The Myth of the Missing Mild Cases

Case fatality rates (CFRs) are calculated by dividing the number of deaths from a given disease by the total number of cases of the disease. For example, if there are 2 deaths out of 100 cases, we would say that the CFR is 2%. Obviously, both the numerator (the number of deaths) and the denominator (the number of cases) are important in calculating the CFR.

The number of deaths is relatively easy to determine, assuming that you have a reasonable assay for the disease of interest. However, when it comes to influenza CFRs, we are often told that we can’t be sure of what the denominator is because we are missing mild cases. Hence, we are told that the disease is probably much less lethal than it appears because there are probably many more mild cases which have not been counted. If they were, the denominator would be larger and the CFR would thus be lower.

The claim for missing mild cases has been made for the new H1N1 virus. Is there any evidence that this is so? The only “data” presented to support this contention are some phone surveys asking people if they had respiratory symptoms during the Spring. Well, I did. I have allergies. So do a lot of other people. Without testing, there is no reason to believe that mild respiratory symptoms represent a dose of swine flu. The people who are contacted in these phone surveys did not complain to their physicians of their symptoms. Hence, they aren’t “cases”.

Several carefully controlled studies with ferrets showed that the new H1N1 virus causes more severe symptoms than seasonal flu. Hence, there is no reason to believe that a person infected with the new H1N1 virus is less likely to come to the attention of a physician than a person with seasonal flu.

When policy and advice is given on the basis of CFR, we are comparing the relative effects of the new H1N1 with respect to seasonal or past pandemic flu strains. To do this, we do not need to identify every single person who has been infected with the new virus. We just need to calculate the ratio of the people with clinically obvious disease who die. This is how the CFRs were calculated for seasonal flu and for past pandemics. During the Spanish influenza, no-one was doing seroprevalance studies or knocking on people’s doors to see if they had mild symptoms. They counted the number of people who came to the attention of doctors, counted how many died, and divided the latter by the former. If anything, we are likely understating the current CFR with respect to the 1918 flu because we are more likely to identify mild cases. People today are far more likely to go to a doctor and complain about their symptoms than in 1918, in my opinion.

If seroprevalence results were released, would we find some cases that had not been identified based on clinical symptoms? Almost certainly we would find some. It is equally likely that some deaths have also been missed. However, we do not need perfect information to compare the effects of the new H1N1 virus with previous pandemic viruses. When we apply the same methods that doctors and researchers used back then, the answer is clear.

Without mitigation, this virus will cause a severe 1918-type pandemic.


2 thoughts on “The Myth of the Missing Mild Cases

  1. What would be helpful is to follow up seroprevalence studies with interviews so that you could chart a severity of illness distribution.

    Out of the those who show antibodies – what percentage remember being sick, did they call a doctor, did they stay home, etc. then plot the results x= number; y= severity, something like, 1 to 10, 1 being had no idea they caught H1N1, 5 called a doctor but stayed home, 10 spent time in the ICU.

    However you assigned numerical weight to severity the distribution, the shape of the distribution, would be very interesting. Particularly if distributions were done for age, underlying condition subgroups.

    Visually it would tell us if the hospitalized were one or two standard deviations to the right of the mean.

    We could also see how far clinically obvious cases extended to the left of the mean, if indeed they did.

    If we had the same data that PFI has managed to piece together only the data was presented on the CDC website and if, in this imaginary world, we had confidence in the CDC, then maybe we could believe the disease is much less lethal than it appears but we don’t, so we don’t. Sigh.

  2. billp, I believe the CDC has already done such a study but has not released the results. My guess is that the distribution will not differ greatly from what we think it is. In a normal flu season, some people are likely infected but have very mild or no symptoms. This is thought to be due to previous exposure to other flu strains which provides at least some immune protection. With panflu, this protection is absent except in the very old. So, one would expect fewer mild or inapparent cases with panflu than with seasonal flu.

    I think we have more than enough evidence to assert that the new H1N1 virus causes much more severe disease than seasonal flu (ferret studies, disease in previously healthy individuals, age range of deaths, etc.). The only reason we have not seen more deaths, imo, is school closures and widespread use of Tamiflu. We know that school closures break the chain of infections. However, I think the role of Tamiflu in stopping transmission is under-appreciated. We think of it as a curative agent, which it is, but it also has greatly decreased the attack rate of the virus, imo. What this means is that countries without access to Tamiflu will have both much higher attack rates and CFRs than countries with it.

    If Tamiflu-resistant strains of the new H1N1 virus become dominant, even rich countries will discover just how dangerous this virus really is.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s