Sexual minority women (i.e., women minoritized for their sexualities) are identified as high risk for mental health and substance use problems; however, there is no consensus on the criteria by which women are categorized as sexual minority. Though there is some evidence suggesting that certain subgroups of women are at higher risk than others based on sexual orientation, different categorization schemes for sexual orientation have yet to be compared within the same sample. Using data from the National Epidemiologic Survey of Alcohol and Related Conditions–III (N = 19,528), we examined how multiple categorization schemes (i.e., identity, behavior, recency of sexual behavior) for categorizing women who have sex with women (WSW) yield different estimates of prevalence of mental health and substance use issues. We used chi-square and logistic regression to analyze the link between sexual orientation categorization schemes and health, categorizing by 1) self-identification only, 2) behavior only, and 3) the combination of self-identification and behavior (recent vs. past). We discovered high prevalence rates of health problems among heterosexual-identified WSW who reported no recent sexual activity with women (i.e., previously had sex with women but not within the past 12 months); this category of women comprised 35% of all WSW. Step by step, we found more detailed information about these women's experiences by moving to the complex categorization scheme (the combination of self-identification and behavior). Heterosexual-identified women who have had sex with women in their past (though not recently) presented as a large group with high prevalence rates of substance use and mental disorders. These women remain invisible to researchers who categorize sexual orientation only by sexual identity or by behavior and ignore the role of behavior change over time—imprecisely categorizing such women as heterosexual or as women who have sex with men. They thus are underserved by health research and represent a significant population for further study and intervention. [ABSTRACT FROM AUTHOR]