Factors associated with transmission of influenza-like illness in a cohort of households containing multiple children

BACKGROUND Household studies of influenza-like illness (ILI) afford opportunities to study determinants of respiratory virus transmission. OBJECTIVES We examined predictors of ILI transmission within households containing at least two children. METHODS A prospective cohort study recorded ILI symptoms daily for 2712 adult and child participants during the 1998 influenza season in Victoria, Australia. Logistic and Poisson regressions were used to explore predictors of household transmission of ILI and the secondary household attack proportion (SHAP). A date of illness onset during the influenza season was used as a proxy indicator of ILI associated with influenza infection (as opposed to other aetiological causes). RESULTS A total of 1009 ILI episodes were reported by 781 of 2712 (29%) participants residing in 157 households. Transmission, defined as detection of ILI in one or more household members following identification of an index case, was observed in 206 of 705 (29%) household introductions. Transmission of ILI was significantly associated with the onset of ILI in the index case during the peak influenza season compared with the remainder of the observation period (37% versus 27%, odds ratio = 1·59, 95% CI 1·09, 2·31, P = 0·017). The SHAP was 0·12, higher if the index case was of secondary school age [incidence risk ratio (IRR) = 1·80, 95% CI 1·08, 2·98, P = 0·022]. CONCLUSIONS Risk of household transmission of ILI was increased during the peak influenza season, indicating an increased burden of disease during the period of influenza circulation. In this cohort, secondary-school-aged children and adults were important transmitters of ILI.