One-dimensional measures of sexual orientation in health disparities research: Who and what are we missing?

BACKGROUND: Sexual minority women face unique health disparities. Research in this area frequently relies on unidimensional measures, either sexual identity or sexual behavior, to categorize these populations. Emerging evidence suggests, however, that greater insight might be gained by applying a multidimensional measure of sexual orientation in health research. METHODS: Using National Health and Nutrition Examination Survey data (2001-2010), we first conducted a unidimensional analysis using sexual behavior to generate categories WSW (women who have sex with women) and non-WSW. We then conducted a multidimensional analysis using a sexual identity/behavior composite measure to partition the WSW category into three subcategories: lesbian, bisexual, and straight WSW. FINDINGS: The multidimensional analysis revealed a more detailed picture of health disparities faced by sexual minority women. While WSW generally had greater odds of uninsurance and negative health status (AOR=1.6 & 1.5), the multidimensional analysis revealed that lesbian and bisexual women, but not straight WSW, had increased risk of uninsurance (AOR=2.0 & 1.7); and only bisexual women, but not lesbian or straight WSW, were more likely to report poor health (AOR=2.1). Further, WSW did not appear to lack a source of routine care, however the multidimensional analysis showed that lesbian women were, in fact, at risk (AOR=2.0). Finally, though WSW in general had greater odds of smoking (AOR=3.4), we found that straight WSW had the greatest risk (AOR=4.4), while lesbian women had the lowest (AOR=1.7). CONCLUSION: A more complete understanding of health disparities is provided by the concurrent consideration of both sexual identity and sexual behavior.