Friday, May 01, 2020

The Importance of Data Clarity

“3.6 Roentgen.
    Not Great, Not Terrible”

     - Anatoly Dyatlov (Chernobyl, HBO, 2019)

So, I posted my last results for Los Angeles to my neighborhood on Nextdoor.  With traditional Nextdoor diplomacy, one of the neighbors said that the data was useless.  While I argued that it was worthwhile to examine the data regardless, as it was the best that we had, I thought this post should show the effect of limited testing on data quality, in particular with the state of California.

The quote I used at the start of this article, of course, is a reference to the amazing HBO miniseries, Chernobyl.  The authorities had been making decisions based on a meter that had, as its highest reading, a radiation dose of 3.6 Roentgen.  Later in the show we see that the true value was 15,000, not 3.6.  What does this mean for us?  I think we experienced a similar event.





As much as I would like to say otherwise, the numbers for Los Angeles since April 20 have been terrible.  On April 20 alone, the number of cases shot up by 1,475 (the previous highest ever single day increase before then was “only” 711).  I tried to tell myself that this was because of a surge of reported cases due to a long-standing backlog.  And, to some extent, this was true.  The new case numbers began to fall, finally getting back to 461 new cases on April 26.  However, they have begun what cannot be disputed as an inexorable climb, resulting in 1,033 new cases reported today, May 1.




How can there be good news in this?  How can we say we’ve even peaked?  Much less that we are declining?  Is there any hope? 

The answer, it appears, is that there is.

I always knew that our testing limitations were affecting the data, but I had no way of knowing how.  I knew that, behind the numbers, was the bias that the amount of testing each day was not constant, and therefore the new cases would be skewed by this number.  It’s very interesting to see how much that was the case.

Let’s zoom out a bit, from Los Angeles, to California as a whole.  Using data on testing, shown here, we observe a few things.  The first is that around April 22, there were over 165,000 new case results (the previous numbers were closer to 20,000 new tests/day).  This caused the massive spike, as indeed, there were over a week’s worth of data coming in in one day.

But, what’s really interesting is if we plot the number of new cases as a percent of the number of new tests.  Every time we see 100% we have our “3.6 Roentgen” moment.  We report that many new cases because that’s how many new tests we had.  The *actual* new cases were much, much more.



We’ve been reading from a pegged-out meter for most of March and April. 

This is why Dr. Fauci (and many others) have been saying that we need more tests.  Only by actually seeing past the infected can we learn about the trends.

So, where’s the hope?  It comes on April 14.  Let’s zoom in to that last part of the curve.



This is the graph to look at.  If we are ramping up our testing faster than our actual cases are falling, we will actually continually see an increase in our cases (until we finally exhaust all of our infected, and then the chart will plunge).  Right now, that seems to be happening.  However, if we look at this graph, we’re basically normalizing the curve to assume a constant number of tests per day.  And when we do *that*, the curve is finally starting to fall.

This data is not dropping particularly quickly, and it’s hard to make predictions from it.  However, at this rate, it implies that we’re on track to have another two weeks or so until we get to zero.  I’m really reluctant to make that kind of statement, given the misadventures with Easter, with Ventura and Orange County beaches, and with our population of tactical armored, Confederate flag-waving, “Patriot” cosplayers.  However, assuming that cooler heads prevail, this is a very important piece of the puzzle, and evidence that data quality is a thing.

Friday, April 17, 2020

The Case for Staying Home - Now it gets hard

I’ve been tracking LA closely this week.  And, unfortunately, we’re not seeing the numbers we want to see.

We are now entering the time when people’s patience for staying indoors is wearing thin.  Protests are happening demanding that we open everything up again.  Disgusting displays, including blocking an ambulance in Michigan, and protests here in Huntington Beach (“Liberate Huntington Beach”, echoing the inflammatory rhetoric of Trump) are starting to happen.  House parties are still going.

And yet, in theory, if nobody came near anybody for a period of two weeks, we’d see new cases crash to zero.  But that’s not happening.  And, it’s not happening because people don’t understand what’s happening, and they feel that social distancing doesn’t apply to them.  We aren’t doing enough - and we certainly can’t afford to slide backwards.

I was quite excited earlier this week.  It really looked liked LA would hit zero new cases before the end of April.  We’d then monitor to keep those cases at zero, and then we’d be done.  I was looking forward to celebrating that event.


That day is looking further and further away.  Today it was announced that in the last 24 hours we reached 11,391 confirmed cases in LA County, and added another 537 new cases - that’s where we were April 3.  At the rate we’re falling, instead of hitting zero new cases by the end of April, we’re now looking at the third week of May (assuming that there isn’t yet another resurgence from people gathering around Easter).  I’m hopeful that with each passing day people treat this with more seriousness, not less.  We need to to get through this.  We must maintain the strength to get through this - by supporting ourselves and supporting each other. 

We have avoided the nightmare of New York, but if you look, Florida just had more new cases in one day than ever before.  The US refuses to peak and decline as regional infection zones begin to grow (we reached over 700,000 cases here in the US today).  Italy, while declining, is doing so at a snails pace and isn’t likely to hit zero new cases until late May, and now we’re seeing Canada backtrack and start to ramp up a second peak. 

This is the cost of complacency.






I also want things to get back to the way they used to be.  But in order for that to happen, we need this number to be zero.  Then, we need this number to stay zero until we are sure that there are no lingering cases.  We need this to happen as quickly as possible.  And it’s not happening fast enough.

All data from Johns Hopkins: https://coronavirus.jhu.edu
 

Wednesday, April 15, 2020

The Case for Staying Home - LA's speedbump

Monday afternoon looked really good for LA.  We reported only 228 new cases, the lowest seen since we hit our peak.  At this rate, we were on track to be zero new cases by April 23 - just 10 days later.

Then yesterday happened and we hit a spike of 627 new cases, which was almost near our previous peak of 711.  Today it came down again with only 449 new cases, but this has pushed the "zero new cases" date into early May.


Now, LA has only 10m people, so the data is going to be much jumpier than when looking at a whole state or country, but I must admit I had allowed myself to get excited by Monday's numbers.  Nevertheless, I'm hoping that we drop back to the levels we were at before.

However, it's been my observation that people are starting to take this less seriously than before.  There's a dual feeling of "this can't continue, so we just need to get out", and "nothing has really happened, so there's no big deal about getting back outside".  If that is the attitude of the population at large, this curve will take a long time to reach zero, or could simply flare back up again.  Now is not the time to be complacent.

Perhaps we shouldn't be surprised.  I had been hoping that LA was doing better, but if we look at Italy, they also are on track for reaching zero new cases in the first week of May as well.  If Easter weekend also created a surge, we won't see that until late next week.  Let's hope people continue to take this seriously, otherwise, it's even longer before we can return to normal.






Thursday, April 09, 2020

The Case for Staying Home - The End of the Beginning

Winston Churchill, 1942

Now this is not the end.

It is not even the beginning of the end.

But it is, perhaps, the end of the beginning.

There are a couple of very important milestones to consider when we think of getting through this Covid-19 crisis.

1) The day that the number of new daily cases peaks, and each day after that we have fewer new cases than the day before.
2) The day that the number of new cases reaches zero.
3) The day that the number of existing cases reaches zero.
4) The day that we are confident that the number of asymptomatic carriers reaches zero.

For Los Angeles, Italy, Australia, and a bunch of other places, we have passed the first milestone.  We are truly at the end of the beginning.  Now, we need to see how quickly we reach the beginning of the end.

The reason that I'm mentioning this, by the way, is that I've seen a number of different reports that talk about a given location "peaking".  Yet, without context, we don't know what peak is being referred to.  The Institute for Health Metrics and Evaluation (IHME) has a brilliant set of analysis, where they talk about "peak hospital resources use" and "peak deaths".  These are both incredibly important metrics, but not what I'm discussing.

Last time I put up a graph which showed the peak of new cases for Italy.  I showed that Italy peaked on March 24, exactly 14 days after their March 10 stay at home order.  However, at the time, we didn't know how Italy's decline would look.  Now we do (apologies for the formatting - I wanted this to be legible):




We can see here that Italy is dropping at a steady rate of roughly 120 fewer new cases each day than the day before.  This doesn't mean that Italy is curing people at the rate of 120 people a day.  It means that, on April 9, Italy had 4204 new cases in the prior 24 hours.  Therefore, on April 10 (tomorrow), they will have only 4084 new cases in the previous 24 hours.  Because of the bounciness of the curve, this exact number is unlikely, but the trend is reasonably good.  If this continues linearly (and I really don't know if that's likely, or if it is set to accelerate or decelerate) then we can predict the day that Italy hits zero new cases.  Currently, that day is May 14 (Edit: with the original figure, this was April 30).  We can also predict the final number of infected, which would be 215,166 (Italy currently has 143,626 (again, all data from Johns Hopkins).  So, even though Italy is past their peak, over 60,000 more people are expected to contract the virus  before they finally get through this first wave (I'll talk more about subsequent waves later).

Given that I live in Los Angeles, I've been tracking LA with specific interest.  I had been hoping to see that we would demonstrate the same peak as Italy.  It took a few days to verify it, but since LA shut down on March 19, we should have seen our peak on April 2.  It looks like we could say it actually happened on April 3.  And, like Italy, there is now enough data to make an early prediction on the rate of decline (and, like Italy, I'm speculating that this decline is linear and I don't know if it will be).  Based on the last week, it looks like LA's new cases are declining at a rate of 21 fewer new cases/day (interesting fact, if we were to decline at the same rate as all of Italy, adjusted for population, that would be 20 fewer new cases/day - so we are seeing an extremely similar type of response).  Based on this projection, Los Angeles county will see zero new cases on April 29 with a total infected count of 12,045.  As of April 9 we are at 7955, so, like Italy, we still have about 1/3 of the way to go, and over 4000 people will become confirmed cases before the end of April.





However, this is all good news.  I think we can safely say that we are past milestone #1.  Now, the US as a whole is in much worse shape and I will continue to track the many states that are just beginning their infection ramp.

So, if this the end of the beginning, what is the beginning of the end?  There are two.  The first is a period after the zeroth new case.  If we assume that it takes 14 days to present possible symptoms, then we would want to see at least 14 days of zero new cases (ideally, much longer).  We want to make sure that all lingering infections have run their course before we lift social distancing measures.  Mayor Garcetti here in LA talked about us staying indoors until the end of May, and based on these numbers, that seems like the right plan.

The second "beginning of the end" is actually the next 12-18 months.  Until a vaccine is discovered and widely deployed, or herd immunity is reached, there will inevitably be flareups.  And, then, unless we act differently, we'll be scurrying back into our houses for another 3 months all over again.  Additionally, because, at least in the US, we have such a large population of states that are treating this virus with widely different levels of seriousness (from "adult and listen to the science" to "childish and criminal" a la Mayor Vaughn of Amity in the movie Jaws) we will have a secondary explosion just from people traveling within the country from one state to another.

So, the solution is twofold.  First, we must do like Germany, and implement widespread national testing for antibodies (I had a link here to FT, but it seems to only work if you go via Google).  This will allow us to know who is immune, and how widespread our herd immunity already is.  This allows us to get the economy restarted.  Second, we must do it right this time (like South Korea) and implement testing and case tracking and isolation to quickly pounce on flareups to allow us to stop the fire before we end up with another inferno.  This period will likely see social norms of continued mask-wearing and fist-bumps, rather than handshakes, well into 2021.  Remember that the Spanish flu of 2018/2019 had three severe waves.




So, we are not at the beginning of the end.  We likely are only 1/3 of the way through our quarantine and we must not relax our efforts to isolate (and we must have the utmost sympathy for the horrors to be witnessed by states that are have only recently entered their quarantine).  But, like Churchill said, we are likely at the end of the beginning.  And there is real comfort in that.


Sunday, March 29, 2020

The Case for Staying Home - The light at the end of the Italian tunnel

From Lord of the Rings, Fellowship of the Ring

Aragorn: Are you frightened?

Frodo: …Yes.

Aragorn: Not nearly frightened enough.  I know what hunts you.

I posted a little over a week ago, and I wanted to update everyone.  I included this quote because not enough people in California are abiding by the restrictions (they *finally* had to shut down the Rose Bowl loop and parks because people just didn’t get it).  And nationally, as of Friday, 27 states still did not have *any* stay-at-home restrictions.  To those governors, this is all a big joke.  So, this post has three purposes - 1) to share some good news out of Italy and to project what that means for LA; 2) to encourage people to continue to follow the isolation policies and to explain how much more there is to come, and 3) to ask you to implore your loved ones who live in the half of the country not under these restrictions to self-isolate regardless.  They are going to be slammed hard.

1) Ok, let’s start with the good news out of Italy.  Many people have been watching Italy to see what it tells us about the path ahead.  Data from China is near useless, and the only other major country (South Korea) to have gotten through this employed massive testing to prevent an explosion.  We did not.  So, the tragedy of Italy (and Germany, Spain, etc) gives us the best look ahead for our fate.

The virus can have a 14 day lag from the time a person is infected until they show symptoms.  In theory, this means that we should expect a 14-day delay from the time of shutdown until the peak of infection (roughly - people don’t follow the shutdown perfectly, and the time to symptoms varies from person to person).  After the peak “new cases”, the number of new cases should begin to drop off (the rate of drop-off being related to the effectiveness of the self-isolation).  Well, Italy initiated their shut-down on March 10, and therefore by March 24 they should be at their peak.  The fantastic news is that is almost exactly what has happened:




Now, Italy’s peak was really lumpy, and it’s possible that it will shoot up next week and all of this optimism is misplaced.  However, I’m going to optimistically project that they are on the other side of this.  They haven’t had a day with more than their peak (6,557 new cases) since March 21, and it really looks like they are on the way down.  Now, we don’t know what the second half looks like.  However, a safe assumption is that they will see at least another 14 days of decline, and then by mid-April will have zero new cases.  Let’s hope that follows through.

The other piece of good news for Italy is the accompanying decline in their infection rate.  This is also shown on the graph.  They’ve managed to get their infection rate down from 25%/day to *5.6%/day*.  It seems hard to believe that Italy isn’t looking at the light at the end of the tunnel, which is great news for us.

2) So what does this mean for Los Angeles?  We initiated a shut-down on March 19.  However, we did it 7x sooner than Italy (at 0.002% infected, rather than 0.015%), so hopefully our peak will be much lower than that of Italy.  If we also see a peak at the 14 day mark, that would put it at April 2 (this Thursday) and possibly even earlier (the chart implies that we might be in it now, but we should see how the next few days play out).  There are a million reasons that this might not happen, but it is tracking that way.  Now, even with that good-news proclamation comes the warning - there are *many* people in LA who still don’t understand the seriousness of this (the Rose Bowl loop was finally closed because people weren’t distancing).  People need to understand that they need to stay indoors to stop the spread.  However, if that’s not enough motivation, then stay indoors to protect yourself.  Just like Frodo, people aren’t nearly frightened enough.  Stay away from each other!

I’ve included the LA chart, just for the completeness of being updated, but it’s really hard to confirm a slowing yet.  However, it is important to note that LA has far fewer people than Italy, and therefore the data will be much noisier (also, we’ve kept this much lower - we’re still 10x lower on a per capita basis than Italy, which is fantastic).




3) What about the US as a whole?  Ok, this is the real horror story.  People are talking about Washington, New York and California.  I don’t want to turn this into a political discussion, but I will say this.  The power of the Republican party is in their ability to stick to a unified set of actions.  It is extremely difficult for any Republican governor to defy President Trump.  Right now, Trump is focusing on a message of blaming others (the China Virus, blaming New York, etc).  This results in 27 states not even starting to self-isolate, while maintaining a narrative that this is a Democrat hoax, no more serious than the flu, we should lift all restrictions by Easter, etc (although, he’s walking that one back).  I implore you, if you have loved ones in a state that has not shut down, to tell them that they need to isolate to save themselves.  Health care in remote, rural areas, is even less capable of dealing with this epidemic than urban areas are.  Georgia has 2600 cases, Tennessee has 1500, Florida has over 4200.  These are numbers California had only a couple of days ago.  On a per capita basis, much of the country is in a much worse position than California.  And yet nobody is doing anything.  Here’s a map from Business Insider from March 27 (it’s a little out of date - my understanding is that Alaska has now shut down).  You can see a huge swath of the country is still not taking this seriously.




Ok, so what does that mean for the rest of the US?  If California (and New York and Washington state, and others) start to peak this week, and then begins to decline in the first half of April, we will still see a horrific explosion in the rest of the states.  The US added over 20,000 cases on Friday.  While the US infection rate is declining, this is because New York is dominating.  Once Florida and Louisiana explode, they will continue to push up that curve.  If you have loved ones who have not sheltered in place yet, in a state that is not responding, for their sake they need to protect themselves.

I’d like to close by sharing a video from Science Insider.  There are two take-aways.  The first is that when people say that 80% of cases of Covid-19 are mild, this is true.  However, a “mild case” is defined as any case that doesn’t require *oxygen*.  That’s the definition of mild.  Anything more severe and you are looking at hospitalization.  Now, what’s the second case?  The pathway for death is something called Acute Respiratory Distress Syndrome (ARDS).  I had a very close family member pass away from ARDS, and we spent a month watching him die, watching his lungs get weaker and weaker as we hoped against hope, until it was over.  This is not “just the flu”, this is not mild, and you don’t want anyone you know to go out that way.




So, let’s recoup.  Italy is our best predictor and they look like they are half-way through.  If we follow that path, we (Los Angeles) will peak this week.  However, people aren’t distancing enough, and that’s because they don’t take this serious enough.  If you don’t care about saving others, save yourself - stay home.  Finally, at least we here in California have solid leadership from at the state and local level.  That’s not happening in over half the country.  If you know or care about someone in one of these states, please have them stay at home for the same reason, or we’ll see horrific numbers just as New York, California, and Washington are recovering.

We can see the light at the end of the tunnel.  It’s a long way off, but I think we can see it.  Stay home and stay safe. 

Remember - we know what hunts us.

The case for staying at home (March 24)

I posted this to a local group on Nextdoor.com back on March 24.  A lot has changed since then, but I'm posting this now to be a background for an update that I will post next.

Hello neighbors. I realize that we're all going through a very difficult time. However, it seems that many people in California don't fully understand the benefit of social distancing, or why we need to stay indoors. I have been tracking data from Johns Hopkins, and I thought I would share. 

The first case to look is Italy. Outside of China, Italy is the most infected country in the world with 63,927 cases on March 23 (let's not get too happy - we're at 46,332). They were growing at a rate of 33%/day (like pretty much every country other than Japan). However, when 0.015% of their population was infected, they implemented a country-wide shutdown. This was on March 10. It's now been two weeks, and we can see the results in the graph. 



Their growth rate started dropping - 25%/day, 20%/day, 18%/day, 15%/day, 10%/day, and then yesterday they grew at 8%/day. The last two days showed fewer new cases than the day before. Two weeks after a national shutdown, they *may* be over the half-way point. If their infectivity keeps dropping, they might only see 15,000 deaths. 

However, if we compare their results to the US, we see a few things. Italy shut down at 0.015%, Spain shut down at 0.014% and LA shutdown (Eaton canyon and beaches aside) at 0.002%. The US as a whole will burst past 0.015% tomorrow, and instead President Trump is talking about *lifting* any restrictions. Ok, so how are those restrictions doing for us? 

The second graph shows Italy, the US, Los Angeles (LA), and Canada and it plots the new cases each day (as a percent of population). You can see Italy's recent downturn. But you can see how the US has exploded nationally (over 11 thousand cases were added *yesterday* - a week ago the entire country only had 4,661). 



So, what does this second graph tell us? Well, the US is where Italy and Spain shut down, and the federal government is still not doing that (although multiple states and cities are acting on their own to do so). Secondly, we're just where Italy was about 10 days ago, but growing much more rapidly. Third, you can see a significant difference between LA and the rest of the US (note, these figures already have population factored in, so don't make the mistake of thinking that LA has fewer people than the US - this has been accounted for). 

So, we *NEED* to shut down. It does work. It is working. We need to stay away from each other. Group with your family, but don't go near anyone else. We must freeze this in its tracks, and the benefits are there. The quicker we shut this down, the shorter we'll be stuck in home. The US as a whole is growing at 35%/day. Unabated, this will be close to ONE MILLION people infected by early April. We cannot let that happen, and this is why everyone is saying this is as serious as it is. I hope this data has been helpful. You can see it yourself at Johns Hopkins (https://www.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6). 

Stay safe. Help your neighbors. Let's get through this.