Comparison of physical, public and human assets as determinants of socioeconomic inequalities in contraceptive use in Colombia - moving beyond the household wealth index.

Background: Colombia is a lower-middle income country that faces the challenge of addressing health inequalities. This effort includes the task of developing measures of socioeconomic position (SEP) to describe and analyse disparities in health and health related outcomes. This study explores the use of a multidimensional approach to SEP, in which socioeconomic inequalities in contraceptive use are investigated along multiple dimensions of SEP. We tested the hypothesis that provision of Public capital compensated for low levels of Human capital.

Methods: This study used the 2005 Colombian Demographic and Health Survey (DHS) dataset. The outcome measures were 'current non-use' and 'never use' of contraception. Inequalities in contraceptive behaviour along four measures of SEP were compared: the Household wealth index (HWI), Physical capital (housing, consumer durables), Public capital (publicly provided services) and Human capital (level of education). Principal component analysis was applied to construct the HWI, Physical capital and Public capital measures. Logistic regression models were used to estimate relative indices of inequality (RII) for each measure of SEP with both outcomes.

Results: Socio-economic inequalities among rural women tended to be larger than those among urban women, for all measures of SEP and for both outcomes. In models mutually adjusted for Physical, Public and Human capital and age, Physical capital identified stronger gradients in contraceptive behaviour in urban and rural areas (Current use of contraception by Physical capital in urban areas RII 2.37 95% CI (1.99-2.83) and rural areas RII 3.70 (2.57-5.33)). The impact of women's level of education on contraceptive behaviour was relatively weak in households with high Public capital compared to households with low Public capital (Current use of contraception in rural areas, interaction p = < 0.001). Reduced educational inequalities attributable to Public capital were partly explained by differences in household wealth but not at all by health insurance cover.

Conclusions: A multidimensional approach provides a framework for disentangling socioeconomic inequalities in contraceptive behaviour. We provide evidence that material circumstances indexed by Physical capital are important socioeconomic determinants while higher provision of Public capital may compensate for low levels of Human capital with respect to modern contraceptive behaviour.