Paradigm Shift Intervention Monitoring
twitter
Live feed of underlying pandemic map data here
Commentary
CDC Concedes Week 12 P&I Exceeds Epidemic Threshold
Recombinomics
Commentary 21:49
April 2, 2010
The above comment is from the revised CDC report for week 12. The initial report used last week’s P&I value of 7.7% in the text and stated in the summary that the P&I death rate was below the epidemic threshold. The week 12 data however was 7.8% and took the death rate above the epidemic threshold, as noted in the corrected version (which was also evident in the original graph).
However, the main point of the P&I rate is its elevation at a time when reported cases of confirmed H1N1 are low. No state is reporting widespread H1N1, and only three states in the southeast are reporting regional levels. Thus, as was seen in January, the P&I death rate is markedly higher than would be expected from the reported levels by individual states.
These inconsistencies raise concerns that the lab confirmed figures grossly underestimate the frequency of H1N1 infections and deaths, the H1N1 is more lethal than indicated in state reports, or both. Traditionally, the P&I is used to estimate the number of flu deaths. An extrapolated level of 36,000 deaths in the US is commonly cited in media reports on the pandemic, and politicians and self described pundits/experts frequently cite the 36,000 seasonal flu deaths as evidence that the current pH1N1 pandemic is “mild”.
However, 90% of seasonal flu deaths are over 65 (mean age at death is 75.7) and deaths in the 1957 and 1968 pandemics also were associared with an older population (mean age at death 64.6 and 62.2 years, respectively). In contrast 90% of 2009 pandemic H1N1 deaths are under 65 (mean age at death is 37.4), which is similar to the 1918 pandemic (mean age of 27.2 years). The older current population is largely spared because the seasonal flu has been crowded out by pH1N1, and most older patients have circulating antibodies to pH1N1 because of exposures to the 1918 virus, seasonal H1N1 circulating prior to 1950, or vaccination by the 1976 swine flu. Thus, if the CDC estimate of 12,000 deaths due to pH1N1 was accurate, the P&I levels would be expected to be markedly lower than previous years.
However, as seen by the P&I levels in January and March, the P&I levels are above the epidemic threshold when reported lab confirmed cases are low. In the fall, when lab confirmed cases were high, the P&I levels spiked well above baseline, but the levels remained high in 2010, indicating the death rate was significant, because there was little contribution from influenza deaths in the older population.
The high level of P&I deaths, when influenza infections in the elderly populations is at a low level and reported lab confirmed cases is low, raises concerns that pH1N1 is producing more severe and fatal cases in 2010. Since the age of teh pH1N1 deaths is uncharacteristically low, release of the age distribution in the P&I deaths would be useful. The ages of the total deaths are given for the 122 cities being monitored, but the P&I deaths are listed as a composite number.
The repeated increases of the P&I death rate to levels that exceed the epidemic threshold, which is not reflected in lab confirmed case reports, continues to be cause for concern.
Media Links
Recombinomics
Presentations
Recombinomics
Publications
Recombinomics
Paper
at Nature Precedings