Latent Class Analysis of Health and Reproductive Risk Behavior of Women in Appalachia

Background and Purpose: Although there has been a public outcry to address the opioid epidemic, the impact of opioid abuse on women in rural Appalachia remains understudied. Further analysis of the risks that this population faces are necessary to end gender based violence, broadly defined, and to address social work’s grand challenge of closing the health gap. In this study, a latent class analysis was used to categorize combined risks for rural, Appalachian, drug-using women who were recruited from rural jails.

Methods: This secondary data analysis used survey data from a larger study that was collected from (n=400) rural, drug-using, Appalachian women recruited from jails. Latent class analysis performed in Mplus was used to estimate a model of heterogeneity in the sample. Seven variables of health and sexual risk behavior were included as indicators of a latent categorical variable. The seven variables follow: (1) pap testing or mammogram in the past year, (2) condom use, (3) any general health screenings received in the past year, (4) sexually transmitted disease history, (5) male partner ever injected drugs, (6) main male partner ever incarcerated, and (7) bartering using sex. Consistent with recommendations, a variety of fit statistics, indlcudeing the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), adjusted Bayesian Information Criterion (aBIC), and Vuong-Lo-Mendell-Rubin likelihood ration test (VLMR-LRT), and its associated p-values, were used to assess model fit.

Results: The women in this sample fit two distinct categories of risks. These categories were defined according to their male partners’ risks and histories, which in turn elevated individual women’s risk status. The group with the lower risk profile was termed “Women whose partner is low risk.” These group members were more likely to have a primary male sexual partner who rarely injected drugs and who did not have an incarceration history. Nearly all of the women in the higher risk profile, which was termed “Women with high risk partners,” identified a primary male sexual partner with high rates of injecting drugs and histories of previous incarceration. The distinction between the two profiles was further buttressed by prominent differences in drug use, most notably in the type of substances and their use during sex as it related to reasons for not using a condom.

Conclusions and Implications: The presence of these two profiles suggests that rural Appalachian women’s risk is more heterogeneous than has been typically understood and exists within a context of systemic poverty and drug use. Findings further indicated that women’s risks were based on the risk levels of their male sexual partners. A harm reduction approach points towards interventions that reduce these risks for women who engage in sexual activity while injecting drugs, by providing increased access to IUDs and STD prevention methods.