Evaluating the outcomes and implementation determinants of interventions co-developed using human-centered design to promote healthy eating in restaurants: an application of the consolidated framework for implementation research

**Background:** Restaurants are an underutilized setting for promoting healthier eating, especially independently-owned, minority-serving establishments. Latin American restaurants are particularly important given their prevalence and the disproportionate burden of diet-related diseases in Latin American communities. This study aimed to examine outcomes of interventions co-developed using Human-Centered Design (HCD) in two Latin American restaurants, assess implementation outcomes (acceptability, fidelity, sustainability), and elucidate determinants using the Consolidated Framework for Implementation Research (CFIR).

**Methods:** The study used a mixed-methods, longitudinal design with data collected pre-, during, and post-intervention. Two Latin American restaurants in New York City participated: a counter-style Puerto Rican restaurant in a food hall (R1) and a full-service Mexican restaurant (R2). The HCD process involved workshops with owners and staff to define problems and co-develop solutions. R1's intervention introduced new healthier menu items (verduras—seasonal non-starchy vegetables with traditional spices—and avocado slices) with menu redesign. R2's intervention focused on social media promotion of existing healthier menu items. Intervention outcomes were assessed through daily sales data analyzed using interrupted time series (STATA BE 17 "itsa" command) and the Nutrition Environment Measures Survey for Restaurants (NEMS-R). Implementation outcomes and determinants were assessed through 14 site visits, 19 staff interviews, 31 customer intercept interviews, social media monitoring, and post-implementation key informant interviews (average 37.2 minutes). Qualitative data were analyzed using directed content analysis with a priori codes based on CFIR.

**Key Results:** Intervention outcomes differed by restaurant. In R1, HMI sales increased by 31 units at intervention onset, followed by a daily decrease of 0.22 units; post-intervention sales were not significantly different from baseline. HMI as a proportion of menu items increased from 15% to 53% in R1. NEMS-R scores in R1 improved from 4 (pre) to 7 (during) to 5 (post). In R2, no significant changes in HMI sales were found; HMI remained at 29% of menu items. NEMS-R scores in R2 were 11 (pre and during) and 8 (post). Social media promotion of HMI in R2 increased from 22% of posts (pre) to 48% (during) but dropped to 11% (post). Acceptability was high among customers and staff in both restaurants, though R1's chef showed initial resistance. Fidelity was high in R1 (new menu items maintained) but lower in R2 (social media posting decreased post-intervention). Sustainability was higher in R1 (verduras kept on menu; avocado side removed due to cost and sourcing issues) than R2 (social media postings decreased after testing). CFIR analysis revealed key facilitators: HCD process fostering owner buy-in, owner innovativeness and tolerance for ambiguity, centralized decision-making, and existing resources/skills. Key barriers included: low tension for change limiting intervention extent, staff capacity constraints, COVID-19-related staffing issues, small restaurant spaces, and administrative burden on owners. Outer setting factors (policies, peer pressure, incentives) had limited perceived influence due to low awareness.

**Clinical Implications:** This study demonstrates that HCD can successfully engage independently-owned restaurants in developing acceptable interventions, but impact on sales of healthier items is mixed and context-dependent. The application of CFIR to restaurant settings reveals that inner setting factors (restaurant structure, implementation climate, centralized decision-making) and individual characteristics (owner innovativeness) are critical determinants of implementation success. Future interventions should address structural barriers like staff capacity, time constraints, and resource needs. The findings also highlight the need for supportive government policies rather than punitive approaches to engage independently-owned restaurants in health promotion. This research expands implementation science to complex community-based settings, informing future interventions to address diet-related health inequities in underserved communities.