Rates of immunization against pandemic and seasonal influenza in persons at high risk of severe influenza illness: a cross-sectional study among patients of the French Sentinelles general practitioners

In France, seasonal influenza immunization is recommended to all persons aged ≥65 years and those with underlying medical conditions that increase their risk for influenza disease or complications [1]. Before 2010, patients with severe bronchopulmonary, cardiac, renal disease, sickle cell disease/ thalassemia, diabetes mellitus, or any immune disorder (HIV infection, immunosuppressive drugs, cancer …) were targeted by French immunization recommendations [2]. Since 2010, pregnant women (at any trimester) and overweight persons (BMI≥30) have been added to the list of patients that should receive the monovalent influenza A(H1N1) 2009 vaccine and subsequently, the seasonal trivalent vaccine yearly [3]. Every year in France, people targeted by seasonal influenza immunization receive an invitation from health authorities to get a free-of–charge vaccine, administered by their General Practitioner (GP). Despite this large national vaccination campaign, rates of immunization against seasonal influenza remain weak: the global vaccination coverage (VC) for the target population was 51% in 2007, 56% in 2008 and 57% in 2009, as estimated by the National Health Insurance [4]. To our knowledge there is no data in the literature on detailed immunization rates among at-risk populations in general practice since the influenza pandemic occurred. The objectives of this cross-sectional study were 1) to describe age distribution and risk factors of severe influenza illness in patients seen on a given day by French GPs, 2) to estimate influenza immunization rates for two consecutive seasons (2009–2010 and 2010–2011), and 3) to identify predictors of seasonal vaccination in target groups.

Our estimates of VC against seasonal and pandemic influenza were generally comparable or higher than that previously reported in certain categories of patients, such as those aged 65 and over, or with chronic cardiac, pulmonary diseases or diabetes mellitus.

Overall rate of immunization against pandemic influenza A(H1N1)2009 was very low in our sample (16%, 95%CI: 15%; 17%), higher than national estimates in the general population (<8%) [21], but close to that (17%) reported in another French cross-sectional study conducted among GPs in spring 2010 (the MOTIVAC study) [22]. These results likely reflect the public perception of low risk from the disease and the negative impact of controversies concerning the safety and quality of the pandemic vaccines [23, 24]. Distinction must be made between GP involvement in a pandemic context and GP involvement in a seasonal context. Indeed, in France in the context of the H1N1 pandemic, their role was relegated to second place in favor of mass vaccination in health centers open for the event. This campaign was considered as a failure because a very small part of the population joined the vaccination while logistical and financial efforts to organize it were particularly important. Hence the importance of the involvement of the specialists in the promotion of vaccination in patients with specific risk factors (such as pregnancy, severe pulmonary, cardiac and renal diseases).

In our study population, 14% of patients were obese, and 11% in those aged <65. More than 50% of them had no other risk factor for severe influenza. Obese poorly vaccinated were those under <65 years for seasonal flu and all subjects of this sample for pandemic influenza. It must be noted that there is no data in the literature on the obese influenza vaccination coverage. Sample sizes were low in other target groups, thus estimates of VC should be interpreted with caution in these categories. Forty four pregnant women could be evaluated for both influenza seasons in our study population. VC for seasonal and pandemic influenza vaccines were particularly low (5-7%). In the United States, were influenza vaccination is routinely recommended in all pregnant women, VC was estimated at 51% for seasonal influenza vaccine and 47% for the pandemic influenza vaccine in 2009–2010 [25]. Similarly, a large number of pregnant women in the Netherlands reported to be vaccinated against 2009 influenza A (H1N1): 63% (2993 respondents/ 14529 patients invited to complete an internet survey) [26]. Influenza vaccination of pregnant women was a focus of public health efforts during the 2009–10 season in France. Before this season, vaccine recommendations didn’t apply to them if they had no identified risk factor for severe influenza. However, the VC reported here suggest a poor adherence to influenza vaccination in this category, as confirmed in another large cohort of French pregnant women in 2009 [27], were 29% of women reported to be vaccinated against pandemic influenza. It might be related to the physicians and patients insufficient information about the safety of the pandemic vaccine [28].

In the 18 patients living with HIV aged 65 or less, 61-67% declared being vaccinated against seasonal and 33% against pandemic 2009 A/H1N1 influenza. In the US influenza vaccination coverage in HIV-Infected patients increased from 29% in the 1990 to 42% in the 2002 season [29]. A French study reported comparable A/H1N1 vaccine coverage (44%) among HIV-infected patients [30]. The analysis of determinants in the target groups showed that VC for seasonal influenza in 2010–2011 was related to previous vaccinations and the number of risk factors for severe influenza illnesses. This significant association between vaccine uptakes was shown previously [31].

Our study has both strengths and limitation that warrant discussion. First, the rate of participation (20%) was slightly lower to those reported in previous studies of the Sentinelles network [32]. Probably, the main reason was that participating GPs were asked to fill the questionnaire for all patients presenting at their practice on a given day, which represented an important effort. In addition, summer holidays covered a large part of the data collection period. However, besides response rate, comparison of respondents to the whole population of Sentinel GPs did not show statistical differences neither on on socio-demographic data nor on type of practice. Second, general applicability of our results might be limited by the fact that the study was conducted in the specific population of French Sentinelles GPs. Comparison of our population to the whole population of French GPs did not show major differences neither on socio-demographic data nor on type of practice. However, the Sentinelles network is composed by volunteer GPs who collect real-time epidemiological data, suggesting a specific interest for public health, in particular for immunization. This limitation could explain the high vaccination rates observed, as compared to studies in general population. Third, since participation rate was relatively low, some target groups were less represented, leading to wide CI for rare target groups, in particular for HIV infection, other immunosupression, chronic renal disease and pregnancy. Patients of these categories probably get medical care from specialists of their chronic disease, hence underlining the importance of conducting specific studies in specialized medical services. And finally, immunization data were self-reported by the patient and/or the GPs and not written-confirmed, leading to some degree of imprecision in evaluation of the immunization status.

Despite these limitations, such studies among Sentinelles GPs can be complementary to the population-based studies. Indeed, our study allowed to collect data rapidly, on a large number of patients, and was resource-saving since all GPs of the Sentinel Network could easily and simulatenously be contacted via an automatic-generated email. If necessary, such study could be renewed yearly to follow variations of the VC from on year to another, for example in case of occurrence of an influenza pandemic. We also specifically described the characteristics of patients attending a medical consultation with French GPs, therefore it could help identify populations that could be counselled by their GPs to get vaccinated. In addition, such design might limit recall biases compared to telephone surveys. Indeed, estimation of VC was based both on answers of patients and on data collected by the GPs in their medical records. Usually, participation biases are in favor of better vaccinated individuals, leading to some over-estimation of coverage in the telephone surveys.

The prevalence of chronic conditions in our study sample was comparable to that estimated in the general population regarding data published by the French Ministry of Health [33] for pregnancy, obesity, diabetes mellitus, Sida-infection and chronic cardiac diseases (prevalence estimates of 1%, 13% 4%, 0.05% and 1% respectively in the general population). Persons aged 65 and over were over-represented in our sample, whereas prevalence of chronic bronchopulmonary and renal diseases were low, maybe due to underdiagnosis. This confirms that for most chronic conditions, involvement of GPs can have a potentially important impact to improve VC. On the other hand, for patients with chronic bronchopulmonary and renal diseases, vaccination promotion limited to GP could be less efficient and specialists must be involved.

This study gives an overview of characteristics of patients seen on a given day in French general practices, as well as estimates of rates of immunization against both seasonal and pandemic influenza in target groups. Our estimates of VC are generally higher than that reported by others in the general population. GPs receive a very large proportion of patients targeted by vaccine recommendations (great variety of chronic underlying conditions and elderly persons). In addition, a lot of these patients are not regularly followed by other physicians who could offer vaccination (patients with obesity, diabetes…). We also report that pandemic influenza VC were lower than seasonal VC in all categories of patients. Contrary to a standard seasonal vaccination campaign, the administration of pandemic vaccine in 2009 in France was not primarily assured by GPs. This underlines the crucial role of GPs in promoting immunization in the community. There are some examples of successful public health measures through GP. Their involvement in colorectal cancer mass screening allows identification of high risk people who can then be managed with a more suitable screening protocol [34]. This study underlines the need to consider the role to be given to GPs in the choice and implementation of immunization strategies.