Responsiveness to Motivational Interviewing Among Latina Ovarian Cancer Survivors Participating in a Large, Well Powered RCT: A Mixed Methods Analysis

Abstract

Background: In the United States, cancer is the leading cause of death among Latinas. It is estimated one in four cancer cases could be prevented with modifiable lifestyle behaviors including diet and physical activity. However, approaches for changing these behaviors may be different based on cultural factors, including diet. Motivational Interviewing (MI) is an evidence-based approach for changing lifestyle behaviors composed of 5 constructs: evocation, collaboration, autonomy, direction and empathy. However, previous research, as well as our own anecdotal observations, suggests that Latinas may be less responsive to usual MI behavior change techniques. Purpose: The present study aims to, assess potential differences in MI responsiveness between Latinas and non-Hispanic white women who recently completed treatment for ovarian cancer. These survivors participated in the intervention arm of a large, well-powered RCT, focused on changing dietary intake. Secondly, it aims to qualitatively describe the 5 MI constructs for behavior change, using qualitative descriptive methods, among Spanish-speaking Latinas and English-speaking Latinas. Methods: Women randomized to the intervention arm received a total of 33 lifestyle counseling sessions with health coaches trained in MI. Recorded coaching sessions were randomly selected from a subsample of Spanish-speaking (SPA) Latina group (n=10), English-speaking (ENG) Latina group (n=10), and age-matched non-Hispanic white (NHW) women (n=10). All Coaching sessions (n=123) were scored using Motivational Interviewing Treatment Integrity 3.0 code, a behavioral coding system to assess MI fidelity by the interviewer (coach). Language preference for either English or Spanish was used as a proxy for acculturation. Differences in Individual MITI constructs were also assessed between groups. Adherence to dietary study goals was assessed with the AFFQ tool. Differences in total MITI scores between groups were evaluated using ANOVA, while individual constructs were evaluated using the Kruskal-Wallis test. For the qualitative descriptive analysis, a sub-sample of calls from the MITI scored sample was used (n=12). These calls were transcribed and coded for the 5 MI constructs in both the SPA and ENG Latina groups. Results: Mean MITI scores were 18.3 (SD ±1.1), 20.6 (SD ±1.0) and 21.0 (SD ±0.3) for SPA Latina group, ENG Latina group and NHW group, respectively. Spanish-speaking Latinas’ MITI scores were significantly lower (p<0.001), with a β-coefficient value of -2.7 (95%CI: -3.5, 1.8) than NHW, while ENG Latina group MITI scores were not significantly different than NHW. Of the 5 MITI constructs, direction and collaboration were significantly different between groups, with SPA Latina group having significantly lower scores than NHW, with a mean of 3.28 (SD ±0.42) for collaboration and 3.88(SD ±0.40) for direction. The SPA Latina group also had higher mean behavior counts for giving information, 7.55 (SD ±2.0), and open-ended questions, 8.78 (SD ±4.9) compared to NHW group. Conclusion: Significant differences between MITI scores were observed in Latina cancer survivors based on language preference used as a proxy for acculturation in this analysis. This secondary analysis supports the need for behavioral interventions tailored beyond language for both race and culture; potentially with special attention to direction and collaboration when using MI for these interventions.