We agree that vaccine has limited ability to prevent influenza-like illness. However, because our study focused on preventing true influenza cases, our model's annual incidence rates were specific for influenza, not influenza-like illness. Likewise, our vaccination efficacy rate of 68% was conservative relative to the 74% found by Demicheli and colleagues (1). We also believe our assumptions regarding the value of lost workdays were fair. We assumed that worker productivity is related to wage and that 1 day of work lost throughout the year should be valued at the average wage, an assumption that is consistent with economic theory when the market is at equilibrium.

In terms of health values, we agree that these should come from a general population. Thus, we did not limit our survey to those with respiratory illness. Regardless of the sample, our model found that treatment decisions were generally unaffected by willingness to pay even if symptomatic benefits equaled $0. In fact, we found that societal treatment costs could theoretically be recouped by earlier return to work of those afflicted (which is not purely a "personal benefit").

Brodkin questions whether our health care system has the capacity to handle all patients seeking influenza treatment. Although this is an interesting point, vaccination, through prevention, may actually reduce the need for acute office visits. In addition, it is unclear whether coverage for antiviral therapies, which are already marketed, would actually induce more visits from persons with influenza, since sick patients often seek medical attention regardless of options for therapy.

The final questions raised had to do with our model assumptions for probability of influenza and workdays saved. Our base-case probability of 15% for influenza-related symptomatic illness was based on rates of serologically demonstrated influenza in a longitudinal epidemiologic study (2). The references cited by Meltzer and Bridges actually show a serologically proven infection rate between 13% and 23% (3). Our estimate of workdays lost was also based on a survey of 411 adults (4) and was further confirmed by a randomized trial of influenza vaccination (5). This trial demonstrated 4.9 workdays lost per episode of illness, making our estimate (2.8 workdays) conservative. We agree, however, that research is needed to confirm that antiviral treatment results in more rapid return to work, and we called for these important studies in our discussion.