Seasonal influenza vaccine uptake and vaccine refusal among pregnant women in France: results from a national survey

ABSTRACT

Pregnant women and infants are at high risk for severe influenza and many countries, including France, recommend annual influenza immunization during pregnancy. We aimed to estimate influenza vaccination and refusal rates and assess associated factors among pregnant women during the 2015–16 season in France. We used data from a national representative sample of women who gave birth in March 2016 and were interviewed before hospital discharge (N = 11,752). In the multivariable analysis, robust Poisson regression models were used to study associations with maternal characteristics and prenatal care characteristics. Influenza vaccine coverage among pregnant women was 7.4% (95% confidence interval [CI]: 6.9–7.9). Only 24.9% (95% CI: 24.2–25.7) of women said that they received a care provider proposal for vaccination and 70.4% (95% CI: 68.7–72.0) of these declined it. Vaccine uptake was associated with low parity (prevalence ratio [PR] = 2.1; 95% CI: 1.4–3.2 for parity 0 vs ≥ 3), high educational level (PR = 2.5; 95% CI: 2.0–3.2), healthcare occupation during pregnancy (PR = 1.8; 95% CI: 1.5–2.1) and preexisting conditions at risk for influenza (PR = 1.7; 95% CI: 1.3–2.2). Women were more frequently vaccinated when their main care provider was a general practitioner. Multiparae women and those with medium or low educational level were significantly more likely than others to decline influenza vaccine after a provider proposal. Influenza vaccine coverage is very low in France, mainly because of infrequent care provider proposals and also frequent women’s refusals. Effective interventions should be designed to promote vaccination among medical professionals and reduce vaccine hesitancy among pregnant women.

KEYWORDS:

• Seasonal influenza
• pregnancy
• maternal vaccination
• vaccine acceptance

Introduction

Pregnant women and their newborns are at high risk for severe influenza-related complications.^1-4 Influenza infection during pregnancy is associated with adverse obstetric and fetal outcomes, such as preterm birth or fetal death.^5,6 In addition, pregnant women are at increased risk of hospitalization for complications as compared with non-pregnant women, especially those with underlying conditions.^7,8

Vaccination against influenza can prevent harm and benefit both women and newborns by reducing the risk of influenza-associated complications and hospitalizations.^9-11 It could also prevent influenza illness in infants under age 6 months because of the transplacental antibody transmission provided by maternal immunization.^12-16 No evidence of adverse pregnancy outcomes after maternal influenza vaccination was reported from systematic reviews and vaccine safety studies, although available safety data are limited, especially for vaccination during the first trimester of pregnancy.^17-20

Since 2012, the World Health Organization (WHO) recommendations have identified pregnant women as a priority group for national influenza immunization programs,²¹ and almost all European countries recommend vaccination against influenza during pregnancy. The same year, French health authorities recommended seasonal influenza vaccination for all women who are pregnant during the influenza season, regardless of the trimester of pregnancy.²² In France, the national health insurance reimburses the cost of seasonal influenza vaccine for high-risk groups identified in national recommendations, including pregnant women.²³ Each year, the national public health agency (Santé publique France) monitors the seasonal influenza epidemic and vaccination coverage. However, influenza vaccine coverage among pregnant women is poorly known at the national level.²⁴

Several studies have found vaccination uptake during pregnancy associated with maternal characteristics and the attitude of both women and healthcare providers playing a major role.^25-29 However, most of these studies focused on the 2009 influenza A (H1N1) pandemic and were performed in North America.³⁰ Available data on seasonal influenza vaccine coverage are scarce in Europe,³¹ including France.^32,33

We aimed to estimate the seasonal influenza vaccine coverage among pregnant women in mainland France for the 2015–16 season and assess factors associated with vaccine uptake and refusal by using data from the National Perinatal Survey performed in 2016.

Methods

Survey methods and participants

The National Perinatal Surveys have been conducted periodically since 1995 and aim to monitor perinatal health, providing data on a wide range of topics. They cover all live births and stillbirths with gestational age at least 22 weeks or birth weight at least 500 g during one week in March in all public and private maternity units. Data are obtained from two sources: (1) an interview with mothers in the postpartum ward before discharge to collect information on maternal characteristics and prenatal care and (2) medical records for complications of pregnancy, delivery, and the newborn’s health status. More details about method, data collection and women’s characteristics are described elsewhere.³⁴

Ethical approval was granted by the French Data Protection Authority (CNIL, authorization 915197) and the INSERM ethics committee (authorization IRB00003888, no. 14–191). In mainland France, 13,147 women delivered during the studied week (March 14–20, 2016). Women were not contacted if they were < 18 years old or had a stillbirth (N = 198). In addition, 1,197 women were not interviewed mainly because of refusal (N = 548), a language barrier (N = 245) and poor health status (N = 144). Finally, 11,752 women were interviewed and influenza vaccine uptake was known for most of them (99.7%, N = 11,712) (Figure 1).

Figure 1. Study flow chart, 2016 French National Perinatal Survey

Studied variables and statistical analysis

Women in mainland France were asked whether they had been vaccinated against seasonal influenza during the influenza vaccine campaign. If the answer was yes, they were asked the qualification of the prescriber (obstetrician-gynecologist [ob-gyn], midwife, general practitioner [GP]) and if no, they were asked why, among the following reasons: “nobody offered it”, “it was offered but I refused”, and “other reason” (please specify). We distinguished several categories of determinants of vaccine uptake: maternal socio-demographic and medical characteristics, and prenatal care. The socio-demographic characteristics included age, parity, family status (cohabiting/single), educational level, country of birth and complementary health insurance. Two broad categories of insurance are used in France for additional costs that are not covered by the national health insurance (e.g., additional cost of medical visits): public insurance for low-income people and private insurance. We also classified women according to their occupation and belonging to priority groups for vaccination: healthcare workers (e.g., doctors, midwives, nurses and nurses’ aides), women in contact with children (e.g., teachers in primary school and daycare staff), women with other occupations and women who were not working during pregnancy.

The medical characteristics collected were the following: (1) preexisting conditions at risk for severe influenza infection (e.g., diabetes, chronic respiratory or kidney disease, immunocompromised state, severe cardiovascular disease, chronic liver disease, and body mass index ≥ 40 kg/m²), (2) other preexisting maternal conditions requiring specialized surveillance by an ob-gyn according to the national medical guidelines,³⁵ and (3) complications during pregnancy requiring hospitalization. The prenatal care characteristics collected were the number of visits and the qualification of the main care provider. Number of prenatal visits were classified according to the recommended calendar of visits and the length of gestation,³⁵ as low (<110% of the minimum expected number), average (110-119%), high (120-149%) and very high (150% or more). Women with no complications can choose care by an ob-gyn, midwife or GP. Complications requiring care by an ob-gyn or hospitalization, care by an ob-gyn or midwife and a high number of visits were considered indicators of numerous contacts with care givers and/or opportunities to follow recommendations for pregnant women.

In our analysis, we first estimated the influenza vaccine coverage during the 2015–16 influenza season and studied factors associated with vaccine uptake. Second, we described the vaccine uptake by qualification of the prescriber (ob-gyn, midwife, GP) and qualification of the main care provider. Finally, we studied factors associated with vaccine refusal. Because we had to choose the groupings for some women’s responses, we repeated the analysis of vaccine refusal by changing these choices to verify the stability of the results.

Univariate and multivariable analyses were conducted by using robust Poisson regression with a robust error variance.³⁶ All statistical analyses were performed with Stata/IC v15.0 (StataCorp LP, College Station, TX). P < .05 was considered statistically significant, and prevalence ratios (PRs) are displayed with their 95% confidence intervals (95% CIs). Missing data were low for maternal and prenatal care characteristics (<1.5%). Women with missing data were excluded in multivariable models.

Results

Inluenza vaccine coverage and determinants

Influenza vaccine coverage among pregnant women was 7.4% (n = 865/11712; 95% CI: 6.9–7.9). Table 1 shows the maternal characteristics associated with influenza vaccine uptake. Coverage was higher for women with age 30–34 years, low parity, preexisting conditions at risk for flu, and high level of education, birth in France, a healthcare occupation, and private complementary health insurance. Pregnant women were more frequently vaccinated when their main care provider was a GP than other care provider. In the multivariable model, influenza vaccination was mainly associated with maternal age (≥ 30 years old), low parity (< 3 children), preexisting conditions at risk for severe influenza infection, high educational level (college and postgraduate) and maternal occupation related to healthcare (all p < .001). The adjusted PR was 2.5 (95% CI: 2.0–3.2) with a postgraduate education versus middle school level or lower education, 2.1 (95% CI: 1.4–3.2) for nulliparity and 1.7 (95% CI: 1.2–2.6) with 1 or 2 previous children compared to other women.

Table 1. Influenza vaccine uptake and associated factors among women in mainland France, 2015–2016 season

	n	Rate (%)	p value^a	PR^b	95% CI	p value^c
TOTAL	11712	7.4
Age (years)			<0.001			<0.001
<25	1502	4.1		1
25-29	3719	6.3		1.17	0.87–1.57
30-34	4011	9.1		1.60	1.20–2.14
35 or more	2480	8.3		1.56	1.14–2.13
Parity			<0.001			<0.001
0	4972	8.3		2.11	1.39–3.20
1-2	5878	7.2		1.73	1.15–2.60
3 or more	862	3.0		1
Medical conditions			0.002			0.001
None	9693	7.4		1
Preexisting conditions at risk for flu	450	11.8		1.69	1.29–2.20
Other preexisting conditions^d	439	6.8		1.03	0.72–1.47
Hospitalization during pregnancy	1130	6.2		0.91	0.72–1.16
Family status			0.21			0.054
Cohabiting	10714	7.5		1
Not cohabiting	975	6.4		1.31	1.00–1.72
Educational level			<0.001			<0.001
Middle school or less	2660	4.4		1
High school	2511	4.5		0.92	0.71–1.20
Some college	2240	6.1		1.14	0.89–1.47
College	2115	9.3		1.61	1.27–2.04
Postgraduate	2087	14.3		2.54	2.01–3.20
Maternal country of birth			0.003			0.86
France	9543	7.7		1
Other countries	2168	5.9		1.02	0.83–1.24
Maternal occupation during pregnancy			<0.001			<0.001
Healthcare worker	1522	13.1		1.76	1.50–2.06
Contact with children	801	9.9		1.17	0.92–1.48
Other	5966	7.3		1
Not working during pregnancy	3406	4.4		0.89	0.73–1.10
Complementary health insurance^e			<0.001			0.39
Public insurance (CMU-C)	1061	4.5		1.25	0.81–1.92
Private insurance	9560	8.1		1.27	0.90–1.80
None	1009	3.8		1
Number of prenatal visits			0.12			0.45
Low	2217	6.2		0.88	0.70–1.10
Average	1921	7.8		1
High	4741	7.8		0.99	0.83–1.19
Very high	2681	7.6		1.04	0.85–1.27
Main healthcare provider			0.015			0.03
Ob-gyn (private)	5771	7.7		1
Ob-gyn (public)	1852	7.5		1.05	0.87–1.26
Midwife (private)	988	6.9		0.77	0.60–0.98
Midwife (public)	1720	5.9		0.90	0.74–1.11
General practitioner	758	9.9		1.29	1.01–1.66
Other^f	523	6.5		1.18	0.85–1.63

CI: confidence interval; CMU-C: universal complementary health coverage; Ob-gyn: obstetrician-gynecologist; PR: prevalence ratio.

The number of missing data varies for each characteristic; n = 11,264 in the final robust Poisson regression model.

^aPearson chi-square test.

^bPrevalence ratios adjusted for all variables in the table and administrative regions.

^cWald test.

^dMedical conditions requiring care by an obstetrician-gynecologist.

^eFor additional costs not covered by the national health insurance; public insurance is for low-income people.

^fMaternal and Child Health Services or several care providers.

For the first 6 months of pregnancy, the main care providers were an ob-gyn (n = 574, 68%), midwife (n = 170, 20%) and GP (n = 72, 9%) (Table 2). The qualification of the influenza vaccine prescriber was often similar to that of the main care provider. However, in each group of care providers, GPs prescribed the influenza vaccine very often.

Table 2. Distribution of prescribers by main care provider during the first six months of pregnancy^a.

		Vaccine prescriber (%)^b
Main care provider	N (%)^c	Ob-gyn	Midwife	GP	Other
Ob-gyn	574 (67.6)	49.5	4.7	25.6	20.2
Midwife	170 (20.0)	5.9	39.4	33.5	21.2
General practitioner	72 (8.5)	1.4	9.7	81.9	6.9
Other^d	33 (3.9)	21.2	24.2	24.2	30.3
Total	849 (100)^e	302 (35.6)	109 (12.8)	271 (32.0)	167 (19.7)

Ob-gyn: obstetrician-gynecologist.

^aPearson chi-square test (p < 0.001).

^bDistribution of the qualification of prescribers.

^cDistribution of the main care providers.

^dMaternal and Child Health Services or several care providers.

^eAmong 865 vaccinated women, the vaccine prescriber was unknown for 13 women and was the main care provider for 3 women.

Non-vaccinated women and factors associated with influenza vaccine refusal

Among the 10,847 non-vaccinated women, 8,172 reported receiving no provider offer, 1,861 refused to be vaccinated after a proposal, 195 reported other reasons for not being vaccinated and 619 did not specify why they did not get vaccinated. The other reasons were beliefs that the vaccine is not necessary or not a priority (n = 57/195), lack of knowledge and lack of awareness about influenza and/or the vaccine (n = 44/195) and medical exemptions (n = 17/195). We hypothesized that women who did not specify why they were not vaccinated did not receive a provider offer and that women who were not vaccinated because of other reasons had an offer but were not in favor of the vaccination and refused it. Therefore, the overall rate of vaccine offers was 24.9% (95% CI: 24.2–25.7).

Vaccine refusal was estimated to be a fraction of the total estimation of women receiving a provider offer, grouped as vaccinated women, women who refused vaccination after a proposal and those reporting other reasons. Thus, 70.4% (n = 2056/2921; 95% CI: 68.7–72.0) of women with an offer refused vaccination. Table 3 presents the vaccine refusal data according to maternal characteristics. Refusal was frequent among women aged < 25 years, who were multiparae, did not have a post-graduate education and who were not working during pregnancy. In the multivariable analysis including socio-demographic and medical characteristics, number of prenatal visits and qualification of the main care provider, only multiparity and less than postgraduate education were associated with vaccine refusal (both p < .001).

Table 3. Vaccine refusal rates according to maternal characteristics in mainland France, 2015–2016 season

	n	Rate (%)	p value^a	PR^b	95% CI	p value^c
TOTAL	2921	70.4
Age (years)			0.007			0.11
<25	259	76.1		1.09	1.00–1.19
25-29	867	73.1		1.07	1.01–1.13
30-34	1118	67.4		1
35 or more	677	69.7		1.03	0.97–1.10
Parity			<0.001			<0.001
0	1260	67.1		1
1-2	1505	71.8		1.08	1.03–1.14
3 or more	156	83.3		1.22	1.11–1.34
Medical conditions			0.2			0.55
None	2383	70.1		1
Preexisting conditions at risk for flu	155	65.8		0.94	0.84–1.06
Other preexisting conditions^d	129	76.7		1.05	0.95–1.16
Hospitalization during pregnancy	254	72.4		1.01	0.93–1.10
Educational level			<0.001			<0.001
Middle school or less	487	76.0		1.31	1.20–1.43
High school	502	77.7		1.34	1.23–1.46
Some college	556	75.5		1.30	1.19–1.41
College	657	70.0		1.22	1.12–1.33
Postgraduate	699	57.2		1
Maternal occupation during pregnancy			0.019			0.28
Healthcare worker	679	70.7		1.00	0.94–1.06
Contact with children	211	62.6		0.90	0.81–1.01
Other	1457	69.9		1
Not working during pregnancy	573	74.0		0.97	0.91–1.04

CI: confidence interval; PR: prevalence ratio.

The number of missing data varies for each characteristic; n = 2,840 in the final robust Poisson regression model.

^aPearson chi-square test.

^bPrevalence ratios adjusted for all variables in the table and family status, maternal country of birth, complementary health insurance, number of prenatal visits and qualification of the main care provider.

^cWald test.

^dMedical conditions requiring care by an obstetrician-gynecologist.

We then considered that the proposal of vaccination was unknown for women who did not give the reasons for lack of vaccination and that the reasons for non-vaccination other than non-proposal and refusal were not close to refusal and could not be included in this category. Under these hypotheses, the overall rate of vaccine offer was 25.0% and the refusal rate was 68.3%. The associations of refusal with the studied determinants were very similar to those found in Table 3 (results not shown).

Discussion

During the 2015–2016 influenza vaccine campaign, the influenza vaccine coverage among pregnant women in France was low, but only about a quarter of pregnant women received a care-provider offer of vaccination (24.9%), and most declined vaccination after receiving the proposal (70.4%). These results highlight the critical role of care providers in the vaccine uptake and prominent hesitancy surrounding influenza vaccine acceptance during pregnancy, especially in women with high parity and low educational level.

The main strength of our study is the use of a large national sample of births. The overall sample (all births during 1 week in March) is representative of annual births in France.³⁴ The proportion of women who agreed to be interviewed was very high. Furthermore, the study design was optimal to assess the vaccine coverage because all women were exposed during pregnancy to both the national influenza vaccination campaign (from mid-October to late January) and the wave of seasonal influenza epidemic (from early February to late April 2016).³⁷ Finally, the NPS provides a wide range of data on maternal sociodemographic characteristics, chronic diseases and complications, and characteristics of prenatal care that could be used to assess the determinants of vaccine uptake.

Nevertheless, our results are subject to several limitations. First, because the NPS covers many topics, we could not focus on attitudes, knowledge and perceptions of pregnant women toward influenza vaccination and explore further misconceptions or barriers to vaccination acceptance. Secondly, women who did not participate may be slightly different from the overall sample; for instance, we found underrepresentation of non-French women (18.7% in our study vs 22.2% in the vital statistics),³⁸ probably because of language difficulties. Finally, some women may have forgotten the vaccination; however, the under-reporting is probably low because vaccination requires special procedures for most women (e.g., vaccine picked up from a pharmacy and administration by a doctor or nurse).

Despite the optimal period of pregnancy covered by our sample, the vaccination coverage was very low in France as compared with that described in other population-based studies – 54% in the United States,³⁹ 45% in Flanders (Belgium),⁴⁰ 42% in England⁴¹ and 23% in Germany⁴² – although methods of data collection varied in these studies.

Our results suggest that both care providers and women play a major role in the situation in France. Their respective roles might be slightly biased because we did not directly ask every woman whether her care providers offered vaccination and whether she was against vaccination. We may have slightly underestimated the proportion of women receiving a care-provider offer of vaccination, because some women may have forgotten the proposal. In addition, the care provider offers may be slightly underestimated, and consequently refusal may be slightly underestimated if providers sometimes refrain from recommending or offering vaccination when they already know the adverse opinion of pregnant women regarding influenza vaccination. However, our results were similar when we excluded women who did not know why they were not vaccinated. The rates of provider offers and women’s refusal are very different from those found in other countries and cannot be explained by our questionnaire alone; for instance, in the 2016–17 season in the United States, about three quarters of women who were offered vaccination were vaccinated.³⁹

In France, the very low vaccine coverage among pregnant women and low attention paid to vaccination by both care providers and women could be explained in part by the fact that national recommendations were relatively recent (3 years at the time of the survey), whereas longer lead times are needed to observe changes in immunization coverage.^25,43 Furthermore, unlike older people and other at-risk groups of severe influenza infection, pregnant women do not receive a voucher for vaccination by the national health insurance, and in 2015–2016 they were not specifically targeted by national media campaigns. However, vaccine coverage for the 2015–2016 influenza season was also low among other at-risk groups: 51% for people ≥ 65 years old and 39% for those < 65 years old with preexisting conditions at risk for severe influenza⁴⁴ as compared with the national target of at least 75% for both groups. These low coverage rates result from controversies surrounding the pandemic 2009 influenza A (H1N1) national vaccination campaign for which vaccination coverage was low among people ≥ 65 years old (10.4%) and pregnant women (29.3%).^24,33 More generally, the French population, including healthcare workers, has a strong negative attitude to vaccines, with strong skepticism about their safety and effectiveness, especially the influenza vaccine.^45-47 This attitude may be reinforced during pregnancy in that mothers are very concerned for their newborn’s safety regarding medication,^30,48 and could be unaware of the severity of influenza infection and the risks associated with influenza for them and their newborn,²⁸ may be because of the rareness of severe infections in both groups.⁴⁹

Our results on the determinants of vaccination and vaccine refusal show that some women have better vaccination coverage and a higher rate of acceptance than other women, mainly nulliparae women, women with a very high education level, women with preexisting conditions at risk for influenza complications and healthcare workers. Other studies of vaccination coverage found some of these maternal characteristics^28,33,39 and they are well-known determinants of other maternal preventive behavior. For instance, it is known that nulliparous women pay more attention to prevention during pregnancy than women with higher parity.^50,51 In addition, multiparae women who could benefit from the cumulative effect of experience and information gained by repeated contacts with care providers often did not receive a vaccine recommendation during previous pregnancies because guidelines for influenza vaccination were recent. However, vaccine coverage never exceeded 15%, including in women with significant co-morbidities or pregnancy complications, which suggests that any campaign or public policy in France should reach the overall population of pregnant women.

Other factors in our survey provide information on the potential role of the health system. First, we found that vaccine coverage was not associated with insurance coverage, which suggests that financial barriers do not contribute significantly to differences in vaccination rates. Second, we found that GPs have a role in the vaccine uptake and management of seasonal influenza: the vaccine coverage was higher when GPs were the main care provider and GPs frequently prescribed the vaccine, even if they were not the main care provider for this pregnancy. However, less than 10% of the women said that their main care provider was a GP. Our results underline the role of family physicians in vaccine coverage but also the low adherence of the main care providers to recommendations, but these professionals examine the women at least every month during pregnancy, and they could have a strong influence on vaccine uptake, discussing and providing relevant information and clear recommendations, as it was pointed out in many studies.^25,28,30,52

Seasonal influenza vaccination among pregnant women illustrates the difficulties in implementing new public health policies, especially routine immunization programs. Several systematic reviews have identified strategies and interventions that may help to improve vaccine coverage during pregnancy.^29,53 The most effective strategies would be (1) communication approaches targeting pregnant women, emphasizing scientific knowledge on the effectiveness and safety of vaccination for both the mother and baby, (2) education and training of healthcare providers to help ensure that they have the best knowledge to apply recommendations and the best information tools to promote them to pregnant women, (3) extended access to vaccination by enhancing procurement and vaccine storage in care-provider offices and maternity units and (4) implementation of alerts about vaccination on electronic medical records. In this perspective, a multicomponent intervention implemented in a maternity unit in Paris, including information given to pregnant women during prenatal visits, on-site vaccination, and active involvement of antenatal care providers, resulted in a significant increase in immunization coverage.⁵⁴

Conclusion

Three years after the release of national recommendations, seasonal influenza vaccine uptake among pregnant women was very low in mainland France. Significant efforts are needed to promote influenza immunization during pregnancy by improving the dissemination of information on its benefits but also implementing effective interventions, targeting both pregnant women and healthcare providers. In this perspective, research is necessary to better understand barriers to vaccination, especially factors influencing vaccine acceptance during pregnancy.

Disclosure of potential conflicts of interests

The authors declare that they have no competing interests.

Acknowledgments

The authors thank the department heads who agreed to perform the survey in their unit, the investigators and all the women who agreed to participate.

Additional information

Funding

The 2016 National Perinatal Survey was developed and implemented by the French National Institute of Health and Medical Research (INSERM), three directorates of the Ministry of Social Affairs and Health and the French National Public Health Agency (Santé publique France).