A questionnaire-based, cross sectional study among a population of West Midlands (WM) Police employees to identify factors affecting intention to receive pandemic influenza A (H1N1) 2009 vaccine, carried out during the winter pandemic of 2009-10.

Following approval from the Occupational Health Department at the WM Police Headquarters, visits to all 21 Operational Command Units in the West Midlands were made from December 2009 to February 2010. Fifteen centers granted permission to distribute the questionnaires, which were left for one week at each station, before being collected.

Twenty questionnaires were distributed to each participating Operational Command Unit, and 80 questionnaires were distributed in the Occupational Health Department at Police Headquarters to be completed only once by any of the staff employed by the WM Police Service. In total 380 questionnaires were distributed. All questionnaires were self administered and anonymous.

The questions were selected after research of similar literature which provided direction towards potentially significant factors. Standard questions were used where available, these were based upon those used in available literature. A pilot study was carried out on 20 volunteers to improve comprehension.

The primary outcome of the questionnaire was to determine the intention to have the pandemic influenza A (H1N1) 2009 vaccine. Participants who had already received the vaccine or those who would receive the vaccine if offered were classified as “intending to accept” the vaccine. In addition to this, the questionnaire also collected information on: sociodemographic factors, job title (later classed into office/non-office based jobs), number of dependents, personal or family illness, history of vaccination, history of pandemic influenza A (H1N1) infection, and general health beliefs, in particular those regarding the influenza A (H1N1) pandemic (See Appendix Figure A1).

Data were analyzed using STATA version 11. Univariate associations were analyzed between intention to accept vaccine and important covariates. Factors found to be statistically significant (p < 0.005) or clinically important were entered into a multiple logistic regression model. Questions on health beliefs were measured on a 5-point likert scale but collapsed into 3 categories. Scores of 1/2 were classified as low, score 3 as medium, and scores 4/5 as high. Sensitivity analyses were undertaken to exclude those who had received the vaccine already.

Between December 2009 and February 2010, 380 questionnaires were distributed and 206 completed (response rate 54.2%). Seven questionnaires were excluded as the respondents did not respond to the question about vaccine acceptance, leaving 199 for analysis (Figure 1).

One hundred and five (52.8%) respondents were female (Table 1), their mean age was 38 (Range: 18–63) and 176 (88.4%) were of white ethnicity. The majority had a non-office based job (n = 149, 74.9%) and 57.1% education to at least A’level standard or equivalent. One hundred and twenty four (62.3%) had never smoked although 22 (11.0%) were current smokers. Thirty-two (16.1%) reported a long term illness and more than two-thirds exercised at least once per week. Of the respondents 86 (43.2%) reported seasonal influenza infection in the past, 90 (45.3%) had ever received a seasonal influenza vaccine (39 (43.3% of these reported side effects)). Twenty-eight (14.1%) reported having had “swine flu” during the 2009 pandemic.

Characteristics of the sample were similar to WM police [13] in terms of the age and ethnicity distribution, although in our sample we had many more female respondents than in the police force as a whole. There were fewer smokers in our sample compared with the WM population [14] (data were not available on smoking in the West Midlands police) (Appendix Table A1).

Eighty-one (40.7%) respondents felt there was a low threat of pandemic influenza A (H1N1) infection to the public, 90 (45.2%) felt there was a medium threat and only 28 (14.1%) felt there was a high threat to the public. Approximately half (n = 94, 49.8%) of respondents believed there was a low probability of catching pandemic influenza A (H1N1) virus and only 12% (n = 24) felt there was a high probability. Correspondingly, most (n = 93, 47.5%) felt the threat to their health was low, while 23.6% (n = 47) felt that the threat to their health was high. The majority of respondents (139, 69.8%) felt the media had overestimated the threat of the pandemic virus (Table 2).

Of the 199 respondents, 14 (7.0%) had already received the vaccine, a further 65 (32.7%) said they would accept the vaccine if offered, and 80 (40.2%) said they would decline. 40 (20.1%) were still unsure if they would accept or not. Overall, therefore, intention to receive the vaccine was 79/199 (39.7%). The remaining analyses are based on the 159 respondents who either stated yes or no.

The most common reasons for accepting the vaccine (Figure 2A) included worry about catching “Swine flu” (n = 42, 53.2%), infecting others (n = 40, 50.6%), and missing work (n = 20, 25.3%) (Figure 2A). Sixteen (20.3%) would accept because they would follow advice from employers/occupational health/Department of Health. The overwhelming reason for declination of the vaccine (Figure 2B) was worry regarding potential side effects (n = 45, 57.0%). However, 9 (11.4%) would/did decline because it was inconvenient, 9 (11.4%) were worried about the vaccine causing “Swine Flu”, 5 (6.3%) would decline because they had already had seasonal influenza vaccine, 7 (8.9%) had doubts about vaccine efficacy and 4 (5.1%) would decline because they had already had pandemic influenza infection that year. Other reasons stated by respondents for declination were: respondents not perceiving themselves to be at risk, simply not wanting to be vaccinated, fate, needle phobia, belief that “Swine flu” is only a threat to those in poor health, or contraindications to the vaccine itself.

On univariate analysis (Table 3), three determinants related to either demographics or past medical history were significantly associated with the intention to accept the vaccine: non office-based employment (OR: 2.24, 95% CI: 1.06–4.72), having received a seasonal influenza vaccine in the past (OR: 4.08, 95% CI: 2.1–7.93), and history of pandemic influenza A (H1N1) infection (OR: 2.81, 95% CI: 1.03–7.68).

Three determinants related to health beliefs and perceptions were significantly associated with the intention to accept the vaccine (Table 4). The belief that “swine flu” posed a high threat to the public was significantly associated with acceptance (OR: 4.64 95% CI: 1.65–13.07). The belief that “swine flu” was a high risk to the respondents’ health was also significantly associated (OR: 3.01 95% CI: 1.36–6.68). In addition, if respondents felt there was a high likelihood of “swine flu” infection, this was significantly associated with intention to accept the vaccine (OR: 4.27 95% CI: 1.42–12.83).

After adjustment (model 1) for age, sex and smoking status, participants who had ever received seasonal influenza vaccine remained significantly more likely to accept the pandemic influenza A (H1N1) vaccine (OR: 7.92, 95% CI 3.38–18.53), with employees over the age of 40 (OR: 0.36 (0.16–0.82), and females (OR: 0.47 (0.23–0.99)) significantly less likely to accept the vaccine, and current smokers more likely (OR: 4.89 (1.05, 22.72)). Additional inclusion of ethnicity and type of job (model 2) suggested a trend towards employees of non-white ethnicity and a job outside the office being more likely to accept the vaccine, although these results were not statistically significant. Excluding participants who had actually received the vaccine (leaving only those stating “intentions”) produced similar results.

Adjustment by age, sex, smoking status and prior receipt of vaccine in a model including determinants related to health beliefs and perceptions highlighted the importance of belief that pandemic influenza A (H1N1) virus was a high threat to the public (OR: 4.44 (1.4, 14.7)), they had a high likelihood of catching the virus (OR: 5.07 (1.44–17.93)) and that “swine flu” was a serious problem to health (OR: 2.86 (1.14–7.15)) to acceptance of the pandemic vaccine (Table 4).

This study showed that 39.7% of the police employees we sampled had already been vaccinated or would accept the pandemic Influenza A (H1N1) vaccine. Those who stated that they would decline the vaccine if offered numbered 40.2%, and the remainder were unsure.

To our knowledge, this is the only study addressing the issue of vaccine acceptance in essential service providers. The characteristics of our study sample were broadly comparable to the overall WM police [13] (or WM population) [14], although in our sample there were many more women. Since women were less likely to indicate acceptance of the vaccine, we may therefore have underestimated likely uptake rates. Social acceptability and interviewer bias were limited through the use of self administered, anonymous questionnaires.

The small sample size and low response rate limited the statistical power of this study. 380 questionnaires were distributed with a response rate of 54.2% accounting for 2.0% of the sample population. Other than the characteristics stated above there was no way of determining the views of non-responders. Responder bias may have caused an overestimate of vaccine acceptance; those not interested in being vaccinated may have been less likely to participate in the study. This could have been further compounded by lack of promotion of questionnaires.

Questionnaire distribution was targeted at Operational Command Units only; smaller stations were not included in this study. There could be differences between the employees of each. We attempted to counteract this by distributing questionnaires to the Occupational Health Department of WM police where all employees have access. Additionally, use of short questionnaires limited the breadth and depth of responses in certain areas of enquiry.

Our study population was limited to the police force; therefore, our results may not be applicable to those in other essential services.

The findings of this study can be used to improve immunization strategies if vaccination of essential service providers is implemented in future influenza outbreaks. It is clear that health perceptions and attitudes play a major role in influencing the decision to be vaccinated. Targeted education programs could be implemented to address the misconceptions held by many individuals about health issues and their management; this is particularly applicable in the case of immunization and for pandemic vaccine the key issue appears to be worry about side effects. Wider dissemination of studies depicting accurate portrayal of the side-effect risks and reassurance of the benefits of vaccination could substantially improve uptake among the 40% “decliners” and the 20% “unsures”. Other strategies that have been shown to increase uptake rates among healthcare workers which could be included include: use of mobile services to provide flexible delivery of vaccine [23], opt-out systems [24], and reminder schemes [25].

We also advise that future research is carried out within both the police and fire service, using a larger sample size; this could be achieved through use of employee distribution lists. Active promotion or incentives could be used to increase response rate. Furthermore, it would be beneficial to obtain qualitative data through interviews of study subjects or focus groups.