BACKGROUND

School-based influenza vaccination campaigns could mitigate the effects of influenza epidemics. A large countywide school-based vaccination campaign was launched in Knox County, Tennessee, in 2005. Assessment of campaign effects requires identification of appropriate control populations. We hypothesized that contiguous counties would share similar pre-campaign patterns of influenza activity.

METHODS

We compared the burden of influenza emergency department (ED) visits and hospitalizations between Knox County (Knox) and eight counties surrounding Knox (Knox-surrounding) during five consecutive pre-campaign influenza seasons (2000-01 through 2004-05). Laboratory-defined influenza seasons were used to measure the weekly incidence of medically attended acute respiratory illnesses (MAARI) attributable to influenza in school-aged children 5-17 years old (campaign target) as well as in other age groups. Seasonal rates of MAARI attributable to influenza for Knox and Knox-surrounding counties were compared using rate ratios.

RESULTS

During five consecutive influenza seasons, MAARI attributable to influenza showed synchronous temporal patterns in school-aged children from Knox and Knox-surrounding counties. The average seasonal rates of ED visits attributable to influenza were 12.37 (95% CI: 10.32-14.42) and 13.14 (95% CI: 11.23-15.05) per 1000, respectively. The respective average seasonal influenza hospitalization rates for Knox and Knox-surrounding were 0.38 (95% CI: 0-0.79) and 0.46 (95% CI: 0.07-0.85) per 1000 children. Rate ratio analyses indicated no significant differences in the incidence of MAARI attributable to influenza between school-aged children from Knox and Knox-surrounding counties. Estimates for other age groups showed similar patterns.

CONCLUSION

Before the Knox school-based influenza vaccination campaign, influenza resulted in an average of about 12 ED visits and 0.4 hospitalizations per 1000 school-aged children annually in Knox County. Since similar morbidity was observed in surrounding counties, they could serve as a control population for the assessment of the campaign effects.