- Health disparity scholars and researchers call to expand the conceptualization of health disparities research beyond the predominant and long-standing race-based analyses. The call requests the inclusion of frameworks and theories that reflect the complex, multi-level and multifactorial social processes that yield health disparities. Intersectionality is a theoretical framework that is a product of Black feminist scholars and activists from the 1970s equipped to answer this call. The framework, in its basic tenets, emphasizes the multiple, mutually constitutive and simultaneous relationships between social categories (gender, race, class, sexual identity, ability) and the positions (advantage and disadvantage), identities and processes (sexism, racism, and other forms of discrimination) that arise from them. The framework calls for the acknowledgment of such complex power relationships in social processes, such as health disparities.
Intersectionality has traditionally been applied via qualitative research methods quantitative applications are in a nascent stage and face a number of methodological challenges. This dissertation identifies an alternative population-based approach that is intuitive and has practical significance while remaining loyal to the principles of intersectionality. The approach identifies groups representing the intersection of social disadvantage, across the categories of gender, race and class, and uses the most socially advantaged group as a referent for comparison. The three manuscripts that constitute this dissertation are empirical applications of this population-based intersectional approach and assess disparities in self-reported health status as captured by the physical component score (PCS) of the SF-12 (a subjective measure of health), cardiovascular disease and stroke diagnoses (objective measures of health), and health care access as measured by insurance coverage and having a usual source of care (structural factors affecting health). The 2010 Household Component of the Medical Expenditure Panel Survey of the Agency for Healthcare Research and Quality was analyzed for the three separate studies.
Self-reported health status disparities were assessed using survey-weighted OLS regression. Disparities were detected; groups of low income women belonging to racial minority groups (Black, Native, Asian and Multiracial) reported significantly lower PCS scores than the White high income male group. Different socially advantaged referents were used, and this changed the magnitude of the decrease of the comparison groups, highlighting the intersectionality tenet that social categories modify one another. This suggests that in health disparities research it is insufficient to use the White group as a referent without paying attention to the gender and social class as well.
Using logistic regression, disparities in cardiovascular disease (CVD) and stroke diagnoses among women were assessed. The models incorporated healthcare access factors and risk behaviors, since both components have an important role in their diagnosis. The specific socially disadvantaged groups that experienced an increase in the odds of being diagnosed with CVD/stroke were contingent upon whether the diagnoses were acute (heart attack or stroke) or chronic (coronary heart disease, angina and other unspecified heart conditions). The study highlights the importance of accounting for multilevel factors when assessing CVD/stroke disparities, as the disparities change contingent upon the factors considered. The lack of improvements in reducing CVD/stroke disparities in the last 30 years may in part be due to the incomplete information that race-based analyses provide.
Healthcare access disparities among groups of different social advantage were assessed using logistic regression and average marginal effects. The odds of being uninsured were higher for all but one group (Multiracial low income) compared to the most socially advantaged referent (White high income male), with Asian low income women having the highest increase in the probability of not being insured. The odds of not having a usual source of care were estimated by stratifying by gender and insurance status. Among insured women the Native low income group had the highest increase in the probability of not having a usual source of care (USOC); for insured women it was the Asian low income group. Among uninsured males, Native low income males had the highest increase in the probability of not having a USOC; for insured males it was the Asian low income group. These results show that belonging to a socially disadvantaged group hampers meaningful access to health care. The disadvantage persists even among the insured, where disparities in having a usual source of care are marked.