Do Current Surveillance Systems Provide Valid and Credible Statistical Information on 2009-H1N1?
Keywords: Disease surveillance, public health surveillance, A(H1N1) influenza , pandemic
The outbreak and rapid world-wide spread of novel A(H1N1) influenza in 2009 came after almost a decade of enhancements to global disease surveillance systems. Yet even with these systems there is reason to question the validity – and the credibility – of the statistical information they provide. Although establishing a standard case definition is a critical step in an epidemiological outbreak investigation, definitions of suspected, probable, and confirmed cases varied from country to country and changed as the virus spread. Some changes reflected an evolving understanding of the epidemiology, limitations in laboratory capacity and requirements of public health practice. For some of the same reasons, case ascertainment processes varied, but typically focused on more severe cases. As a result, there was substantial uncertainty about virulence and transmissibility of the novel viral strain. One of the most commonly held assumptions about the novel A(H1N1) virus is that children and young adults are at especially high risk. The data on which this assumption is based, however, are not reliable, and it is possible that the differential risk for children and adolescents is an exaggeration. In particular, higher age-specific rates for A(H1N1) incidence as well as hospitalizations and deaths, regardless of the source, are biased upwards for children and young adults and downwards for older adults. The degree of this bias is unknown, but comes from a combination of younger patients being more likely to present themselves for medical attention and older patients not having samples sent for laboratory testing. These patterns seem to be due to both patients and physicians responding to what they believe are the facts about the risk of A(H1N1), as well as physicians responding to public health recommendations (based on the same assumptions) regarding whom should be tested. The tendency of public health officials and the media to report cumulative case counts adds to the confusion. Cumulative numbers reflect when cases are reported or confirmed in the laboratory rather than the time of onset. As a result, they reflect patients’ decisions about when to seek care, reporting requirements, laboratory capacity, and statistical processing rather than the incidence of disease in the population. Furthermore, by definition, cumulative numbers can only increase, even when incidence is waning, contributing to a false impression about the pandemic.