Date Published: October 28, 2008
Publisher: Public Library of Science
Author(s): Cécile Viboud, Mark Miller
Abstract: Cécile Viboud and Mark Miller discuss the implications of a new study that used a mathematical model to simulate influenza transmission in nursing homes.
Partial Text: Most countries in Europe and North America target influenza immunization to persons at highest risk for complications, including seniors 65 years and older, those with certain chronic illnesses, and young children. Despite increases in vaccination rates in these high-risk groups over the last few decades, morbidity and mortality from influenza remain high.
Seniors suffer about 90% of the influenza seasonal mortality burden, and rates of hospitalization and death are increasing as the population ages . Studies of national trends in the United States and Italy showed that even after adjusting for population aging and pathogenicity of circulating influenza viruses, vaccine uptake in seniors was not associated with a decline in influenza-related mortality ([2,3]; Figure 1). These disappointing experiences can be explained by the phenomenon of immune senescence, whereby immune response to vaccines declines in the last decades of life . These results—along with a growing understanding that the expected effectiveness of vaccination had been greatly overestimated in seniors [5,6]—have fueled debate over the best strategy to protect high-risk populations [1,7].
The Ontario province of Canada launched a unique influenza vaccination campaign in October 2000, offering free influenza vaccine to everyone over six months of age through extensive delivery in nontraditional settings, including community centers and shopping malls. In this issue of PLoS Medicine, Kwong and colleagues report their evaluation of the impact of this intense and costly strategy on a variety of influenza-related outcomes . The authors compared changes in influenza burden in Ontario before (1997–2000) and after (2000–2004) universal vaccination was implemented—relative to changes in surrounding Canadian provinces that had lower vaccine uptake. Electronic health databases and influenza laboratory surveillance data were used to estimate trends in influenza-related mortality, hospitalization, and visits to emergency departments and doctors’ offices.
The authors found that the influenza burden decreased as vaccination rates increased across all age groups in Ontario, with a 49%–59% decline for various outcomes (p < 0.002), relative to other provinces. Comparisons across provinces showed that there was also a convincing coverage–response relationship among people under 65 years, in that the largest vaccine uptakes corresponded to the greatest reductions in disease burden. Influenza disease burden must be inferred indirectly through statistical methods (; Figure 1), in the face of substantial interannual variations in pathogenicity of circulating influenza viruses, vaccine mismatch seasons, and geographical variability. Thus Kwong and colleagues  met great challenges to evaluate the effect of Ontario's population-level experiment. By comparing trends between Ontario and other Canadian provinces and conducting extremely thorough sensitivity analyses, the authors cleverly attempted to control for factors unrelated to vaccine uptake. But even so, lack of statistical power is a key limitation of observational studies comparing disease trends across time periods and geographical regions with different intervention strategies. Based on the experience in Ontario  and elsewhere, there is now solid evidence for the effectiveness of influenza vaccination in people under 65 years. However, the direct and indirect benefits of universal immunization remain unclear for seniors, and Kwong and colleagues' study  was likely underpowered to evaluate such benefits. Source: http://doi.org/10.1371/journal.pmed.0050216