Abstract C78: The effect of navigation on mammography uptake among Latinas: Effect modification by facility and neighborhood characteristicsedit
Purpose. Improving mammography among Latinas may improve ethnic disparities in stage at diagnosis, morbidity, and quality of life. In response, multiple interventions have been developed and assessed, including patient navigation. Little is known about how efficacy varies across socioenvironmental contexts. Such work is warranted for future comparative effectiveness research and clinical implementation. Objective. The study examined the interactive effects of patient navigation with contextual factors (facility and residential neighborhood characteristics) on mammography uptake among a Chicago-based sample of Latinas. Methods. The larger trial, “Patient Navigation in Medically Underserved Areas”, is a five year project to assess the effectiveness of primarily phone-based navigation services within 3 hospitals in South Chicago. The current study’s analytic sample includes 715 women who: 1) identified as Latina; 2) were referred for mammography; and 3) had age, insurance status, type of mammogram, and zipcode information documented in their medical records. We classified participants using an as-treated operationalization (navigated or not). The facility variable of interest was accreditation as a Breast Imaging Center of Excellence (BICOE; 1 yes and 2 no). The two neighborhood characteristics of interest were median household income and percent of Latino residents (%Latino) based on American Community Survey 2007-2011 data. The outcome of interest was days to mammography uptake, defined as days between randomization and attendance at the referred mammography appointment. Results. First, Cox regression survival models were conducted that adjusted for patient age, insurance status, and type of mammogram (screening or diagnostic) and incorporated main effects of facility and neighborhood characteristics. There were significant interaction effects of navigation with neighborhood %Latino (p<.0001) and facility BICOE status (p = .01). There was also a non-significant interaction with neighborhood median income (p =.08). The analysis next conducted models, wherein we stratified by tertiles for %Latino and by facility BICOE status (yes/no). For these models, it included covariates and the main effects of non-stratified contextual variables (e.g., neighborhood median income). Among residents of neighborhoods with the least %Latino, non-navigated women had a greater number of days to mammography uptake relative to navigated women, HR =2.8, 95%CI[1.6, 4.9], p<.0001. Navigation effects were not significant for residents of neighborhoods with greater %Latino (ps=.19-.23). Among BICOE patients, non-navigated women had a greater number of days to mammography uptake relative to navigated women, HR=2.5, 95%CI[1.5,4.5], p =.001. Navigation effects were not significant for patients attending non-BICOE facilities (p=.55). Conclusions. The study suggests navigation may be particularly effective within more resourced facility settings and for Latinos living outside of ethnic enclaves. Future studies are warranted to confirm our findings and assess the potential of these services across diverse clinical settings.