Table 1 Variable categories for the intermediary logistic regression models

Socio-demographic variables

Ethnic group, religion, recent immigration, years of education, monthly income, type of house and employment status

Contact variables

Number of children in the household, contact with children of different ages, with or without an URTI, number of adults in the household, size of house, hours of childcare, sharing a room with children when sleeping, contact with a case of pneumonia, recent travel beyond the immediate residential vicinity, use of public minibuses for transport (matatu), frequenting social places and contact with selected animals, working and living in different locations

Drug use

Quantified alcohol consumption and use of traditional brews (matingas, busaa, changaa, mnazi, muratina), quantified present and past cigarette smoking, use of ground tobacco (snuff), use of Khat (miraa).

Indoor air pollution

Exposure to cooking fuel smoke in the home, use of mosquito coils, ventilation in the cooking room, use of air conditioning, passive smoking and occupational exposure during welding

Clinical variables

Chronic bronchitis, recent history of viral URTI (<14 days), previous history of pulmonary TB, previous pneumonia, body mass index, mid upper-arm circumference, anemia, splenomegaly, malaria parasitemia, sickle cell status, HIV sero-positivity, ABO blood group, Glycosylated hemoglobin (HbA1c) and current pregnancy status.