Predicting the total number of deaths in a pandemic would seem to be an easy thing to do assuming you know two things: the clinical attack rate (CAR) and the case fatality rate (CFR). The first number tells you what percent of the people are sick with the virus, the second tells you what percent of these people will die. One might think that it would be possible to simply multiply the CAR by the total population and then multiply that number by the CFR to get the total number of expected deaths. However, as I pointed out in my previous blog about CFR, this number varies depending on the age of the person. So, it is necessary to calculate the expected number of deaths for each age group separately and then add them together to get the total number of expected deaths.
The CAR can also vary by age group. In a normal flu season, the CAR for children is much higher than for adults. This is thought to be due to partial immunity of at least some of the adults due to past exposure to related flu strains. However, during a pandemic, no-one (except perhaps the very old) will have any immunity to the novel virus. Thus, it is assumed that the CAR will be more similar across age groups. For the calculations below, I will assume a CAR of 30% for all age groups.
The CFR used for the calculations below come from the CDC Director’s brief that was apparently leaked and uploaded to Cryptome.
The table below contains the following information: age; number of people in the cohort (in the US); number of expected clinically ill people; CFR; expected number dead in cohort.
- 0 – 4 years old; 21,005,852; 6,301,756; o.17%; 10,712.
- 5 – 24 years old; 82,693,215; 24,807,964; 0.22%; 54,578.
- 25 – 49 years old; 106,312,569; 31,893,771; 1.5%; 478,406.
- 50 – 63 years old; 55,178,372; 16,553,523; 3.33%; 551,232.
- 65+; 38,869,716; 11,660,915; 5.24%; 611,031.
Total expected dead: 1,705,961.
This number obviously indicates a potential severe pandemic. Also, note that most of the dead would be expected to be under the age of 65.
The numbers presented here assume an unmitigated pandemic. There are several interventions that have the potential to bring them down. First, a well-matched vaccine in sufficient quantities to vaccinate most Americans would be expected to reduce these numbers considerably. Unfortunately, it is unlikely that such a vaccine will be available before the next large wave of infections begins in the United States. Tamiflu has been very effective in decreasing deaths if given within 48 hours of infection. However, since most testing for the new H1N1 virus has stopped, fewer and fewer people in the US are receiving Tamiflu within the recommended time period. Finally, closing schools greatly diminishes the CAR. This is especially effective in reducing deaths because flu normally spreads first among children (in schools), then to teachers, parents and other middle-aged people and finally to the elderly. By breaking the chain of transmission early in this process, closing schools has the potential to greatly reduce the number of deaths. Unfortunately, the CDC is strongly opposed to closing schools. Schools are already starting to open in the US. Thus, the cycle of infection will begin shortly.
We have some Tamiflu, which some people will get. There will be some vaccine at some point for some people. At some point, schools will close either due to parental concern or because of overflowing ICUs. Thus, I do not expect that we will see the total number of deaths projected in the table above. However, it is disappointing that the one intervention which could save many lives, keeping schools closed, is being taken off the table by the agency which has access to the same numbers used in my calculations, the CDC.
What are they thinking?