The impact of European vaccination policies on seasonal influenza vaccination coverage rates in the elderly

Abstract

Despite strong recommendations, seasonal influenza vaccination coverage rates (VCRs) remain limited in Europe, even in high-priority groups. There is a need for understanding the impact of vaccination-related policy elements and barriers toward vaccination. We aimed at assessing essential elements of vaccination policies and the influence of policy-related driving factors on VCRs among elderly. Sixteen European National Vaccine Industry Groups (NVIGs) were included in a survey to make an inventory of vaccination policies implemented at national level (2009). The questionnaire was structured around four topics: management of vaccination programs; influence of health care workers (HCWs); role of information/ communication campaigns; and access to vaccine. The information retrieved was put in relation to current VCRs among the elderly (≥65 y). Correlation coefficients between policy elements and vaccination rates were calculated. Several policy elements may be suitable to increase influenza vaccination uptake in the elderly, but only few countries make use of all alternatives. Countries with good monitoring systems regarding vaccine uptake rates (Spearman's rho = 0.639, p = 0.010) or sending personal letters offering free vaccination (Sp = 0.728, p = 0.002) showed on average higher coverage among the elderly than countries with less developed vaccine management systems. The presence of additional policy elements (setting national objectives, HCW incentives, vaccination reimbursement systems, awareness campaigns and clear VCR objectives) led to numerically increased VCRs. The presence of several elements of vaccination policies at national level, including broad information and reminding systems, strong official recommendations and good access to the vaccine may help to achieve improved influenza vaccine coverage rates among elderly.

Keywords: :

• Influenza
• policy
• vaccination

Introduction

Seasonal influenza epidemics remain a considerable burden in the adult population. According to the European Centre for Disease Prevention and Control (ECDC), approximately 40,000 Europeans are dying each year due to influenza.¹ According to the ECDC and the World Health Organization (WHO), vaccination remains the most effective way to prevent influenza or severe outcome from the disease;^1,2 however, some authors regard the available evidence in favor of influenza vaccination as poor.^3-5 Despite strong and reiterated official recommendations, influenza vaccination rates remain limited in many countries, also in targeted populations. The 2003 World Health Assembly set the goal of achieving at least 50% vaccination coverage among the elderly population ≥65 y of age by 2006, and 75% by 2010. In 2009 this recommendation was adapted by the Council of the European Union and the target was re-set to 75% coverage among the elderly by 2014/15. The same goal applies for all defined target groups, e.g., people with chronic conditions and health care workers (HCW), in addition to the elderly.⁶ In most countries, seasonal influenza vaccination programs are offered to the population but there may be room for policy optimization. Furthermore, it is poorly documented how vaccine programs are introduced and vaccine policy decisions are made.

A recent global study assessed the immunization policy making process across 140 Member States of the WHO in the African, American, Eastern Mediterranean, South East Asian and Western Pacific regions.^7,8 The authors concluded that securing funding, disease surveillance at national level, and coordination of government and stakeholders were the main challenges for decision making on immunization policies. There was also a clear consensus that WHO global and regional vaccination recommendations are of great importance.

Improved understanding of the impact of vaccination-related policy elements and barriers toward vaccination is essential as a basis for enhancing seasonal vaccination coverage rates (VCRs) in European countries.^9-12 Several sources emphasize the value of enhancing vaccination rates among HCW as well as their important role as providers of vaccination, facilitators of access to vaccine, and educators improving the knowledge and awareness toward influenza vaccination in the population.^13,14 There is evidence that HCW-related policy elements influence vaccine coverage rates.^15-18

The objective of this analysis was to study the use and impact of various key vaccination policy elements that may influence vaccine uptake rates in European countries. The ultimate aim of this study is to better understand critical determinants of seasonal influenza vaccination uptake at national levels and to share best practices or propose solutions to ensure successful and sustained improvement of influenza vaccine uptake rates. We do not address the question of appropriateness of recommended target populations and coverage goals. Policy information with regard to influenza vaccination was collected in 15 countries and linked to actual VCRs in the population ≥65 y of age because official recommendations were similar and vaccine coverage data were available across countries for this age group. Our analysis focused on four main areas: official recommendations and vaccination programs, structured cooperation with HCWs, organized information and education of the general public on influenza and influenza vaccination, and access to vaccine.^10,19 Factors associated with high vaccine coverage among the elders were to be determined and this evidence was to be used for assessing possible challenges and opportunities in the immunization policy decision-making process.

Results

First, we describe implemented policy elements by country. In a second part, we address the relationship of implemented policy elements with actual vaccine coverage rates among the elderly.

Implementation of policy elements across countries

In Table 1 implemented policy elements are presented on a by-country basis. Table 2 provides the frequency of implementation of each policy element, across all countries included in the survey.

Table 1. Implemented policy elements to increase influence vaccination coverage rates in 16 European countries, in 2009

Table 2. Implementation of policy elements and correlations with vaccination coverage rates in the elderly (n = 15)

	Program implementation (yes/no)	Number of countries (%)	Average VCR in the elderly	Spearman's rho° Correlation Coefficient (p-value)
Management of seasonal influenza vaccination programs
Recommendation for all people aged 50–55 or 60 to 64 (nationally or regionally)	Yes No	6 (40%) 9 (60%)	54% 58%	0.047 (0.867)
National objectives adopted for risk-groups	Yes No	9 (60%) 6 (40%)	64% 46%	0.489 (0.064)
Year-specific national objectives for flu	Yes No	3 (20%) 12 (80%)	65% 54%	0.135 (0.631)
Monitoring VCR (By HA and/or NVIG)	Yes No	11 (73%) 4 (27%)	65% 39%	0.639 (0.010*)
Influence of health care workers (HCWs)
HCW have an objective to achieve in high risk groups (GP and/or Specialist)	Yes No	5 (33%) 10 (67%)	61% 54%	0.279 (0.314)
HCW financial incentive	Yes (partially or total) No	8 (53%) 7 (47%)	59% 53%	0.263 (0.343)
Role of information and communication to the public (Awareness campaigns)
Radio and tv	Yes (most/some regions) No	9 (60%) 6 (40%)	59% 54%	0.221 (0.429)
Press adverts	Yes (most/some regions) No	11 (73%) 4 (27%)	55% 58%	-0.052 (0.853)
Flyers in medical waiting rooms	Yes (most/some regions) No	12 (80%) 3 (20%)	56% 58%	0.058 (0.837)
Website for public	Yes (most/some regions) No	11 (73%) 4 (27%)	54% 62%	-0.035 (0.902)
Press/media conference for pub	Yes (most/some regions) No	13 (87%) 2 (13%)	56% 57%	0.114 (0.687)
Access to Vaccines
Reimbursement of vaccine	90–100% < 90% or not at all	11 (73%) 4 (27%)	60% 46%	0.402 (0.137)
Letter for free flu vaccine	Yes (most/some regions) No	7 (47%) 8 (53%)	68% 46%	0.728 (0.002**)

°Correlations with presence of policy element on average VCR of the elderly; *significant at the 0.05 level (2-tailed); **significant at the 0.01 level (2-tailed); VCR, vaccination coverage rates; HA, Health Authority; NVIG, National Vaccine Industry Group; GP, general practitioners.

Management of seasonal influenza vaccination programs

In all countries assessed, health authorities recommended the vaccination of elderly above 65 y. On average, 40% of all countries used official recommendations with lower age cut-offs (50+ years, 60+ years). The WHO 2010 target was adopted in 53% of the countries as a national objective. In three countries (Netherlands, Switzerland and Finland) there were “intermediary” year-specific targets for influenza VCRs. In eight countries, health authorities assessed the vaccine uptake rates and communicated the results by target group yearly (Table 2). In France, such surveys were also conducted by the National Vaccine Industry Groups (NVIG).

Influence of health care workers (HCWs)

In many countries (5/15), general practitioners or specialists were given a target to achieve in regard to influenza VCR of their patients. In some countries, targets were set regionally for general practitioners. Nearly half of the countries (8/15) provided a financial incentive or revenue for HCW if they immunized their patients (Table 2).

Role of information and communication to the public

All countries used awareness campaigns, but some countries only used a limited range of media (e.g., Italy or Bulgaria). The annual awareness campaigns made use of different kind of media such as TV/ radio (9/15), press advertisements (11/15), flyers in medical waiting rooms (12/15), public websites (11/15), or press/ media conferences (13/15) (Table 2).

Access to the vaccine

In 73% of all countries, vaccination was free or reimbursed to a large extent (90–100%) for the elderly. In some countries personal invitation letters or vouchers to the target population were sent nationwide (3/15) or regionally (4/15). In 4/16 countries, doctors advised the targeted population by letter or electronic mail to get immunized (Table 2).

Program implementation and elderly vaccination coverage rates

Correlation of policy elements with vaccination rates in the elderly

Stand-alone policy elements

The correlations of implementation of policy elements with vaccine-uptake rates among the elderly are shown in Table 2 for the 15 countries with coverage rates available. Annual surveillance and communication of VCR as well as sending personal letters or vouchers for free vaccination appeared to be elements contributing to high vaccine uptake rates in the elderly. Noteworthy, the item ‘personal letter’ includes the element of reimbursement, but is also a reminder to the target population about influenza vaccination.

In multiple regression analysis, only sending a personal letter (t-value 2.26, p-value: 0.050) and national objectives for influenza (t-value 1.86, p-value: 0.096) were independent significant or near-significant predictors of higher VCRs, respectively (Table 3). These two elements were also the only significant parameters in the univariate Spearman correlations (Table 2). Objectives adopted for risk groups indicated a negative impact, which may be a statistical artifact. Given the limitations of our data source, namely small sample size, these regression results need to be interpreted with caution. Hence, interactions between policies could not be assessed using regression analysis.

Table 3. Mutivariate regression analysis

Variable	Parameter Estimate	Pr > |t|
Monitoring VCR (By HA and/or NVIG)	0.194	0.111
Patients receive personal letter/voucher for free flu vaccine	0.131	0.050
National objectives	0.195	0.096
National objectives adopted for risk-groups	-0.122	0.420
90–100% reimbursement of vaccine	0.201	0.105

VCR, vaccination coverage rates; HA, Health Authority; HCW, Health care worker; NVIG, National Vaccine Industry Group

Set of policy-elements

Vaccination policies are mostly implemented in a set and not as stand-alone element. Despite small sample size, we calculated Spearman correlations of sets of two or three policy elements with VCR in the elderly, which could be interpreted as an informal assessment of for assessing interaction between policy elements (Table 4).

Table 4. Implemented policy elements compared against vaccination coverage rate in the elderly (single or combinations of two or three programs)

Not all combinations showed a numerical impact on the vaccination behavior among the 65+ population. Although monitoring of annual VCR represents a strong policy element by itself, its influence can apparently be enlarged by combining it with a second item, e.g., reimbursing influenza vaccination or awareness campaigns in radio/TV or flyers in medical waiting rooms. This implies that vaccination campaigns through the media may be an important element in educating the public on influenza and influenza vaccination.

Similarly, patients receiving a personal letter or voucher as an invitation for free vaccine or vaccination was positively correlated with high VCR in the elderly and the combination of this policy element with flyers, radio/ TV vaccination campaigns, or press and/ or media conferences seemed to increase vaccination coverage further.

Moreover, combining national objectives with incentives for health care professionals, sending letters for free vaccine, reimbursing the vaccination and communication campaigns through radio or television appears to have a positive impact. In addition, we assessed combinations of three different policy elements (Table 4). In the presence of three policy elements (e.g., flyer, letter and reimbursement), the VCR was increased for the elements presented in Table 4. Also, combinations with predetermined targets for HCWs were associated with a positive influence on influenza vaccination uptake rates.

In summary, key elements showing the most substantial correlations with VCR in the elderly were the monitoring of vaccination rates as well as the distribution of personal letters or vouchers for free vaccination. The impact of those elements may be enhanced by other policy elements: setting national objectives, providing financial incentives to HCW, implementing vaccine reimbursement, or launching influenza awareness campaigns through powerful media (TV, radio).

Discussion

Despite strong national and international recommendations on seasonal influenza vaccination, only few countries have achieved a high vaccination rate among targeted groups. Vaccination-related policies vary across European countries. We tried to identify practices that can contribute to a successful and sustained improvement of seasonal influenza vaccination coverage in the elderly. There appear to be several possibilities to increase vaccination rates in priority groups, but only few countries take advantage of all possibilities. Our study identified policy elements implemented at the national level and put them in relation to actual influenza vaccine uptake in the elderly to identify possible contributors to a high vaccination rate.

Strong influenza-related activity seemed to have a positive impact on VCRs. Key elements associated with high VCR were reminding people to get vaccinated, providing the vaccination at no cost and being aware of the national influenza vaccine uptake rates. These elements are consistent with a recent review article from the USA, aimed at describing an extensive conceptual framework in terms of four possible determinants of influenza vaccination.²⁰ The study suggests that VCRs might be enhanced by multi-dimensional policies covering issues such as the epidemic level, influenza related mass media reports, vaccine and vaccination reimbursement or vaccine supply. Also, Tsutsui et al. assessed in a recent US study the willingness to get vaccinated against influenza and how coverage rates might be enhanced. The authors considered elements such as the dissemination of information, the access, and the reduced costs of the vaccine as helpful instruments to improve vaccine uptake rates.²¹

The positive correlation of VCR with vouchers for free or reimbursed vaccination may indicate substantial sensibility to personal invitation (in the form of a letter) in addition to the financial nature of these actions. Strong national recommendations combined with effective implementation of programs appear to be a pre-requisite of high VCRs. Furthermore monitoring uptake rates was identified as a factor for good VCR among the elderly. This is in line with recent recommendations by the WHO and the Council of the European Union (2009/1019/EU), which perceived a distinctive surveillance system for seasonal influenza vaccination as a major mean of enhancing vaccination rates.⁶

Mass information and communication is a further element to achieve high VCRs. Yoo et al. showed a positive association between influenza-related media coverage (in newspaper, television and wire service) and influenza vaccination rates among the elders in the US.²² Three influenza seasons were studied and a significant increase in annual vaccine uptake rates of up to 8% was found. Similarly, we identified a positive relationship between high media coverage of annual awareness campaigns and vaccination rates. It seemed to be crucial to reinforce the disease and vaccine awareness of the public in order to increase vaccination rates. This can be done by e.g., the media, websites or information provided in waiting rooms.

Thomas and colleagues found evidence from randomized controlled trials that rising community demand (e.g., by reminders), increased access (e.g., home visits to elderly), and provider-based interventions (as e.g., reminders to physicians or education of HCW) and system-based interventions (as e.g., financial incentives by enhancing preventive interventions), could effectively increase influenza coverage rates among US seniors.¹⁹ Studies of interventions at the societal level (i.e., programs such as incentives for HCW) were not identified by the authors.^19,23

Elderly individuals may tend to believe and follow the advice of their family doctor or nurse. In fact, several studies have shown that physicians’ recommendations have a large, positive influence on uptake rates.^10,11,13,24 In our analysis, we identified combinations with policy elements regarding the role of HCWs as important correlates of VCR among the elderly population.

Apart from HCWs' role in advising vaccination to their patients, there is a general demand for enforcing and developing programs to improve yearly influenza vaccination among medical staff. Nevertheless, HCWs generally indicate a very low willingness to get vaccination themselves, despite the facts that they would personally benefit because they are at risk of catching influenza from their patients and that HCWs' vaccination would provide an indirect benefit to the most vulnerable population by preventing HCWs from infecting some of their patients if they are ill themselves.¹⁸ However, this topic was not studied in our survey, which did not collect data on HCWs' flu vaccination status and did not analyze its impact on vaccine uptake in the elderly patient population.

Our analysis is subject to some limitations. First, the number of countries included was limited, hence possibilities for statistical data analysis were restricted. Spearman correlation coefficients were regarded as suitable to indicate probable directions of effects but not to provide reliable estimates of effect size. Therefore, the strength of the correlations shown should be interpreted with caution. In addition to this, the level of detail of the survey data was restricted. They only provided an approximation of country-specific influenza vaccination policy characteristics at one specific time-point. For example, the implementation of different policy elements may have occurred at different points in time. This implies that in some countries several policies may not have been put into practice long enough to develop their full potential impact, or may have to be optimized. Furthermore, we did not take into account the duration and timing of the annual awareness campaigns. Therefore, media coverage and pressure may differ across the countries surveyed and this may influence the impact of the campaigns. In addition, cultural and religious beliefs and socioeconomic factors may also influence the impact of a specific immunization policy. The cross-sectional design of the survey did not allow taking this aspect into consideration and trend analyses were not possible.

Those collecting the data had a national industry background. However, the information retrieved was based on the best available national sources in a pre-pandemic situation. Bias due to direct financial interest in vaccine production cannot be ruled out, but is not specifically expected for this research question. The questionnaire covered issues on hard facts on national immunization policies in place and no sensitive questions based on subjective assumptions.

Based on earlier knowledge and published literature, we tried to include all potential policy elements with a possible influence on elderly’s VCRs.^10,25-28 Nevertheless, non-differential misclassification might occur and other factors not covered by our study might have an additional influence in particular countries. But for the sake of consistency, we have chosen a consistent approach in terms of data collection and analysis for all nations.

This might also be one cause for possible deviations from the results of the Vaccine European New Integrated Collaboration Effort (VENICE) survey.²⁸ In addition, differences in definitions used, in survey questions on specific issues or in time frames could have occurred.²⁹ As an example, the definition of media campaigns may have been a different one in the VENICE questionnaire. National government subsidies reducing the cost of the vaccine and its administration may also have been assessed differently. The included vaccine coverage data stem from two different sources, given gaps in data availability in the one or the other study.³⁰ Different survey methods may have influenced the resulting VCR but the data included were the most suitable to the best of our knowledge.

We also investigated into a possible influence of the gross domestic product (GDP). Policy activity per county was compared with its GDP.³⁰ No correlation was identified between these two factors (Spearman's rho = 0.105, p = 0.700).

This work focuses on possibilities to increase VCR. Questions around vaccine efficacy are not addressed. However, we acknowledge a lack of evidence on the magnitude of vaccine effect from randomized placebo-controlled trials, which are run over several seasons.³ In addition, externalities such as mis-match between B-strain components in vaccines and circulating B lineage or potential herd immunity can affect the effectiveness of the seasonal vaccine in both directions.^31,32 Given that implementing new vaccination policies is associated with considerable costs, there is a need that national policy-making bodies continue to combine the most medically effective strategies with appropriate policy elements, as assessed here, to achieve the highest possible public health impact.³³

Conclusion

By way of conclusion, the assessed policy elements differ in importance. Some policies seem to have a stronger impact on the vaccination rates in the elderly population than others. This implies that it is of importance to implement in priority those policies with the highest impact on VCRs. Our results indicate that monitoring vaccine uptake rates and distributing personal letters and/ or vouchers for free vaccination may be of such particular importance. Enhancing annual influenza vaccine uptake rates will contribute to an improved preparedness for upcoming seasonal influenza outbreaks and can hence reduce the burden of disease especially in the most vulnerable populations. More emphasis should be placed on the implementation and evaluation of potential best practice policies.

Methods

Overview

Data collection was coordinated by the European Vaccine Manufacturers (EVM). Chairs of 16 European NVIGs provided data on the situation in their respective countries. In general, the role of NVIGs is to promote vaccination as an essential element of improving public health, as well as to represent the interests of the vaccine industry. The selection of the countries included in the survey was based on EVM membership. Countries were excluded if there was no VCR available for the elderly population (e.g., Bulgaria), or no NVIG active during 2009 (e.g., Austria, Greece, Norway, Denmark).

The survey questionnaire was based on an extensive literature search to identify policy elements with a potential influence on VCR. NVIGs received the electronic questionnaire to document the presence of vaccination policy elements that were regarded as potential determinants for a successful implementation of elderly seasonal influenza vaccination policies.^10,19 The questionnaire contained closed questions (yes/ no), but adaptations were allowed (e.g., information could be added if policy implementation was restricted regionally) and details could be provided by the interviewee. Questionnaire items were categorised into different categories (Appendix 1).

1. Management of seasonal influenza vaccination programs;

2. Influence of health care workers;

3. Role of information and communication to the public

4. Access to vaccine

The questionnaire was sent out in June 2009 and completed between July and September 2009.

Analysis

Analysis was primarily descriptive. Questions answered with “Yes,” indicating the presence of a specific policy element whatever the level of implementation, were awarded one point. For example, if a country had fully or partially implemented a particular policy element (e.g., monitoring annual VCRs), it was scored 1, and in the case of no implementation, it was scored 0. Missing data lead to the exclusion of the respective country from the analysis for the applicable element. We thus avoided inappropriately deflating averages across countries by interpreting non-responses as negative responses. Therefore, in the calculations, the denominator for each average proportion changed according to the number of included countries, as a function of questionnaire completion status and also availability of data on vaccine coverage rates. Resulting scores described the presence or absence of the policy elements covered by the questionnaire. For the purpose of analysis, the different policy elements were set in relation to actual vaccination coverage rates seen in the countries, for the 65+ elderly population, by calculating correlation coefficients. The focus on this recommended group was mainly chosen for two reasons: first, all countries had at least a recommendation available for the 65+ population and second, vaccine coverage rates for this target group were available from the majority of countries (15/16). In an additional step, associations between implemented policy elements and vaccine coverage rates were assessed by Spearman's correlation coefficient (Spearman's rho). Correlation coefficients can range from –1 to +1 for a perfect negative relationship and a perfect positive relationship, respectively. A value of 0 indicates no correlation. The correlation coefficients were analyzed on the basis of 0 or 1 variables (representing absence or presence of a given policy element, respectively). When the effect of two or three programs was assessed, the programs were interlinked with “and,” meaning that a country was only scored 1 if both or all three policy elements, respectively, were implemented. If this was not the case, the country was scored 0.

Given the fact that the dependent variable (VCRs of the elderly) was roughly normally distributed, multiple regression analysis was performed. Candidates for potential predictors of VCR were determined by univariate Spearman correlation analyses and confirmed by simple linear regression analyses (not shown). The final model covered five covariates (Table 3).

For all statistical tests, two-sided p-values ≤ 0.05 indicated statistical significance. For the selection of candidates for potential predictors of VCRs in univariate analyses, a two-sided p-value of ≤ 0.02 was assumed as appropriate. Given the predominantly descriptive nature of the analysis and risk of spurious results due to small sample size, no corrections for multiple testing were made.

Data source for vaccination coverage rates

Data on vaccine coverage rates in the elderly were available from the VENICE survey²⁹ and our own vaccination coverage rate surveys for season 2007/08 as described elsewhere (Fig. 1).¹⁰

Figure 1. Influenza vaccination coverage rates of elderly population (≥65 y) among 15 European countries (season 2007/08)

Technical implementation

The questionnaire was provided in English. The data were analyzed using Microsoft Excel® 2008. SPSS version 16 for Windows and SAS 9.3. (Cary, NC, USA) were used for the statistical evaluations.

Authors’ contribution

PB performed the data analysis, contributed to the data interpretation and wrote the manuscript. MS contributed to data interpretation and the final version of the manuscript. TS contributed to data interpretation and supervised the development of the project. All authors contributed to the design of the study and have read and approved the final manuscript.

Acknowledgments and Funding

This study was made possible by an educational grant from the European Vaccine Manufacturers Group of the European Federation of Pharmaceutical Industries and Associations (EFPIA), Brussels, Belgium.