To our knowledge, this is the first study that measured changes in physicians' KAP regarding A(H1N1) pandemic influenza and its prevention by vaccination before and after the approval of the vaccine and start of the vaccination campaign. Previous studies among health care workers assessed acceptability of the A(H1N1) vaccine before its official approval and program implementation or used hypothetical pandemic vaccination scenarios [18
]. In these studies, intention to be vaccinated against A(H1N1) pandemic influenza varied from 48% to 80% among healthcare workers, comparatively to 84% of paediatricians surveyed before the beginning of the vaccination campaign in our study. Similarly to our results, a study done in Mexico reported that 72% of healthcare workers would recommend the vaccine to their patients, and they were more likely to do so when they had the intention to get vaccinated themselves [25
Results of this national survey among paediatricians indicated an important increase in paediatricians perceptions of the burden of A(H1N1) pandemic influenza and support for A(H1N1) vaccination after the beginning of the vaccination campaign. Respondents' endorsement of almost all items regarding A(H1N1) pandemic influenza and its prevention by vaccination increased after the start of the vaccination campaign. This is not surprising given the fact that the first A(H1N1) vaccine available in Canada used a novel adjuvant (AS03) for which limited information regarding the safety and immunogenicity was available. The proportion of physicians who reported they had received sufficient information about A(H1N1) vaccine also increased by 42% after the start of the vaccination campaign. This increase may be attributable to the important educational efforts done at the beginning of the vaccination campaign along with the official recommendations by expert groups and professional associations that were released in early November [27
Health professionals' knowledge about vaccines has been previously shown as a main determinant associated with their own vaccine uptake and their intention to recommend the vaccine to their patients [30
]. The association between physicians own vaccination behaviours and their recommendations to their patients was previously established [12
]. It appears to hold true in a pandemic context, as shown by our results and those previously reported in Mexico [25
Our results highlight the positive change in paediatricians' knowledge and level of support of the A(H1N1) vaccines throughout the pandemic vaccination campaign. This change may be attributable to increased education efforts and the very rare vaccine associated adverse events [36
], but may also reflect the intense media attention focused on the vaccination campaign. A recent UK study has shown that healthcare workers were more willing to accept stockpiled H5N1 vaccine during a period of high media coverage of a H5N1 outbreak in a poultry farm than 6 months after (63.4% vs. 51.9%, p = 0.009) [18
]. The increased exposure of paediatricians to severe cases of A(H1N1) disease, as observed in our results, may also have enhanced paediatricians' acceptability of A(H1N1) pandemic vaccine.
In logistic regression analysis, paediatricians' intention to get vaccinated against A(H1N1) pandemic influenza themselves was the most significant factor associated with the intention to recommend A(H1N1) pandemic vaccine to patients. Results obtained by the multiple correspondence analysis (MCA) are consistent with results from the logistic regression analysis. The negative levels were also more discriminatory than the positive ones. This is consistent with results of previous studies that have shown that knowledge and behaviors regarding seasonal influenza influenced A(H1N1) vaccination status: individuals who were not vaccinated against seasonal influenza were less likely to have the intention to receive the pandemic vaccine [37
]. Intention to be vaccinated against A(H1N1) pandemic influenza was also higher than vaccine uptake against seasonal influenza among healthcare workers usually reported in Canadian studies [39
], which was estimated at 64% in 2006 [41
]. In our study, a significant proportion of paediatricians who disagreed with the usefulness to protect children with the seasonal influenza vaccine did not intended to recommend the A(H1N1) vaccine to their patients. However, results obtained by MCA should be interpreted with precautions as only 13.6% of the variability is summarized by the first and second principal components. This is principally due to the uneven distribution of response levels.
Our study has several limitations. First, the study wasn't initially designed for a "before-after" analysis. The increase in the willingness to be vaccinated against A(H1N1) pandemic influenza observed among paediatricians may result from a response bias of respondents having more doubts about pandemic vaccination before the vaccination campaign actually started. Nonetheless, respondents' demographic and professionals characteristics as well as their attitudes toward vaccination in general were very similar, thus suggesting the two subsets of participants were comparable. Second, the dichotomization of the dependant variable ("strongly agree" and "agree" versus all others) was a conservative choice and physicians who answered "somewhat agree" were considered as having a neutral opinion, not a positive one. Third, the repartition of the answers to the dependant variable in the subset of data collected after the beginning of the vaccination campaign did not allow us to perform two multivariate analysis for the two subset. However, the model was adjusted to take into consideration the time period when the paediatricians completed the survey (before or after the initiation of the campaign). Finally, the response rate was 50% and a non-participation bias cannot be excluded. However, the response rate remains satisfactory for a mail-based survey with physicians [42
]. In addition, socio-demographic characteristics of respondents are comparable to those reported in other surveys conducted among Canadian paediatricians [42
]. Socio-professional characteristics of respondents also allow us to suppose a good representativeness.
Results of this study indicated a high level of paediatricians willingness to be vaccinated against A(H1N1) and to recommend the vaccine to their patients. Lack of knowledge on A(H1N1) vaccine, belief that A(H1N1) was not a severe disease as well as concerns over A(H1N1) vaccine safety and usefulness were barriers to paediatricians' intention to recommend it. This is consistent with barriers to seasonal influenza vaccination among healthcare workers reported in the literature [31
]. Public health interventions to promote seasonal influenza vaccination among healthcare workers should include the delivery of evidence-based information regarding influenza vaccines' safety, efficacy and usefulness. Educational campaigns should also stress out the threat posed by seasonal influenza to healthcare workers and the patients.