Background: Low health literacy is a significant problem affecting our country.  While the associations between low health literacy and poorer health outcomes have been well documented (Berkman et al., 2011), the literature lacks evidence of effective strategies to address health literacy in the context of health behaviors such as diet and physical activity (PA). Likewise, few interventions have reported on how health literacy status influences performance and engagement in the intervention. Two potential intervention strategies include the teach back method or teach to goal approach and interactive voice response (IVR) technology. These strategies hold promise as a means of improving health literacy and reaching vulnerable, low health literate populations, but these strategies have not been widely explored in the literature (Paasche-Orlow et al., 2005; Baker et al., 2011; Schillinger et al., 2009; Bennett et al., 2012; Piette et al., 1999).

Primary Aims: This research was embedded in a larger trial, Talking Health, which is a 6-month, 2 group randomized controlled trial to determine the effects of a health behavior intervention on reducing sugar-sweetened beverage (SSB) consumption in Southwest Virginians. The primary aims of this study were to examine the associations between health literacy status and 1) number of rounds of teach back needed to reinforce key concepts, 2) proportion of correct answers on the first round of teach back, 3) level of intervention engagement (i.e., completion rates for teach back call, IVR calls, and small group classes), and 4) perceptions of the intervention components. Methods: The data reported represent the first 3 cohorts of the Talking Health trial including participants in Lee, Giles, and Pulaski Counties. Eligibility requirements included being 18 years or older, English speaking, consuming at least 200 calories per day from SSB, able to participate in moderate intensity PA, and having reliable access to a telephone. Data were collected at baseline and at the 6-month follow-up assessment. Health literacy was assessed using the validated Newest Vital Sign. Participants were randomized to a behavioral intervention aimed at decreasing SSB consumption (SipSmartER) or to a matched-contact control group targeting PA (Move More). Both groups participated in 3 small group education sessions, received a live teach back call, and 11 supportive IVR calls. Participants completed a summative evaluation at the 6-month follow-up, which captured their perceptions of the intervention components. ANOVAs were used to measure differences in outcomes by health literacy status, randomized condition, and interactions. Results: Of the 125 enrolled participants, 92.0% were Caucasian, 76.8% were female, 29.6% had d high school education, 64.0% had <$25,000 annual household income, and 32.8% had low health literacy skills. Eighty-five participants (68.0%) completed the teach back call. The overall model when looking at the degree to which health literacy status and randomized condition predicted the number of rounds of teach back needed to reinforce key concepts was significant (F= 8.323, p < 0.001). Out of 3 possible teach back attempts, participants in the low health literacy category required a significantly higher number of teach back attempts as compared to those with high health literacy (F= 16.769, p <0.001), and participants randomized to Move More required a significantly higher number of teach back attempts compared to SipSmartER participants (F=7.296, p= 0.008). Similarly, the overall model when looking at the degree to which health literacy status and randomized condition predicted the proportion correct on the first round of teach back was significant (F= 9.836, p<0.001), such that those with higher health literacy status  (F= 19.176, p< 0.001) and those randomized to SipSmartER condition answered a significantly higher proportion of questions correct (F= 9.783, p= 0.002). Intervention engagement including completion of the small group education sessions, the live teach back call, and the IVR calls did not vary significantly across randomized condition or literacy levels. Low health literate participants had a significantly higher overall perceived satisfaction with the IVR, as compared to high health literate participants (F= 5.849, p= 0.020). However, perceptions of other intervention components (e.g., small group sessions, teach back call, personal action plans, drink diaries/exercise logs,) were similar among participants with low and high health literacy status and across randomized conditions.

Conclusion: These data confirm the importance for multiple teach back opportunities and additional exposure to health information to ensure participant comprehension of key intervention content"in particular for those with lower health literacy. This research also supports that IVR is an effective approach to reaching vulnerable, low health literate populations. Future research should investigate the efficacy and cost-effectiveness of utilizing teach back methods delivered using automated technologies. Future research also is needed to determine how teach back performance are related to other study factors such as retention, engagement, and health outcomes.