Friday, November 20, 2009

Perfect Storm?

Originally posted at Rhiza:

D225G popping up throughout Europe.
H274Y in a cluster in Scotland.
Low yield of vaccine, and large numbers of people reportedly won't take it when it becomes available.

Will E627K start showing up?

We could be in for rockier times soon should the various concerning changes become dominant, or at least widespread. And no chance of herd immunity for the foreseeable future.

Wednesday, November 18, 2009

Rethinking a previous post.

This blog has been stale for too long, but I have had RL to deal with, including one known case of swine flu in my house and another suspected one. I will need to blog on those another time. But first:

My previous entry described a speculation I had about how an increase in transmissibilty could lead to a more severe pandemic. But now I have some new thoughts on that.

The biggest news in the pandemic lately has been the rapid and apparently severe outbreak in Ukraine. Deep in that long thread you will read about an apparently large number of deaths involving severe lung disease. We now have a possible explanation of that. New sequences from Ukraine released at GISAID indicate a relatively new polymorphism (for swine flu) showing up in the lungs of deceased patients: D225G.

I do not (at this time) know the exact significance of this, but of interest is the following:

You have the general concept right, but the specifics wrong. The general concept is mixing and matching of genetic changes via recombination. Thus, a given change like H274Y, which confers Tamiflu resistance, jumped from one seasonal flu H1N1 background to another until it found a powerful driver. Thus, H274Y wasn't the powerful driver. It was the hitchhiker and for seasonal H1N1, A193T was the driver. Such drivers are frequently receptor binding domain changes. A similar event happened with another viral resistance marker, S31N on M2 of seasonal H3N2. Once again the viral resistance (Adamantane) marker was the hitchhiker, and receptor binding domain changes were the driver. For S31N in H3N2 that was position 193 again (S193F) as well as 225 (D225N).

Thus, position 225 has been in the "driver" position, which helped lead to the widespread distribution of S31N in seasonal H3N2. That same change is in swine H1N1, which is why all swine H1N1 is adamantane (S31N) resistance.

However, in addition to hitchhikers jumping from one background to another via recombination, drivers can also jump from one background to another to create a combination that offers a strong selective advantage. On this thread I described D225G, one of three changes at position 225 in swine H1N1 (the other two are D225N and D225E).

So far I have described D225G hoping onto a China genetic background, as well as a jump in Australia (anto a Singapore/Japan backbone.

I had also previously described its appearance on yet another backbone in Sau Paulo, where it was in samples from lungs of fatal cases.

The recent jumping of D225G onto so many backgrounds raise concerns that it has done the same in Ukraine, leading to transmission or tropism changes and the associated lung destruction and death.

It appears that D225G may lead to an easier infection of the lungs. I had previously thought that the apparent prevalence of primary viral pneumonia in this pandemic was due to the presence of E627 from avian influenza, and not E627K from human influenza. This particular change affects the virus' thermostat, E627 to E627K would lower the ideal replication temperature to one that matches the temperature in the human nose versus the intestines of a bird. Therefore, I used to think that the spread of E627K ( would make the virus easier to transmit but less virulent. Now I am met with the possibility of increasing the transmissibility of E627K and increasing the ability of the virus to attack the lungs via D225G. The spread of these two polymorphisms therefore warrants additional vigilance.

Tuesday, September 1, 2009

Does swine flu need to become more lethal?

A news story,discussing the possibility of the pandemic H1N1 becoming more virulent or transmissible via reassortment with circulating human influenzaviruses is spreading around the internet and flublogia now.

I think they may be looking in the wrong direction. The "common knowledge" of 1918 is that there was a relatively mild form of the virus that broke out in the late spring of 1918 and spent the next several months spreading around the globe, to be replaced, essentially, with a more lethal strain for the second and third waves of fall and winter 1918-1919. Unfortunately there are really no solid samples of the 1918 virus, and little to nothing to indicate how it evolved after introduction (see Wikipedia So is there any evidence the 1918 virus reassorted in the late summer of 1918 to produce the more lethal strain? It would seem that any such evidence would be tenuous at best.

So what of the mild spring/summer wave in 1918 and what might it tell us about 2009, which also had a "mild" summer wave? From Wikipedia:

Deadly second wave

The second wave of the 1918 pandemic was much deadlier than the first. During the first wave, which began in early March, the epidemic resembled typical flu epidemics. Those at the most risk were the sick and elderly, and younger, healthier people recovered easily. But in August, when the second wave began in France, Sierra Leone and the United States,[39] the virus had mutated to a much more deadly form. This has been attributed to the circumstances of the First World War.[40] In civilian life evolutionary pressures favour a mild strain: those who get really sick stay home, but those mildly ill continue with their lives, go to work and go shopping, preferentially spreading the mild strain. In the trenches the evolutionary pressures were reversed: soldiers with a mild strain remained where they were, while the severely ill were sent on crowded trains to crowded field hospitals, spreading the deadlier virus. So the second wave began and flu quickly spread around the world again.[41] It was the same flu, in that most of those who recovered from first-wave infections were immune, but it was now far more deadly and the most vulnerable people were those who were like the soldiers in the trenches - young, otherwise healthy, adults.[42] Consequently, during modern pandemics, health officials pay attention when the virus reaches places with social upheaval, looking for deadlier strains of the virus.[41]

The first thing of note is that it is already established that the 2009 virus has a predilection for the young and healthy. But beyond that, we have two pandemic viruses that started as relatively mild spring/summer outbreaks, which many would not be able to distinguish from a seasonal human influenzavirus. Furthermore, we now have the WHO reporting as follows: New York Times

WASHINGTON (Reuters) - Doctors are reporting a severe form of swine flu that goes straight to the lungs, causing severe illness in otherwise healthy young people and requiring expensive hospital treatment, the World Health Organisation said on Friday.

Some countries are reporting that as many as 15 percent of patients infected with the new H1N1 pandemic virus need hospital care, further straining already overburdened healthcare systems, WHO said in an update on the pandemic.

Now, I would submit that there is nothing new here, this virus has always had the ability to hospitalize those numbers and kill a significant portion of them. So what would make this virus worse? Perhaps no added lethality is necessary. Perhaps it needs to transmit just a little better:

US News

Their conclusions: The 2009 H1N1 virus is spreading at a rate comparable to the 1957 and 1968 flu pandemics -- the most recent pandemics prior to this year's swine flu -- and to the SARS (Sudden Acute Respiratory Syndrome) outbreak, which surprised the world in 2003.
"Even if flu has a reproduction only a little bit above 1 [in this case 1.5], it has its effect because, in a susceptible population, it can start jumping from person to person within one to two days," said study co-author Dr. Nathaniel Hupert, co-director of Cornell University and Weill Cornell Medical College's joint Institute for Disease and Disaster Preparedness. He is also director of the Centers for Disease Control and Prevention's Preparedness Modeling Unit.
"What that means is, three days later, you've got an additional half a person infected. In three days, each of those new people have infected an additional half person, and it's like compound interest. It's the same calculation that lets you grow $1,000 into a $1 million 20 years later."

Suppose the first wave of 1918 was of such a mild nature, not because of the lethality of the virus, but because it couldn't transmit at a high rate?

Study on Transmission Rates for 1918

Recurrent outbreaks of the avian H5N1 influenza virus in Asia represent a constant global pandemic threat. We characterize and evaluate hypothetical public health measures during the 1918 influenza pandemic in the Canton of Geneva, Switzerland. The transmission rate, the recovery rate, the diagnostic rate, the relative infectiousness of asymptomatic cases, and the proportion of clinical cases are estimated through least-squares fitting of the model to epidemic curve data of the cumulative number of hospital notifications. The latent period and the case fatality proportion are taken from published literature. We determine the variance and identifiability of model parameters via a simulation study. Our epidemic model agrees well with the observed epidemic data. We estimate the basic reproductive number for the spring wave R1;=1.49 (95% CI: 1.45-1.53) and the reproductive number for the fall wave R2;=3.75 (95% CI: 3.57-3.93). In addition, we estimate the clinical reporting for these two waves to be 59.7% (95% CI: 55.7-63.7) and 83% (95% CI: 79-87). We surmise that the lower reporting in the first wave can be explained by a lack of initial awareness of the epidemic and the relative higher severity of the symptoms experienced during the fall wave. We found that effective isolation measures in hospital clinics at best would only ensure control with probability 0.87 while reducing the transmission rate by >76.5% guarantees stopping an epidemic.

This study has the transmission rate for the spring of 1919 at ~1.5, the same as for the 2009 pandemic. With the tremendous increase in transmissibility in the second wave, did the virus really need to be that much more lethal, or would it simply have become obvious?

Saturday, August 29, 2009

Lies, Damn Lies, and Statistics

Originally posted on :

So we know the CDC, media, pandemic pooh-poohers like to trot out the estimated 36,000/year deaths in the US as gospel fact, well known. Those who are long time followers of this board know this is garbage and meaningless when dealing with a pandemic that strikes the young instead of the old. But what if we use their statistics against them? Here is the study that came up with 36,000 deaths/year:

The study includes a mortality by age table. Ages over 50 account for over 97.5% of flu deaths, according to the study, and children under 1 account for 0.25% Ages 5-49 are lumped together and account for the remaining ~2.25%. This makes it difficult to determine the percentage of mortality for those under the age of 18. For this analysis I will assume just under half, so all pediatric mortality will account for ~1.35% of total flu deaths.

Thus far, there are about 55 pediatric deaths associated with the pandemic virus. With these deaths accounting for ~1.35% of total mortality, that yields total US mortality so far ~4074. I challenge anyone to tell me why my analysis is any less valid than those who compare 36,000 estimated deaths (with <2000 actually identified in any given year) to the ~550 reported thus far, which is 500% higher than any other year where the CDC has reported officially counted deaths. I don't even know if gsgs could argue my statistics. ;)

Please use this information as needed.

Thursday, July 30, 2009

Musings on the gradual development of a highly lelthal strain and simultaneous emergence in multiple locations.

This is related to my observations below about more countries seeming to be listing their first deaths. It is taken from a discussion on in a member's area so I can't link into it. I'll edit it a little bit to remove references to the discussion it was part of.

One thing I am noticing, qualitatively, not quantitatively, is that deaths worldwide seem to be on the increase. Now there's certainly many reasons why this might be, and a large part of it could simply be due to an increase in the number cases, more cases will necessarily be more deaths. But is that all there is to it? Has much of the rest of the world reached whatever "critical point" in number of infections is necessary to start seeing an increase in the number of deaths? Or is there something else going on?

Apparently the CDC still thinks there is no change in the virus versus the vaccine consensus, but as we have now seen, as long ago as a month an apparently poorly understood but potentially very important change was discovered in Brazil. If this change was fit then we have no way of knowing how far it spread since then, and apparently we are not seeing sequences released in a timely manner anywhere, hence we really don't know what we are looking at now.

The lethal second wave struck three ports on the Atlantic Coast of three continents on or about August 22nd (Boston, MA, Brest, France, and Freetown, Sierra Leone). Shipping traffic can't have delivered the lethal strain to them all simultaneously, there is a fair chance that whatever ship(s) involved would have had a significant portion of their crew killed if they carried the lethal strain. But here is a different view.

Twentieth century pandemics (and possibly the Russian Flu of 1889) had a mild first wave that saw the virus get seeded around the world. In the case of 1918, the much more lethal second wave seems to have hit multiple places at roughly the same time. Perhaps this is what happened. As it spread around the world, the 1918 pandemic virus would have encountered other flu viruses and had chances to acquire various traits. As troops and war materiel moved around the world, these different strains would have spread back to other places, meeting up with different strains of the pandemic virus, or still other flu viruses, picking up additional traits. One of these, still not terribly lethal, but almost there, then got spread to these port cities. Three different ships from a single source could have done it, and perhaps it could have been South America, that being one location where the travel time would have been close to equal for all three destinations. With perhaps only one or two relatively small changes left, normal evolutionary pressures could have produced the more lethal strain simultaneously in different locations, and the new strain was much more fit, allowing it to spread.

So is Brazil pointing us that way? Perhaps they are the source of the new infections in Southern Mexico and Central America. Perhaps the "Sao Paulo strain" isn't the "final" more lethal version, but it could be one more step on the way thee. But with airline travel, the final version could easily crop up in multiple locations at the same time. We are looking for an immediate source, to areas where H5N1 is widespread, for instance. But whatever offspring H1N1 and H5N1 might produce will, perhaps, not be distinguishable from the current pandemic strain until a few more key changes crop up, and then it may already be worldwide.

The above was purely speculation, and I may be totally off-base.

Wednesday, July 1, 2009

HIPAA And disclosure of underlying conditions

I recently posted this on

Most announcements of deaths in the US have been light on discussion of what constitutes an "underlying condition." Generally, privacy laws are cited. In the US this is predominantly HIPAA (Health Insurance Portability and Accountability Act of 1996). But what exactly does HIPAA protect and it is really being used to hide some unsavoury facts. I am citing Wikipedia in this discussion, but what I see in the cited article more or less matches my understanding as a hospital employee in a non-direct patient care area; I am trained in this yearly for regulatory reasons. Other who work in health care can hopefully add to this discussion. First, HIPAA itself:

The Health Insurance Portability and Accountability Act (HIPAA) was enacted by the U.S. Congress in 1996. According to the Centers for Medicare and Medicaid Services (CMS) website, Title I of HIPAA protects health insurance coverage for workers and their families when they change or lose their jobs. Title II of HIPAA, known as the Administrative Simplification (AS) provisions, requires the establishment of national standards for electronic health care transactions and national identifiers for providers, health insurance plans, and employers. It helps people keep their information private.
The Administration Simplification provisions also address the security and privacy of health data. The standards are meant to improve the efficiency and effectiveness of the nation's health care system by encouraging the widespread use of electronic data interchange in the U.S. health care system.
Privacy Rule
The Privacy Rule took effect on April 14, 2003, with a one-year extension for certain "small plans". The HIPAA Privacy Rule regulates the use and disclosure of certain information held by "covered entities" (generally, health care clearinghouses, employer sponsored health plans, health insurers, and medical service providers that engage in certain transactions.) It establishes regulations for the use and disclosure of Protected Health Information (PHI). PHI is any information held by a covered entity which concerns health status, provision of health care, or payment for health care that can be linked to an individual.[10] This is interpreted rather broadly and includes any part of an individual's medical record or payment history.
Covered entities must disclose PHI to the individual within 30 days upon request.[11] They also must disclose PHI when required to do so by law, such as reporting suspected child abuse to state child welfare agencies.[12]
A covered entity may disclose PHI to facilitate treatment, payment, or health care operations,[13] or if the covered entity has obtained authorization from the individual.[14] However, when a covered entity discloses any PHI, it must make a reasonable effort to disclose only the minimum necessary information required to achieve its purpose.[15]
The Privacy Rule gives individuals the right to request that a covered entity correct any inaccurate PHI.[16] It also requires covered entities to take reasonable steps to ensure the confidentiality of communications with individuals.[17] For example, an individual can ask to be called at his or her work number, instead of home or cell phone number.
The Privacy Rule requires covered entities to notify individuals of uses of their PHI. Covered entities must also keep track of disclosures of PHI and document privacy policies and procedures.[18] They must appoint a Privacy Official and a contact person[19] responsible for receiving complaints and train all members of their workforce in procedures regarding PHI.[20]
An individual who believes that the Privacy Rule is not being upheld can file a complaint with the Department of Health and Human Services Office for Civil Rights (OCR).[21][22] However, according to the Wall Street Journal, the OCR has a long backlog and ignores most complaints. "Complaints of privacy violations have been piling up at the Department of Health and Human Services. Between April 2003 and Nov. 30, the agency fielded 23,896 complaints related to medical-privacy rules, but it has not yet taken any enforcement actions against hospitals, doctors, insurers or anyone else for rule violations. A spokesman for the agency says it has closed three-quarters of the complaints, typically because it found no violation or after it provided informal guidance to the parties involved."[23]
Now, there is a segment in there that describes protection of PHI (Protected Health Information). This would be the key part we are concerned with here. Using broad strokes, PHI is any of the data necessary to tie information back to an individual, either medical record information (such as name, dates of treatment) or payment history.

An expansion of PHI:

Protected health information (PHI), under the US Health Insurance Portability and Accountability Act (HIPAA), is any information about health status, provision of health care, or payment for health care that can be linked to an individual. This is interpreted rather broadly and includes any part of a patient’s medical record or payment history.
List of 18 Identifiers according to HIPAA 1996:
1. Names;
2. All geographical subdivisions smaller than a State, including street address, city, county, precinct, zip code, and their equivalent geocodes, except for the initial three digits of a zip code, if according to the current publicly available data from the Bureau of the Census: (1) The geographic unit formed by combining all zip codes with the same three initial digits contains more than 20,000 people; and (2) The initial three digits of a zip code for all such geographic units containing 20,000 or fewer people is changed to 000.
3. dates (except year) for dates directly related to an individual, including birth date, admission date, discharge date, date of death; and all ages over 89 and all elements of dates (including year) indicative of such age, except that such ages and elements may be aggregated into a single category of age 90 or older;
4. Phone numbers;
5. Fax numbers;
6. Electronic mail addresses;
7. Social Security numbers(SSN);
8. Medical record numbers;
9. Health plan beneficiary numbers;
10. Account numbers;
11. Certificate/license numbers;
12. Vehicle identifiers and serial numbers, including license plate numbers;
13. Device identifiers and serial numbers;
14. Web Universal Resource Locators (URLs);
15. Internet Protocol (IP) address numbers;
16. Biometric identifiers, including finger, retinal and voice prints;
17. Full face photographic images and any comparable images; and
18. Any other unique identifying number, characteristic, or code (note this does not mean the unique code assigned by the investigator to code the data)
Here is where it gets interesting. Note number 2 in the list. If HIPAA is what is really being used to "protect" the data then all announcements we have are in violation of it since we have more information than state or the other narrow exceptions listed. For all of these deaths we have city and/or county, both of which are protected.

Furthermore, statistical information is reported to, for instance, the CDC and they publish it publicly, and this includes various "underlying conditions." Statistical information is not, apparently, protected, and that is really what we are after here. Some statistical information is available, but parts are being withheld, and I am having an increasingly difficult time reconciling that to the (real) need to protect privacy. There are so many cases where information that really is protected has been released (patient names and/or hospitals), so the decision not to release relevant information is extremely puzzling.

Tuesday, June 23, 2009

Tracking US Deaths versus Deaths in the Healthy

Much is made in official pronouncements as well as in the media about how much increased risk of severe illness and death there is from people who have pre-existing medical conditions if they also contract swine flu. That might be true with seasonal human influenzaviruses. Problem is, this is not a human influenzavirus.

In the US, seasonal influenza is said to kill 36,000 people every year (need source), of which roughly 90% are elderly. It is already borne out by the CDC and the WHO that deaths from swine flu are much higher in the young category than is typical in flu season.

Data from the US also show an increase in deaths from the healthy. I have put together a Google spreadsheet tracking US deaths over time as well as the percentage of total deaths that are in people who do not have underlying medical conditions.

There are some problems with this. I used as the source for my data a different spreadsheet mentioned in a thread on (link to original needed) It seems rather complete. The issue is that firm death dates are not given. When no death date was given I tried to use the lag time from other reports from that state to determine where to set a death date. If I couldn't make a determination, I made an assumption of 3-4 days lag. So other people might come up with slightly different graphs, but the trend and final number should be a good match.

To determine if a person was healthy I looked at the remarks. A null comment is assumed to be healthy. If an underlying condition is marked disputed I assumed a healthy individual. Given the risks of pregnancy in this pandemic I assumed pregnancy to be a valid underlying condition.

Here is my spreadsheet:

And here is a chart showing the increase in the death rate in healthy individuals:

There is clearly an upward trend, and we may soon break 50% of all US deaths being from healthy individuals. It is probably true that an elderly person (who is most at risk from seasonal flu) has one or more underlying conditions, so the deaths that made up that graph are probably in young and middle aged people.