Household-level risk factors for influenza among young children in Dhaka, Bangladesh: a case–control study Saumil Doshi 1,2 , Benjamin J. Silk 1 , Dhiman Dutt 3 , Moshtaq Ahmed 3 , Adam L. Cohen 1 , Thomas H. Taylor 1 , W. Abdullah Brooks 3 , Doli Goswami 3 , Stephen P. Luby 1,4 , Alicia M. Fry 1 and Pavani K. Ram 5 1 Centers for Disease Control and Prevention, Atlanta, GA, USA 2 Howard University, Washington, DC, USA 3 International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh 4 Stanford University, Stanford, CA, USA 5 University at Buffalo, Buffalo, NY, USA Abstract objectives To identify household-level factors associated with influenza among young children in a crowded community in Dhaka, Bangladesh. methods We conducted a case–control study using existing active surveillance for respiratory illness. Cases were children aged 12–59 months with laboratory-confirmed influenza. Controls were children frequency-matched by age group with no respiratory illness in the prior 6 months. We interviewed caregivers and observed household handwashing behaviour. Soap consumption was estimated by summing weight differences of three bars of soap sequentially left in each household. We measured concentrations of airborne particulate matter <2.5 lg in diameter (PM 2.5 ) in a subset of households. We used logistic regression to estimate adjusted odds ratios (aOR) and 95% confidence intervals (CI). results We enrolled 145 cases and 341 controls between March 2009 and April 2010. Case and control household members were observed to wash hands with similar frequency during a 5-h period (mean, 0.64 events vs. 0.63, P = 0.87), and similar daily soap consumption per capita (mean 2.92 grams vs. 2.93, P = 0.92). Case households were more likely than controls to have crowded (≥4 persons) sleeping areas (aOR = 1.67, CI: 1.06–2.63) and cross-ventilated cooking spaces (aOR = 1.75, CI: 1.16–2.63). Case and control households had similar median 24-h geometric mean PM 2.5 concentrations in the cooking (69.2 vs. 69.6 lg/m 3 , P = 0.45) and sleeping (65.4 vs. 67.4 lg/ m 3 , P = 0.19) spaces. conclusions Handwashing with soap was practiced infrequently and was not associated with paediatric influenza in this community. Interventions aimed at crowded households may reduce influenza incidence in young children. keywords influenza, human, handwashing, prevention and control, Bangladesh, crowding Introduction Acute lower respiratory infections (ALRI) are a leading cause of childhood morbidity and mortality worldwide [1–3]. Influenza viruses accounted for approximately 21 million ALRI episodes among young children in 2008, and 99% of deaths from influenza-associated ALRI in young children are estimated to occur in low-income countries [4]. The burden of influenza among children <5 years old in Kamalapur, a low-income community in Dhaka, Bangladesh, was estimated at 102 episodes per 1000 child-years [5]. Influenza caused 14–16% of ALRI or febrile respiratory infections and was associ- ated with 10% of pneumonia cases among children <5 years [6, 7]. Influenza vaccines, the best prevention tool against influ- enza, are not widely available in low-income countries. The 2009 pandemic highlighted the need for non-pharmaceuti- cal interventions during the lengthy process of vaccine pro- duction for novel influenza strains [8]. Because many young children spend the majority of their time in the household environment, identifying household-level risk factors for influenza infection can guide implementation of non-phar- maceutical interventions for this vulnerable group. Handwashing has been promoted as a feasible interven- tion against influenza[9,10]. Although studies supporting © 2015 John Wiley & Sons Ltd 719 Tropical Medicine and International Health doi:10.1111/tmi.12475 volume 20 no 6 pp 719–729 june 2015