COVID-19

[jacob]

@JL13 - First, the key number of the basic reproduction number or R0. That tells you how many new people a single person manages to infect on average. If R0=4, then from a single patient (the index patient or patient zero), the next generation will have 4, the one after than 16, and so on. Of course one patient might infect two and another six and a superspreader might infect a hundred, so R0 is the average number.

This number can be determined statistically or thoroughly detailed contact tracing (counting the branchings in the infection tree and taking the average) or even theoretically.

If people change their behavior and infect less people the reproduction number goes down. It is thus a combination of both the innate qualities of the virus but also the host behavior. Some diseases even cause host behavioral changes that increase their chances of spreading, for example, rabies that makes the animal more likely to bite, or toxoplasmosis that makes the animal more likely to be biten. Humans are more interesting in that we're technically capable of altering our culture and thus change how diseases spread.

When it comes to infectious diseases, there are five kinds of humans from the perspective of the disease.

1) People who are susceptible to being infected.
2) People who are infected but not infecting others.
3) People who are infected and infecting others.
4) People who are recovered and immune.
5) People who are dead.

A person can change designation several times (except of course those who end up in number 5 ). For example, if immunity doesn't last, someone group 4 can go back to group 1.

Herd immunity happens when the disease can no longer reach enough new group 1 people to keep growing. Recall, to keep growing, it has to infect >1 people for each new generation. If it infects 0.9, say, then the next gen will have 0.9, then next 0.81, and so on .. thus eventually dying out.

Given herd immunity, that is, a sufficient number of people who can not be infected, how large does the reproduction number have to be to find at least one susceptible person?

I find it easiest to think in examples. Lets say R0 is 4 and you're the virus looking to infect 4 random new guys. If 3 of those are already in groups 2, 3, 4, or 5, then there's only one left. When does this happen? It happens once 3/4 of the population is in groups 2, 3, 4, or 5.

Generalizing, the equation for simple herd immunity is thus (R0-1)/R0.

So lets say that herd immunity is reached in this example. 75% of the population is either infected, immune, or dead. 25% still are susceptible. If the virus is introduced from a single traveler, the traveler might (=will on average) infect 1 other person who in turn will infect another... and so on. But soon enough, the odds will decrease to 76%/24% and the new generation will see 0.99 ... the next 0.99^2 and so on .. and the virus will lose traction and die out.

Now lets try multiple travelers and lets say R0 stays at 4 but the number of susceptible people has dropped to 20%. Those travelers still have a 20% chance of meeting a susceptible person and infecting them. However, the infected resident also only has a 20% chance of propagating the disease and so it will die out.

Just to emphasize, that just because the herd has immunity, that doesn't mean that individuals in the herd can't be killed off. It's only the herd that can no longer be killed off.

It's important to distinguish between the reproduction rate and the attack rate. The attack rate is the chance that someone exposed to the virus (because they inhale it, inject it, ingest it, ...) gets infected. It's part of what goes into the reproduction number which also includes how many people a person meets, how they meet, how long they meet (dose), ... and so on.

Animal reservoirs are just like the multiple traveler scenario except these travelers come from the animal kingdom. Obviously social distancing to humans doesn't work here but distancing and avoiding contact with animals by e.g. not building housing in their habitats or using screen doors against insects or outright killing infectious animals work. Basically same idea.

A disease becomes endemic if group 1 (the susceptibles) somehow keep increasing. Childhood diseases in the pre-vaccine era are good examples of this. Here group 1 increases die to new people being born. Measles has a herd immunity of 95% and will pretty much wash through a susceptible population very fast until 95%+ have been infected. Then it dies out there while moving elsewhere. However, after a few years with a 2% birthrate, the number susceptible people has grown from ~5% to maybe 12%. Thus if an infected traveler (from group 3) comes to visit, the infection will spread until the 95%+ threshold is reached again.

Insofar vaccines exist and it's possible to trace down infections all over the planet, it's possible to ensure that it's no longer possible for travelers to reintroduce the disease. The disease has thus been eliminated from the planet. This was done with smallpox. It was almost done with measles...

Alternatively, one could prevent travelers from entering unless they can prove immunity (by vaccination or prior recovery). Many colleges test for diseases before allowing admission. So does the porn industry. Countries could do something similar (a health certificate) for travel and many already do for immigration. Another method would be quarantine. The US used to be well protected simply by virtue of having two large oceans on either side, jungle/desert to the south and ice to the north. For example, the invention of steam powered ships made travel became fast enough for people on the ship to stay in groups 2 and 3 as they made landfall and thus introduce a bunch of new diseases to the north American continent that wasn't possible under sail where transition times were long enough to send the crew into groups 4 or 5. Of course air travel opened a big can of worms in that regard. Perhaps quarantine will become normal?

It's important to distinguish R0 from Rt. The reproduction number changes as the population works towards herd immunity; or change their social behavior. It even changes as the virus mutates and changes its behavior. If herd immunity is possible, then Rt will ultimately converge on 0 as the disease fails to find new people to infect.

With CV19, what appears to be happening in western countries is a complex adaption in behavior from both people and governments that has Rt converging on or rather oscillating around 1. My hypothesis (and it's just mine) is that this is driven by the news cycle and the fact that humans seem to tolerate a certain amount of fear/pain while weighing this against their personal behavior. Thus if people are mainly hearing bad news (from media, colleagues, politicians, friends, ...) they will modify their behavior and bring Rt down. If people are hearing good news (the disease is under control, we're over the worst, ... ) they will modify their behavior and Rt will go up again. Since immediate news is determined by the current disease burden, this acts to keep the disease burden constant. And a constant disease burden happens at Rt->1. Again, this is just my theory for the cultural adaption dynamics. The consequence of keeping Rt around 1 is that the disease will slowly wash through the entire population until either herd immunity or a vaccine is found but it will do so w/o overwhelming the medical system.

In ASEA countries, the behavior has instead been more collective. Individuals have not made their own choices to go out and engage in behavior that increases the reproduction number. Governments have not relaxed restrictions to get back to business as usual either and therefore Rt->0. Since they don't have herd immunity, they therefore have to restrict travelers.

[jacob]

It begs the question, how many in the 30-39 y/o age group experienced mild or no symptoms, fought the infection and never produced antibodies?

It is certainly frustrating that we cannot know these numbers. It must be a significant number, but at least we know that we cannot know. In other words, they are known unknowns.

I'll give you that. This number has to be very large to make a difference in the argument that we're close to herd immunity now. If the Chinese serology vs PCR test shows that 30% of those who show up at the hospital fail to produce antibodies, then that just changes the Spanish number if population infection from 5% to 5%/(1-0.7)=7.1% and the IFR from 1.17% to 1.17*(1-0.3)=0.82%. This doesn't move the needle much. Point being ... that 30% number has to be a lot higher to make a material difference. We're talking 90%+ in order to generate hope.

To get at this known unknown, a massive PCR testing is required. It's practically mindblowing that the world hasn't figured out how to do this yet. It's been almost half a year since this [shitshow] started. Trillions has been spent on bailouts but apparently randomly (keyword: random) testing 10-100,000 people is too hard.

Up to now those doing the modeling have been pretending that they don't exist.

Ehh.. I wouldn't go that far. Pretty sure that those who are doing the modelling are well aware of these shortfalls and doing the best they can with what information they have.
Three points.

First, their models are substantially more advanced that what armchair quarterbacks like MDs or ex-physicists can do with back of the envelope calculations. Also, they have a better idea of what they don't know than we (non-experts) do. A lot of us are learning as we go along---as unknown knowns become known-unknowns and maybe known-knowns. Experts on the other hand are observing and confirming as they go along turning known unknowns into knowns. This is very different.

Second, they have better and more data than we do. We (amateurs) rely on what summary data filters through from twitter, journalists, or ... real models have much better granularity than e.g. "30-39yos of a 46M country". For example, I have five groups in my explanation to JL13 above. A real model might have hundreds of groups also showing county location, gender, age, health issues, ... This is the reason why I stopped making estimates once R0 started changing. I knew that the simplifying (exponential) assumptions that held in the beginning were no longer valid.

Third, and worst or saddest. Laymen and experts have come to have an odd relationship over the past several decades. On one hand laymen expect experts to have precise answers (while in reality experts strive for accurate answers) but on the other hand, laymen like to override any advice with their own (second or third) opinions. You see the same dynamics in climate change debates (remember the threads). Laymen mostly lack the coherent framework that experts have. In particular they don't know what they don't know and thus laymen aren't dealing with known unknowns as much as they're self-inflicted victims of unknown-knowns. This leads to grasping for factoids hoping to disprove expert conclusions in order to arrive at the desired "opinion". In practice, it's like doing battle with google. It's always possible to find someone on the internet with another opinion. My point is that we're in the unfortunate situation where 1) most people can't distinguish opinion from fact; and 2) worse, that most people don't have a coherent framework for facts and don't trust that experts have it either. This results in a shitload of incredulity with laymen stating that "they don't know so they don't think anyone can know". This is what it is. The problem arrives when the experts have to give "effective advice". Very few experts are actually very good at this because they're not experts in dealing with intellectually recalcitrant politicians or civilians. It's thus common to regress to parent-child type interactions and just go with a "mommy and daddy knows best" (rather than "mommy and daddy are only 90% sure, but still much more likely to be correct about this") type of messaging when the "child" is playing with matches. As I stated above, I've come to believe that most humans only want to learn complex lessons the hard and personal way. I no longer think complex(*) problems can be mitigated in any way. We'll just have to muddle through the consequences when they hit us ... like we're doing now.

(*) Problems that are undetectable by literal "common sense" (impressions that are common).

[JL13]

@Jacob

To see if I'm understanding this correctly:

1.) Assuming everyone manages to keep Rt around 1, how long would it take for it to spread through the whole population? If each person infects just one more person, and we start with, say 1% of the population infected. Then every 14 days another 1% of the population gets infected? So 1,400 days?

(Not actual math, but just conceptually)

2.) Each Rt has it's own herd immunity level, so in general, can we expect everyone to modify their behavior to keep Rt around 1, and once herd immunity is reached at that Rt, then Rt will drop below 1, and we'll increase our interactions, putting Rt above 1 again until we reach herd immunity at the new level and so on?

3.) If the immunity achieved after infection is like SARS1, 24 months, then eventually would we expect to see, very roughly, 1/24th of the population being infected, always?

The same way that chicken pox infection rate tended to follow the birth rate, would the Covid-19 infection rate tend to follow 1/immunity period?

[jacob]

@JL13 -

1) Mostly correct. Keep in mind that it will be hard (sans travelers) to keep Rt at ~1 once herd immunity is reached. If the characteristic time of a generation is 14 days and herd immunity happens at 70%, the estimate would be more like 14*70 = 980 days as a constant disease load works its way through enough of the population.

R0 unfortunately does not say anything about the rate of infection and so doesn't tell you how long a generation lasts. In the SIR model, R0 is set by the infection rate and the recovery rate and it's "easy" to derive the infection rate, since we know the recovery rate from hospitals and we know R0 from contact tracing.

(Knowing the attack rate from laboratories, it's also possible to figure out the infection rate with access to an agent-based computational model. Simpler versions of this are found in the papers estimating e.g. how many hours in the company of a breathing person it takes to get infected.)

Conceptually, this is therefore not complicated, but it's hard to do precision science.

I give out 2-3 decimals in order to illustrate the math for those who want to follow along and check the math. Not to imply any kind of precision.

2) Right. Lets say we modify our herd behavior so R0 = 1.1. Then herd immunity is achived at (1.1-1)/1.1 = 9%. If we maintain our behavior, the disease will die out after 9% have been infected. However, if we slack our behavior after that---increasing R0 to 1.5, say, some traveler or pocket will restart the epidemic until the new herd immunity is achieved at 1.5-1/1.5 = 33%. And so on ...

If the virgin R0 was 3.5 corresponding to normal pre-CV19 behavior, and people suddenly developing a elevatorbutton licking behavior or turn coronaparties into a trend, then Rt could correspondingly become much higher.

3) Insofar people roughly take one month to get reinfected and immune(*) again, then yes, but if and only if 1-1/24=95.8% is below the herd immunity threshold. This is likely not the case as SARS-CoV2 has a lower herd immunity threshold. Therefore in this case, the immunity lasts long enough for the disease to die out (once enough people have been infected once).

However, suppose immunity existed for 3 months (it's probably longer), so 1-1/3 = 66% remain susceptible (baring a vaccine). Suppose also that herd immunity is 70%, then the conditions are there to make it endemic. The disease would reappear as waves every several months or years. Like measles did and now do again.

Basically, very roughly(!) since it ignores the time dynamics (things get iffy away from the asymptotic boundaries)
If susceptible% > 1-herd immunity% => endemic
If susceptible% < 1-herd immunity% => occasional unsustainable breakouts that die out on their own

(*) What really matters is the time it takes between someone starting to infect others and being taken out of the equation (group 3 -> group 2). Here it's a problem if many are asymptomatic. If displaying symptoms after 2-12 days, people could self-quarantine in which case, the characteristic time is 2-12 days.

[steveo73]

Third, and worst or saddest. Laymen and experts have come to have an odd relationship over the past several decades. On one hand laymen expect experts to have precise answers (while in reality experts strive for accurate answers) but on the other hand, laymen like to override any advice with their own (second or third) opinions. You see the same dynamics in climate change debates (remember the threads). Laymen mostly lack the coherent framework that experts have.

I reckon this is so wrong it isn't funny. Please note that this is a subjective opinion just like your opinion. I think that the problem is that some laypeople think science equates to certainty when it doesn't. I think most people do lack the correct framework from which to understand scientific discussions but it's because they give too much credence to the science. I'm hopeful now that people understand science a little better. I'm less confident that scientists learn from this but I've heard lots of great scientists speak on the COVID-19 issue. I think often scientists have a really poor framework from which to assess issues like COVID-19 or climate warming. They are alarmist by nature and portray certainty when it doesn't exist. These are among the worst people to listen too when making decisions for society. On the flip side you don't want to listen to people who don't understand an issue at all and cannot see any risk at all. To me though it's extremely clear that we cannot trust those scientists who do not have the proper framework including humility to discuss issues. I hope these people's viewpoints are now sidelined by the vast majority of people. Sure they can speak but we just state they cannot predict the future with any accuracy and they are too pessimistic.

We have a classic example right now in relation to COVID-19. Some scientific predictions have been shown to be too pessimistic and more problematically not explained with the proper frame work to understand what is being stated. There is too much certainty for whatever reason when the science is evaluated by laymen and therefore people trust the science too much when they shouldn't. Then scientists put numbers out and quote them with certainty in order to push their policy agenda. I'm not sure if it is the scientists fault or the media or people's understanding of science or a combination of all of these factors.

We are still discussing herd immunity, R numbers and the potential impact and trying to put numbers to that. There may be some value in that but we won't get certainty when it comes to predicting how this virus develops. Does immunity even work let alone herd immunity ? Is it partial immunity ? We do not have complete certainty in relation to our understanding of this virus. Our knowledge improves over time so we are getting better at it but let's be clear that we are assessing this in real time and there is still plenty of uncertainty and complexity in relation to how we manage this virus.

We need a risk management approach to making decisions where the process is complex. We have to move away from trusting pessimistic but uncertain science.

One poster a while back commented on something being statistically significant and it was a really interesting point. Some people believe when something is statistically significant it will hold true across the population. In other words it is a proven verifiable fact that cannot be disputed. That isn't necessarily the case in reality. These are the big types of issues that we need to state clearly when it comes to discussing issues. Be clear on the assumptions within the science. Be open about the flaws in the science. Do not promote a false sense of certainty. Predictions of the future fail all the time.

[George the original one]

We have to move away from trusting pessimistic but uncertain science.

On the other hand, we also have to quit listening to optimistic pipe dreams from the Trump administration.

Apr 19, USA was at 40,000 deaths. IHME issued a cockamamie forecast saying there will only be 60,000 deaths and Trump seized on it for his political agenda. It was obviously hogwash and just looking at the trends as a layman, I predicted 60,000 deaths would be reached by the end of April and 100,000 deaths by the end of May... missed my April prediction by two days and we're still on target for 100,000 by end of May.

Meanwhile, IHME & Trump have had numerous revisions. Trump administration are now forecasting 147,000 deaths by August 4. Back at the beginning, in mid-March, Trump was congratulating himself if deaths were only 100,000, so you can see he's failing by his own standards.

[7Wannabe5]

@steve073:

Sounds good. When will the risk management based model you are working on be ready for presentation?

[steveo73]

@steve073:

Sounds good. When will the risk management based model you are working on be ready for presentation?

It's very simple and we have it now. It's just a standard approach to risk management. You have a probability of an event occurring and the impact of that event occurring. You then need to understand the implications of taking that action and assess the potential impact there.

The problem is twofold:-
1. We've listened to scientists that are too pessimistic and believed them because they are scientists.
2. We've listened to morons who are way too optimistic without any facts to back their optimism up.

Let's be clear though - this is not a difficult concept to grasp.

[steveo73]

On the other hand, we also have to quit listening to optimistic pipe dreams from the Trump administration.

I want to be clear that this is a twofold approach and it's designed to get the loonies away from running the asylum.

It's not about accepting all risk and just putting your head in the sand. It's about evaluating risks to society in a professional manner.

When people state stating the science says we have to do this they are probably wrong. When people state it's not an issue they are probably wrong. We need to get the extremists out of the debate.

[Riggerjack]

It's very simple and we have it now. It's just a standard approach to risk management. You have a probability of an event occurring and the impact of that event occurring. You then need to understand the implications of taking that action and assess the potential impact there.

The problem is twofold:-
1. We've listened to scientists that are too pessimistic and believed them because they are scientists.
2. We've listened to morons who are way too optimistic without any facts to back their optimism up.

Let's be clear though - this is not a difficult concept to grasp.

So what we just need to do, is use information we don't have to project the range of possibilities. And then do just the right thing at the right time. Brilliant!

Next, maybe we can look at portfolio management. Just buy before the price goes up, and sell before the price drops. Management, at it's finest...

.........

Look, I don't know what your sciencey sources are, but you need to participate in a better class of audience. If a scientist sounds certain about a developing situation, then you are listening to a scientist who is having what Jacob describes above as a parent-child conversation. And you are in the child role.

If you want to participate as an adult, even if only as an adult observer, you have to go where those conversations are. Hint, there are no journalists at those conversations.

For a simple laypersons breakdown, Bret Weinstein and Heather Heying have been doing a very nice set of podcasts on the virus as things unfold. They are not virologists, they are biologists. They do a very good job of explaining the concepts at a freshman level, where the uncertainties are, and what this means. And they, like anyone else with half a brain, are not certain about much of anything about our current situation. Bret has a YouTube channel, if you are interested.

But if you have been getting your dose of science from the guy in a lab coat talking to the news anchor, you are right, you shouldn't be listening to him.

If your science comes from tweets from an orange man, well... Damn, I don't know how to help.

[J_]

@Jacob Thanks again. Every time when it becomes difficult to understand the workings of a plague like Sars-Covid2, you take the time to enlighten the situation and explain what is still unsure/unclear/unknown.

Especially now, due to new insights and due to social and economic reasons restrictions are (slowly) lifted, one can easily become more sloppy in avoiding to be contaminated.

There is no (herd) immunity, the virus can be very dangerous for some of us. So I have mixed my life with doing a lot outside, be cautious when I go to shops and meet friends. But I am not locking myself up. In our rowing-club one persons boats are allowed to row again. I volunteered to be one of the overseers to look/help sport members to follow the Coronona-avoiding-protocols.

[steveo73]

So what we just need to do, is use information we don't have to project the range of possibilities. And then do just the right thing at the right time. Brilliant!

I don't understand this attitude. I'm stating the exact opposite approach. I'm stating when we don't have certainty we need to accept that and do our best to make rational decisions. We need to understand that the science is typically not clear cut and we won't make perfect decisions. When we make those decisions we need to think through the implications of our actions.

Can you imagine if this approach was taken to COVID-19 or even better viruses in general. We may have been prepared for this event and had processes and procedures in place to handle it.

I should add that I recognize that there are good scientists out there. Scientists that have the ability to look at their models and data critically. I've listened to a bunch of them during this COVID-19 issue and on other issues. At the same time the alarmists have got this virus wrong and maybe they have caused a lot of damage. We need to learn from this. Even great scientists get it wrong.

“There is not the slightest indication that nuclear energy will ever be obtainable. It would mean that the atom would have to be shattered at will.” — Albert Einstein, 1932

[jacob]

Meanwhile, IHME & Trump have had numerous revisions. Trump administration are now forecasting 147,000 deaths by August 4. Back at the beginning, in mid-March, Trump was congratulating himself if deaths were only 100,000, so you can see he's failing by his own standards.

IHME is a phenomenological model that fits US data points on curves from other countries that are further ahead in the infection. Phenomenological models are used to get results when the inner workings of a system's dynamics are poorly known. It's success is entirely dependent on the US curve or experience matching up with those other countries. That seems increasingly unlikely and the shifting numbers is to due metaphorically speaking trying to fit a straight line to a parabola, say. (Round peg in a square hole.)

Trump is ... uhh ... an ESTP (SeTiFeNi stack) who therefore tends to focus on whatever is right in front of him (dominant Se) while struggling with contingencies and hypotheticals (inferior Ni). A 100% tactics only operator and also the type most likely to genuinely believe in their own BS. Stereotype: "used car salesman".

Will note that around end of March, when Trump went along with the lockdown recommendations, the White House adults (Fauci and Birx) projected 100-240,000 US deaths. For the purpose of governing, it's best to give a range that reflects the uncertainty and the ballpark, here "low six figs". The IHME does come with wide uncertainty bands (right now 115-230k as far as I can eyeball it) and an exact expectation value (now 147,040). Unfortunately, I don't remember the extent of the uncertainty bands six weeks ago.

From an expert perspective, giving the expectation value line (the dotted line) is useful because it indicates the shape of the distribution between the uncertainty bands which either gives a standard deviation or some p-value. Here it clearly skews towards lower numbers for all times going forward. However, the layman perspective tends to ignore the distribution and just look at "the number" focusing on how precise it is (how many decimals), how often it changes, and whether there are other sources who give a different number that they might like better. In short, they have their own heuristic interpretation of the uncertainty which may or may not match the actual uncertainty.

[7Wannabe5]

Maybe we can do a book group to improve the level of discussion?

I am thinking either “An Introduction to Mathematical Modeling of Infectious Disease” by Li or “An Introduction to Infectious Disease Modeling” by Vynnycky and White. They’re both kind of pricey, but this is an important topic.

[Riggerjack]

I'm stating when we don't have certainty we need to accept that and do our best to make rational decisions. We need to understand that the science is typically not clear cut and we won't make perfect decisions. When we make those decisions we need to think through the implications of our actions.

I can agree with that. At the individual level and possibly at the small, tight community level.

But at a governmental level, this isn't possible. To expect it, is to court dissatisfaction. There are too many different needs and situations for one set of decisions to work poorly, let alone well.

Let me give a specific example.

Here in WA, the county hardest hit is Yakima county. It is a rural, agricultural county. At over 0.5% of the population having cases of coronavirus. This is because it is agricultural, and the virus is burning through the migrant worker community.

Currently, we are in phase 1 of the lockdown, "full protection mode". But among the Governor's proclamations, by which this crisis is being handled, was a proclamation that bunk beds for migrant workers was just fine.



Yup. Gotta keep everyone at home and safe, except for the most infected community. They should travel, settle in for a few nights in an enclosed environment, sleeping in bunk beds, then move on to do it again.

So why is that?

Because Yakima is red, and the governor is blue?

Because migrant workers don't get to vote, for the most part?

Because nursing the virus in Yakima keeps pressure on to maintain the lockdown elsewhere, and the governor can't let go of the power he grabbed?

Because we have a representative government, and the governor didn't think he could enforce a massive cultural shift, like forcing farmers to provide tents for migrant workers?

Honestly, we will never know. In my cynical view, the governor probably compared the outcomes of the news showing unharvested crops vs dead migrant workers in November, and made a career decision. After all, there are going to be plenty of dead people, so migrants will just blend in; but unharvested crops at a time of potential shortages and an election is a career killer.

That is one decision, about one part of how this virus is being handled. Inslee is pumping out these proclamations at a few per day.

At the state level, no decisions can be made that are good for everyone. This is even worse at the federal level. The distance between decision maker and the situation on the ground is too great.

The government can't fix this for you. They can't make good, rational decisions as you would. Their interests are not your interests. Their toolkit is not your toolkit. Their values are not your values.

Now, if you want to apply your toolkit, and make better choices, the real question is why didn't you?

Go back and read the Ebola thread. Is there anything we should have done to prepare for coronavirus that wasn't covered in that thread? If you read that thread 6 years ago, why would this piddly little virus be any kind of a concern? This should have been a solved problem for EREers.

And if you did read it, and didn't prepare for today, why would you be surprised that some administrators did pretty much the same thing?

Why would you expect them to do better next time?

Their interests are not your interests...

[7Wannabe5]

@Riggerjack:

I largely agree with your take, but there is a limit to personal preparedness. For instance, when you find yourself shitting blood, running a fever of 102, with a resting pulse of 124, and your doctor informs you that your choice is risk of sepsis without IV antibiotics vs. risk of exposing yourself to the Covid suspect realm of the emergency room. That’s when your mitigation plan gets multiplied by the rest of society’s average tendency towards wiping their nose on somebody else’s shirt.

[Riggerjack]

I'm certainly not suggesting that a public health issue is entirely an individual responsibility.

What I am trying to get at, is that this is not a situation that CAN be managed. It can be weathered. Some folks will need more help than others. And the best preparation is only preparation.

But to expect the government to manage the situation is a complete failure.

Look at the daily numbers of new cases, in any region. They are all over the place. Just counting takes days to get an accurate day's case count. One day's count being double the previous day's count is common.

At what point does a "leader" decide that you have a real trend, not noise? And how does she know what restrictions are appropriate, given that you have spotted a real trend? And what are the odds that she picks the right restrictions, and are able to enforce said restrictions?

This ignores simple complications like special interests, counter reactions from people who don't want to comply (people like me.), or don't believe in the dangers (cuz old bottles of Lysol said something about coronavirus back in the 50's so it's all a scam), or for simple political affiliation ("The Man" is wearing the wrong shirt, we should defy her on principal!)

How long does it take to make those restrictions stick? How much has the infection rate changed in all this time? Were the right restrictions 2 weeks ago still the right restrictions, now?

AND SERIOUSLY, WHY ALL THE FLIP FLOPPING?!?

......

I'm not a guy with any sympathy for politicians. I'm not saying to go easy on your elected officials. I'm saying that to expect impossible results, is to court dissatisfaction. And while some folks are looking for a new crop of dissatisfaction, by habit or inclination, that still leaves them unprepared for the world as it currently exists.

And that unprepared state makes them virus bait, and our current situation worse.

[jacob]

I am thinking either “An Introduction to Mathematical Modeling of Infectious Disease” by Li or “An Introduction to Infectious Disease Modeling” by Vynnycky and White. They’re both kind of pricey, but this is an important topic.

That is some serious dough at ~$60 a piece. Do you have access to them?

The Li book seems to have everything we're looking for. Yet, I can't tell if it's analytical treatment or a computational treatment. He sticks to simple models (SIR) and stability analysis which suggests the former. That's mathematically interesting but less interesting in practical terms.

Vynnycky also begins with models based on diff eqs like SIR but also adds stochastic/discrete models. There's more focus on demographics and infection between age groups. The book is restricted to human transmission, so no malaria, but lots of HIV modelling. I'm thinking this is the mathematically gentle one. It also looks to have more meat than Li(?)

Amazon also lists https://www.amazon.com/dp/3319974866/ as a "frequently bought together" book. The front cover looks familiar and if true, I've read a similar R book in another field, which I liked. Looking inside it's about my speed. Using R means not having to DIY the numerical algos. Conversely, not building your own algos might miss out of some pedagogical points.

A more hardcore book might be https://www.amazon.com/Modeling-Infecti ... 012HTV65E/ but that's also @#$@#$ $$$$. Now is likely the worst time to buy books about infectious diseases. Even the Laurie Garrett books are pricey---they used to be <$1 used + shipping.

[7Wannabe5]

@jacob: I think the Vynnycky and White book would be best. I did the first course in R for Data Science from John Hopkins on Coursera, but do not consider myself competent. Unfortunately, I don’t know of any way to get a cheaper copy, but the kindle edition is only $39.99. I am currently attempting to read Paul Romer’s paper on endogenous technological growth, so I figure it can’t be any tougher than that.

Also, what !fun! it would be to make my own grim little model projecting deaths next month in my zip code based on traffic volume this week.

[steveo73]

The fact that geniuses make mistakes is not a license to promote your or my judgment over that of experts where it is convenient to you or me, or is it?

Definitely not. I think it is though completely obvious that we cannot trust scientists predictions of the future. They get it wrong all the time. They get it wrong more often than they get it right. It's really about being rational and not emotive. It definitely shouldn't be about you or me. It should be evidence based and transparent.

I think we need to be really careful when it come to abdicating our ability to critically evaluate the science and we need to be really careful about abdicating our ability to make rational policy decisions. If anything I'm stating let's take a scientific approach to evaluating the science. I'm also stating we need to have the proper framework in mind when making societal decisions and it's pretty clear we need a risk management based approach rather than an abdicate all responsibility and just trust the scientists or let's not listen at all to the scientists. The scientists are actually a diverse group with differing viewpoints. The science typically isn't so clear cut and deterministic.

How many people on here have promoted the line that COVID-19 progresses as per an exponential model, These people I'm pretty sure were stating it's the science and it's the math and they were dead wrong. Why not admit they got it wrong. That is science. You observe and adjust your opinion/conclusion to represent observations of reality.