Dalarna University's logo and link to the university's website

Background: Rural communities in the Horn of Africa face high exposure to interconnected health risks, such as zoonotic diseases, food and water safety issues, and antimicrobial resistance, that benefit from an integrated One Health (OH) approach. However, these populations remain underserved by conventional top-down health interventions. We assessed whether community conversations (CCs), a participatory and action-oriented dialogue method, could improve OH-related knowledge, attitudes, and practices (KAP) among (agro-) pastoralists. Methods: A pre-post study was conducted in 19 communities in Ethiopia, Kenya, and Somalia from June to October 2024. Community members (n = 358) participated in four CC sessions over 2-3 months, guided by trained facilitators using a standardised facilitation manual. KAP outcomes were assessed using validated psychometric models with fixed-parameter scoring: a bifactor 2-PL IRT model for knowledge, a bifactor graded response model for attitudes, and a network-based composite score for practices. Within-person changes were assessed using paired Wilcoxon or McNemar tests, and predictors of KAP change were examined using mixed-effects regression models accounting for community clustering. Results: Participation in the CC intervention was associated with significant improvements in OH knowledge (mean theta scores pre- and post-intervention: 0.04 vs 1.39; p < 0.001), attitudes (0.07 vs 1.07; p < 0.001), and practices (mean network-weighted composite scores pre- and post-intervention: 1.06 vs 1.66; p < 0.001). Positive gains were observed across all three countries and most participant subgroups, with shifts in knowledge and attitudes emerging as significant predictors of improved practices. Conclusions: CCs were effective in improving OH KAP in the study communities. The findings support CCs as a scalable, community engagement model with potential to strengthen health behaviours in resource-limited contexts.