This is a case–control study with vaccination status as the ‘outcome’ and personal or external environmental factors as ‘exposures’. A case–control study design was chosen because of a low prevalence of eligible cases. A street intercept interview method enabled the interviewers to screen and approach a larger number of people, according to the outward appearance of their age. This probably lowered the rejection rate and enabled a greater control in completing the questionnaire. It was estimated that a larger number of people would have had to be approached should a telephone or postage survey been used. The low response rate (41.7 %) was attributable to the difficulty in finding cases, as the excess controls approached by the interviewers were counted as non-responders. Moreover, the interviews were conducted in summer time when the street temperature was >30 °C, the streets were crowded and no incentive was offered.
Multi-dimensional factors have contributed to people’s choice of whether or not to receive vaccination. These factors comprise of social, environmental and economic dynamics in a specific context. The factors were put in a multiple logistic regression model and statistically adjusted for age, employment status, in receipt of social security, and all independent variables. Before statistical adjustment, most of these factors had statistically significant crude odds ratios. The variables affected each other and many became non-significant after adjustment. There would be a confounding effect between variables.
Discussion on study results
The majority of the cases (80.8 %) and controls (93.9 %) were not aware that they were in a group recommended by the health authority to receive influenza vaccination. Among the controls, a higher percentage (71 %) deemed vaccination to be ‘unnecessary’. This revealed a failure of DH and health professionals in communicating the message that ‘vaccination is recommended’ to this age group. Given that there was an association between ‘knowing oneself to be in the recommended group for flu vaccine’ and vaccination, better communication of the risks might have improved the vaccination rate. A health promotion strategy on empowerment and enhancement of knowledge on this issue needs to be planned and supported by health-care policy.
Studies suggested that previous influenza vaccination was a predictor for subsequent vaccination (OR 1.62-5.40) [19–22]. However, past behaviour does not provide an insight into the reasons why a person chooses to be vaccinated.
The vaccination coverage rate is price sensitive. This was demonstrated in this study and in countries which provided vaccine reimbursements to users [23, 24]. To receive influenza vaccination, most (95 %) people aged 50–64 years in the general Hong Kong population had to pay out-of-pocket. In this study, the odds of the cases being ‘eligible for free government vaccine’ were 6.4 times the controls. Among the cases, half (52 %) of them attended a private clinic or hospital and paid the vaccination fee. Many study cases and controls expressed they were willing to receive the vaccine if it was free or subsidised. Such a vaccination service could possibly increase the vaccination rate.
There was only a mild association between chronic disease(s) and vaccination and the association was insignificant after the OR was adjusted (OR1.13, 95 % CI 0.65-1.96, p = 0.67). This result contradicted the findings of many studies that indicated that the presence of chronic diseases was one of the most persistent factors associated with vaccination [19, 20, 25–30].
‘Accept advice by health professional’ was moderately associated with vaccination (OR2.67, 95 % CI 1.19-5.99, p = 0.02). Several other studies have shown that doctors’ and health professionals’ advice was associated with influenza vaccination [17, 31]. Health professionals had a duty to recommend vaccination to high-risk groups in order to protect them from influenza and severe complications.
‘Had family member received flu vaccine’ was associated with people’s uptake of the vaccination, but ‘accept advice from relatives and friends’ was not. In Japan, advice from health professionals, family and/or close friends was strongly associated [17]. In the USA and other western countries, advice from family and/or close friends was not a significant factor in acceptance of influenza vaccination [32, 33]. This could possibly be due to the differences in cultural backgrounds between individuals in these countries.
This study showed no association between vaccination and smoking and drinking. It is uncertain whether people were consistent in their health behaviours. Studies have proven that smoking is not associated with vaccination [21, 34]. No data was found on other health behaviours, such as drinking or frequent exercise, having a link to vaccination.
Given past experiences of infectious disease epidemics in Hong Kong, people may be more inclined to receive vaccination to protect themselves in anticipation of the occurrence of a disease epidemic such as SARS or swine influenza.
Discussion on vaccination policy
Previous research has suggested that newly issued recommendations are not quickly embraced by the majority of citizens. In the US, government National Health Interview Survey data did not show a marked increase in vaccination rates among adults aged 19–49 and 50–64 years after the US Advisory Committee on Immunization Practices expanded its recommendations to these subgroups in 2000 and 2010, respectively [31, 35].
This vaccination policy limited the government vaccination free service to those suffering economic hardship and chronic diseases among 50–64 year-olds. Although the price of receiving an influenza vaccination constitutes a minute percentage of monthly income, this does not necessarily mean socio-economically deprived groups who are ineligible for free vaccination would be willing to pay for the vaccine. Subsidised vaccination would attract those who are willing to pay at a discounted price. Health providers could be engaged, with or without incentives, to promote the benefit of vaccination. In addition, DH should consider health promotion messages addressing factors with strong associations to encourage payment by the individual. These factors included ‘the perception of having severe or moderate symptoms when contracting flu’, ‘knowledge of being in the recommended group for flu vaccine’ and ‘good vaccine protection for healthy adults’.
Discussion on study strength and limitations
A case–control design enabled the measurement of many different exposures at once and for the combined effects of exposures to be examined. In addition, data were collected within a short time-frame. One of the important limitations of this case–control was the temporal sequence and reverse causality. It is difficult to interpret the time sequence of the exposures and the outcomes. For example, it is uncertain whether perception of the safety of the influenza vaccine was a cause or a consequence of vaccination. Other limitations of this case–control include the information and recall bias of the respondents, and the inability to estimate the coverage of vaccination in this age band.
One limitation of using the street-intercept method would be the possibility that the interviewers approached those who looked 50–64 years and, potentially missed a number of younger and older looking individuals; the extent of this bias is difficult to assess. Another bias would be due to the sampling of respondents from different locations, e.g., on public and private estates, in train stations and shopping malls. A comparison of the demographic characteristics of the samples collected in different locations, and those of the relevant population, would be useful to identify potential bias.
The study results have important implications for the general population aged 50–64 years in Hong Kong. There would be considerable differences between cultures, beliefs, norms and external environments - such as health systems and service provision - which have to be taken into consideration when applying the results to other populations. Further studies on the local vaccination policy and the views of health professionals would provide a comprehensive account of the low vaccination coverage in this age group.
