Oct 29, 2009 (CIDRAP News) – A report from the Centers for Disease Control and Prevention (CDC) estimates that the true number of H1N1 influenza cases in the first 4 months of the epidemic may have been as high as 5.7 million, or more than 100 times the official case count.
The report also estimates that as many as 21,000 people might have been hospitalized for H1N1 in that period, more than four times the official 5,000, and that there may have been up to 1,300 deaths instead of the recorded 302.
CDC officials have said repeatedly that the actual number of cases is probably far more than the number of known, confirmed cases. Today's report in Emerging Infectious Diseases appears to be the agency's first methodical attempt to estimate the true numbers.
The number of lab-confirmed H1N1 infections from the virus's emergence in April until Jul 23 was 43,677, says the report by Dr. Carrie Reed, several other CDC researchers, and Marc Lipsitch of Harvard University. This total included 5,009 hospitalizations and 302 deaths.
But the official case count is believed to be a gross underestimate, for several reasons: some sick people don't seek medical care and have specimens collected, not all specimens are sent to a public health lab for confirmatory testing, and some specimens yield false-negative results because of timing or quality problems.
To estimate the real number of cases and hospitalizations, the authors used a probabilistic multiplier model similar to one that has been used to estimate the extent of foodborne disease in the United States.
For each step, they identified a range of values derived from previous studies and from recent surveys and investigations of H1N1 outbreaks. They adjusted for the fact that confirmatory testing was done much more often in the first few weeks of the epidemic than it was later, after physicians were encouraged to reserve testing mainly for hospital patients.
Using this method, the researchers estimated that each known H1N1 case represented a total of between 47 and 148 cases, with a median of 79 cases. This yielded an estimated total of between 1.8 million and 5.7 million cases, with a median estimate of 3.0 million, in a nation with a current estimated population, according to the Census Bureau, of about 307 million.
Patients hospitalized with H1N1 are much more likely to be tested and have their infection confirmed than outpatients are, the authors note. They estimate that each confirmed hospital case represented between 1.9 and 4.3 cases, with a median of 2.7 cases. Nationally, this translates into between 9,000 and 21,000 cases, with a median of 14,000.
The report also hazards an estimate of H1N1 deaths, though with more caveats than the other estimates.
Among confirmed hospital cases through Jul 23, 6% of the patients died. Applying this percentage to the hospitalization estimates from the model yielded a range of estimated deaths from 550 to 1,300, with a median of 800, the authors report. But they say this approach has several limitations, and more sophisticated models are being developed in an effort to come up with better estimates.
The report also offers estimates of H1N1's impact in different age-groups, which align with the general view that children and young people have been hit hardest. The median estimated incidence of cases is lowest in people aged 65 and older: 107 per 100,000. At the other end of the scale are those between 5 and 24 years old, with an estimate of 2,196 cases per 100,000.
The incidence of hospitalization was estimated to be highest in children under 5 years old, with a median of 13 cases per 100,000. Elderly people (65 and up) had the lowest estimated hospitalization rate: a median of 1.7 per 100,000.
The authors call their estimation method "a relatively quick and simple approach" that can be used while more complete data and more rigorous studies are awaited. They note that a spreadsheet version of the model is available online so health officials can use it and adapt to their local circumstances.
The CDC has not come up with estimates of total cases and hospitalizations for the period since Jul 23, CDC officials said today. Reed, lead author of the study, told CIDRAP News that the agency is receiving reports of confirmed H1N1 hospitalizations and deaths from about 30 states, and officials are looking at ways to generate more current estimates with it. But she said she is unsure how well the multipliers used in the study apply to the current situation.
She noted that the agency stopped trying to count total cases months ago because of the burdens that testing imposed on the medical system.
At a news briefing today, Dr. Anne Schuchat, head of the CDC's National Center for Immunization and Respiratory Diseases, made similar comments. "We believe that many millions of people have already contracted this virus here in the US and that we have had probably by now well more than 20,000 hospitalizations," she said.
Reed C, Angulo FJ, Swerdlow DL, et al. Estimates of the prevalence of pandemic (H1N1) 2009, United States, April-July 2009. Emerg Infect Dis 2009 Dec (published online Oct 28) [Full text]
CDC Oct 28 Q&A on article
CDC FluView weekly flu surveillance reports