Saturday, May 02, 2009

Swine flu

Like I wasn't going to make a post about this? =)

I don't have sage advice to offer; I'm not a trained medical professional nor do I specialize in emergency preparedness. I am not advocating panic in the least, since the strain seems to be currently able to be effectively treated in the United States with antiviral medications.

But I do have a brain, and I thought that since enough time had passed that it was good to offer some thoughts on this...

I want to analyze this outbreak from a numbers perspective. Bear with me on this. I am using data from the CDC report from 2007-2008.

Seasonal influenza in the United States is tracked by the CDC. Not all suspected cases are tested because there are simply too many. Last year, for the 2007-2008 flu season, the CDC reports the following:

225,239 specimens tested
39,827 of those specimens were positive
18% of specimens tested were positive. That's a fairly large sample size, but not the entire sample size; this is only the sample that actually underwent laboratory testing at a CDC certified viral laboratory.

We will take a tiny, tiny leap here, and assume that the number of pediatric influenza deaths the CDC reported for the 2007-2008 season is based on specimens that underwent laboratory testing.

The number of pediatric deaths from influenza for 2007-2008 is 83.
83 deaths/39,827 positive specimens=0.21% chance of death for children with seasonal flu.

Kids are fairly susceptible to the flu, but we can get a better estimate on normal mortality of the flu for EVERYONE from the CDC's statement that the epidemic threshold was exceeded using their 122 Cities Mortality system. It peaked last year at 9.1% mortality (or cause of death in the people that died during flu season), but remember, that's not based on actual laboratory testing--it's based on a survey of death certificates in those 122 cities and which ones note the cause of death as either pneumonia or influenza. Not the most reliable of methods, but it's the data we've got.

Let's assume though, that normal chance of dying from the flu isn't that much. Other pages on the CDC site state that about 16 million people each year are diagnosed with influenza-like illness (not verified all the time by lab tests), and of those, using our fun mortality systems, we get that approximately 36,000 die each year.

Rough, rough chance of dying from the flu then is 36,000 deaths/16 million cases=0.225% overall. Remember that number, since we're going to come back to it shortly. Please note that it's statistically not that far off the pediatric mortality rate for last year, which we calculated at 0.21%.

Now let's hop over to our neighbor to the south, Mexico, which has a large number of suspected cases and some number of confirmed cases. We're not going to even bother with "suspected" cases here, we're going to use the numbers that have actually undergone laboratory analysis, according to the World Health Organization. See the NY Times article here for more information about this set of numbers.

In Mexico, they tested 908 cases for swine flu.
Of those 908 cases, only 397 cases turned out to have the particular influenza A H1N1 variant we're calling swine flu.
397 positive specimens/908 specimens tested=43.72% infection rate

Please compare that number to our regular seasonal flu infection rate of 18%.

On top of that, we also get the numbers that there have been 16 deaths out of those 397 specimens.

16 deaths/397 positive specimens=4.03% mortality rate in Mexico

Now, granted, we're dealing with a much smaller sample size than the total number of cases that were tested in the USA for the 2007-2008 seasonal influenza season. The mortality rate is likely to go way down, if they test more cases and figure the deaths against that.

We also have to take into account that Mexico doesn't really have a great infrastructure, and many of those deaths may have gone undiagnosed initially. So the medical care isn't as good there, compared to other places.

BUT... and this is the big "but" in all of this analysis.

Right now, it appears that swine flu is spreading like regular seasonal influenza. It's got a higher infection rate, but that's also based off of a smaller sample size. Statistics are hugely important here. Let's say for the sake of argument, that eventually the swine flu infection rate is found to be 18%. Common infection rate as we've established. And let's say also, that the swine flu mortality rate in Mexico is strongly influenced by their lack of medical care in certain areas of the country. Let's say, that if we didn't know about swine flu at all, and it developed in the US first, that it might have a 1% mortality rate given our much, much better medical care system infrastructure. So far we have no deaths, but that doesn't mean that there won't be deaths in the US and people are currently being uber-vigilant about any influeza symptoms.

16 million cases of influenza a year in the US x 1% mortality rate=160,000 deaths from swine flu, as opposed to the standard 36,000 deaths each season. In fact, if we're extra conservative, we can even say that maybe the US mortality rate will only go to 0.5%, which would mean 80,000 deaths instead of 36,000 deaths.

I'd like to point out that 80,000 is still a much larger number than 36,000. Those are still lives that can be potentially lost. Which is pretty much why the health authorities are ever-so-slightly concerned about swine flu....

.... because of the last thing, that might happen, which they always worry for.

Genetic mutations!

Flu viruses mutate as they pass from person to person. In some seasonal flu years, the viruses undergo more mutations, and that's why a flu year may be worse than normal. The mutations do a variety of things, including change how well the body's cells can defend themselves against the virus (the 1918 flu pandemic had a doozy of a mutation), and how easily communicable the strain of the virus is.

So far, no mutations. Which is a good thing. =)

But if there were to be mutations, they could do a lot of unexpected things to the strain of swine flu virus. Not all of them good. Most of them bad. Including, as mentioned above, making the virus affect people with better immune systems (thus causing cytokine storms like in 1918) or it could make the virus pass from person to person more easily or it could just make it stick around in our bodies longer and be a bitch to cure.

We don't know what will happen--this may be nothing now, and then come back to hit us next year. It could be nothing forever. It could stay dormant in the flu virus strains for years, and then come back thirty years from now, as is the nature of diseases.

So it's nothing to worry about now--not unless there are mutations to the virus strain or we start having more deaths in the United States or the antiviral medications stop helping alleviate the symptoms. It's not a cause for panic, but I think it is necessary for you to understand the math behind it all, and to make your own informed decisions.

A 0.225% mortality rate isn't much. A 1% mortality rate still isn't great. The 1918 flu pandemic had a 2.5% mortality rate, when all was said and done, with the majority of those deaths being young healthy adults because of the mutation that caused the cytokine storms. And so I look at that 4.03% mortality rate in Mexico and I wonder... =)

Feel free to weigh in with your own thoughts.

13 comments:

PurpleClover said...

We've had one death in the US in Texas. But I think the reason we haven't had more deaths is because it hasn't reached a larger immunosuppressed population (hospitals or nursing care facilities).

There are two anti-virals that are working in the US currently that Mexico most likely doesn't have access to. But these injections will only work as long as the virus hasn't mutated. :)

So it has the potential of being a pandemic for sure. All we can do is hope that we can reduce the numbers by spreading the facts and information.

Jenny Rae Rappaport said...

I wasn't counting the one death in the US since it was a Mexican toddler who died while visiting across the border in Texas. And while I am saddened for his family, for it surely must be terrible to lose a 23 month old... it's not really applicable to the statistics right now.

And the cases are slowly spreading, so I think it will eventually hit a larger population here. My husband is sick now, but he insists it's not the flu. I remain slightly skeptical and feed him chicken soup. (Btw, any flu, swine or regular seasonal is sucky for me, since I have fibromyalgia which weakens my immune system. Yay for auto-immune diseases! =)

P. Bradley Robb said...

According to two independently conducted computer simulations, the worst case scenario for the spread of this particular flu strain is 1700 infections within the United States. These studies took into account normal human travel patterns.

http://news.yahoo.com/s/livescience/20090501/sc_livescience/worstcasescenarioforfluestimated

If we also consider that Type A H1N1 is lacking an important amino acid which contributes to the speed with which it can reproduce, this number becomes much more realistic.

If we look at a real world case - the dead toddler in Texas - and realize that the rest of his family has not tested positive for the flu at all [ http://www.honoluluadvertiser.com/article/20090429/BREAKING/90429008/-1/RSS01?source=rss_breaking ], and it would seem that despite Type A H1N1 achieving the technical qualifications for a pandemic (human-to-human transmission in three or more countries is more than 2 continents), the spread of the disease seems to fall short of the traditional definition - affecting a large geographic and large amount of the population.



In other words, it would appear that our end of the world superbug has been drastically overrated.

Jenny Rae Rappaport said...

P. Bradley Robb, that is interesting about the toddler's family being fine.

What interests me is that if you listen to the data they used for the simulation, they're not totally sure of the accuracy of their data. They did say that without any measures taken there would be 1700 cases for the worst-case scenario by the end of May, but in the Yahoo article, the researcher at Northwestern also says that the numbers get substantially worse if you run the worst-case scenario out through a full four month spread.

I'm not saying the superbug is going to kill us all, but it is interesting, which is why I'm following it.

P. Bradley Robb said...

@Jenny

True, but this was from another AP article pulled off the wire about an hour ago:

"The virus has been circulating for over a month in a city of 20 million of high population density. It could have been much worse," said CDC epidemiologist Marc-Alain Widdowson.

Waterman agreed that the virus does not appear to match the ferocity of past killers. "Most people think it is unlikely this is going to be as virulent as the 1918 epidemic. From what we know so far, it doesn't seem like it is as virulent," he said."

When the data is viewed through a lens of exponential growth factor, it would seem that the current spread rate is far less than X^2.

The healthiness of the toddler's family is especially telling, as they were traveling and living with the child in a confined space during the entire incubation and symptomatic periods. This adds evidence, at least anecdotal evidence, that though the disease is potentially deadly, it has a hard time being transmitted from person to person.

As it seems now, the Swine Flu is just a story that has significant legs within the media community. The rather heavy pollen count we're getting here in Virginia is no doubt helping to fuel that fire.

KLo said...

I just posted a blog addressing Swine Flu but approaching from a different direction (namely, its impact on workplace employees). Here's the link, if you or your readers are interested:

http://philosophyofklo.blogspot.com/2009/05/stay-home-if-swine-flu-is-possibility.html

Thank you for the great facts presented here : )

Martha Flynn said...

I was all set to not panic until you got to the part about it coming back in 30 years possibly stronger than ever. :)

Jenny Rae Rappaport said...

Martha Flynn, it's sort of just the nature of diseases. We thought we eradicated smallpox, but there's tiny pockets here and there, sometimes. Flu strains regularly change and mutate from year to year. We had a pretty good fight against TB and Staph, etc, and now we're back with "superbugs" because of antibiotic-resistant bacteria. It's what makes them so neat to track and learn about, I think. And we won't even get started on the resurgence of polio or common childhood diseases because people are not vaccinating their children...

P. Bradley Robb, I agree, it is overblown in the media. And I like the statement you quoted off the wire, plus I went and looked at the other simulation and I like their data better. =) They seem to be using a really interesting method and they have a whole website about it. It's called GleamViz.

I just like waiting to see and if things mutate... it's my latent genetics love popping up. =)

Iapetus999 said...

Right now it's just a race between the disease spread and vaccine production. Once people are vaccinated against H1N1, all those rates you mentioned will drop like a rock. Of course, mutations can render the vaccine useless, or some other virus can step forward while we focus all our efforts on H1N1.

PurpleClover said...

You're right about the boy in Texas. I knew he was Mexican but thought he was a resident. My mistake.

As for the flu - in order for the Tamiflu injection to be effective you must have it within 48 hours of the onset of symptoms. Otherwise, it will not be able to effectively decrease the viral load. So I strongly urge everyone to not hesitate to go to their doctors...even if you are healthy it doesn't mean your spouse, children, friends, parents, or siblings are!

Okay I'll stop being a student nurse. Now I'm a blog follower:

Thanks for the stats and info!!!

Alexander Field said...

I'm all about simplistic answers when it comes to something like this. In other words don't panic, and: http://www.doihavepigflu.com

suelder said...

During the 1918 influenza, Philadelphia did not take precautions and lost 13,000 citizens to the disease in a matter of weeks. (During the month of October.)

St. Louis had one-eigth the peak mortality rate. They both instituted the same public health precautions, but St. Louis implemented them much more quickly - two days after the first cases appeared. Philadelphia waited a couple of weeks.

http://www.nih.gov/news/pr/apr2007/niaid-02b.htm

And this "Spanish" flu leapt the species from birds to humans in Kansas (that's what they believe now), was sent to Europe with troops during WWI and came back, mutated and more virulent.

I'm glad that this strain doesn't appear that deadly, but I approve of the measures taken. I would much prefer to stop it now, before it mutates into a killer.

People will say that the media hype over H1N1 is overblown. But if it works, we'll never know if the precautions kept it contained or it wasn't that big of a risk.

Someday, I'm going to write a plague novel. This is just too fascinating.

Authoress said...

EXCELLENT post, Jenny!

I might add that the majority of deaths from the 1918 pandemic were actually from secondary infections (strep, mostly) and not from the flu itself.

You deserve some chocolate for this well thought out post! =)