negative test results seems to be holding at around 10%... which is strange since they basically only test people who are presenting as ill and sometimes only those presenting as seriously ill. And in some statistics I've seen, the excess mortality exceeds positive tests by a factor of six or so.

It seems to me that there are a lot of false negatives. Either that or there's some other reason for excess mortality, and that doesn't pass the smell test.

The tests are either not sufficiently sensitive or are being administered too soon--or too late!--to detect the virus. Or there's something else going on like selection bias. Anyway, it's hard to draw any conclusions from the testing data. The data that seem incontrovertible are hospital/ICU admissions and fatalities... even if these can't always be tied to COVID-19.

It's hard to see what can be learned at all from negative test results at this point, especially considering how inconsistent reporting and data collection has been.
posted by sjswitzer at 8:12 PM on March 28, 2020

Oregon reports positives and negatives both.
posted by SLC Mom at 10:22 PM on March 28, 2020

I know for a fact that the death numbers for Nevada are not even close to accurate. I have no idea why or what the issue is but I’m pretty bummed about it. People should get accurate info to help them make better choices
posted by yodelingisfun at 10:43 PM on March 28, 2020

« Older Cononavirus ethics filter: visit my mom or no?   |   Another "what is this book?" question Newer »
This thread is closed to new comments.