22 DECEMBER 2009 | GENEVA -- Efforts to assess the severity of the H1N1 influenza pandemic sometimes compare numbers of confirmed deaths with those estimated for seasonal influenza, either nationally or worldwide. Such comparisons are not reliable for several reasons and can be misleading.
Numbers of deaths for seasonal influenza are estimates. They use statistical models designed to calculate so-called excess mortality that occurs during the period when influenza viruses are circulating widely in a given population.
The models use data, as recorded in death certificates and medical records, indicating mortality from all causes, and compare the number of deaths during epidemics of seasonal influenza with baseline data on deaths during the rest of the year. The assumption is that infections with influenza viruses contribute to the “excess mortality” observed during the influenza season.
During epidemics of seasonal influenza, around 90% of deaths occur in the frail elderly, who often suffer from one or more chronic medical conditions. Although influenza can worsen these conditions and contribute to death, testing for influenza viruses is not done in most cases, and deaths are usually attributed to an underlying medical condition.
Methods for estimating excess mortality were introduced in the 19th century to capture these influenza-associated deaths that would otherwise be missed. Such estimates have helped counter assumptions that influenza is a mild illness that rarely kills.
In contrast, numbers of deaths from pandemic influenza, as notified by national authorities and tabulated by WHO, are laboratory-confirmed deaths, not estimates. For several reasons, these numbers do not give a true picture of mortality during the pandemic, which is unquestionably higher than indicated by laboratory-confirmed cases.
As pandemic influenza mimics the signs and symptoms of many common infectious diseases, doctors often do not suspect H1N1 infection and do not test. This is especially true in developing countries, where deaths from respiratory diseases, including pneumonia, are common occurrences. Moreover, routine testing for pandemic influenza is costly and demanding, and beyond the reach of most countries.
When testing confirms H1N1 infection in patients with underlying medical conditions, many doctors record these deaths as due to the medical condition, and not to the pandemic virus. These cases are also missed in official statistics.
As recent studies have shown, some tests for H1N1 infection are not entirely reliable, and false-negative results are a frequent problem. Accurate test results further depend on how and when samples were taken. Even in the best-equipped hospitals, doctors have reported seeing patients with distinctive and virtually identical disease profiles, yet only some have positive test results.
Moreover, in a large number of developing countries, systems for vital registration are either weak or non-existent, meaning that most deaths are neither investigated nor certified in terms of the cause.
Comparisons of deaths from pandemic and seasonal influenza do not accurately measure the impact of the pandemic for another reason. Compared with seasonal influenza, the H1N1 virus affects a much younger age group in all categories – those most frequently infected, hospitalized, requiring intensive care, and dying.
WHO continues to assess the impact of the influenza pandemic as moderate. Accurate assessments of mortality and mortality rates will likely be possible only one to two years after the pandemic has peaked, and will rely on methods similar to those used to calculate excess mortality during seasonal influenza epidemics.