The role of DSM-5 diagnoses and COVID-19–related cognitive appraisals in vaccine refusal: a SCID-5–based cross-sectional study

Background Vaccine refusal remains a major public health challenge. Although previous studies have linked psychiatric symptoms to vaccine hesitancy, nearly all relied on self-report scales rather than clinical diagnoses, and few have jointly examined psychiatric diagnoses alongside pandemic-related cognitive appraisals. This study compared vaccinated and unvaccinated individuals with respect to DSM-5 diagnoses established through a structured clinical interview and their perceptions and attitudes toward COVID-19, aiming to identify the independent predictors of vaccine refusal. Methods In this cross-sectional study, 215 adults attending a family medicine outpatient clinic between July and September 2022 were enrolled: 101 vaccinated (≥ 3 doses) and 114 unvaccinated (0–1 dose). Psychiatric diagnoses were established using the SCID-5-CV, administered face-to-face. Perceptions and attitudes toward the pandemic were assessed with a validated scale. A hierarchical logistic regression was conducted, entering demographic variables, psychiatric diagnoses, and illness perception subscales in sequential blocks. Results The overall prevalence of psychiatric diagnoses did not differ between groups (48.5% vs. 51.8%, p = .635), and the psychiatric diagnosis block did not significantly improve model fit (Δχ² = 13.6, p = .058). In contrast, illness perception variables markedly improved the model (Δχ² = 37.2, p < .001). Stronger COVID-19 conspiracy beliefs (OR = 1.877, 95% CI 1.332–2.643, p < .001) and lower perceived dangerousness of COVID-19 (OR = 0.486, 95% CI 0.306–0.771, p = .002) were the strongest independent predictors of vaccine refusal. Younger age remained a significant predictor across all models (OR = 0.921, p < .001), whereas the association of female sex was attenuated after adding perception variables. The comorbidity of depressive and anxiety disorders was numerically more frequent in the unvaccinated group but remained at the margin of significance ( p = .054). Conclusions Psychiatric diagnoses established through a structured clinical interview were not independent predictors of vaccine refusal, whereas modifiable cognitive appraisals, particularly COVID-19 conspiracy beliefs and low perceived dangerousness of COVID-19, were the strongest predictors. Interventions to reduce vaccine refusal should target attitudinal factors rather than psychiatric diagnoses.