Trends in influenza vaccination coverage in Portugal from 1998 to 2010: effect of major pandemic threats

The public health burden of annual influenza epidemics represents ongoing vulnerability to pandemic influenza and highlights gaps in bioterrorism preparedness and response efforts [1, 2]. The recent emergence of the pandemic influenza A (H1N1) 2009 virus is a good example of how influenza can impact health systems around the world [3].

Currently, vaccination is the key measure available for prevention of influenza and associated complications. Strategies that focus on providing routine vaccination to persons at higher risk for influenza complications have long been recommended, although coverage among most of these groups remains low [4]. Concurrently, there is a need to increase seasonal vaccine use through clear immunization policies as a way to stimulate industry to boost production capacity [5].

During the 2009/10 season, the influenza pandemic led to reinforcement of influenza vaccination recommendations in European countries, including Portugal. The pandemic also led to renewed interest in influenza vaccination surveillance.

Continued annual monitoring is needed to determine the effects of vaccine supply, changes in influenza vaccination recommendations, changes in groups targeted for vaccination, and other factors, on vaccination coverage among adults and children [4]. Vaccine coverage rates constitute the basic measure for evaluation of public health programs designed to improve vaccination uptake and for estimation of how the vaccination program affects the rate of disease.

Circumstantial factors (e.g., personal reasons) may affect the rate of vaccination [6]. In the last 12 years three major pandemic threats, not all caused by influenza virus, affected public and healthcare professionals’ perceptions of the need for influenza vaccine (IV) uptake. These pandemic threats were severe acute respiratory syndrome (SARS) in 2003, “avian influenza” (influenza A (H5N1)) in 2005, and the recent influenza A (H1N1)2009 pandemic. The rapid worldwide dissemination of SARS in 2003 was a rehearsal for the next influenza pandemic [7]. In 2005, the World Health Organization (WHO) declared that the emergence and persistence of influenza A (H5N1) in birds and the associated human fatalities were a public health threat and fostered early response strategies to contain the pandemic [8]. In April 2009, the spread of novel swine flu origin influenza A (H1N1) prompted the declaration of a pandemic by WHO in June of the same year [9]. Whether this raised awareness had a positive or negative effect on IV uptake varies according to country and to situation [6, 10, 11].

Every year, based on surveillance of clinic and laboratory data provided by a global network of Influenza Surveillance Centers, the WHO recommends the composition of the vaccine that will be used the next season [12]. The General Directorate of Health (DGS) in Portugal issues an information guide to Ministry of Health and private sector doctors in September/October of each year. The guide describes vaccine specifications for the current season and indications for vaccine uptake. Each flu season, the main objective of the DGS has been to increase IV coverage in high risk groups (i.e., individuals >65 years old and patients with specific chronic diseases) and priority groups of health professionals [13, 14]. Table 1 describes how the recommendations have changed over time.

In 2004, WHO established a 75% IV coverage target in the elderly (>65 years) that was in effect until 2010 [12]. In 2006, DGS established an interim target of 50% vaccine coverage among individuals aged 65 years and over for the 2006/07 season [14].

Despite the occurrence of a pandemic due to a new strain of influenza virus in 2009, the DGS has maintained the same seasonal IV recommendations for the major risk groups [15]. Concurrently, the recommendation expanded to include vaccination against pandemic influenza A (H1N1) 2009 for children from 6 months to 12 years. Older individuals were included only if they had a specific chronic disease [16].

Since the 1998/99 season, the Department of Epidemiology (DEP) of the National Institute of Health Doctor Ricardo Jorge (INSA; previously the National Health Observatory (ONSA)) has monitored IV coverage. This system is the only one in the country that estimates IV coverage in the general Portuguese population and in subgroups. This information cannot be obtained from a count of vaccine sales or from vaccine administration data.

This article describes national trends in seasonal IV coverage in Portugal from 1998 and 2010 in the general population and in age groups presents an analysis of progress towards WHO 2010 coverage goals. It also addresses the effects of the major pandemic threats (SARS in 2003, influenza A (H5N1) in 2005, and the influenza A (H1N1)2009 pandemic) on vaccination trends for the last 12 years.

The ECOS panel consisted of a random sample of Portuguese families with a landline telephone and of families with landline and mobile phone since the 2009/10 sample (dual sample frame). The sample was stratified and was evenly distributed to represent the five health regions of the country. Landline phone households were selected by simple random selection from the national telephone directory. Mobile phone households were selected by random digit generation. All households received a letter from INSA with an invitation to participate in the ECOS panel and provide informed consent. Telephone contact was then used to formalize participation and record each household member’s demographic data. The households included in each panel were renewed approximately every 3 years. The ECOS panel of families was approved by the Portuguese Data Protection Authority, which is in charge of ethical issues and protection of individual data collection in Portugal.

The seasons included in the surveys were the winters from 1998/99 to 2010/11, except for the winter of 2000/01. No investigation was carried out in 2000/2001 season owing to lack of financing.

All surveys used the same questionnaire. The questionnaire was presented to one individual (≥18 years of age) in each household using CATI (Computer Assisted Telephone Interview) technology. This individual provided information on his/her vaccination status and information on the household. The terminology “percent of vaccinated” used in reporting results refers to individuals who reported being vaccinated or on which the respondent said they were vaccinated.

The 1998/99, 1999/00, and 2001/02 surveys used the sample set formed in 1998. The surveys from 2002/03 to 2005/06 were carried out with a sample that was selected in 2002. The investigation in the 2006/07, 2007/08, and 2008/09 seasons was conducted with a sample formed in 2006. The 2009/10 and the 2010/11 surveys used a sample that was selected in 2010 (Table 2).

A detailed description of the ECOS methodology can be found in a published report [17]. Permission to use the data for this study was obtained from the National Institute of Health.

IV coverage was analyzed for the entire sample and for specific groups defined by age and sex. Information on demographic questions was collected during the initial survey done at household recruitment time.

For the surveys conducted in 1998/99 to 2008/09, all IV coverage estimates were adjusted by health region using the 2001 Portuguese population census data (2001 Census data from the National Institute of Statistics). For the surveys of the 2009/10 and 2010/11 periods that were conducted using the dual sample frame, the IV coverage estimates were adjusted by health region (Census population 2001) and for cell phone and landline phone coverage of Portuguese households using the methodology described in Brick[18] and Kennedy[19]. Weighting factors were adjusted by post stratification for population age and sex distribution.

To test the association (or independence) with disaggregation variables, we used the modified F-statistics of the second order adjustment of the Rao-Scott Chi-squared test [20] whose properties are presented in Rao and Thomas[21]. A 5% significance level was used for the statistical tests, and the null hypothesis was rejected when the probability of test significance (p-value) was <0.05. We also calculated 95% confidence intervals for all proportions.

We used a meta-regression model to test the linear trend in IV coverage throughout the study period, and the effect of the SARS, A (H5N1), and A (H1N1)2009 pandemics. Each survey estimate was weighted by the inverse of the variance in the logit scale. The model fitted to the logit of IV coverage included three dummy variables (one for each event) and a sequence of numbers from 1 to 12 years to measure and test the time trend effect. The model was adjusted to IV coverage in the general population and age groups.

All analyses were performed using the statistical programs SPSS [22] or STATA SE [23].

From 1998/99 to 2010/11, seasonal IV coverage in Portugal varied between 14.2% and 17.5%. No significant increasing trend was observed in the Portuguese general population during this period.

There was a clear, increasing trend in IV coverage in the elderly population (p < 0.001). Coverage increased from 31.3% in 1998/99 to 48.3% in 2010/11. After 2006/07, however, IV coverage in the elderly has remained near 50%, which suggests that there has been a slowdown in the growth trend. Thus, in 2010, the 75% WHO target in this major risk group had not been met. This result indicates that there is a need to improve the influenza vaccination strategy in Portugal to comply with the ambitious coverage goals proposed by WHO.

The major pandemic threats of the last decade had an effect on seasonal influenza vaccination. There was a significant increase in vaccine uptake in the general population in 2003/04 (p = 0.032) and in 2005/06 (p = 0.018). The 2009/10 seasonal vaccine coverage in the general population was not significantly higher when accounting for the baseline trend, but the 12.9% of vaccination occurred in individuals <15 years old. Although not the highest of all the seasons, this result was significant when accounting for vaccination trend in this age group during the past decade.