Communication-related vulnerability at the last mile: a mechanism-based framework

Social vulnerability is partly shaped by communication barriers that limit people's access to, understanding of, and response to hazard information, so improving disaster risk outcomes depends on strengthening systems that reduce these communication gaps. We propose a mechanism-based framework for understanding how communication-related vulnerabilities shapes last-mile outcomes in low-trust, low-connectivity settings. Drawing on risk governance, social vulnerability, behavioural decision-making, and people-centred early warning principles, the framework identifies five mechanisms shaping last-mile communication: trust, message design, access, information integrity, and social meaning-making. The framework is applied in four flood-prone districts of Khyber Pakhtunkhwa, Pakistan (Peshawar, Charsadda, Dera Ismail Khan, and Nowshera) to understand how flood risk information is received, interpreted, and acted upon. Findings from a household survey (n = 210) reveal limited receipt of official warnings and widespread dissatisfaction with the clarity, accessibility, and practical usefulness of information. Households rely mainly on television and interpersonal networks, reflecting institutional reach gaps and a persistent digital divide. Key barriers, including low literacy, language mismatches, digital exclusion, affordability constraints, and mistrust, further limit comprehension and action. We propose a hybrid, equity-oriented communication model combining mass media, low-bandwidth mobile alerts, trusted community and faith-based institutions, and trained intermediaries, supported by feedback mechanisms to enhance credibility, accessibility, and response. The study demonstrates how inequalities in communication systems translate into differential access to warnings, interpretation of risk information, and capacity to act, and contributes a mechanism-based framework alongside practical pathways for reducing communication-related vulnerabilities in low-resource, low-trust settings.