Flu Vaccine For Everyone: When the Science Contradicts the Science

[Skeptic Ginger]

Originally Posted by kellyb

SG is trying to play "U R ignorant, I R authority" while reading/encountering for the first time stuff (fairly complicated stuff, to be fair) that kellyB and Este have been arguing about in detail with each other (off and on) for almost a decade now.

To be fair, you posted a number of incorrect things. Maybe you should look at why that is, instead of worrying about how well the pissing is going.

As for the condescending nature of your comment, pot meet kettle.

[kellyb]

What did I post that was incorrect?

[kellyb]

Originally Posted by Skeptic Ginger

The science isn't that complicated.

Wrong:

http://jid.oxfordjournals.org/content/206/5/625.extract

Quote:

Estimation of the mortality burden of winter-seasonal pathogens is notoriously difficult.

Not that you'll listen to Lone Simonsen, or know who she is, or not just knee-jerk write her off as a nut.

[MarkCorrigan]

I'm well aware you're a nurse. You've used that one before to set yourself up as an authority.

[Estellea]

Originally Posted by Skeptic Ginger

You've provided no evidence for the bolded claim, only assertion and a misunderstanding of graphs you are looking at.

Here is that quote:

Quote:

Originally Posted by Estellea View Post
ILIs are used as indicators of flu activity and no, not a reliable indicator given the numerous other respiratory pathogens co-circulating.

I have and you have provided that evidence; you are simply resistent to the idea so let me explain more thoroughly (or dumb it down in your words). The CDC collects respiratory illness data from sentinel sites although a passive system. Virologic specimens are collected as well as ILI outpatient data. They then use the ILIs and lab confirmations, which are not population based and have limited demographic information and ILIs are a proxy to estimate influenza burden. Now what do you suppose they use for the denominator? The EISN does the same. I don't know why you are so resistant to the fact that flu epidemiology is messy business, partly inherent to the pathogen of interest and partly due to collection methods. I would think that you, "an expert" would see this and be cautious about interpreting these data and advocate for better collection methods. Even the CDC admits to the gaps and is trying to improve the system.

Quote:

You'd have to research the EU's surveillance system for their timeline. I've been looking at those flu culture proportions for more than a decade. In-office flu screening tests are a more recent development (if you call 5+ years recent). But not the sentinel culture monitoring in the US. And now that we have wider use of in-office flu tests, they correlate with past data.

1999-2000 surveillance data.In that same summary:Of course you don't believe we actually understand that.

The WHO has had a formal sentinel flu monitoring system in place since 1952. They sample specimens from worldwide sites and culture them to monitor the shifting and drifting genetics of circulating strains.
WHO Interim Global Epidemiological Surveillance Standards for Influenza (July 2012)

Yes, we know and it still falls back to poor methods to estimate flu morbidity and mortality.

Quote:

I've not seen any evidence in this thread from you or kellyb showing ILI spikes above epidemic threshold that didn't have an influenza component. You showed graphs of viral spikes, but you don't understand what you are looking at.

Funny because you were the one who provided some very nice graphs in your own state demonstrating just that. Kelly also provided some good data demonstrating the burden of non-influenza ARI disease in other age groups. And now you want to dismiss your own data and pretend that others just don't possess your intellectual capabilities to know what we are looking at? Convenient.

Quote:

You, like kelly, are misunderstanding how to compare two different data sets. Just like P&I are not derived from the same data set as ILIs, neither are the viral cultures you two are looking at derived from the same set.

Take kelly's link: "Weekly laboratory test result data" showing RSV. That is not the same data set as ILIs or the ILI surveillance where a subset of ILI specimens are tested for influenza.

I guess you aren't aware that both are collected from the NRVESS and have the same sentinel sites? And even if they aren't the exact same specimens that they can provide a comparable estimation of lab-confirmed ARIs.

Quote:

I don't know how else to explain it to you, neither of you seems interested in actually understanding why you are making the mistake you are making.

That's okay, I think we can muddle through.

Quote:

A subset of viral cultures ordered by doctors on patients that are more seriously ill is not the same subset as the sentinel surveillance of ILI specimens.

Um yes they are overlapped.

Quote:

A huge increase in RSV does not cause a huge increase in total ILIs because infants make up a small proportion of the total ILIs. ILI rates and RSV rates are simply not derived from the same sub-population of patients.

Again, you and Kellyb both provided data that would indicate otherwise and also your ignorance that they are all within the same surveillance network.

Quote:

In 20 years I've never seen an ILI spike above the epidemic threshold in the winter that didn't have a substantial influenza component. Yes, there are plenty of ILIs that are not influenza. And of all the people with ILIs, the majority typically don't have influenza. Those elevated rates of ILIs are built in to the epidemic threshold which is higher in the winter.

And you would know this how? You mean like right now that:

Quote:

Nationwide during week 4, 4.2% of patient visits reported through the U.S. Outpatient Influenza-like Illness Surveillance Network (ILINet) were due to influenza-like illness (ILI). This percentage is above the national baseline of 2.2%.

Outpatient ILIs are above the baseline (epidemic) yet the majority of those ILIs are not influenza?

Quote:

Every september-october in the US we see a community spike in URIs (upper respiratory infections) that are not due to influenza. It happens when the kids go back to school. Most of those illnesses are not included in the ILI reports because most of those people do not have a fever >101F, nor do the bulk of them seek medical treatment.

Yea so?

Quote:

But even if the ILI threshold is exceeded when influenza is not circulating, it doesn't negate the usefulness of monitoring ILIs. Your whole premise is that we count all ILIs as flu cases AND WE DON'T DO THAT! For whatever reason, you and kellyb can't seem to let go of that misinformation.

And that would be yet another strawman. Neither me nor kellyb has said that all ILIs were counted as flu cases nor did we say that ILI monitoring was useless. Could you please let go of that misinformation?

Quote:

Predicting the Epidemic Sizes of Influenza A/H1N1, A/H3N2, and B: A Statistical Method
I.e. ILI data is useful. Imagine that.

Not only have you said that ILIs were not used for estimating influenza morbidity (they are) but you now use that as some kind of sabre and you really should slow down and read your own cites instead of quote-mining. The authors explain that their model is very limited due to only two seasons audited and how numerous (unrealistic) conditions would have to be met for their model to work consistently so have to rely upon a proxy measure, of which ILIs are part. Although they did find some interesting stuff about interantigenic competition. Something I believe you dismissed kellyb about? Kind of like Maurice Hilleman being "bizarro".

Este

[Estellea]

Originally Posted by Skeptic Ginger

What evidence do you have that the circulating flu strains are substantially different within the northern hemisphere between continents during the same flu season?

Odd, I didn't use the word "substantially" now did I? And here is a difference that would absolutely effect a survey's external validation:

Quote:

Few 2009 H1N1 viruses have been detected in the United States, but the virologic picture in Europe is different, with the virus accounting for close to a third of influenza A viruses that have been subtyped.

Quote:

While some strains emerge locally thus vary temporarily (like the 2009H1N1v when it first emerged in Mexico) or are confined to small regions due to great efforts (like the HPAI H5N1 which remains confined at the moment) the majority of what circulates are the same strains across the n or s hemispheres during a single epidemic.

Strain differences and dominant strain differences will absolutely effect vaccine efficacy survey's generalisability to other countries.

Quote:

But not only that, the strains are closely monitored and genetic drift is well documented in real time.

And a big so what when you are evaluating flu vaccines used in other countries that have differences in influenza epidemiology and trying to apply that to a different set of circumstances.

Quote:

And some of the vaccines differ, but not all of them.

Oh FFS could you stop being so bloody contrary just for the sake of being contrarian. Most EU vaccines are different than what is available in the U.S. That and usage are enough to make comparisons very difficult. And more importantly, it's something the U.S. should have done adequately prior to making recommendations.

Este

[Estellea]

Originally Posted by Skeptic Ginger

Or yours does.

I'm adhering to the best available evidence, not worshipping authority.

Quote:

It wasn't a straw man it was an analogy. Bottom line, Jefferson makes a false charge that any knowledgeable professional is confusing ILIs with raw flu counts.

Here is a previous quote of yours regarding Jefferson's "false charge":

Quote:

The Debate About Flu ShotsIt's an outright lie. He might be conflating what the public and news media confuse for flu, but not the health care community. The more I read from Jefferson, the more I think he's full of his own confirmation bias when it comes dismissing the morbidity and mortality of influenza. I can't find any other explanation for his claims.

And another from you which sheds light on why you think this:

Quote:

What about the 7-15%? Well, if he is talking about during flu season, which is the only relevant number, then he's contradicting years of collected data that is specifically CULTURE CONFIRMED INFLUENZA. I more than substantiated that claim in my last page of posts.

So what's the deal? Why is he wrong? Because he clearly is. Did he cherry pick his data? Did he completely ignore the EU and CDC's weekly culture confirmed influenza? I can't figure it out. But I do know the evidence in all the culture confirmed data overwhelmingly demonstrates he is wrong. He doesn't cite the source of that 7-15% number.

Jefferson is not wrong with his charge although he does overstate it somewhat. But but but you believe that the lab-confirmed specimens represents flu prevalence and you would be wrong so very wrong. So tell me, what is the annual flu prevalence range if you think he's wrong. And for the last time (I can only hope), it isn't "culture-confirmed", it's "laboratory-confirmed".

Quote:

If onset is past the initial 24 hours, antivirals like Tamiflu are not useful. If a child is hospitalized for RSV, they are likely to get ribavirin. If/when a secondary bacterial pneumonia develops an antibiotic is indicated.

So "treated the same" is something you might tell a patient on the phone avoiding a lengthy explanation. Not all people with influenza need to be seen by a doctor. So?

Way to miss the point. I'm not interested in your "medical expertise" nor is it relevant. The point was to relay an anecdote of how the medical community will often jump to a flu "diagnosis" just because it's "flu season" in spite of your assertion that your lot don't make those kinds of judgements. There were no less than three different parents who relayed a similar story to me two days ago.

Este

[Skeptic Ginger]

Originally Posted by MarkCorrigan

I'm well aware you're a nurse. You've used that one before to set yourself up as an authority.

I'm a nurse practitioner. I see and treat people, I prescribe medications. I know my job. This thread is specifically within the field of medicine I've been practicing for 20+ years.

But I stand by my posts, not my authority.

[Estellea]

Originally Posted by Skeptic Ginger

Then what's your problem with ILIs?

Both of you claim without evidence that ILIs are not a good flu correlate.
Both of you claim without evidence the morbidity and mortality flu burden is overestimated because it is based on ILIs and P&I mortality. ILIs and P&I mortality is only one of many data sets that goes into determining the morbidity and mortality flu burden.

Actually, you have provided most of the evidence of the inherent difficulties residing with ILIs as a proxy for flu activity and P and Is as a proxy for flu deaths; you just don't or can't accept/understand it. All of the statistical voodoo in the world isn't going to change that. We need better surveillance and additions to ICD coding for starters to provide more accurate estimates of influenza morbidity and mortality.

Quote:

Then you can learn something about how you've been incorrectly comparing different data subsets.

Or yanno, you can stop embarrassing yourself by reading your own material and playing flu epi goddess. Kellyb has provided (and vetted) some good material here; it's too bad you would rather try and gaslight her instead of carefully considering the evidence and flaws in our current understanding of some disease epidemiology.

Quote:

Both of you claim incorrectly that RSV cases are being counted as flu cases, and that seasonal RSV epidemics are a significant cause of ILIs exceeding the epidemic threshold in winter.

They are as per your own cites and additional ones by kellyb.

Este

[Dymanic]

Originally Posted by Skeptic Ginger

I see and treat people, I prescribe medications.

And yet you are not familiar with the guidelines for use of Tamiflu. That's troubling.

[Estellea]

Originally Posted by Skeptic Ginger

Here's a study with some important findings:
Sentinel Surveillance of Influenza-Like Illness in Two Hospitals in Maracay, Venezuela: 2What this indicates is detecting viruses isn't easy. Many of the 85% of specimens that no virus was found are false negatives. (We know this because antibody increases are found more often than antigen.)



Jefferson's claim that influenza is a minor pathogen is not evidence based. I'll get back to his paper tomorrow.

You're basing this on one study in a small geographical region with a lot of limitations as noted by the authors? Let's take a look at some of those statements:

Quote:

While influenza viruses were observed to be the most prevalent viral pathogen, adenovirus, parainfluenza virus, and RSV were important causes of ILI at the two hospitals (Table 2). In Venezuela, the NESS reported RSV as the most frequently detected non-influenza respiratory virus followed by parainfluenza virus and rhinovirus [17]–[21]. The study performed in Zulia, Venezuela, also reported RSV as the most common detected virus, followed by adenovirus, parainfluenza virus and influenza viruses [22]. As previously mentioned, the differences in the detection rates may be attributed to the different procedures used by the NESS, other Venezuelan researchers. [22] and us. These viral pathogens were also frequently detected in other surveillance studies in Central and South America [14]–[16].

And

Quote:

Our study had limitations worth noting. Data collection at only two hospitals in an urban area limits our ability to generalize our findings to the population. The low virus detection rates may be attributable to variations in the skills of the health staff employed to collect the respiratory specimens. The sampling method may have a significant effect on the proportion of respiratory viruses identified, and nasopharyngeal washes may yield higher sensitivity over nasopharyngeal or oropharyngeal swabs [32]. Study participants were exclusively seen in the outpatient setting, thus limiting our ability to examine the impact of respiratory viruses in hospitalized cases. A large proportion of ILI cases were not associated with any pathogen, and the impact of bacteria on this clinical syndrome cannot be determined from this study, since the respiratory samples were not cultured for bacteria. Two methods of detection were used for identification of influenza (PCR and culture) whereas only one method of detection was used for the other viruses (culture). Twenty-two percent (9/40) of the respiratory samples which were positive for influenza viruses by RT-PCR were negative by viral isolation illustrating that viral detection by culture underestimated the true prevalence. Viral culture may not be the ideal way of isolating organisms such as RSV, hMPV, HBoV and rhinoviruses leading to significant underestimation of their detection rates [33]. Despite these limitations, our study contributes information on the distribution and etiology of ILI at two hospitals in Maracay, Venezuela. This knowledge can serve as a baseline for future, more expansive population-based surveillance studies of influenza and other respiratory viruses in this region.

Just to head you off of the inevitable strawman, this survey is far from useless but demonstrates the vast differences in results without standardised, population-based surveillance.

Este

[Estellea]

Originally Posted by Skeptic Ginger

You don't know what you're talking about. Antibodies and the reagents to test for them come in a gazillion flavors.

Here's one example of a very general test I can order on a patient. Quest is the main lab I use in my practice:Only if the titer is between 1:8 and 1:32 is a second specimen needed.

Here's a variety of tests one can use for research purposes:
World's largest influenza research reagent manufacturer 300+ influenza research tools: protein, antibody, ELISA kit, gene cDNA clone

She does know what she is talking about here. She is correct in her description of sero-surveys that can be very useful (moreso than what we are currently using) to determine influenza morbidity. So what that there are a multitude of antigens, are you even familiar with serological assays that oh say, allow us to test "a gazillion' samples in unison? And I've not only "looked up" the various antigens, antibodies, reagents, plates, strips, and conjugates, I've purchased them and used them and even managed to simultaneously test multiple sero samples, antigens and conjugates. Who knew?

Este

[kellyb]

Originally Posted by Estellea

You're basing this on one study in a small geographical region with a lot of limitations as noted by the authors? Let's take a look at some of those statements:

And

Just to head you off of the inevitable strawman, this survey is far from useless but demonstrates the vast differences in results without standardised, population-based surveillance.

Este

Did you look at the tables?

They got a whopping ONE viral positive for ANY virus in the 60+ age group (which had a whopping 7 total participants.)

It seems to me that the survey is very close to useless? (Especially when it comes to extrapolating anything to N America?)

[Estellea]

Originally Posted by Skeptic Ginger

I don't know what you mean by, "an authority". I'm a nurse practitioner who specializes in infection prevention and occupational infectious disease hazards. I give a ~1,000 flu vaccinations/yr and am a paid consultant to a number of health care facilities where promoting health care worker flu vaccinations has been a key component in infection control since 2005.

So it isn't as though you have a very personal vested interest in maintaining your appearance of authority in your profession.

Quote:

I started the thread because one team of Cochrane Reviewers (an important evidence based medicine source) headed by Dr Tom Jefferson has published a number of meta-analyses that suggest flu vaccinations are not useful in healthy people. However, a very large body of experts in the field do not agree with the findings and Dr Jefferson has very strong opinions that could be resulting in confirmation bias in his work.

But you haven't produced any compelling evidence that Jefferson is wrong with his review conclusions nor methods. Granted you actually have to read the full studies. You are merely appealing to authority, not evidence. And shall I remind you that Dr. Osterholm found the same results and the same poor methodological quality of searched studies?

Quote:

This is a useful place to start: The Cochrane Collaborative’s Tom Jefferson makes the huge mistake of appearing on Gary Null’s show

Of course, when you don't feel like or can't do the heavy lifting involved with critiquing the work, just ad hominem attack. We get it; he's kind of a jerk. Will you please move along?

Quote:

Este, kellyb and I believe Professor Yaffle as well, find the Cochrane Review to be authoritative while I find the rest of the medical community with expertise in influenza to be a more reliable source given Dr Jefferson's less than objective opinion about influenza and influenza vaccine.

There is good reason why I consider the Cochrane Reviews on flu vaccine effectiveness authoritative. I first encountered the review years ago on healthy children and being somewhat suspicious of some of Dr. Jefferson's claims I read it with a very critical eye. But it is what it is and I couldn't find any glaring problems with the methodology nor conclusions. Public health policy is being based upon no good evidence of flu vaccine effectiveness in children (and other groups for that matter) and I take exception to that. Think on it and why you are so resistant to accept replicated evidence.

Quote:

If I reply to a post that has incorrect information in it, like the posts misinterpreting surveillance data or the most recent one with incomplete information about influenza antibody testing, I have tried to support my position with evidence. If someone wants to whine that is acting as a know-it-all, I refer to my posts with the very specific reasons and sources for why I am calling something misinformed.

Your hubris precludes you from even knowing what incorrect information is but dig away.

Este

[Estellea]

Originally Posted by kellyb

Did you look at the tables?

They got a whopping ONE viral positive for ANY virus in the 60+ age group (which had a whopping 7 total participants.)

It seems to me that the survey is very close to useless? (Especially when it comes to extrapolating anything to N America?)

Yes I know but given the state of affairs in some countries, it's a start but in no way valid to say the EU or U.S.

Este

[Skeptic Ginger]

Originally Posted by Dymanic

And yet you are not familiar with the guidelines for use of Tamiflu. That's troubling.

What do you base that on? What incorrect prescribing guidelines for Tamiflu do you think I misstated?

[kellyb]

Originally Posted by Skeptic Ginger

What incorrect prescribing guidelines for Tamiflu do you think I misstated?

Both of them:

Originally Posted by SG

(3 days in the antiviral drug does very little if anything)

It's after 48, not 72 hours when the 0 (or even negative) correlation pops up.

Originally Posted by SG

If onset is past the initial 24 hours, antivirals like Tamiflu are not useful.

Again, 48, not 24.

You're all over the place, hitting every marker besides the one that's correct.

[Skeptic Ginger]

Originally Posted by Estellea

Here is that quote:

I have and you have provided that evidence; you are simply resistent to the idea so let me explain more thoroughly (or dumb it down in your words). The CDC collects respiratory illness data from sentinel sites although a passive system. Virologic specimens are collected as well as ILI outpatient data. They then use the ILIs and lab confirmations, which are not population based and have limited demographic information and ILIs are a proxy to estimate influenza burden. Now what do you suppose they use for the denominator? The EISN does the same. I don't know why you are so resistant to the fact that flu epidemiology is messy business, partly inherent to the pathogen of interest and partly due to collection methods. I would think that you, "an expert" would see this and be cautious about interpreting these data and advocate for better collection methods. Even the CDC admits to the gaps and is trying to improve the system.

You keep ignoring the fact you are talking about apples and oranges.

What flu burden? Specifically?

The estimate of the number of deaths and hospitalizations caused by flu (AKA flu burden) is based on MULTIPLE DATA SETS.

The ILIs are useful for public health planning, they are an aid to diagnosing a patient.

Comments that we could do better do not tell you the current system is grossly unreliable.

Originally Posted by Estellea

Yes, we know and it still falls back to poor methods to estimate flu morbidity and mortality.

Define "poor methods."

And explain why you keep cherry picking a single indicator and claiming the estimate which is based on multiple data sets is "a poor method".

[Skeptic Ginger]

Originally Posted by kellyb

Both of them:


It's after 48, not 72 hours when the 0 (or even negative) correlation pops up.



Again, 48, not 24.

You're all over the place, hitting every marker besides the one that's correct.

It would help if you read my post.


It works best in the first 24 hours, after 72 it likely does very little. Do you suppose it might be 17 hours for one patient and 52 for another? Or do you think 48 is magical dividing line?

When you see a patient, do you think all patients are good historians? Think they come to you knowing the exact hour their symptoms began?

You demonstrate one of the reasons a person cannot always read a drug insert and know what they should do.

The reason I said 72 hours is that's the cutoff I recall from the study that looked at who died and who didn't.

[kellyb]

Originally Posted by Skeptic Ginger

The reason I said 72 hours is that's the cutoff I recall from the study that looked at who died and who didn't.

I would be very, very interested in seeing that link.

[Skeptic Ginger]

Originally Posted by kellyb

I would be very, very interested in seeing that link.

I think the Jain et al study was one of those I read but the full text isn't free online. It's discussed in this review article that does have full text access.

Quote:

In one prospective study, McGeer et al.²³ demonstrated that oseltamivir treatment (even after 48 h admission) improved 15 day survival in 327 adult patients hospitalized with influenza between 2005 and 2007. Similarly, several different Asian studies demonstrated improved survival among hospitalized patients who were treated with oseltamivir. In the first, Lee et al.^24,25 showed that initiation of oseltamivir within 4 days of onset of influenza-associated symptoms was associated with improved survival in multivariate analysis (n = 356). ...

... Jain et al.³⁹ described a relationship between time from onset of symptoms to initiation of antiviral therapy (almost exclusively oseltamivir) and outcome. The median delay in therapy initiation among all hospitalized patients was 3 days (range 0–29 days; 39% patients <48 h); among those requiring ICU admission, 7 days (range 0–24 days; 23% <48 h) and among deaths, 8 days (range 3–20 days; 0% <48 h).

23 McGeer A, Green KA, Plevneshi A, et al. Antiviral therapy and outcomes of influenza requiring hospitalization in Ontario, Canada. Clin Infect Dis 2007;45:1568-75. doi:10.1086/523584.
Abstract/FREE Full Text
24 Lee N, Cockram CS, Chan PKS, et al. Antiviral treatment for patients hospitalized with severe influenza infection may affect clinical outcomes. Clin Infect Dis 2008;46:1323-4. doi:10.1086/533477.
FREE Full Text
25 Lee N, Chan PKS, Kin WC, et al. Factors associated with early hospital discharge of adult influenza patients. Antivir Ther 2007;12:501-8.
MedlineWeb of Science
39 Jain S, Kamimoto L, Bramley AM, et al. Hospitalized patients with 2009 H1N1 influenza in the United States, April-June 2009. N Engl J Med 2009;361:1935-44. doi:10.1056/NEJMoa0906695.
CrossRef

What that hi-lighted data says is no one who died got Tamiflu within the first 3 days while among the patients who did get Tamiflu within the first 3 days there were no deaths.

Reducing Occurrence and Severity of Pneumonia Due to Pandemic H1N1 2009 by Early Oseltamivir Administration: A Retrospective Study in Mexico

Quote:

Even when administered >48 hours after symptom onset, oseltamivir showed considerable potential for reducing pneumonia. Application of these results would benefit patients affected by future influenza pandemics.

Antiviral Therapy and Outcomes of Influenza Requiring Hospitalization in Ontario, Canada

Quote:

One important difference between our study and others is that the patients in our cohort appeared to benefit from antiviral therapy initiated >48 h after symptom onset. This does not contradict data demonstrating that, in healthy adults, antiviral therapy must be started sooner than 48 h after symptom onset to be of benefit [20, 21]. In otherwise healthy adults, influenza virus is cleared promptly by the immune response. Viral load begins to decrease 24–48 h after symptom onset, and late antiviral therapy is unhelpful [22]. However, patients with severe immunocompromise may not control viral replication for many days [23, 24], and little is known about the time course of viral load in older patients at risk of influenza complications. In our cohort, all treated patients were shedding virus immediately prior to treatment (88% had a positive direct antigen test result), so that specific antiviral therapy might have been expected to be of benefit. Although we did not detect an effect of earlier therapy relative to later therapy, the power of this study to detect such an effect was very limited, and our results should not be interpreted to mean that timing of therapy is not important.

[Skeptic Ginger]

Originally Posted by Dymanic

And yet you are not familiar with the guidelines for use of Tamiflu. That's troubling.

I'm waiting your reply. You are making a serious charge and I think you should follow through.

[kellyb]

I'm only going to address the study that looks at mortality, since that's the endpoint you honed in on.

Quote:

What that hi-lighted data says is no one who died got Tamiflu within the first 3 days while among the patients who did get Tamiflu within the first 3 days there were no deaths.

You're misreading it. From the full text:

Quote:

The median time
from the onset of illness to the initiation of antiviral therapy was 8 days (range, 3 to 20); none of
the patients who died received antiviral therapy
within 48 hours after the onset of symptoms.

And...

Quote:

In a multivariable
model that included age, admission within 2 days
or more than 2 days after the onset of illness,
initiation of antiviral therapy within 2 days or
more than 2 days after the onset of illness, and
influenza-vaccination status, the only variable that
was significantly associated with a positive outcome was the receipt of antiviral drugs within
2 days after the onset of illness.

That 0% you highlighted is saying that 0% of the death cases got Tamiflu within 48 hours or less.

The 3 you're looking at there is the earliest day antivirals were given on among the death cases.

[Skeptic Ginger]

Originally Posted by kellyb

...
You're misreading it. From the full text:
...
That 0% you highlighted is saying that 0% of the death cases got Tamiflu within 48 hours or less.

The 3 you're looking at there is the earliest day antivirals were given on among the death cases.

AMONG the deaths: RANGE 3-20 DAYS AND none were <48 hours.

The reason they added that is they had more than one variable and <>48hrs was a specific variable, as was the range of days to treatments.

And yes, I am reading the data forward and backward. None of the people in this study who died got Tamiflu in the first 3 days, therefore, of all the people who did get Tamiflu within the first 3 days, NONE were among the dead.


Not to mention I added additional links about the >48 hours. I'm not even positive the one I linked to was the study I read, it was 4 years ago. No matter, they all support the same case, 48 hours is not a magic number.

[kellyb]

Do you or do you not still stand by this:

Originally Posted by SG

What that hi-lighted data says is no one who died got Tamiflu within the first 3 days while among the patients who did get Tamiflu within the first 3 days there were no deaths.

Because that is a totally different interpretation than this (your words here, but my [correct] interpretation that I had to spell out for you):

Quote:

AMONG the deaths: RANGE 3-20 DAYS AND none were <48 hours.

[kellyb]

Actually, you're still misreading it.

It says:

Quote:

Of the 19 patients who died, 90% received antiviral drugs, and all received antibiotics. The median time from the onset of illness to the initiation of antiviral therapy was 8 days (range, 3 to 20); none of the patients who died received antiviral therapy within 48 hours after the onset of symptoms.The median time from the onset of illness to the initiation of antiviral therapy was 8 days (range, 3 to 20)

You say:

Quote:

None of the people in this study who died got Tamiflu in the first 3 days, therefore, of all the people who did get Tamiflu within the first 3 days, NONE were among the dead.

The highlighted is a non sequitur. We know that none who died got Tamiflu within 48 hours. And we know that among the dead, none got Tamiflu until 3-20 days after symptoms started showing.
But none of that can be interpreted to mean "of all the people who did get Tamiflu within the first 3 days, NONE were among the dead."

In fact, that 3-20 range means at least one of the deaths was for someone who got Tamiflu within 3 days (but after 48 hours.)

[Dymanic]

Originally Posted by Skeptic Ginger

I'm waiting your reply. You are making a serious charge and I think you should follow through.

Summary of Influenza Antiviral Treatment Recommendations

Clinical benefit is greatest when antiviral treatment is administered early, especially within 48 hours of influenza illness onset.

http://www.cdc.gov/flu/professionals...clinicians.htm

Your statement was:
"If onset is past the initial 24 hours, antivirals like Tamiflu are not useful."

You were just flatass wrong, and I think you should admit it.

[Skeptic Ginger]

Getting back to the incorrectly interpreted data which seems to be the nature of a number of the posted claims, (virology lab reports, P&I mortality, ILI reports and sentinel specimen sampling that is tested for influenza):

UW virology reports, 1990 to present
National P&I mortality, ILI reports and sentinel specimen sampling that is tested for influenza

Look at 3 years of data sets, 1999-2000, 2000-2001, and 2009-2010.

First a caveat re the virology lab data, it should be noted that until 2009, not a whole lot of specimens were submitted for influenza culture. That's because RSV in infants requires a specific diagnosis. Using Ribavirin is not without its own difficulties and you don't usually treat infants with it empirically. Whereas with influenza, even in a hospitalized patient, you can treat empirically and the bacterial cultures looking at secondary pneumonia are more important for choosing treatment.

In the 2009 pandemic a lot more attention was paid to specifically identifying the pandemic strain and more patient cultures were submitted to virology labs from clinicians.


Virology lab data
The other data on one page
In the '99-2000 season there was a sharp spike in P&I deaths around January.
RSV had a high rate (~175 from ~Jan to Mar.) Influenza A numbers from the virology lab peaked in Jan at ~100.
The ILI reports were high and peaked in week 52.
Cultures from sentinel samples showed ~30% flu+ (that's high).

The ILI peak, P&I peak and the 30%+ cultures all correlate.

It might look like the RSV spike correlates but if you look closer, the P&I peak comes earlier and the RSV remains elevated longer. P&I and RSV do not correlate.

Now look at 2000-2001 data.
Virology lab
RVS has the same very high peak of 175 for a shorter time while influenza A counts from the virology lab are low ~25 at the peak.
CDC numbers (you need to read the numbers as the images aren't opening for me)

Quote:

From October 2000 through May 2001, WHO and NREVSS laboratories tested 88,598 respiratory specimens for influenza viruses and 9,962 (11%) were positive.

Despite the high RSV numbers, P&I mortality doesn't exceed the epidemic threshold. (You can see P&I in this graph of a 4 year span.)
ILI reports are unremarkable. (You have to look here at figure 6.)
Flu+ cultures from sentinel sites was only 11%.

The ILI peak, P&I peak and the 11%+ cultures all correlate.
While RSV does not.

It's not just the numerical nature of the denominators, it's the sample they are drawn from.

Now look at 2009-2010 numbers when we had a pandemic that hit young people hard (a smaller proportion of the flu morbidity), and spared the elderly (a large proportion of typical flu morbidity).

The epidemic threshold for P&I mortality is barely breached mid winter. But notice the earlier spike. It's above the epidemic threshold but because it isn't above the mid winter peak (that includes all the mid winter viruses which are not flu) the excess mortality is less obvious.

But ILI reports peak high and early (~week 30-50).
The % of sentinel samples that are flu+ peak high and early (~week 25-45). At one point >40% of the sampled specimens were flu+.
[url=http://depts.washington.edu/rspvirus/documents/VD2009-10.pdf]UW virology shows the huge increase in flu+ results that correlate with the early peak in P&I fatalities and the ILI.
RSV rates have a low slow peak much later in the year, Feb-Apr.

Lab confirmed hospitalizations for flu is telling. The hospital cases are very high in kids under age 4. But they are a small total number of typical flu morbidity. So they don't impact the P&I threshold.

Bottom line, P&I, ILIs, %flu cultures for flu, all correlate. RSV, no correlation at all. And why all the flu cultures at the virology lab? It's because there was an effort to differentiate the 2009H1N1v cases by MDs seeing patients in clinics and hospitals.

If we only looked and P&I, we'd be wondering why the EDs were overflowing. Total flu deaths weren't revealing an important key factor: that this strain was hitting young people hard but sparing the elderly.

At the same time, we did have high ILIs which tells the medical community something's going on. RSV was not an important component of the ILIs exceeding the epidemic threshold.

[Skeptic Ginger]

Originally Posted by Dymanic

Summary of Influenza Antiviral Treatment Recommendations

Clinical benefit is greatest when antiviral treatment is administered early, especially within 48 hours of influenza illness onset.

http://www.cdc.gov/flu/professionals...clinicians.htm

Your statement was:
"If onset is past the initial 24 hours, antivirals like Tamiflu are not useful."

You were just flatass wrong, and I think you should admit it.

You took my comment flatass out of context.

Originally Posted by SG

Originally Posted by Estellea

Just yesterday a neighbour informed me that one of her children is ill and her physician informed her that it "was probably flu" but no need to bring him in to be tested since that or any other respiratory illness was treated the same.

If onset is past the initial 24 hours, antivirals like Tamiflu are not useful. If a child is hospitalized for RSV, they are likely to get ribavirin. If/when a secondary bacterial pneumonia develops an antibiotic is indicated.

As far as the doc telling Este's neighbor the kids didn't need tx, one reason could be there was no benefit at that point putting the kids on Tamiflu.

Obviously since I'm arguing with kellyb that 48 hours is not a magic cutoff for very ill patients, I was not referring to 24 hours as an absolute applicable in all situations.


The belief in the magical 48 hour Tamiflu cutoff in this thread is rather anal. It's a bit more complicated than that.

[Skeptic Ginger]

Originally Posted by kellyb

.....
The highlighted is a non sequitur. We know that none who died got Tamiflu within 48 hours. And we know that among the dead, none got Tamiflu until 3-20 days after symptoms started showing.
But none of that can be interpreted to mean "of all the people who did get Tamiflu within the first 3 days, NONE were among the dead."

In fact, that 3-20 range means at least one of the deaths was for someone who got Tamiflu within 3 days (but after 48 hours.)

I've already answered this but if the range was 3-20 days, who died in that study that got Tamiflu between 48 and 72 hours?

[kellyb]

Originally Posted by Skeptic Ginger

I've already answered this but if the range was 3-20 days, who died in that study that got Tamiflu between 48 and 72 hours?

They don't say (I looked and triple checked). All we know is that zero who got it within 48 hours died, and at least one died who got it after 48 hours but before day 4.

[Dymanic]

Originally Posted by Skeptic Ginger

The belief in the magical 48 hour Tamiflu cutoff in this thread is rather anal.

To say nothing of the 24 hour cutoff point.

Quote:

It's a bit more complicated than that.

Indeed, as the CDC link I provided indicates. Nowhere among those recommendations, however, is anything that supports the idea of 24 hours as an absolute applicable in any situations. The continued arm-waving is not useful to your cause.

[Skeptic Ginger]

So summing up the 48 hours nonsense, kellyb's arguing >72 hours isn't right and Dymanic's arguing <48 hours isn't right. How did I end up being wrong on both counts?

[MarkCorrigan]

It's impressive that you did actually.

Like I say, I remember you being a nurse practitioner (sorry, forgot a word before) because you've used it before to try and support a completely incorrect assertion on this forum.

[Dymanic]

Originally Posted by Skeptic Ginger

So summing up the 48 hours nonsense, kellyb's arguing >72 hours isn't right and Dymanic's arguing <48 hours isn't right.

That "summary" is of rather limited coherence.

The CDC and the W.H.O. guidelines are in agreement that the best clinical outcomes are obtained when antiviral treatment is administered within 48 hours of onset. Neither treats that as an absolute, and both point to studies indicating that treatment may be effective even if initiated later. Both agree that treatment should be started "as soon as possible", which obviously means that within 24 hours would be preferable to within 48 hours (and it would logically follow that 12 hours would be better yet, etc). What neither of them say is what YOU said: that "if onset is past the initial 24 hours, antivirals like Tamiflu are not useful".

Besides simply being factually wrong, that wording suggests a rigidity in thinking which many would regard as a less than desireable quality in a person responsible for providing medical treatment -- and you apparently concur, as indicated by the haste with which you have attempted to distance yourself from your erroneous statement, accusing others of being "anal", and noting that "it's complicated" (implying that it's too complicated for anyone with less than your level of expertise to grasp). I cannot help but wonder how many times you may have withheld antiviral treatment from patients who might have benefitted from it, but whom you judged to be beyond the point where it would be "useful" based on an erroneous understanding of where that approximate point typically occurs.

That enough follow-through for ya?

[Estellea]

Originally Posted by Skeptic Ginger

Getting back to the incorrectly interpreted data which seems to be the nature of a number of the posted claims, (virology lab reports, P&I mortality, ILI reports and sentinel specimen sampling that is tested for influenza):

Well thank goodness we have a nurse-practitioner with zero research experience here to explain this to us dunderheads.

Quote:

UW virology reports, 1990 to present
National P&I mortality, ILI reports and sentinel specimen sampling that is tested for influenza

Look at 3 years of data sets, 1999-2000, 2000-2001, and 2009-2010.

First a caveat re the virology lab data, it should be noted that until 2009, not a whole lot of specimens were submitted for influenza culture. That's because RSV in infants requires a specific diagnosis. Using Ribavirin is not without its own difficulties and you don't usually treat infants with it empirically. Whereas with influenza, even in a hospitalized patient, you can treat empirically and the bacterial cultures looking at secondary pneumonia are more important for choosing treatment.

In the 2009 pandemic a lot more attention was paid to specifically identifying the pandemic strain and more patient cultures were submitted to virology labs from clinicians.

Thanks for the load o' bollocks apologetics but what does that tell you about drawing any conclusions the way you are doing?

Quote:

Virology lab data
The other data on one page
In the '99-2000 season there was a sharp spike in P&I deaths around January.
RSV had a high rate (~175 from ~Jan to Mar.) Influenza A numbers from the virology lab peaked in Jan at ~100.
The ILI reports were high and peaked in week 52.
Cultures from sentinel samples showed ~30% flu+ (that's high).

The ILI peak, P&I peak and the 30%+ cultures all correlate.

It might look like the RSV spike correlates but if you look closer, the P&I peak comes earlier and the RSV remains elevated longer. P&I and RSV do not correlate.

Try looking at the national RSV data; that had two spikes, one being at the same time as influenza. You're trying to pound a square peg into a round hole. I can't say with certainty that RSV is responsible for the remainder of P&I deaths but there is more going on than influenza as the underlying cause. Yes 33% is a high spike but what of the other 67% that you ignore?

Quote:

Now look at 2000-2001 data.
Virology lab
RVS has the same very high peak of 175 for a shorter time while influenza A counts from the virology lab are low ~25 at the peak.
CDC numbers (you need to read the numbers as the images aren't opening for me)
Despite the high RSV numbers, P&I mortality doesn't exceed the epidemic threshold. (You can see P&I in this graph of a 4 year span.)
ILI reports are unremarkable. (You have to look here at figure 6.)
Flu+ cultures from sentinel sites was only 11%.

The ILI peak, P&I peak and the 11%+ cultures all correlate.
While RSV does not.

You can't say that the P&I peak and "cultures" correlate because there is a spike above baseline at ~week 35 which cannot be explained by influenza activity. RSV activity is increasing at that time but again, there is no way I can draw any conclusions and more importantly, neither can you. I should point out that that Dushoff et al. P&I mortality model you trot out as superior has the average mortality at 41,400 inclusive of the 2000-2001 season. How in Hades do you reconcile that?

Quote:

It's not just the numerical nature of the denominators, it's the sample they are drawn from.

WTF? This is so stupid I can't even mock it accordingly. So I'll show you the "sample they are drawn from" and how (in spite of your assertion to the contrary) ILIs are used to estimate flu burden (I can't even bring myself to say prevalence since it isn't). http://www.plosone.org/article/info%...l.pone.0048806

Quote:

From September 1, 2009 through August 30, 2010, CDC sponsored a BRFSS module on influenza-like illness (ILI). ILI cases were determined by affirmative responses to two questions: “During the past month, were you ill with a fever?”, and “Did you also have a cough and/or sore throat?” Additional ILI questions were included in the BRFSS module, but the number of affirmative responses and/or number of eligible respondents were too small for meaningful interpretation in Washington State [14].

And: http://www.ncbi.nlm.nih.gov/pubmed/22952308

Quote:

To assess the amount of illness in the community, we conducted a door-to-door community survey for ILI, defined as more than 1 day of illness that included fever plus either cough or sore throat within the prior month. The in-town suburb of metro-Atlanta we surveyed during September 26, 2009–October 10, 2009 covers 4.2 square miles (6.7 km2) and has a population of 18,147 (12). A simple random sample of 400 households allowed us to estimate the proportion of residents with ILI within 5%. In order to account for ineligible addresses (i.e., vacant buildings, businesses, dormitories, and buildings containing institutionalized populations), we approached 500 randomly selected addresses from a master list of all residential addresses within the in-town suburb’s city limits. If no one from the household was available after at least 3 attempted contacts on separate days, we made 3 additional attempts to contact the household by phone.

A standardized questionnaire was given to the initial respondent, who answered for all household members after we obtained verbal consent. We estimated an average
weekly percentage of ILI for children and adults by dividing the percentage of ILI by 30 to obtain the average daily ILI percentage and then multiplying by 7. If dates within the month prior to the interview date fell outside of the investigation time frame, we weighted the response appropriately when calculating the weekly ILI percentages.

There's your "numerical nature of the denominators, it's the sample they are drawn from."

Quote:

Now look at 2009-2010 numbers when we had a pandemic that hit young people hard (a smaller proportion of the flu morbidity), and spared the elderly (a large proportion of typical flu morbidity).

The epidemic threshold for P&I mortality is barely breached mid winter. But notice the earlier spike. It's above the epidemic threshold but because it isn't above the mid winter peak (that includes all the mid winter viruses which are not flu) the excess mortality is less obvious.

But ILI reports peak high and early (~week 30-50).
The % of sentinel samples that are flu+ peak high and early (~week 25-45). At one point >40% of the sampled specimens were flu+.
[url=http://depts.washington.edu/rspvirus/documents/VD2009-10.pdf]UW virology shows the huge increase in flu+ results that correlate with the early peak in P&I fatalities and the ILI.
RSV rates have a low slow peak much later in the year, Feb-Apr.

Lab confirmed hospitalizations for flu is telling. The hospital cases are very high in kids under age 4. But they are a small total number of typical flu morbidity. So they don't impact the P&I threshold.

Scroll down and look at the P&I mortality figure. There is a peak of 8.1% at ~week 43 and another of 8.2% at ~week 3. Again, cannot be explained by influenza activity alone.

Quote:

Bottom line, P&I, ILIs, %flu cultures for flu, all correlate. RSV, no correlation at all. And why all the flu cultures at the virology lab? It's because there was an effort to differentiate the 2009H1N1v cases by MDs seeing patients in clinics and hospitals.

Nope, you cannot make that claim at all. Did you think I wouldn't actually look at the data and even then some? Blimey, you didn't even consult the NREVSS for national RSV data.

Quote:

If we only looked and P&I, we'd be wondering why the EDs were overflowing. Total flu deaths weren't revealing an important key factor: that this strain was hitting young people hard but sparing the elderly.

Wrong. The mortality rate per 100K population (lab-confirmed pH1N1) was nearly the same in the elderly (>65) as it was in adults 25-49. The hardest hit age bin was 50-64 (more than 1.5 times as much as the other two age bins) so I don't see how you can make that claim.

Quote:

At the same time, we did have high ILIs which tells the medical community something's going on. RSV was not an important component of the ILIs exceeding the epidemic threshold.

You can't say that because you can't even parse your own links correctly. I can't say that it is RSV; kellyb and I merely raised the spectre of other pathogens, such as RSV, contributing to the P&I mortalities because influenza alone cannot explain them.

Este

[Estellea]

Originally Posted by Skeptic Ginger

You took my comment flatass out of context.

The belief in the magical 48 hour Tamiflu cutoff in this thread is rather anal. It's a bit more complicated than that.

Ah yes, the "out of context" and "it's too complicated" arguments. At least three of us "misread" your statement. What are the chances of that?

Originally Posted by Skeptic Ginger

So summing up the 48 hours nonsense, kellyb's arguing >72 hours isn't right and Dymanic's arguing <48 hours isn't right. How did I end up being wrong on both counts?

How about simply you are wrong? You know it isn't the end of the world to just admit it and move on. More attention is drawn to errors that are compounded with nonsense.

Or what Dymanic said.

Este

[Skeptic Ginger]

Re two RSV peaks in the 99-00 season, taking Este's premise this somehow correlates with the P&I mortality, it's even less of a correlation.

Respiratory Syncytial Virus Activity --- United States, 1999--2000 Season

According to figure 1, labs in the north and south regions peaked earlier while an equal peak occurred later in the west and midwest regions. If RSV represented a large proportion of the P&I mortality, you would then expect to see two peaks in P&I deaths. If RSV accounted for the large spike in P&I deaths, you would expect any year with a similarly high peak in RSV cultures to produce an equally high peak in P&I. But you do not see that in other years with relatively high RSV rates.

As for what did I mean by: "It's not just the numerical nature of the denominators, it's the sample they are drawn from", I'll give you an example.

If you take a population at high risk of a disease, then tests on that population will give much different results than you'd get in a population at low risk. It's not just total numbers, it's also proportions that are going to differ.

I noted the biggest problem trying to match RSV data to P&I and ILIs, the RSV cultures are not coming from a random sample of people with ILIs, and in fact, you would expect more RSV in the lab data because all seriously ill infants with an ARI are going to get an RSV culture. You need it to guide treatment. This is not true for all seriously ill adults with an ARI.

The virology lab data shows the proportion of influenza and RSV in seriously ill infants, but compared to all seriously ill patients with ARIs, infants are going to be a small subset. Since RSV is a mild disease in most people, you don't see a large impact on ILI surveillance. The small subset of very ill infants, even if you add in a proportion of elderly, has a small impact on total P&I mortality.

More importantly, RSV doesn't account for surpassing epidemic thresholds for P&I, instead, RSV is part of the baseline winter P&I burden.

[kellyb]

It's kind of tragic, SG, that you can't approach this subject with genuine curiosity.

Dymanic, Este, and I have always (as far as I know) been primarily motivated by an inner drive to really understand this weirdness ( your contention that there is nothing confusing to the initiated not withstanding). And flu/winter respiratory illness and deaths have a HUGE "unknown" factor (or series of "known unknown" factors).

I can cherry-pick seasons that show that RSV is the driving force behind the highest peaks in P&I, too.

I actually started to make a spreadsheet with images and everything, before realizing I would only be arguing with you, and you are not an intellectually honest person/debater.

Have fun in that self-prescribed ghetto, arguing with with your own robo-ilk, mooching opinions from the US-only CDC/ACIP consensus, and dissing the consensus of almost every other public health authority on planet Earth.

[kellyb]

About this, SG (neverminding your inability to even FIND, much less read, and MUCH MUCH less understand actual original research)

This?

Originally Posted by SG

"If onset is past the initial 24 hours, antivirals like Tamiflu are not useful."

At some point, even if it's not true, you're better off just saying you "mis-spoke" or something like that.