Covid-19 testing in New Jersey

Hey people, what is up with this?

Via John V.  And in the meantime, the virus has now affected 70% of New Jersey’s long-term care centers.


Maybe people want to go to see their families for Easter. Getting tested negative allows them to do that.

Without Donald Trump's China ban, NJ would be in much worse shape. The Dems tried to call him a racist for making our country safe. Losers.

You would be better off blaming New Jersey's poor performance on the state government as well as the governments of New York City and New York State. As it stands, entire countries with bigger populations and closer proximity to China have done a better job of containing the virus. South Korea has only slightly more confirmed cases than Bergen County alone, and the Bergen County figure is almost certainly an underestimate.

Highly descriptive post, I enjoyed that a lot. Will there be a part 2?|

The non sequitur of a truly addled mind.

Just out of curiosity, do you think things would be better or worse if trump didn't spend weeks leading up to this minimizing it as another "flu" and implying it was a democratic hoax? + or -? asking for a friend.

i love the people who out themselves by repeating mistruths ("hoax), either as an idiot (because they don't know they're wrong) or malicious (because they know it's a mischaracterization but they just want to score political points. either way, makes clear nothing you say should be taken seriously. so are you an idiot or evil? asking for a friend.

So Trump says that this is the Democrats' new hoax (which could be interpreted as "new thing they'll use to try to hurt me politically") but then goes on to downplay the virus for a couple of weeks and says it's like the flu and will go away like a miracle... I.e., he casts doubt on whether this is really something serious... like, maybe, he thinks it's a hoax. Sorry, maybe that's too much reading between the lines. But if that makes me malicious, then I'm guilty as charged.

i think you're wrong -- it makes you both malicious and an idiot. because trump obviously barred travel from china, took other early steps. so he pretty clearly was saying what you said he was saying (another hoax to try and make him look bad). but let's see you double down and include faucial among the "hoaxers," given his early statements this wasn't a concern for USA.

Trump said anybody who wants a test gets a test but that is not true. They only give testing to those that are seriously sick.

Why do people keep giving in to TDS. On March 6, our president said this - “Anybody that needs a test gets a test…. the tests are all perfect like the letter was perfect. The transcription was perfect. Right? This was not as perfect as that but pretty good.”

Only people trying to get the president, like the impeachment hoax, keep talking about how since March 6, any American is not getting a perfect test. Read the transcript.

The fact that Senate Republicans (except Romney) are scared of losing the Trump voters does not make impeachment a "hoax".

It wasn't a hoax. It was a contrived fraud. I'm sure that's so much better.

I dipped in to see the madness. I see it. I'm out.

Trump's short-sightedness has botched international relations across the board. The wonder is that the Senate Republican didn't throw him out when Pelosi gave them the off-ramp. Possible explanations are 1. They figure that they can repair international relations behind Trump's back. 2. They are as short-sighted as Trump. 3. They are cynical as Trump.

In the priors game I win always.
The entire mess is the result of California unions trying to raise commercial property taxes on my passive investment.

Never underestimate my priors, my priors beat your priors every time.

Adam and Eve make better priors. But even they don't make Trump's impeachment (nor the Mueller report) into a "hoax".

From the website you posted... Trump: "Anybody that wants a test can get a test. That's what the bottom line is." at 27:55

Will you admit the president was wrong?


Talking about our beautiful tests is a derail concerning Covid-19 testing in New Jersey? Call me skeptical that quoting Trump talking about testing is a derail on a post about testing.

Anybody who is still a Trumpist at this point is beyond redemption, but so too is anyone who wants to have a 'rational' discussion of this national emergency without mention of the President of the United States.

Neither group can engage constructively with the factual universe.

Hammer: TDS. Nail: the testing pattern in NJ.

“They called me mad. They called me insane. They called me looney. They were right!”

The poor guy doesn't really understand what DNA is.

"So now the Fake News @nytimes is tracing the CoronaVirus origins back to Europe, NOT China. This is a first! I wonder what the Failing New York Times got for this one? Are there any NAMED sources? They were recently thrown out of China like dogs, and obviously want back in. Sad!"

They're going to disappear. One day, it’s like a miracle, they will disappear.

Shouting: OK, boys! Not in here! Take it outside!

I would expect essentially 100% of long term care facilities will become infected. By their nature, isolation is not realistically feasible.

Yes. I was struck from the first that nobody who was meant to be planning for an epidemic seemed to have given much thought to Care Homes and, indeed, to hospitals and their staff.

Having infectious disease clinics and wards within general hospitals seems mad to me. People in the days of TB were far more intelligent about hospital architecture.

They were more precise about hygiene in hospitals generally. No antibiotics.

I think they knew there was a problem. The first US infections were in an nursing home in WA state that had a birthday party. However there was not that much that could be done. Some of these places are well run, others not so much. They started isolating these people in place and forbidding visits. Tests were scarce, so people going in and out could not be vouched for. Some health care workers worked in several nursing homes and spread it around.
These people are frail and some of them have dementia/Alzheimer’s . Isolating them and stopping them from moving much will also have negative consequences.

I have beaten this horse to death over the last month on different blogs. The sheer incompetence of the state/city/district health departments is stunning. You are right- it is absolutely mad to admit people with an infectious disease to a hospital, especially for an infectious disease that spreads like the common cold, is potentially lethal, and has no vaccine.

The very first step should have been to set up completely separate facilities for such patients, and give those facilities dedicated staff- that don't serve any other patient population. As for the long-term care facilities, they should have closed their doors to visitors and their staff should also have been dedicated solely to those facilities. Instead, what the various governors and their appointed health secretaries did was close low risk businesses in a virtue signalling orgy. As far as I can tell, no state in the union has really taken seriously the problem of keeping infected people out of the general hospitals and long term care facilities. It is a clusterfuck of incompetence.

Watch some of the youtube videos taken from the hospitals in New York City. What you will see are the infection "control" methods in place to protect the rest of the hospital clientele and staff. They consist solely of signage warning staff and visitors that various wings are off limits or elevators are off limits, or that staff moving from one section to another must change their PPE here and wash their hands. It is like these people never once heard of Murphy's Law.

My have a former work colleague who has a daughter in a long term care facility, since she was 8 months old. She was brain damaged during a heart operation due to a mistake on the part of the surgeon. They have spent the last 25 years having a weekly telecon with her in the care facility, which lasts anywhere from 5 to 30 minutes, depending on how long it takes until she has a seizure. Honestly, I can only pray that she gets infected and dies so that they can be freed of that torture.

I left out that she also has the mind of a 1 year old and cannot walk or speak. Only smile and wave at the camera and play with a few baby toys.

Population trends cannot be meaningfully generalized from rationed tests.

Day-over-day change in ever-hospitalized is currently the best metric.

Good point, and what is the link for that?

I snapshot this daily, since we are bunkered in Brooklyn :

Snapshotting daily is essential. For example, the April 1st hospitalizations changed from 1,162 to 1,439 yesterday. And that was not a recategorization from another day. Other days increased too.

People at various health facilities are too busy with more pressing matters than reporting data in the middle of a pandemic's first wave.

Yes, that’s why we also tally how many sirens we hear each day.

"Day-over-day change in ever-hospitalized is currently the best metric."

But what if the number of seriously ill people keeps rising -- but the hospitals are at capacity? And tell people to stay home unless they're critical?

In that scenario infection rates would be soaring higher than ever but hospitalization rates would level off.

Yup, hence the "currently" qualifier.

We also tally the sirens we hear each day. There are enough to be fit for the purpose of benchmarking the official stats.

I live in New Jersey. As everywhere, only those who display symptoms get tested. And a lot of people who have symptoms don’t bother to get tested, because if you don’t have it, you might get it at the hospital/office where you get tested. And if you do test positive, you get sent home anyway. Only the very sickest get medical treatment.
So there are really no incentives to get tested.
The fact that most people display mild symptoms means that the number of people with the virus is far, far larger than the number shown in any state. And since maybe one in four with the virus display no symptoms, the undercount is even greater.
Carry that out a bit and one might conclude that we are far more likely to reopen society too soon rather than too late.

Just to add that I used to estimate the number with the virus by multiply deaths by 100 - assuming a 1% death rate (you might use a different assumption, which is fine with me). But the spike in deaths at home in New York makes me think even that is too low.
It will be a couple of years before we have a good grip on how much deaths spiked.
The lack of testing has seriously impinged our ability to understand and fight the virus.

Someone has posted deaths per month in NY from 2000-current. Not sure of the source. It gives us some hint about how much deaths spiked.

You can get cdc deaths by week: have a look yourself. compare NYC states with nationwide states, you'll be surprised at what you'll find.

i meant to say "stats" not "states"

Fair point. During the killer smog of December 1952 in London, the authorities acknowledged 3.5-4k more deaths in December, about three times more than normal, due to the smog, but then, after the fog lifted, in an apparent cover-up, said that all the spike in deaths after December were due to the "flu" and "unexplained", despite experts saying these deaths were aftereffects from the December killer smog. Same as it ever was.

Multiplying deaths is a good way to estimate the number currently infected, but simply multiplying it by 100 or a number close to the reciprocal of the death rate only tells you how many people were infected 2-3 weeks ago because of the lag in how long it takes from someone to show symptoms and then die. For that reason, multiplying the number of dead by 1000 offers a clearer picture of how many are currently infected. Neil Ferguson, an epidemiologist at University College London specifically mentioned this.

Agreed, we are massively undercounting in NJ. Worse than in any other state.

COVID-19 growth is exponential. COVID-19 testing is not.

Two things I can think of. One is it’s still very tough to get tested in NJ, so testing levels remain low. But why do positive test % remain elevated?!?! We’re supposedly on lockdown here in NJ. I don’t know why, but lots of assisted living centers here in NJ, and Covid seems to still be running through those places. Also, we know poorer communities can’t comply to stay at home standards the way wealthier people can. And we’ve got lots of low income urban communities in NJ.

if you only test people with symptoms your positive rate will remain high

Right. And as you get better as recognizing the symptoms of coronavirus, your positive rate will approach 100%. Really this rate of
"tested positive/tested" is the most meaningless of all those used in the discussion about coronavirus.

In February, at the height of flu season, it was a fairly useful metric in a number of ways, particularly in showing a growing trend in covid19 infections compared to normal flu.

Not convinced. If we’re doing so well almost 4 weeks into lockdown, why are we getting so many positives? People have symptoms of all sorts of things. Half of those with symptoms have Covid!?!?

"Also, we know poorer communities can’t comply to stay at home standards the way wealthier people can."

Aren't you confusing wealth with compliance?
Those ethnicities/people who are orderly and compliant of COVID-19 guidelines in general will fair better than those who are disorderly and non-compliant.

There’s no leadership and the healthcare system is fragmented. All the system problems that we refused to fix for decades and fought over and over again, and no doubt will continue to fight over, they have culminated in a perfect storm. That’s the reality.

“Thousands of coronavirus tests are going unused in US labs”

So the solution is to centralize the "fragmented" system when it was the central government bureaucracy that created the biggest obstacle to testing?

You are right that the central government bureaucracy created the testing disaster. But I would say American healthcare has somehow got itself the worst of both worlds: a rotten federal bureaucratic structure that is unable to react to a crisis and at the same time a fragmented, mish-mash of a private healthcare system downstream to that, which is unable to work cohesively to face a national problem like a pandemic. It is not necessarily a public vs private problem, though that is the sort of framing we have become used to in our political discourse.

It is that we have healthcare system which has neither its public nor its private parts well suited to tackling a national crisis.

Virus is widespread (10% of population), not containable, and detection is limited by testing capacity.

These are all people being tested because they are sick. It is interesting that so many people are getting sick with anything after being locked down for three weeks. Coronavirus must be of comparable prevalence to colds and flu at this point.

Several counties in NJ are up to 1% of the population with confirmed positive tests. So is definitely possible that 5-10% of the population in certain counties in NJ have been infected, depending on how extensive the under testing has been.

For example, 390 deaths in Bergen County so far. Assuming 0.5% IFR, that’s 78,000 cases. Population in Bergen County is 930k. That would imply infection rate of at least 8% infection rate.

Mortality rate is in the 3-4% range. For Bergen, that would suggest something like 15,000 - 20,000.

Where are you getting such a mortality rate from?

I have not seen a rate that high for anywhere that has extensive testing. If testing is inefficient the mortality rate is overstated because of bias towards the most serious cases.

Most IFR estimates I've seen fall in the range of 0.2--2%.

There’s some time lag. Incubation’s mean is 5 days and some people don’t start to feel bad until 10 days after infection. In addition if you share a home, you can get infected from other occupants even after lockdown.
That’s also the nursing home situation. You’re locked down but other nearby may be infected.
From data from other countries it takes roughly 12 to 17 days after lockdown for the new cases to peak ( assuming testing is not a constraint )

New Jersey only made wearing masks and face coverings in stores mandatory on April 8th, so one might expect the state's infection rate to peak around April 18th, or 10 says after exposure. However, if infection rates continue to rise, the state will impose more severe guidelines and penalties.

Paul Romer wants 20 million tests a day. That's about 8 billion tests a year.

1. Find what price you need to offer to get 8 billion tests a year.
2. Solve for X, the amount the Feds should borrow to implement the needed testing regime.

Test, serology style are $10. Not much. But if you add in my previously planned bailouts, the total becomes 5 trillion.

Romer wants 20 million tests per day forever? Or for a year? Or until the entire population has been tested? Or what?

Testing the entire population at $10 each is about $3 billion dollars.

The entire US population every 2 weeks. He proposes a different test, one he says is much faster.
Read today’s assorted links ( #2)

The lock downs are not very effective. Soon, Italy and Spain will offer an opportunity to compare what happens when one place with massive covid19 problems eases up its lockdown measures compared to another which has absolutely no plans to do anything but the smallest changes to its current lockdown.

actually the lockdowns DO seem to be effective. Remember the R0 for this bug implied 60-70% of the population would get it fairly rapidly-and that was the much of the problem. Flu tends to spread to less than half as many people as this bug- so even if the COVID bug "only' has the same death rate we were looking at TWICE as many deaths as the flu-it it is likely COVID-19 is more lethal than the flu AND has a lot higher hospitalization rate. Current data indicates probably 10-15% of the population has caught coronavirus probably b/c the lockdowns have slowed the spread. No lockdown & the result would have almost certainly much worse.

'actually the lockdowns DO seem to be effective.'

Sure, they are more effective than doing nothing. The question, as can be seen with Italy and Spain, is just how hard a lockdown needs to be to be to prevent hundreds of deaths and thousands of new cases a day. I was talking to someone who teaches in Virginia Beach yesterday, and she said that gatherings with 10 people are allowed there. That number has not been legal in Italy, Spain, France, Germany, etc. for several weeks.

Look to Australia, the Nordic countries, and South Korea for effective curve flattening. Even Malaysia, Philippines and Thailand seem to be starting to wresting control from unmitigated spread. Likewise for China.

But is the cost worth it and is it sustainable?

Most of Hubei lifted its lockdown on March 25, and Wuhan on April 8. I guess 1 or 2 weeks from now we'll see.

Lockdowns are effective, and needed if you didn’t squash the epidemic at the beginning when few are infected, using targeted quarantines and contact tracing.
We missed that opportunity.
They’re a blunt tool ,and could become economically ruinous but they reduce Ro no doubt.

They're not totally ineffective, but mostly the problem is that they're futile. How much good is hospitalization really doing? We can't keep this up for years until there's a vaccine.

Everyone seems hung up on the hope that people will gain immunity after one infection, but the recent antibody study in China suggests that any immunity is weak and transient.

Seems like nursing homes would have been revealed to be the anterooms to death before now, every year during flu season. Or maybe they were, but in numbers spread out enough to fall under the rubric of mercy.

They are. The Life Care Kirkland , WA nursing home that was seeing 15 deaths a week in March had a background of 3 to 7 deaths a week before Covid-19

In Spain, assisted living facilities are colloquially known as "morireros", death camps.

it is indeed spread further than most people think, current lockdowns aren't doing much or have happened too late, the epidemic is closely following it's unmitigated curve, positive rates are more closely correlated to actual prevalence at this point. [I fully admit this could be (is likely) wishful thinking]

John V here. Two striking facts about COVID-19 testing in New Jersey:
- NJ has the highest positivity rate (48%) of any state in the US. The US average positivity rate is 19%.
- NJ has a higher positivity rate than any country in the world other than the Philippines.

Much more on Twitter:

Vainly, but inevitably, the thought recurs that New Jersey is also the most Italian state in the Union.

Possible reasons NJ testing is flat and % positive so high:

1. Doctors really good at guessing who’s positive
2. Many negative tests not reported
3. Only most severe cases tested
4. Huge backlog of tests
5. Capacity underutilized
6. Those going to get tests are self-selected to just the symptomatic who’ve also come in contact with someone positive.

#1 is unlikely. Why would NJ doctors be better at guessing who’s positive than other states?

#2 also unlikely. NJ claims 95% of total tests being reported, from 7 largest labs. That group of tests are 44% positive.

#3 very likely. They are limiting testing to the symptomatic.

#4 very possible. Supposedly have over 50 testing sites now. And results taking a week or more in many cases to report back. Maybe a surge in test results is coming?

#5 is possible, but anecdotally the public drive thru sites at least are filling up every day. Perhaps private and hospital tests are not using full capacity?

#6 also possible. People in NJ may assume it’s very hard to get tested, so aren’t seeking tests. So it’s possible only those seeking tests are people highly confident they are positive.

Bottom line, NJ needs to take several actions:
- Massively increase testing by 5-10x
- Loosen restrictions on who can get tested. Any symptoms whatsoever? Get tested. Someone you know or met tests positive? Get tested.
- Start random sampling the population to track true spread.

Much more coverage of NJ testing on my Twitter:

You're not going to see widespread testing until there are home tests because, as things stand now, the phrase "going to get tested" = "going to the place you're most likely to be infected."

This. People are only going in to get tested if they are really symptomatic. If there's a good chance you might NOT have COVID-19 don't go to the place where you're most likely to to get it. Stay home and wait until symptoms are severe enough that you really need to know.

The NJ testing numbers aren't much different from the national numbers, as illustrated by the COVID Testing tracking project (see Calculated Risk blog, which does a nice summary of testing at the national level.) That said, So I would guess that the lack of testing (in both NJ and, to a lesser extent, the broader U.S.) is largely due to some combination of 2 factors.

First, there are definitely bottlenecks in the testing process. Some combination of shortages of testing machines, reagents (i.e., specialized consumables that are used in the machines), technicians and testing facilities. My guess is that the maind shortages are in the reagents and the facilities. I believe that the number of companies that produce the reagents is fairly limited, and although there are lots of facilities, most of these are either centralized (e.g., Lab Corp, Quest Diagnostics) or inside of a hospital. Thus, these facilities are not in the ideal location, in the context of the current crisis. It would be better if more facilities were located outside hospitals to reduce the likelihood of disease transmission.

The second issue may be reimbursement, although this is just speculation on my part. Reimbursement drives much of what sort of healthcare procedures are, and are not, performed in the U.S. and much of the world. I don't know how reimbursement works for paying for COVID tests, but I can imagine a scenario in which it is just pretty lousy, or just not that great, and thus there is little incentive to perform tests. Because of the current crisis, the actual cost of performing the COVID tests are probably much higher (let's just say 10x or 20x) compared to a standard flu test. If the reimbursement isn't keeping up, then many providers will be dragging their feet when it comes to performing tests. As a practical matter in the U.S., much of the reimbursement is determined at the national level (e.g., CMS), and other payors such as insurance companies often follow the federal lead. On this subject, I will poke around the internet and see what I find.

Just my $0.02

Okay, so according to a March 16 article, CMS is reimbursing about $36 to $51 per COVID test. See this link:

This is just....well I can't use civil language here, but this makes absolutely no sense in the circumstances. The costs of running these tests must be hundreds of dollars under the current crisis conditions. But this is typical, as CMS often underpays or overpays for stuff.

Okay, so according to a March 16 article, CMS is reimbursing about $36 to $51 per COVID test. See this link:

This is just....well I can't use civil language here, but this makes no sense in the circumstances. I would think the "marginal" costs of running additional tests must be hundreds of dollars under the current crisis conditions. And additional marginal tests are exactly what we need right now. But this is typical, as CMS often underpays or overpays for stuff.

There are probably bottlenecks in reagent supplies at various times during a week in New York and New Jersey, but I believe the real bottleneck is much simpler than that, and applies across the entire nation- the sampling itself. Have you watched a video of the process? Have you watched a video of the waiting in line to have the sample taken? Have you watched a video of what the samplers are supposed to do after taking the sample (change all their gear).

Simply stated- the test have plateaued because the numbers of people willing to be sampled has plateaued- they don't take the test unless they are forced to at the ER or clinic. This is one reason why I have mocked all the proposals for testing and tracing- the idea that we were ever going to take even a million samples a day was ludicrous on its face, and yet the so-called experts continue to promote these silly ideas. Even Tyler is doing it here by wondering why New Jersey's testing rate is so flat. People are ignoring the actual process involved.

70% of long-term care facilities have outbreaks in NJ, you say. That’s even worse than Canada. But I seem to recall just yesterday a WaPo article declaring the Swedish approach reckless because of reports that fully one third of Stockholm’s nursing homes had outbreaks of covid.

NJ tests are on a 10-14 day lag per the Governor to get results.

NJ doesn't have a lot of testing capacity. They are testing the sick in hospitals, and they have drive through centers with very strict criteria that typically fill up for the day before 10 am.

I was tested for coronavirus at Quakerbridge Mall near Princeton on Friday.

Here's how the system works. Your physician writes a prescription and sends it to the testing folks. They schedule an appointment with you, in my case, for two days later.

You drive to the test facility, which is in the parking lot of the mall. I was in line with perhaps 80 other cars. It was still early in the morning, just before 9 am. The police then got the line going, and we snaked to the testing tests, which had two testing lines.

You pull up, open the window, and they stick a swab up your nose. In this case, it was a cute nurse who looked absolutely fetching in her hazmat suit -- not the kind you change after every use. The whole thing took ten seconds. No one asked for any insurance documentation or anything beyond showing my appointment number on the dash and my driver's license pressed up against the window.

I would guess that the facility has a capacity around 1,000 / day, which could easily be doubled with two more cute nurses in hazmat suits.

Everything was very smooth, very professional. It was running like an oiled clock. I was in and out, including waiting, in thirty minutes.

There was no visible overwhelming demand, and given that I was seen within 36 hours of the request, I do not see any really credible evidence of some massive unfilled backlog.

At present, to get into a doctor's office or a hospital more generally, you have to be tested for the virus. Those afflicted in Princeton are treated and held in a separate facility at the hospital there.

I agree with the notion that testing and sequestration started too late. New Jersey, like New York, is a very international place and it was freely imported because the White House chose to treat the whole thing like a joke until events got out of hand.

Thanks, that's a fascinating firsthand report. Which suggests that there might not be a bottleneck, and there's plenty of testing available in NJ.

Or maybe not? You needed a prescription to get into that testing location, maybe the prescriptions are being rationed out in low numbers (and perhaps only to those with COVID-ish symptoms, explaining the high positive rate)?

The various reports and numbers for NJ do create a puzzling pattern.

P.S. Without trying to get personal (so treat this as a rhetorical rather than a real question): why do you think you were given the prescription to get a test? If due to symptoms, all I can do is wish you good luck.

I had a bit of a skin condition, which absolutely exploded in the last two weeks. Could be just stress, but the coronavirus has a capacity to make the immune system go crazy, and really the damage is done by the immune response, not the virus itself. My doctor wanted to be in a position to see me in person, and I need a test to do that. I don't know if it's coronavirus or not, but I would not be surprised if it were.

I'd note that I am otherwise asymptomatic. No fever, cough, respiratory distress. So we'll see. Results on Monday.

Things could be rather different in more crowded parts of New Jersey such as Newark, Jersey City etc. Maybe lines do fill up there and there are capacity issues.

Testing is a shit show in the US. Basically you can get a test if you're showing bad symptoms but are not yet in a hospital. Otherwise people have died who almost certainly had the virus but don't count, for now, if they weren't tested.

The only good testing is doing now, it seems, is if you have a bad fever but you test negative, you may not need to spend the next 14 days locked in your bedroom isolated from the rest of your family since that's likely just a cold or flu.

Testing at this point I feel is useless. Trump let that boat sail. Hopefully once the curve is bent down we can get another shot at mass testing to keep cases contained.

I live in NJ and from what I've read about testing here, it's a mess: not enough people being tested, test results taking anywhere from 3 days to 2 weeks to arrive, etc. So here's a question about Tyler's graph on the top (and many other such graphs we have been seeing): the labeling says that the graph shows "daily" tests, but what Day is being used? For example, if someone is tested on March 15 and the result comes on March 29, is this included in the March 15 number or in the March 29 number? My guess is that the result is being included with the March 29 numbers, not with the March 15 numbers. That's sort-of ok if there's a consistent 14-day lag in test results, but with a variable lag (better tests that give results more quickly, or labs that are ever-more backed up), you can easily see how the data becomes such a jumble that day-to-day fluctuations have more to do with what's going on with the testing procedures than it has to do with trends in the number of people getting infected.

And just to add another similar item: our local town provides a daily report of the number of cases and deaths in the town. At first, I naively assumed that the death numbers were pretty cut-and-dried and meaningfully showed trends. But then they started reporting the actual date of death for each person, and it turned out that in many cases the person died a week or two before. Not surprising that the medical examiners are running behind in ascribing causes of death, but it does make it very difficult to interpret the statistics.

Two comments on the graph.

1. Stricter criteria for who gets tested, resulting in higher percentage of positive tests.

2. Graph not useful without knowing number of daily discharges.

Comments for this post are closed