Author manuscript, published in "Vaccine 2010;28(15):2743-8"
DOI : 10.1016/j.vaccine.2010.01.027
Positive attitudes of French general practitioners towards A/H1N1 influenza pandemic
vaccination: a missed opportunity to increase vaccination uptakes in the general public?
Michaël Schwarzinger1,2,3*, Pierre Verger1,2,3, Marc-André Guerville4, Catherine Aubry5,
Sophie Rolland1,2,3, Yolande Obadia1,2,3, Jean-Paul Moatti1,2
1
INSERM, U912 (SE4S), Marseille, France
2
Université Aix Marseille, IRD, UMR-S912, Marseille, France
3
ORS PACA, Observatoire Régional de la Santé Provence Alpes Côte d'Azur, Marseille,
inserm-00636175, version 1 - 26 Oct 2011
France
4
URML PACA, Union Régionale des Médecins Libéraux de Provence Alpes Côte d'Azur,
France
5
URML Pays de la Loire, Union Régionale des Médecins Libéraux des Pays de la Loire,
France
* Corresponding author:
Dr Michaël Schwarzinger
INSERM, U912 (SE4S)
23 rue Stanislas Torrents, 13006 Marseille, France
Tel: +33 496102860
Fax: +33 491598924
michael.schwarzinger@inserm.fr
-1-
Summary: Attitudes of general practitioners (GPs) towards A/H1N1 pandemic vaccination
are unknown. We conducted a cross-sectional survey with computer-assisted telephone
interviewing in the French Regional Panel of General Practices from June 16 to September
22, 2009. Of 1,434 respondents representative of GPs in four French regions, 885 (61.7%)
were willing to accept A/H1N1 pandemic vaccination for themselves. The personal history of
seasonal flu vaccination was the strongest independent predictive factor of willingness to
accept A/H1N1 pandemic vaccination (P52). GPs practicing exclusively in hospitals or long term care facilities, GPs practicing
exclusively alternative medicines (such as homeopathy or acupuncture), GPs who were not
practicing at time of the survey due to sick leave or retirement, and GPs planning to move out
of their present region in the next six months were excluded. GPs received a compensation
equivalent to 2 consultation fees for their participation to each cross-sectional survey.
Participant GPs, who dropped out in a following cross-sectional survey, were randomly
replaced according to their stratum.
Of the 2,498 eligible GPs invited to participate to the French Regional Panel of General
Practices in March 2007, 1,451 (58.1%) gave their agreement to participate. The 1,047 GPs
who refused to participate did not differ from participants according to practice location and
gender, but they were older (P=.02). Lack of time was their main reason for refusal. Results
presented in this paper are based on the 1,434 panel GPs who participated to the fifth cross-
-5-
sectional survey conducted between June 16 and September 22, 2009 (response rate of
1,434/1,451; 98.8%).
Procedure and questionnaire
The survey was conducted among GPs by professional investigators with computer-assisted
telephone interviewing. The questionnaire was pilot-tested for clarity, length and face validity
among 20 GPs.
Respondents were asked whether or not they would accept the A/H1N1 pandemic vaccination
inserm-00636175, version 1 - 26 Oct 2011
for themselves as soon as it became available. In addition, they were asked to assess the
importance of other protective measures against A/H1N1 2009 influenza virus in the practice
on a 10-point scale from 1 “not important at all” to 10 “of utmost importance”: 1) to wear a
facial mask during each clinical encounter with patients; 2) to provide facial masks to all
patients and having them wear it in the practice; 3) to recommend patients with influenza-like
illness to stay at home and to visit them at home for medical care. A total score of importance
of A/H1N1 protective measures was also computed by aggregating the answers to these three
items [min=3 to max=30].
Moreover, respondents were asked what was the maximum number of additional consultations
they would be ready to provide per day at the peak of the influenza-pandemic to care for
patients with flu or to vaccinate the population. Respondents were asked how many times they
had been vaccinated against seasonal influenza in the prior three years (each year, twice, once,
never); whether they had read the “national plan for influenza-pandemic preparedness and
response” (yes/no); and whether the primary objective of pandemic vaccination should be “to
protect individuals at higher risk for influenza complications” or “to mitigate the transmission
of influenza virus in the whole population”.
-6-
The questionnaire also collected data on individual and occupational characteristics of
respondents: gender and age; solo or group practice; being on call for emergencies; working
part-time in long term care facilities or in hospitals in addition to their ambulatory practice;
practicing some alternative medicine such as homeopathy; participating in Continuing
Medical Education during the previous year (CME); payment scheme for consultation (Social
Security fixed fees or free pricing). Observational data on GP’s activity in 2007 and 2008 was
obtained, in parallel to the survey, from the Social Security exhaustive reimbursement
database that includes for each GP the total number of consultations and home visits per year,
and the age distribution of the GP’s clientele according to four age categories (less than 16; 16
inserm-00636175, version 1 - 26 Oct 2011
to 59; 60 to 70; more than 70).
The survey was approved by the National Data Protection Authority (Commission Nationale
Informatique et Libertés/ CNIL) which is in charge of ethical issues and protection of
individual data collection in France.
Statistical analysis
The main outcome variable was GP’s willingness to accept A/H1N1 pandemic vaccination for
themselves (yes versus no or don’t know). Univariate associations between respondent’s
willingness to accept A/H1N1 pandemic vaccination and other variables were tested using
logistic regression. Explanatory variables, that were related to willingness to accept A/H1N1
pandemic vaccination at the p≤.15 level, were subsequently introduced in a backward
multivariate logistic model (p exit>.05) to identify independent predictive factors for such
willingness. Region, location of general practice, gender, and age were forced in the final
regression model since these variables had been used to stratify the sample. Adjusted odds
ratios and 95% confidence intervals were presented for the main findings. Data were analysed
using SAS 9.1.3 (SAS Institute, Cary NC).
-7-
Results
Individual and occupational characteristics of the 1,434 GPs who participated in the survey, as
well as characteristics of their practice, are detailed in Table 1. A total of 885 respondents
(61.7%) declared their willingness to accept A/H1N1 pandemic vaccination for themselves as
soon as it became available, while an even higher proportion (70.6%) had systematically been
vaccinated for seasonal influenza in the prior three years (Table 2). On average, GPs were
ready to increase their workload in response to the pandemic by undertaking 11.6 (sd = 6.5)
consultations per day in addition to their usual practice activity. A large majority of
respondents agreed that the main objective of pandemic vaccination was the public health
inserm-00636175, version 1 - 26 Oct 2011
goal to mitigate the transmission of the influenza virus in the whole population and had some
knowledge of the national preparedness plan against an influenza-pandemic. Respectively,
36.8%, 39.5%, and 38.9% declared that it was “of utmost importance” (score = 10) that the
GP wears a facial protective mask with each patient; orders patient to wear facial masks in the
practice; and recommends influenza-ill patients to stay at home. Table 2 also presents the
scores of importance that GPs associated with these three measures of protection against
A/H1N1 2009 influenza virus in the practice.
Univariate analyses showed that willingness to accept A/H1N1 pandemic vaccination did not
vary significantly according to the four individual characteristics used for stratification in the
survey: region, location of general practice, gender, and age. Several occupational
characteristics of GPs and their practice (working part-time in hospitals in parallel to their
ambulatory practice; participating in CME; and type of payment scheme) were not associated
with acceptability of A/H1N1 vaccination. Age distribution of GP’s clientele was also not
associated with their personal intention towards vaccination.
By contrast, as shown in Table 3, some GPs’ characteristics were found to be significantly
related to their willingness to accept A/H1N1 pandemic vaccination even after adjustment
-8-
through multivariate logistic analysis. GPs more willing to get vaccinated were those:
working in group practices; being on call for emergencies; working part-time in long term
care facilities (usually for the elderly) in parallel to their ambulatory practice; and having the
highest workload in practice (more than 4000 consultations per year). Quite logically, GPs
who were more ready to increase their workload in response to the pandemic, who adhered
the most to implementing other protective measures in their practice, and who believed that
mitigation of transmission of the influenza virus in the whole population was the primary
objective of pandemic vaccination were also more willing to accept A/H1N1 pandemic
vaccination for themselves. The multivariate logistic analysis presented in Table 3 highlights
inserm-00636175, version 1 - 26 Oct 2011
that history of seasonal flu vaccination in the prior three years was the strongest predictive
factor of willingness to accept A/H1N1 pandemic vaccination. This willingness increased
significantly with the number of seasonal flu vaccines received in the prior three years
(Cochran-Armitage test for trend: P 70 years old in 2007-2008, mean (SD) 15.2 (7.4)
* Values are numbers (percentage) of respondents, except where stated otherwise.
- 20 -
Table 2. Beliefs, attitudes, and opinions of French GPs toward the A/H1N1 influenza-pandemic (N=1,434).*
Characteristics
Willingness of GPs to accept A/H1N1 pandemic vaccination for themselves
Yes 885 (61.7)
No 474 (33.1)
Don't know 75 (5.2)
Had been vaccinated against seasonal influenza in the prior three years
Every year 1,013 (70.6)
inserm-00636175, version 1 - 26 Oct 2011
Twice 110 (7.7)
Once 90 (6.3)
Never 221(15.4)
Has read the national preparedness plan against an influenza-pandemic
Yes 1,238 (86.3)
No 196 (13.7)
Primary objective of influenza-pandemic vaccination
To limit the transmission of the virus in the whole population 1,071 (74.7)
To protect individuals at higher risk for influenza complications 363 (25.3)
Score of importance of protective measures against A/H1N1 pandemic, mean (SD), (1-10) score†
Wearing a facial mask with each patient 7.0 (3.2)
Ordering patients to wear a facial mask in the practice 7.1 (3.2)
Recommending patients with flu to stay at home 7.2 (3.1)
Total score of importance of 3 protective measures against A/H1N1 pandemic, mean (SD), (3-30) score 21.3 (6.6)
Additional workload to respond to A/H1N1 influenza pandemic, mean (SD), max number of consultations
11.6 (6.5)
per day
* Values are numbers (percentage) of respondents, except where stated otherwise.
†Score ranging from 1 (not important at all) to 10 (of utmost importance)
- 21 -
Table 3. Factors associated with willingness of GPs to accept A/H1N1 pandemic vaccination for themselves: univariate and multivariate logistic models (N=1,434
GPs)
No (%) willing to
accept pandemic Unadjusted Adjusted odds
Variables and variable levels† vaccination odds ratio p value ratio (95% CI)‡ p value
Practice type
Solo 428 (55.9) 1 [reference] 1 [reference]
Group 457 (68.4) 1.71 4,000 total per year 621 (65.7) 1.63 20 590 (69.0) 2.14 10 consultations per day 313 (67.2) 1.41 .002 1.39 (1.10 to 1.77) .007
* Continuous variables were also significantly associated to willingness to accept A/H1N1 pandemic vaccination in the multivariate logistic model and were dichotomized at
the median value for presentation of the final multivariate model
† Only variables associated with willingness to accept A/H1N1 pandemic vaccination at the p≤.15 level in univariate analysis are shown
‡ Adjusted odds ratio were also controlled for region, location of general practice, gender, and age used to stratify the sample; these variables were not significant in
univariate analyses (p=.377; p=.835; p=.266; and p=.281, respectively). Hosmer-Lemeshow test: P=.78, suggests that the goodness of fit was adequate for the final
- 22 -