The Case Fatality Rate (CFR) for the swine flu

Comment by TMN
June 5, 2009

    If you read the newspapers and listen to the TV you will hear from time to time their favorite swine flue story that the H1N1 flu is leveling off as the summer approaches but we should still worry about the possibility of the pandemic hitting in the fall. They also point out that the flu is mild and kills only people with underlying health conditions.
    Since May 10th, the reported cases in the USA (CDC) and in the world (WHO) appear to be growing with a doubling time of 11 or 12 days. This is a rate of increase of about 6 to 7% per day. If you look at the graphs of the USA and World-Wide cases of H1N1 put out by the CDC and WHO, it is clear that the flu is NOT leveling off in the USA or in the world as a whole.
    There are many indications that many counties in many US states have stopped reporting cases due to loss of interest combined with budget difficulties. CDC and WHO officials believe that the cases of H1N1 are underreported by a factor of 20 (I love it when people make models of things they don't have data about).
    The flu symptoms are not severe or life threatening for many cases so it is easy to believe that the reported flu rates are a gross underestimation of the number of actual cases. It is possible that even people that die with H1N1 are not being reported. For example, one of the first women to die in Canada with H1N1 was not diagnosed until after she was already dead and relatives were diagnosed with the disease. It also seems likely that many unreported cases must exist in countries where the citizens have poor access to health care.
    While we can be pretty sure that the reported cases from the CDC and WHO are low estimates of the number of people who have contracted the H1N1 virus, a reasonable scientific assumption is that the totality of cases are roughly proportional to the reported cases. This assumption allows us to use the CDC and WHO numbers to try to learn something about the spread of the disease even though we know they are not 100% accurate. The things that we might try to learn from these numbers are the spread of the disease and the danger posed by the disease.
    It is clear from the graphs of H1N1 cases in the USA indicate that H1N1 flu is not only increasing in the USA but the rate of growth is also increasing. At the same time that more and more states and counties have stopped testing for H1N1, the total cases as of June 5th, 2009 are still growing at about 6-7% per day. The world cases are growing at a suprisingly very similar rate.
    Another interesting question that one can hope to answer from the available data is the danger posed by the disease. If you look at the currnet CDC numbers today the USA has 13,217 reported cases and 27 deaths. Before doing long division to find the death rate, it is important to realize that there is a significant time lag between the diagnosis of H1N1 and when the death from the reported illness can occur. If we assume that on average it takes about one doubling time of 11 or 12 days to die of the disease, we have to use about 1/2 of the current number of cases for the calculation of the case fatality rate (CFR). These assumptions provide an estimate for the Case Fatality Fate of CFR=0.4%. The World Health Organization (WHO) numbers for World-Wide reported cases of H1N1 are 21,940 cases and 125 deaths. Since the doubling time of 11-12 days we can again make the assumption that we have to divide the current number of cases by 2 we get a CFR=1.1%.
    These rates are similar to the reported case fatality rates that studies have found for reported snakebites. See for example, Mortality due to snakebite envenomation in Costa Rica (1993–2006) by Pablo Fernándeza and José María Gutiérrez in June 2008. Remember that this study about snakebites has the same problem we have finding out the CFR of H1N1 due to unreported cases. Clearly, many snakebites especially non lethal ones are not reported. It may, therefore, be a reasonable analogy that if you get the swine flu, you have about as much chance of dying as if you get bit by an arbitrary snake in Costa Rica. The danger of being killed by a snake in Costa Rica will not deter many people from taking a beautiful vacation there. On the other hand, most people will try very hard to avoid being bitten by a snake on their Costa Rica vacation. Moreover, if someone is bitten by a snake while they are enjoying the natural beauty of Costa Rica and the symptoms look very severe, I would advise them to immediately seek medical attention. The same level of respect should be given to H1N1 flu.

    In conclusion, I would definitely NOT tell people that they should be overly fearful of the swine flu. However, I would no more tell someone that the H1N1 virus is mild and not to worry about it than I would advise them to start playing with random snakes in Costa Rica! I wish the news media would simply tell American's the truth, "We don't know what will happen in the future but the current data is cause for concern.