Abstract
<title>Abstract</title> <p>Background Traditional one-size-fits-all public health communication often fails to account for individual psychological barriers, potentially triggering psychological reactance and backfire effects. To address this, a Precision Public Health approach is required to categorize populations into tailored communication paths based on their unique trust profiles. Methods Utilizing data of a cross-sectional study of 457 Iranian adults, we developed a computational framework to map objective demographic and geographic signatures onto three psychometric segments: Accepting, Ambivalent, and Resistant. The framework utilized a theory-grounded rule-based mapping (HBM and 3C models) to define target classes, followed by a Safety-First machine learning pipeline (Ensembled Logistic Regression and HistGradientBoosting). To address selection bias inherent in our highly educated, healthcare-heavy cohort (46.4% medical professionals), an internal fold-based preprocessing pipeline was strictly enforced. To prioritize ethical safety, a confidence threshold (τ = 0.53) was implemented as a Safety Valve to re-route statistically ambiguous classifications into a human-centered, dialogue-based intervention path. Results While the model achieved a modest accuracy of 47.28%, it represents a 42% relative improvement over the random baseline (33.33%).. Explainable AI (SHAP) analysis revealed that Province of Residence and Age were the primary determinants of segment membership. The τ threshold re-routed approximately 20% of uncertain cases to the Ambivalent segment, significantly reducing critical errors and ensuring the framework defaults to engagement under conditions of uncertainty. Conclusion This study provides a validated descriptive blueprint for risk-stratified health communication. By shifting the paradigm from raw accuracy to Safety-First triage, the framework offers a scalable method for health authorities to identify and address specific psychological barriers within the Iranian population, enhancing future pandemic preparedness through empathetic, data-informed intervention.</p>