BACKGROUND & OBJECTIVE: This study examines the association between household attributes and perceived morbidity within resource-poor house holds (HHs) in India at five locations. This presents an innovation compared to most epidemiological studies, which focus on associations between the incidence of an illness and characteristics of the ill person. METHODS: Perceived morbidity was represented by a variable called "Incidence of illness in a HH" (IIH) = the number of self reported illness episodes during three months preceding the survey, divided by household size. Variables were analyzed through bivariate correlation and multivariate linear regression. The evidence was based on a HH survey conducted in 2005 in Maharashtra, Bihar, and Tamil Nadu. Data yield reflected responses of 3,531 HHs, representing 17,323 individuals and 4,316 illness episodes. RESULTS: Analysis showed that incidence of illness among women was higher; the under 5 yr olds and elderly (+55) were particularly vulnerable. However, in the multivariate linear regression model, gender ratio within HHs became an insignificant explanatory variable. Age distribution had a small but significant effect. Household size and the level of education in the HH were negatively and significantly associated with IIH. The regression analysis showed that income had a modest positive effect, but improved housing was associated with reduced IIH. Large differences were noted in IIH across locations. INTERPRETATION & CONCLUSION: Our findings showed that attributes of the unit household, including type of house, income, education and size, have significant effects on IIH; variability in IIH cannot solely be explained by age and gender of HH members.
BACKGROUND & OBJECTIVE: This study examines the association between household attributes and perceived morbidity within resource-poor house holds (HHs) in India at five locations. This presents an innovation compared to most epidemiological studies, which focus on associations between the incidence of an illness and characteristics of the ill person. METHODS: Perceived morbidity was represented by a variable called "Incidence of illness in a HH" (IIH) = the number of self reported illness episodes during three months preceding the survey, divided by household size. Variables were analyzed through bivariate correlation and multivariate linear regression. The evidence was based on a HH survey conducted in 2005 in Maharashtra, Bihar, and Tamil Nadu. Data yield reflected responses of 3,531 HHs, representing 17,323 individuals and 4,316 illness episodes. RESULTS: Analysis showed that incidence of illness among women was higher; the under 5 yr olds and elderly (+55) were particularly vulnerable. However, in the multivariate linear regression model, gender ratio within HHs became an insignificant explanatory variable. Age distribution had a small but significant effect. Household size and the level of education in the HH were negatively and significantly associated with IIH. The regression analysis showed that income had a modest positive effect, but improved housing was associated with reduced IIH. Large differences were noted in IIH across locations. INTERPRETATION & CONCLUSION: Our findings showed that attributes of the unit household, including type of house, income, education and size, have significant effects on IIH; variability in IIH cannot solely be explained by age and gender of HH members.