Variations in influenza vaccination coverage among the high-risk population in Sweden in 2003/4 and 2004/5: a population survey

Influenza continues to be a significant health risk for elderly persons and persons with chronic conditions (e.g. cardiovascular conditions, respiratory conditions, diabetes mellitus, renal failure, reduced resistance due to disease or treatment). All European countries have recommendations for influenza vaccination for these high risk groups [1]. WHO recommends that in 2006 50% of the high-risk population should be vaccinated, and that this rate should be increased to 75% in 2010. Despite relatively uniform recommendation policies, influenza vaccination coverage rates (VCRs) vary largely between different European countries [2, 3]. In some countries (UK, Spain, The Netherlands) the 2010 target has already been reached for the elderly[4, 5]. In some other countries the 2006 recommendation has been reached (Germany, Sweden). For the diseased, however, the situation is worse. In Germany, Poland, Sweden, and Spain, for instance, the 50% target has not yet been reached [4].

To increase the VCR, it is important to find out why uptake is low. Variation in VCR may be caused at three different levels. Firstly, personal preferences and characteristics of the potentially vaccinated persons may play a role. Secondly, different medical opinions of physicians that administer or recommend vaccinations may cause variation. Thirdly, the organisation of the health care system and especially the vaccination campaign may stimulate or frustrate vaccine uptake.

Ad 1 Personal reasons mentioned can be divided into beliefs and perceived barriers. Beliefs concern the perception of the severity of flu, perceived own susceptibility for influenza, ignorance of eligibility for the vaccination and forgetfulness [6–12]. Perceived barriers may be bad experiences in the past, practical problems (like timing of the appointment for vaccination and distance to administrator), financial barriers and (a negative) attitude towards vaccinations in general [4, 7, 13, 14].

Ad 2 The role of the physician is quite important, since people that are informed about influenza vaccinations or receive a personal invitation of their physician are more likely to be vaccinated [15–21]. Research into the effect of reminders for physicians to immunize their patients was inconclusive [22, 23].

Ad 3 Financial incentives for the physician may be the remuneration system. Here we assume that extra remuneration for influenza vaccinations will lead to higher uptake rates [2]. Research into the effect depending on which health professional administers the 'shot' is scarce. The British GP Willis, however, argues that this should happen in general practice [24]. The vast majority of studies into strategies to increase VCR has been conducted in General Practice settings [11, 16, 19, 25–35], whereas few studies address interventions targeted on primary care nurses [36–38] or secondary care [38, 39]. We assume that the chance of being vaccinated will increase if more different health professionals are involved, because of the wider choice and availability. Studies into VCR so far have addressed only one or two of the levels of variation. In this study we aim to combine all three levels.

Sweden did not yet reach the WHO-recommended vaccination uptake target for 2010. However, it is possible that under the current policy conditions, the VCR will increase gradually. In that case, we would expect an increase in VCR from the 2003/4 season to the 2004/5 season. This leads to our first research question:

What is the development in VCR between the 2003/4 and the 2004/5 season in Sweden?

In Sweden the responsibility of the vaccination campaigns is delegated to the individual counties, which has resulted in different vaccination campaigns in the counties. The arrangements differ, for instance, on the following aspects: the health worker responsible for administering the vaccine (e.g., GP or public health worker), the way the administrators are remunerated (salaried or fee-for-service) and in out-of-pocket contributions for patients. This leads to our research question:

What is the influence of personal preferences of the population, medical opinion of administrators and health care system characteristics on VCRs in Sweden?

The VCR of the Swedish elderly has almost reached the WHO-target for 2006. No difference in VCR was found between the two seasons. Sweden remained far below the WHO-target for the high-risk population due to disease. The fact that no change was found may be reason for concern, since without increase, the target of 2010 will certainly not be met.

High-risk persons refraining from vaccination were mainly guided by misconceptions. These misconceptions have increased in importance over the seasons. Perceived resistance and perceived non-qualification may be solved by information campaigns. Receiving a personal invitation for a vaccination remains an important way to increase the VCR which confirms already existing research evidence [11, 16, 25, 27, 30, 34, 41–43].

The different health care arrangements in the counties appeared to contribute, to some extent, to differences in VCR for the elderly. In counties where the administrator received extra remuneration, the elderly appeared to be more likely to be vaccinated, although the relationship was not significant. Despite of little research on this subject, the existing papers addressing the subject suggested a positive effect[19, 27, 29]. Interestingly, a larger number of different health professionals that could administer influenza vaccinations led to lower VCRs. An explanation may be that when more professionals are involved, none of them will feel responsible for the vaccination. There was no evidence that elderly were vaccinated less often in regions where out-of-pocket payments were requested, which is surprising in the light of other research findings suggesting that making influenza vaccination free of charge will increase the VCR [34, 38].

The pattern found in Germany [4] and the Netherlands [13] that persons with cardiovascular disease and diabetes were vaccinated more often compared to those with pulmonary disease was not valid for Sweden. This may be due to the low VCR for the diseased in Sweden compared to Germany and the Netherlands.

Of course, this study has its limitations. The data on vaccination uptake and chronic conditions are based on self-reported data by the respondents. Although research into self-reported data compared to data from medical records revealed a satisfactory reliability of the self reported data, over-reporting as well as under-reporting could occur [4, 13, 44–48]. In Sweden in both years, 0.15 chronic condition per person was reported in the total sample. A second limitation concerned the fact that interviews were held in March and April. It would have been better to have a shorter interval between the vaccination period and the survey. However, in both years, the data collection took place in the same period, therefore, no systematic variation between the two years will occur due to different periods of data collection. A third limitation concerns the small number of respondents per county. This resulted in relatively large confidence intervals. However, due to practical limitations, a larger sample size was not possible.

Our study did not address the practical organisation of the vaccination campaign within GPs' offices. For instance, according to the literature, vaccination uptake may increase by offering vaccination clinics, offer influenza vaccination at every GP visit in the vaccination season, independent of the reason of the visit, and administer vaccines under standing orders (e.g. by practice nurses) [35, 37–39, 49–51]. However, the design of our study did not include practices as unit of observation and as a result we were not able to include this issue.

The most important finding of this study is the lack of increase in VCR, which indicates that Sweden may not be able to meet the 2010 WHO-recommendation if no further action is undertaken concerning vaccine uptake. Personal invitations highly increase the chance of being vaccinated, so it seems to be important to persuade and facilitate administrators to invite those at risk. For the elderly, restricting the number of different professionals involved in vaccination may influence the VCR positively. Financial disincentives for the elderly themselves (out-of-pocket payments) did not influence the VCR. Personal characteristics (being older and female) were associated with higher VCRs. These characteristics cannot be influenced, but the results may instigate to target information campaigns towards the younger and male persons at risk.