Letters Section Editor: Stephen J. Lurie,
MD, PhD, Senior Editor.
In Reply: Mr Gay and colleagues suggest that
we used an incorrect statistical model to analyze our data. We agree that
it is important to confirm that analyses are appropriate for the data. We
compared the effects of using a Poisson regression model with a log-link vs
with a linear-link as suggested by Gay et al and found that the results were
highly correlated (R2>0.95). Furthermore,
our models had lower mean square errors.
Dr Simonsen and colleagues compare our model estimates to estimates
from their model and suggest that our models overpredict deaths during the
1990s due to an increase in the percentage of influenza specimens testing
positive during the study period. Simonsen et al suggest that the increase
in deaths seen in their model estimates were solely due to the aging of the
Thompson WW, Shay DK, Weintraub E, Brammer L, Cox N, Anderson LJ, Fukuda K. Estimating Deaths Due to Influenza and Respiratory Syncytial Virus. JAMA. 2003;289(19):2499-2502. doi:10.1001/jama.289.19.2499-a