Background Research on intersecting social inequalities constantly challenges traditional assumptions about the operation of gender, economic class, caste and other social markers in health and health care. Evidence from such research reveals non-linear patterns that sometimes upend what research on a single inequality may predict. However, our knowledge of the mechanisms through which these intersecting inequalities operate is relatively thin. We delineate three inequality-producing mechanisms within households and communities through an analysis of empirical data from Koppal district, Karnataka, South India. We present some illustrations of the ways in which these mechanisms operate.

Methods Our 2002 household survey included a stratified sample of 1,920 households from 60 villages in Koppal district. In earlier work, we had developed an innovative intersectional methodology for the analysis of large datasets.1 We analyse survey data on healthcare use for long-term ailments using this methodology. We also analyse qualitative data on care seeking during pregnancy and obstetric emergencies from verbal autopsies of 33 maternal deaths in 67 villages (for the period 2008–2011). The data were gathered using a rigorous methodology that engages with the subjectivities of death reporting.2

Findings The role of gender, caste and economic class varied by the outcome being measured. Gender interacting with economic class produced inequalities in the uptake of purchased health care, while caste did not. On the other hand, caste interacting with economic class shaped the uptake of pregnancy-related care.

Inequality at the upper end of a multidimensional socioeconomic spectrum stemmed mainly from the imperative to consolidate authority within the prevailing social order. This translated into ‘politics of honour’, entailing tight control over social reproduction. Rich or upper caste women's experiences of pregnancy and illness were framed by gender relations which afforded only restricted mobility, virtually no say in decisions concerning education, marriage and childbearing, and a lack of acknowledgement of their health needs while sick or pregnant. This lack of acknowledgement means that, when such women had long-term ailments, they were less likely than the men of their households to seek (or receive) any form of care. Concern for the family's honour also inhibited upper-caste women with serious obstetric risk from approaching outsiders for help when their families were indifferent to their needs.

In the middle of the spectrum, a process of leveraging3 helped poor men continue treatment for long-term ailments. Poor men leveraged gender power to pay for treatment, while poor women dropped out prematurely: they appropriated the household's resources (to the detriment of poor women) or took loans. This enabled poor men to achieve levels of continuation that were similar to those achieved by non-poor women. Among groups at the bottom of the spectrum where there were no resources to be leveraged, the poorest men fell to the level of the poorest women in what we have called a ‘perverse catch up’.

Discussion Three mechanisms working simultaneously at different points in a multidimensional socioeconomic spectrum indicate that social inequalities in health are more complex than mainstream research would suggest. These mechanisms are not mutually exclusive. They feed into each other. For example, leveraging by poor men gave rise to gender-biased household rationing.4 Recognising which mechanism works at what level is necessary, if there is to be conceptual clarity and theoretical advancements in the field of health equity.

The strong role played by caste in shaping the experiences of pregnancy-related care versus its insignificance as a predictor of treatment seeking for long-term ailments suggests that a priori assumptions cannot be made about the operation of gender, caste and economic class. Even so, all of the mechanisms described above inevitably worked through the gender system. This means that gender power is a force that drives inequality at all levels of the socioeconomic spectrum.

As an approach to studying health inequalities, intersectionality can seem complex. However, it captures the many-dimensioned lives that people live more fully than one-dimensional approaches ever can. Knowledge gained from the application of such an approach can usefully inform public policy in health.

Grant funding (over the years, the Gender and Health Equity Project received financial support from the Swedish International Development Cooperation Agency (Sida), the John D. and Catherine T. MacArthur Foundation, the Rockefeller Foundation, the Ford Foundation and the International Development Research Centre) for research, but no other competing interests.