Is the swine flu going to be pandemic?

Looking into Debora MacKenzie's credentials who wrote that New Scientist article I found she is a "science writer" and writes volumes on multiple topics from the economy to the KGB to bird flu. I have not yet found her credentials but they look to be strictly journalism. OTOH, she is proscience even if she doesn't get all the facts right.

A reader's comment regarding another article DMacK wrote about H5N1: "This is lopsided piece, based on flimsy evidence"

Looks like she may be the same DMacK that has a news cast on the BBC.
 
Last edited:
I was just playing. See what I quoted below:



Just for giggles I took "one" to be "future" rather than "antibody test" as you intended.

ETA: Chillzero caught it. I really should catch up completely before posting.
And I really try hard to get those pronoun objects correct.

I might have figured it out if I wasn't so tired and annoyed at the time. Sorry.
 
Skeptigirl -

Let me run this back to you to see if I understand what you're saying in a general sense.

Suppose I write a computer virus that sends e-mails to 50% of the people in the address book on an infected computer. Everybody who opens that e-mail is infected with the virus. In turn the virus sends out e-mails to 50% of the people in the address books of those people.

A week after I launch my virus, 10 people are confirmed to have been infected by my virus including me (just keeping it very simple). As we examine the address books of those 10 people, we learn that each and every one of them was either in my address book or in the address book of someone in my address book. In a loose sense we can say with some level of confidence that we know the rough extent of the infection.

However, suppose we examined those 10 people and found 3 that could not be traced in some way to the other 7. Those 3 people tell us that our level of infection must be greater than 10. Somewhere there are people we don't know about who got infected and spread the virus to those 3 people.

Again, keeping it simple, suppose we accept my numbers as accurate:
* 50% of the people in the address book get the e-mail.
* Only some percentage X of the people who get the e-mail actually open it.
* It takes N days on average for those people to open their e-mail.
* Everyone who opens the e-mail gets infected.
* There is some duplication of names in the infected address books since some percentage Y of the friends and business asscociates listed in any given address book know each other.
* There are 10 known infections.
* 3 of the infections are not linked to the others.

We could then create a model to estimate the actual number of infections, which by definition must be greater than 10.

Am I following your point at least in a rudimentary manner?
 
These findings, along with other suggestive surveillance data presented below...
Since this appears to be about as close as you're going to get to answering my question, "which other data?" -- let's have a look at some of this "other suggestive data":

Syndromic surveillance data from hospital emergency room chief complaints suggest that an increasing number of individuals are being seen in emergency departments for influenza-like illness (ILI). The largest increase has been seen in patients complaining of fever in the 18-44 year old age group. It is not known whether this increase reflects changes in care-seeking behavior or a true increase in community ILI.

Okay, so we have patients complaining of fever. That's data. Is that data about what this virus is doing? Not without making some as yet unproven assumptions.

You have a limited concept of what 'data' is
More limited than yours, clearly. I will continue to argue that by its very nature, "data" is a narrow concept. It has nothing to do with me. You are, of course, free to use the term any way you please.
...and what conclusions are possible to draw from epidemiological data.
I'm quite well aware of what conclusions may be drawn from the data. And it is perfectly reasonable to do so. It's just that I find it important to recognize which are the data and which are the conclusions. A fundamental difference we had in another (fairly) recent thread is that you seem to feel that at some point it is reasonable to treat the conclusions as if they were the data, while I prefer a more cautious approach.

If you have hundreds of students in a school sick with similar symptoms in a very short window of time and you find 80% of 50 tests done were positive for a single virus strain, that is data most of them have the same thing.
No, that is inference. I'm not saying it is not valid to make that inference, but to insist that it be called "data" is ridiculous.

What is pure speculation is to say you can predict how many severe cases you will or will not see in an infection which has only so far spread to several hundred people at the very beginning of an epidemic.
Now there's something we agree on. Can we predict that there should be at least a significant increase in clinical cases if the thing has been going on long enough that the true number of infected people is in the "tens of thousands"?

We know that because that hypothesis has been tested many many times and confirmed to be a valid assumption for an epidemiologist to make.
I will also argue that "assumptions" are not data. And I've found real epidemiologists to be quite reserved about using language like "confirmed to be a valid assumption to make" -- especially when it comes to influenza. On the contrary, they invariably go to great lengths to emphasize influenza's unpredictability. I find it a valid assumption to make that when you see overconfidence and arrogance, it's a safe bet that you're dealing with a person who doesn't know as much about influenza as they think they do.

I highly recommend if you are truly interested in learning something here rather than just confirming what you think you already know, that you read the whole NYPHD page and see all the evidence both objective and subjective, as well as established epidemiological models that these public health epidemiologists use to assess contagious disease outbreaks.
Woah, the whole page? Skeptigirl, you are talking to a guy who is absolutely obsessed with influenza surveillance methodology. I have spent so many hours out of the last six or seven years pouring over influenza related literature that I am frequently the butt of jokes among people who know me. I can clear out a room just by mentioning the subject.

Here's something I'm truly interested in learning: what exactly is the difference between, say, the HA molecule of an H5 subtype and that of an H1 subtype? Want to impress me with your expertise? Answer that one.
 
UncaYimmy,

I think you have it, but I found Skeptigirl's explanation far sexier. ;)
Sorry.

BJ
 
Regarding inferences...

Dymanic,

I'm not going to argue with you because I am a mere novice here, but:

You cannot get confirmed cases in dozens of distant geographic locations all over the world in a single week from the small number of confirmed cases reported in Mexico. You cannot make that epidemiological model work.
You cannot have a virus contagious enough to infect over 200 students in a school in 4 days which genetically matches that confirmed in Mexico City crowded with millions of inhabitants and expect to only see 1,000 of those citizens [in Mexico City] infected. You cannot make that epidemiological model work.


Strictly, they may be inferences, but they are inferences you could stake your life on. Wouldn't you agree?

BJ
 
Skeptigirl -

Let me run this back to you to see if I understand what you're saying in a general sense.

Suppose I write a computer virus that sends e-mails to 50% of the people in the address book on an infected computer. Everybody who opens that e-mail is infected with the virus. In turn the virus sends out e-mails to 50% of the people in the address books of those people.

A week after I launch my virus, 10 people are confirmed to have been infected by my virus including me (just keeping it very simple). As we examine the address books of those 10 people, we learn that each and every one of them was either in my address book or in the address book of someone in my address book. In a loose sense we can say with some level of confidence that we know the rough extent of the infection.

However, suppose we examined those 10 people and found 3 that could not be traced in some way to the other 7. Those 3 people tell us that our level of infection must be greater than 10. Somewhere there are people we don't know about who got infected and spread the virus to those 3 people.

Again, keeping it simple, suppose we accept my numbers as accurate:
* 50% of the people in the address book get the e-mail.
* Only some percentage X of the people who get the e-mail actually open it.
* It takes N days on average for those people to open their e-mail.
* Everyone who opens the e-mail gets infected.
* There is some duplication of names in the infected address books since some percentage Y of the friends and business asscociates listed in any given address book know each other.
* There are 10 known infections.
* 3 of the infections are not linked to the others.

We could then create a model to estimate the actual number of infections, which by definition must be greater than 10.

Am I following your point at least in a rudimentary manner?
With certainty you could say the infection was wider spread than what was visible, and your contact tracing missed a line(s) of infection somewhere.

In epidemiology we would be working backward most of the time, however, not forward.

So you would send out the virus. Then you would survey the population. For each infected computer you would then trace it back to its origin. When you find unrelated cases now you have a problem. They didn't get infected through the air. There has to be infected computers in the chain that you have not detected.

To calculate how many exist that you didn't detect requires lots of information you didn't include here. So you could make that model, but you need more information.

There is the rate of infection. Maybe you have enough data to figure that out.

There is the population you surveyed, how big was it? What percentage of the population was infected?

Somewhere there had to be a population you did not survey.

And so on. But you have the right idea.


I did a really rough calculation (totally guessed numbers based on seasonal flu stats) on the cases in Mexico and came up with needing at least 120,000 cases of flu to see 150 deaths given the population of Mexico City, the average attack rate for pandemic flu and the typical # of deaths per case of flu you would expect.

That suggests we need to see a few thousand cases before we'll know what kind of fatality pattern we can really expect here. And there is no way a couple hundred cases is enough to go on.
 
2009 H1N1 is not going to be a pandemic, nor will any other one based on the sound logic and obvious wisdom in this thread from another forum.

GodGunsGuts said:
The CDC is a waste of money. Science and medicine should be returned to the private sector where they belong. Otherwise, we will continue to be treated to government created panics year after year over ever impending pandemics that never materialize.
DannyTN said:
The funny thing is, if the CDC does it’s job and prevents pandemics, then you will always be right. They will always be predicting pandemics that don’t materialize.
GodGunsGuts said:
They don’t prevent anything. The standard of living that free market capitalism has given us prevents the types of plagues that we used to see in previous centuries.
 
Since this appears to be about as close as you're going to get to answering my question, "which other data?" -- let's have a look at some of this "other suggestive data":
You've attributed the quote from the NYC health department to me. I just want to correct that. And you've taken it out of context. But, moving on....

Okay, so we have patients complaining of fever. That's data. Is that data about what this virus is doing? Not without making some as yet unproven assumptions.
Unproven assumptions? Are you suggesting the NYCPHD needs to prove to you they know what they are looking at? Because I know what they are looking at and there are no unproven assumptions involved. This is an example where you just don't know what you are doing, you only think you do. You don't have a broad enough knowledge base in epidemiology or infectious disease, there's no nicer way to put it.

Think about it. You are trying to critique expert epidemiologists. I am not suggesting argument from authority, only that if you are going to challenge people with an exponentially higher level of expertise than yourself, your criticisms are very likely going to be naive.

The assumptions here you think are unproven: that the ED doesn't recognize an increase in patients presenting with fever. All EDs I've ever worked in and worked with keep logs and do recognize patterns. In fact, EDs and schools are part of the sentinel sites specifically charged to report certain trends that are known indicators of influenza, newly emerging infections, and bioterrorism events.

If a school has more than 10% absenteeism, for example, it is a mandatory reportable event. The school must notify public health if absenteeism exceeds 10%. Typically this occurs during influenza outbreaks and indicates widespread community infection.

EDs have specific training and plans to notice unusual patterns of disease and reporting rules for when to notify public health. A slew of patients reporting to the ED with fever at the end of flu season would be unusual and very noticeable. Just like 222 kids reporting with fever to the school nurse was noticeable and was reported to public health. It's part of the job to notice and to report such occurrences.

More limited than yours, clearly. I will continue to argue that by its very nature, "data" is a narrow concept. It has nothing to do with me. You are, of course, free to use the term any way you please.
I have run into more than a few people who have a limited view of data because they are less aware of scientific fields outside their own. A few people in fields like physics or astronomy don't realize that other fields of study use different kinds of data and different methodologies but the science is just as successful as the so called 'hard' sciences. I suspect that may be part of your problem here. (But part of it is also misreading, see below.)


I'm quite well aware of what conclusions may be drawn from the data. And it is perfectly reasonable to do so. It's just that I find it important to recognize which are the data and which are the conclusions. A fundamental difference we had in another (fairly) recent thread is that you seem to feel that at some point it is reasonable to treat the conclusions as if they were the data, while I prefer a more cautious approach.
What ever you think you remember about a past conversation, find it or I have to assume you remember incorrectly because I know the difference between data and conclusion.


skep said:
If you have hundreds of students in a school sick with similar symptoms in a very short window of time and you find 80% of 50 tests done were positive for a single virus strain, that is data [which supports the conclusion] most of them have the same thing.
No, that is inference. I'm not saying it is not valid to make that inference, but to insist that it be called "data" is ridiculous.
I said that in a confusing way but you misread it. "That" refers to the stuff behind it. You are misreading it to refer to the thing in front of it. Try re-reading it with the inserted clause that clarifies what I was saying. The fact they all had the same symptoms, the fact they had those symptoms in close proximity in time and place, the fact the sampled kids had 80% positive tests for the same virus --- that was the data. The conclusion was most of them likely had swine flu. I wasn't clear in how I said it but I did not confuse data and conclusion. You have a confirmation bias that led you to assume I don't know what I am talking about. It likely contributed to you misreading what a pronoun referred to.


Now there's something we agree on. Can we predict that there should be at least a significant increase in clinical cases if the thing has been going on long enough that the true number of infected people is in the "tens of thousands"?
Not sure what you mean. The tens of thousands are cases in Mexico. We will have tens of thousands of cases here in another week or two, maybe 3 at the most if the current contagion pattern holds.

I will also argue that "assumptions" are not data. And I've found real epidemiologists to be quite reserved about using language like "confirmed to be a valid assumption to make" -- especially when it comes to influenza. On the contrary, they invariably go to great lengths to emphasize influenza's unpredictability. I find it a valid assumption to make that when you see overconfidence and arrogance, it's a safe bet that you're dealing with a person who doesn't know as much about influenza as they think they do.
Again with your confirmation bias. You don't make much sense here because it's a non sequitur to anything I posted.

I haven't confused data with conclusions. You seem to be confusing the concept of a fact. All data are not facts and many conclusions are. Erroneous data is still data even though it is not factual. By the same token, a conclusion one is fairly certain of is often called a fact.

You need to go back and reread what I've posted without that confirmation bias that is leading you to misread it.


Woah, the whole page? Skeptigirl, you are talking to a guy who is absolutely obsessed with influenza surveillance methodology. I have spent so many hours out of the last six or seven years pouring over influenza related literature that I am frequently the butt of jokes among people who know me. I can clear out a room just by mentioning the subject.
Well if that is the case, why do you have such little knowledge of the epidemiology here? How can you be obsessed with influenza surveillance and not know about EDs reporting sentinel events or what confirmed cases indicate or the fact it was reported in multiple CDC and WHO reports that Mexico's numbers were only serious cases while the US and Canada were reporting all cases?


Here's something I'm truly interested in learning: what exactly is the difference between, say, the HA molecule of an H5 subtype and that of an H1 subtype? Want to impress me with your expertise? Answer that one.
The genetics and proteins involved here are interesting especially now that research goes on at the molecular level. But unless you follow the details like a geneticist or virologist would, a more general explanation is usually sufficient to get the gist of it.

Essentially there are various proteins on the surfaces of different cells in the respiratory tracts of humans and other animals. The hemagglutinin proteins have a best match with specific cell proteins so that they attach most readily to the cells they are adapted to.

The surface proteins of the influenza virus are easily mutable and 1-16 hemagglutinin types represent genetic differences resulting in protein differences at the sites where the virus binds to the cell surface to get in. So H5, for example currently doesn't bind well to the upper airway cell proteins in humans. People with the infection have virus replication in the lung. It's highly lethal but doesn't result in lots of viral shedding out the airway.

If only 1 or 2 nucleic acid substitutions occur on a specific gene, however, the resulting protein change will allow the virus to more easily enter cells higher up in the human bronchial tree. In chickens the H5 protein enters multiple cells in multiple organs so there must be some more universal protein on multiple cells but that is just a guess on my part.

In ducks there is some mechanism that specifically interacts with the flu virus so that upon passing through the duck the virus emerges attenuated for ducks. I don't know the mechanism but I found that fact fascinating.

The neuraminidase OTOH is an enzyme. Have you seen the video that was used without permission in the movie, Expelled? The video they ripped off is a really interesting visualization of what proteins and cell membrane interactions are like.

http://multimedia.mcb.harvard.edu/media.html

Click on any of the top 3 versions of Inner Life and watch the complexity of the interacting of the cell proteins and structures. While it is not influenza or the H or N functions it will give you an idea what they are all about.

There are a couple other videos at the site I don't have the FireFox plugins to play but they look interesting if you do. Try the ones on antibodies and analyzing proteins though I don't know if they are cool animations like the "inner life" is.
 
Last edited:
I just hope this is overrated and will flunk like sars did.
Flunk?

This is the kind of post that boggles the mind. From the WHO

There were 1706 health care workers who got SARS out of the 8096 total people infected. 774 people died of which ~200 of them were health care workers.

Before that I was aware of and disturbed by a case of Lassa Fever that resulted in the deaths of 29 health care workers in the 70s. This single outbreak not only infected and killed all these health care workers, several of them killed their family members when they inadvertently brought it home to them.

Just because it didn't make it to a city near you thanks in part to the incredible response by those health care workers and the public health agencies involved is no reason to think it was no big deal. It was a very big deal.
 
State Lab Swamped With Swine Flu Samples
Only a fraction of the nearly 2,500 nasal swabs sent to the state health department lab have been tested.

The Austin lab had tested 181 of the 2,492 nasal swabs received from counties around the state by Friday, Department of State Health Services spokesman Doug McBride told the Austin American-Statesman for Saturday's editions.

The lab's single testing machine became overwhelmed Wednesday when samples began flooding in from around the state. Three more machines are being added, more staff is being trained to run on two more shifts and the lab will work seven days a week to catch up, McBride said.

That hospital's emergency room received twice as many patients Thursday as usual. On Friday, the hospital set up three air-conditioned tents with cots to treat overflow patients in a covered parking area.

Swine flu swamping state labs
MILWAUKEE -- A hundred cases of swine flu in the U.S.? Health officials say there are likely more. Just how many is not important, they say.

As the world faces a potential pandemic, swamped labs are not testing all possible cases. Getting an exact tally has taken a back seat to finding new outbreak hot spots or ways to limits its spread, health officials said.

"The specimens are coming in faster than they can possibly be tested," said Dr. Jeffrey P. Davis, state epidemiologist in Wisconsin, where a lab helped spot the nation's first known case, in a 10-year-old boy from San Diego.

New York, which has more cases than any other U.S. location, also has had to limit the samples it tests, said Dr. Don Weiss of that city's health department.

"Sure, we'd want to diagnose every case, but we don't have that resource," he said. Instead of trying to confirm every sign of the virus, "we're focused on where else is it going and how do we prevent it."

Confirms exactly what I said about why the WHO is not confirming every case but rather would be prioritizing confirming locations the virus had spread to.

And overwhelming the hospitals, just wait until next week and lets see how many serious cases are being reported then. We're still on week 2.
 
Last edited:
Confirms exactly what I said about why the WHO is not confirming every case but rather would be prioritizing confirming locations the virus had spread to.

So, once it's established that a particular location has the virus, why bother collecting more swabs from new cases? Are they really going to eventually check every last one of them?

And overwhelming the hospitals, just wait until next week and lets see how many serious cases are being reported then. We're still on week 2.
All quiet in Australia so far, with no confirmed cases.


But from 13 hours ago:

http://www.abc.net.au/news/stories/2009/05/03/2559192.htm?section=world

Mexico's health minister said authorities might, on reflection, have overestimated the danger of the virus when it first appeared.
Researchers in the United States are also playing down the threat posed by the virus.




"What we have found by looking very carefully at the sequences of the new H1N1 viruses is that we do not see the markers for virulence that were seen in the 1918 virus,"
 
Well, flunk might be the wrong word.

We got little paper masks to hand out to everyone coming onboard in Hong Kong, and it were made up to be the next spanish flu.
It was loose in Hong Kong, a very densely populated place with poor hygiene.
How many died from regular flu in the same time period?

I would not call that succeding as the next great plague.
 
We will have tens of thousands of cases here in another week or two, maybe 3 at the most if the current contagion pattern holds.
Ok, here's mine:
We continue to see new cases and new clusters here and there throughout the summer, but interventions (both pharmacological and non-pharmacological) are sufficient to keep the R0 at a low enough level that it fails to produce community-wide outbreaks... until the onset of the regular flu season, when all bets are off. At that point -- IF it reappears at all -- it might fail to obtain enough traction to compete with the seasonal strains, or it might reassort with them, producing a whole new virus. Most likely one with resistance to Oseltamivir. Let's wish the vaccine developers and manufacturers the very best of luck in their endeavors.

Well if that is the case, why do you have such little knowledge of the epidemiology here?
Because I don't concur with all of your conclusions? That's an interesting metric. Do you often find that the epidemiologists themselves are always in firm agreement? Because I don't.

Essentially there are various proteins on the surfaces of different cells in the respiratory tracts of humans and other animals...
Woah, woah. You're talking about antigens. But each subtype is capable of shifting its antigens across a wide enough range that you can have one virus (swine H1N1) which is so antigenically different from another (seasonal H1N1) that we call it a "novel" virus. But it's still the same subtype. Understand the question now?

Have you seen the video that was used without permission in the movie, Expelled? The video they ripped off is a really interesting visualization of what proteins and cell membrane interactions are like.
Of course. I love that video. But don't talk down to me. My idea of a good time is to step up to a chalkboard and diagram the structure and functions of the hemagglutinin, neuraminidase, and M2 proteins for a rapt audience (which I find to be rather few and far between, for some strange reason).
 
I'm not going to argue with you...
Aw, but I came here for an argument!

Strictly, they may be inferences, but they are inferences you could stake your life on. Wouldn't you agree?
I'd prefer not to stake my life on inferences, but as one seldom enjoys the luxury of having perfect information, there doesn't seem to be much choice. It's ultimately about deciding whether to err on the side of daring or on the side of caution -- a decision which involves calculating the relative costs of being wrong. Break into a run every time you hear a rustle in the bushes, and you waste a lot of extra calories. Sit tight until you know for sure that it's a tiger, and you may pay for that certainty with your life.
 
Regarding predictions, I don't think any of us can make any until we know more about the attack rate.
If it really is very close to "regular" H1N1, it might die out completely over the summer (here...who knows what might happen in NZ and OZ over the next several months. It might mix and mingle with the 1950/1977 H1N1 over there and pop out a hybrid that is antigenically very different, but clinically fairly similar to "regular" H1N1, to circulate over here next winter.)

If that's what happens, similar things have probably happened many times before, and has been a silent factor in seasonal flu antigenic shifts that have happened over time.

Or, maybe we'll just have a summer flu season this year, in addition to the one that's going away right now. But that would be extremely unusual, to the point where I can't really critisize dymanic for finding it highly unlikely.
But this is flu, so you never know.

According to the 33 year average cycle, we are "due" in 2010 for a pandemic of "something new". But a new H1N1? That's just not in line with any of the previous thinking on flu virology and epidemiology.
But - this is flu, again. So who knows?
 
So, once it's established that a particular location has the virus, why bother collecting more swabs from new cases? Are they really going to eventually check every last one of them?

All quiet in Australia so far, with no confirmed cases.


But from 13 hours ago:

http://www.abc.net.au/news/stories/2009/05/03/2559192.htm?section=world
Typical news reporting of particular aspects of the unfolding events. The Mexican government has every reason to claim the epidemic isn't so bad after all. Tourism likely has come to nearly a standstill.

OTOH, while I'm sure it's BS there are really less cases in Mexico than at first reported (see the last 10 pages of posts) they probably are correct in that the case fatality rate is not as high as first thought. It's a funny paradox here. If you understand the disease, less cases in Mexico is actually worse. That could mean a higher case fatality rate and a more dangerous disease. Much better to have many unrecognized but otherwise mild cases.

As for not having the genetics of the 1918 virus, that's been obvious for a while now.
 

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom