Strong negative associations with intention were found for having an omission bias, holding naturalistic views, for the disbelief in scientific

evidence that influenza vaccination is effective, BMS907351 and the disbelief in the relevance of the flu shot. Results of the multinominal logistic regression are shown in Table 4. HCP were more likely to have no intention to get vaccinated vs. not having made a clear decision when they reported a negative attitude towards influenza vaccination and high feelings of autonomy, when they showed a stronger omission bias, a lesser sense of personal responsibility to protect patients by getting vaccinated, when they reported high self-protection motives, and lower frequency of influenza GW786034 vaccinations in the past. When comparing having a high intention vs. not having made a clear decision, we found that HCP with a positive attitude towards influenza vaccination and a higher frequency of influenza vaccinations in the past were more likely to have a high intention

vs. not having made a clear decision. No other significant unique contributions to the prediction of having a high intention were found. The variables in the regression model explained 80% of the variance in intention (pseudo R2 = .80), with a classification accuracy of 82%. In an exploratory manner we excluded the most influential variable, attitude, from the multinominal analysis, because we hypothesized that it might overrule the (indirect) influence of other variables on intention. Only one additional significant predictor appeared Oxymatrine in this analysis: higher sense of personal responsibility significantly predicts a high intention to get vaccinated as opposed to an unclear decision when attitude is excluded. We next tested whether attitude mediates the relationship between personal responsibility and high intention vs. an unclear decision. To test for mediation, we used the SPSS macros that Preacher and Hayes [28] provide for a binary logistic regression with bootstrapping technique. The bias corrected and accelerated

(BCa) confidence intervals were set at .95 with 5000 resamples. The mediation analysis revealed that there is a meaningful indirect effect of attitude on the relationship between personal responsibility and intention (b = 1.29, BCa 95% CI [.874; 1.856]), only for participants in the categories high intention vs. no clear decision (N = 274). The fact that zero falls outside this interval indicates a significant mediation effect. For the regression coefficients for the relationship between personal responsibility and intention (high/unsure) as mediated by attitude, see Fig. 1. Table 5 shows that amongst the HCP that got vaccinated against influenza, the majority had reported to have a high intention to get vaccinated at baseline (N = 68, 73.9%). The percentage of participants that were vaccinated differed by intention, χ2 (2, N = 458) = 224.42, p < .001. Of the HCP who participated in the follow-up survey (N = 458), 90 (19.