Targeting vulnerability and policy conflict: Evaluating the impact of integrated safety nets and nutrition-sensitive agriculture on food security and dietary diversity in Northwest Ethiopia

Background Ethiopia's persistent food insecurity and stunting necessitate a shift from emergency aid to resilience-oriented programs, such as the Productive Safety Net Program (PSNP) and Nutrition-Sensitive Agriculture (NSA). This study assessed how integrating PSNP and NSA interventions affects household food security and dietary diversity. Methods A cross-sectional mixed-methods study was conducted among 586 mother–child pairs (children under two years old) in Northwest Ethiopia, comparing an intervention district (Ebinat Woreda) with a non-intervention district (Farta Woreda). Data collection was conducted using validated tools, including the Household Food Insecurity Access Scale (HFIAS), Food Insecurity Experience Scale (FIES), Household Dietary Diversity Score (HDDS), and qualitative interviews. Outcome variables included household food security, household dietary diversity, and child dietary adequacy. Ordered logit was used to assess the key predictors of the outcome variables of interest. Results Food insecurity affected 40.3% of households. While intervention groups (PSNP + NSA) showed higher dietary diversity, they also reported more frequent crop failures and hunger. Dietary diversity was alarmingly low, with only 17.1% of women and 35.8% of children meeting Minimum Acceptable Diet standards. Regression analysis revealed a "smallholder paradox": households with < 1 ha of land and those in lower wealth quintiles were significantly more likely to be food secure than those with > 2 ha or higher wealth (P < 0.001). Male-headed households were 4.51 times more likely to be food secure (P = 0.00). Key positive drivers included agricultural extension services (AOR = 2.65 for food security; 2.12 for dietary diversity) and paternal literacy (AOR = 2.02).Conversely, low income (< 10,000 ETB) and program participation were linked to higher vulnerability, reflecting the targeting of distressed populations. Qualitatively, a wheat-farming initiative was found to penalize crop diversification, directly undermining nutrition-sensitive objectives. Conclusion Integrated PSNP and NSA interventions show promise but face challenges. Their effectiveness is hindered by targeting challenges that expose the most vulnerable and by key policy misalignments that weaken nutrition-focused goals. Policy implications point to the need for stronger alignment between social protection and nutrition‑sensitive agriculture strategies, with improved targeting mechanisms to ensure that the most food‑insecure households benefit.