Unexplained variations in general practitioners’ perceptions and practices regarding vaccination in France

European Journal of Public Health, Feb 2019

Given geographical disparities in vaccination coverage (VC) and the crucial role general practitioners (GPs) play in vaccination in France, we sought to: 1) determine the existence of geographical variations in GPs' perceptions of vaccines, their trust in information sources, and the frequency of their recommendations to patients by comparing data from southeast (SE), central-west (CW), northwest (NW), and the rest of France; and 2) identify individual and contextual factors associated with regional variations in GPs' recommendations.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://academic.oup.com/eurpub/article-pdf/29/1/2/27596989/cky146.pdf

Unexplained variations in general practitioners’ perceptions and practices regarding vaccination in France

Abstract Background Given geographical disparities in vaccination coverage (VC) and the crucial role general practitioners (GPs) play in vaccination in France, we sought to: 1) determine the existence of geographical variations in GPs' perceptions of vaccines, their trust in information sources, and the frequency of their recommendations to patients by comparing data from southeast (SE), central-west (CW), northwest (NW), and the rest of France; and 2) identify individual and contextual factors associated with regional variations in GPs' recommendations. Methods This cross-sectional observational study in 2014 collected data from a panel of 2586 French GPs in private practice: 3 specific regional samples and a fourth sample for the rest of France. We calculated a composite score summarizing GPs' vaccine recommendation frequency for 6 vaccine situations and used a five-step hierarchical linear regression to study the score's links with practice location and individual and contextual factors. Results GPs' vaccine recommendation frequency score was highest in NW France and lowest in the SE. The low SE score was explained by GPs' greater doubts about vaccine utility and risks and lower trust in information sources. The high NW score was partially explained by greater adherence to guidelines by GPs there. The contextual factors studied did not explain regional differences. Conclusion The geographical variations in GPs' vaccination-related attitudes and practices suggest that vaccine hesitancy among GPs differs in prevalence between regions. These variations coincide with north/south trends in population VC. Intervention strategies to restore confidence in vaccines should target GPs and must be adapted to each regional context. Introduction Vaccination coverage (VC) for routine childhood vaccines and seasonal influenza vaccine can differ substantially within countries.1–3 It does so in France, with a downward gradient from its northern to southern regions of VC against measles, meningococcal C (menC) and human papillomavirus (HPV) in the general population4,5 and of VC against hepatitis B and pertussis among general practitioners (GPs).6 These variations raise questions about the distribution of vaccine hesitancy (VH) within countries, which has been shown to be highly context-dependent7 and is especially marked in France.8 Geographic clusters of underimmunization have already been observed in various countries, particularly in infants, and often correspond to pockets of VH among parents.9,10 These observations justify the investigation of the determinants of VH at local levels, as already reported for parts of the USA and Europe (e.g. for HPV vaccination).11,12 Healthcare professionals influence their patients’ decisions regarding vaccination.13,14 These professionals may also be vaccine-hesitant, as found for significant percentages of GPs in France and elsewhere.15,16 Studies controlling for patients’ characteristics regularly show geographical variations of medical practices in various fields.17,18 Data from previous national and regional studies in France (grey literature) of GPs’ practices report north/south differences, for example in the management of pharyngitis in children and the use of rapid streptococcus tests. Several reasons may explain these differences. They may result from variations within a country in the content/modalities of initial training (medical school programmes and clinical internships), the availability and variety of continuing medical education (CME), or the epidemiologic context. A second line of explanations lies in the adaptation of physicians’ practice styles (i.e. their preferences and habits regarding prescriptions) to their environment (socio-cultural context, constraints, and advantages, such as local health policies, demographics and density of various healthcare professionals).19 Better understanding of within-country geographical variations of medical practices concerning vaccination may improve our understanding of the causes of within-country variations of VH and enable us to design tailored interventions to address it more appropriately. Using data from a representative panel of 2586 GPs in private practice in France composed of 4 separate, mutually exclusive samples (three specific regional samples and another multiregional for the rest of France),20 we sought to: (i) compare, between these samples, GPs’ perceptions of the utility and risks of some vaccines recommended by French health authorities, their trust in official sources about vaccination, and the frequency of their vaccine recommendations to their patients, and (ii) identify the individual (professional characteristics, experience related to vaccination and vaccine-preventable diseases (VPD), and perceptions of information sources about vaccines) and contextual factors (epidemiological context of some VPDs and local density of physicians) associated with regional variations in GPs’ vaccine recommendations to patients. Methods Population Primary care in France is mostly provided by GPs in private practice, paid according to a fee-for-service system. A supplementary payment-for-performance (P4P) programme (‘Rémunération sur objectifs de santé publique, P4P-ROSP’) began in 2012, funded by the National Health Insurance Fund to induce GPs to work towards public health objectives.The panel was designed to collect information about GPs’ medical practices, working conditions and opinions about public health policies. The method used to set up the panel has been described in detail elsewhere.20 Briefly, between November 2013 and March 2014, GPs were randomly selected from the exhaustive French database of healthcare professionals (62 000 GPs in total) and included into one of four different samples: three separate regional samples, covering southeastern (SE), central-western (CW) and northwestern (NW) France, and a fourth multiregional sample covering the rest of France (map in Supplementary appendix S1). Sampling was stratified for gender, age, number of office consultations and house calls in 2012 (workload), and for the density of each GP’s municipality of practice. The National Authority for Statistical Information (Commission Nationale de l’Information Statistique) approved the panel. Procedure and questionnaire The first cross-sectional survey took place from April to July 2014 and focussed on GPs’ vaccination perceptions and practices.15,20 GPs received compensation equivalent to one consultation fee for their participation in this survey. Professional investigators conducted the interview with the help of computer-assisted telephone interview software. After a review of existing literature and qualitative interviews on the topic with 10 GPs, we developed a standardized questionnaire (see Supplementary appendix S2), which was then pilot-tested for clarity, length, and face validity among 50 GPs. It collected information about: (i) GPs self-reported vaccine recommendation frequency (1 = never to 4 = always) in six specific vaccine situations, chosen because their current VC rates in France do not meet official objectives: MMR for nonimmune adolescents and young adults, MenC for 12-month-old infants, MenC for ages 2–24 (catch-up), HPV for girls aged 11–14, hepatitis B for adolescents (catch-up), and seasonal influenza for adults under 65 with diabetes; (ii) GPs’ trust in the reliability of information about the overall benefits and risks of vaccines provided by various sources (1 = no trust to 4 = strong trust); and GPs’ opinions about (iii) their perception of the likelihood of links between potential severe adverse effects and certain vaccines (1 = not at all likely to 4 = very likely); and (iv) the utility of vaccines generally (1 = strongly disagree to 4 = strongly agree).15 GPs were also asked about their professional characteristics (including CME participation and practice of complementary medicine, such as homeopathy and acupuncture) and experiences related to vaccination (table 1), and their attitudes towards the official vaccine recommendations (table 2). Table 1 Characteristics of the study population by geographic sample (panel of French GPs, weighted data)a % Rest of Franceb (ref.) (n = 1289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Response rates 93% 94% 90% 88% Stratification variables         Age—years (tertiles)             <50 (ref.) 29.9 34.9 33.1 28.1         50-58 32.0 31.9 30.7 33.0         >58 38.1 33.2 36.2 38.9**     Gender             Male (ref.) 70.4 67.5 68.9 70.5         Female 29.6 32.5 31.1 29.5     GP density in the municipality of practice (Min-Q1/Q1–Q3/Q3-Max)             <−19.3% of national average (ref.) 27.0 20.6 17.9 9.4         −19.3% to +17.7% of national average 47.3 63.5** 55.9* 64.8**         > +17.7% of national average 25.7 15.9** 26.2* 25.8**     2012 workload (Min-Q1/Q1-Q3/Q3-Max)             <3067 consultations/visits (ref.) 22.3 22.3 19.3 34.0         3067–6028 consultations/visits 51.4 52.6 48.7 48.4**         >6028 consultations/visits 26.3 25.1 32.0 17.6** Professional characteristics         Practice             Solo (ref.) 43.5 34.0 29.5 47.8         Group 56.5 66.0* 70.5** 52.2*     Occasional practice of complementary medicinec             No (ref.) 87.2 90.6 90.2 85.7         Yes 12.8 9.4* 9.8 14.3     CME completed last year on vaccination/infectious diseases >1 session             No (ref.) 56.0 59.5 51.4 63.4         Yes 44.0 40.5 48.6 36.6** Practice population characteristics         Proportion of patients aged under 16 21.0 24.0 20.6** 18.8**     Proportion of patients aged over 60 24.3 23.1** 28.5* 27.8** Experience related to vaccination         Any patients with at least one VPD in the past 5 yearsd             No (ref.) 10.9 14.0 8.0 5.8         Yes 89.1 86.0* 92.0 94.2** Any patients with a potential vaccine adverse effect             No (ref.) 83.8 79.8 83.3 81.4         Yes 16.2 20.2 16.7 18.6  % Rest of Franceb (ref.) (n = 1289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Response rates 93% 94% 90% 88% Stratification variables         Age—years (tertiles)             <50 (ref.) 29.9 34.9 33.1 28.1         50-58 32.0 31.9 30.7 33.0         >58 38.1 33.2 36.2 38.9**     Gender             Male (ref.) 70.4 67.5 68.9 70.5         Female 29.6 32.5 31.1 29.5     GP density in the municipality of practice (Min-Q1/Q1–Q3/Q3-Max)             <−19.3% of national average (ref.) 27.0 20.6 17.9 9.4         −19.3% to +17.7% of national average 47.3 63.5** 55.9* 64.8**         > +17.7% of national average 25.7 15.9** 26.2* 25.8**     2012 workload (Min-Q1/Q1-Q3/Q3-Max)             <3067 consultations/visits (ref.) 22.3 22.3 19.3 34.0         3067–6028 consultations/visits 51.4 52.6 48.7 48.4**         >6028 consultations/visits 26.3 25.1 32.0 17.6** Professional characteristics         Practice             Solo (ref.) 43.5 34.0 29.5 47.8         Group 56.5 66.0* 70.5** 52.2*     Occasional practice of complementary medicinec             No (ref.) 87.2 90.6 90.2 85.7         Yes 12.8 9.4* 9.8 14.3     CME completed last year on vaccination/infectious diseases >1 session             No (ref.) 56.0 59.5 51.4 63.4         Yes 44.0 40.5 48.6 36.6** Practice population characteristics         Proportion of patients aged under 16 21.0 24.0 20.6** 18.8**     Proportion of patients aged over 60 24.3 23.1** 28.5* 27.8** Experience related to vaccination         Any patients with at least one VPD in the past 5 yearsd             No (ref.) 10.9 14.0 8.0 5.8         Yes 89.1 86.0* 92.0 94.2** Any patients with a potential vaccine adverse effect             No (ref.) 83.8 79.8 83.3 81.4         Yes 16.2 20.2 16.7 18.6  VPD, vaccine-preventable disease; CME, continuing medical education. a Multivariable regressions on unweighted data, adjusted for stratification variables, were used to test for regional vs. ‘Rest of France’ (reference) differences. b Sample, not including any of the GPs in the three regional samples considered here. c Complementary medicine: e.g. homeopathy and/or acupuncture. d Five VPDs were mentioned in the questionnaire: measles, acute or recently diagnosed chronic hepatitis B, bacterial meningitis, cervical cancer, and complicated seasonal influenza requiring hospitalization. * P < 0.05. ** P < 0.01. Table 1 Characteristics of the study population by geographic sample (panel of French GPs, weighted data)a % Rest of Franceb (ref.) (n = 1289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Response rates 93% 94% 90% 88% Stratification variables         Age—years (tertiles)             <50 (ref.) 29.9 34.9 33.1 28.1         50-58 32.0 31.9 30.7 33.0         >58 38.1 33.2 36.2 38.9**     Gender             Male (ref.) 70.4 67.5 68.9 70.5         Female 29.6 32.5 31.1 29.5     GP density in the municipality of practice (Min-Q1/Q1–Q3/Q3-Max)             <−19.3% of national average (ref.) 27.0 20.6 17.9 9.4         −19.3% to +17.7% of national average 47.3 63.5** 55.9* 64.8**         > +17.7% of national average 25.7 15.9** 26.2* 25.8**     2012 workload (Min-Q1/Q1-Q3/Q3-Max)             <3067 consultations/visits (ref.) 22.3 22.3 19.3 34.0         3067–6028 consultations/visits 51.4 52.6 48.7 48.4**         >6028 consultations/visits 26.3 25.1 32.0 17.6** Professional characteristics         Practice             Solo (ref.) 43.5 34.0 29.5 47.8         Group 56.5 66.0* 70.5** 52.2*     Occasional practice of complementary medicinec             No (ref.) 87.2 90.6 90.2 85.7         Yes 12.8 9.4* 9.8 14.3     CME completed last year on vaccination/infectious diseases >1 session             No (ref.) 56.0 59.5 51.4 63.4         Yes 44.0 40.5 48.6 36.6** Practice population characteristics         Proportion of patients aged under 16 21.0 24.0 20.6** 18.8**     Proportion of patients aged over 60 24.3 23.1** 28.5* 27.8** Experience related to vaccination         Any patients with at least one VPD in the past 5 yearsd             No (ref.) 10.9 14.0 8.0 5.8         Yes 89.1 86.0* 92.0 94.2** Any patients with a potential vaccine adverse effect             No (ref.) 83.8 79.8 83.3 81.4         Yes 16.2 20.2 16.7 18.6  % Rest of Franceb (ref.) (n = 1289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Response rates 93% 94% 90% 88% Stratification variables         Age—years (tertiles)             <50 (ref.) 29.9 34.9 33.1 28.1         50-58 32.0 31.9 30.7 33.0         >58 38.1 33.2 36.2 38.9**     Gender             Male (ref.) 70.4 67.5 68.9 70.5         Female 29.6 32.5 31.1 29.5     GP density in the municipality of practice (Min-Q1/Q1–Q3/Q3-Max)             <−19.3% of national average (ref.) 27.0 20.6 17.9 9.4         −19.3% to +17.7% of national average 47.3 63.5** 55.9* 64.8**         > +17.7% of national average 25.7 15.9** 26.2* 25.8**     2012 workload (Min-Q1/Q1-Q3/Q3-Max)             <3067 consultations/visits (ref.) 22.3 22.3 19.3 34.0         3067–6028 consultations/visits 51.4 52.6 48.7 48.4**         >6028 consultations/visits 26.3 25.1 32.0 17.6** Professional characteristics         Practice             Solo (ref.) 43.5 34.0 29.5 47.8         Group 56.5 66.0* 70.5** 52.2*     Occasional practice of complementary medicinec             No (ref.) 87.2 90.6 90.2 85.7         Yes 12.8 9.4* 9.8 14.3     CME completed last year on vaccination/infectious diseases >1 session             No (ref.) 56.0 59.5 51.4 63.4         Yes 44.0 40.5 48.6 36.6** Practice population characteristics         Proportion of patients aged under 16 21.0 24.0 20.6** 18.8**     Proportion of patients aged over 60 24.3 23.1** 28.5* 27.8** Experience related to vaccination         Any patients with at least one VPD in the past 5 yearsd             No (ref.) 10.9 14.0 8.0 5.8         Yes 89.1 86.0* 92.0 94.2** Any patients with a potential vaccine adverse effect             No (ref.) 83.8 79.8 83.3 81.4         Yes 16.2 20.2 16.7 18.6  VPD, vaccine-preventable disease; CME, continuing medical education. a Multivariable regressions on unweighted data, adjusted for stratification variables, were used to test for regional vs. ‘Rest of France’ (reference) differences. b Sample, not including any of the GPs in the three regional samples considered here. c Complementary medicine: e.g. homeopathy and/or acupuncture. d Five VPDs were mentioned in the questionnaire: measles, acute or recently diagnosed chronic hepatitis B, bacterial meningitis, cervical cancer, and complicated seasonal influenza requiring hospitalization. * P < 0.05. ** P < 0.01. Table 2 GPs’ attitudes regarding official vaccine recommendations and GPs’ guideline adherence indicator by sample (panel of French GPs, weighted data)a % Rest of Franceb (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Considers that a GP’s role is always to encourage vaccination         No (ref.) 9.4 11.3 6.3 13.7     Yes 90.6 88.7 93.7 86.3* Relies on her/his own judgment rather than the official vaccine recommendations         No (ref.) 65.7 71.4 74.6 67.6     Yes 34.3 28.6* 25.4* 32.4 Has consulted the new (2013) vaccination schedule         No (ref.) 2.2 1.1 1.7 2.1     Yes 97.8 98.9 98.3 97.9 Degree of guideline adherencec         Q1 (ref.) 24.6 17.6 23.6 39.2     Q2 24.6 22.4 29.7 25.1*     Q3 25.7 24.1 25.4 23.9**     Q4 25.1 35.9** 21.3 11.8**  % Rest of Franceb (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Considers that a GP’s role is always to encourage vaccination         No (ref.) 9.4 11.3 6.3 13.7     Yes 90.6 88.7 93.7 86.3* Relies on her/his own judgment rather than the official vaccine recommendations         No (ref.) 65.7 71.4 74.6 67.6     Yes 34.3 28.6* 25.4* 32.4 Has consulted the new (2013) vaccination schedule         No (ref.) 2.2 1.1 1.7 2.1     Yes 97.8 98.9 98.3 97.9 Degree of guideline adherencec         Q1 (ref.) 24.6 17.6 23.6 39.2     Q2 24.6 22.4 29.7 25.1*     Q3 25.7 24.1 25.4 23.9**     Q4 25.1 35.9** 21.3 11.8**  Q, quartile. a Multivariable regressions on unweighted data, adjusted for stratification variables, were used to test for regional vs. ‘Rest of France’ (reference) differences. b Sample, not including any of the GPs in the three regional samples considered here. c GPs in the fourth quartile are those who adhere most closely to guidelines. * P < 0.05. ** P < 0.01. Table 2 GPs’ attitudes regarding official vaccine recommendations and GPs’ guideline adherence indicator by sample (panel of French GPs, weighted data)a % Rest of Franceb (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Considers that a GP’s role is always to encourage vaccination         No (ref.) 9.4 11.3 6.3 13.7     Yes 90.6 88.7 93.7 86.3* Relies on her/his own judgment rather than the official vaccine recommendations         No (ref.) 65.7 71.4 74.6 67.6     Yes 34.3 28.6* 25.4* 32.4 Has consulted the new (2013) vaccination schedule         No (ref.) 2.2 1.1 1.7 2.1     Yes 97.8 98.9 98.3 97.9 Degree of guideline adherencec         Q1 (ref.) 24.6 17.6 23.6 39.2     Q2 24.6 22.4 29.7 25.1*     Q3 25.7 24.1 25.4 23.9**     Q4 25.1 35.9** 21.3 11.8**  % Rest of Franceb (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) Considers that a GP’s role is always to encourage vaccination         No (ref.) 9.4 11.3 6.3 13.7     Yes 90.6 88.7 93.7 86.3* Relies on her/his own judgment rather than the official vaccine recommendations         No (ref.) 65.7 71.4 74.6 67.6     Yes 34.3 28.6* 25.4* 32.4 Has consulted the new (2013) vaccination schedule         No (ref.) 2.2 1.1 1.7 2.1     Yes 97.8 98.9 98.3 97.9 Degree of guideline adherencec         Q1 (ref.) 24.6 17.6 23.6 39.2     Q2 24.6 22.4 29.7 25.1*     Q3 25.7 24.1 25.4 23.9**     Q4 25.1 35.9** 21.3 11.8**  Q, quartile. a Multivariable regressions on unweighted data, adjusted for stratification variables, were used to test for regional vs. ‘Rest of France’ (reference) differences. b Sample, not including any of the GPs in the three regional samples considered here. c GPs in the fourth quartile are those who adhere most closely to guidelines. * P < 0.05. ** P < 0.01. Next, the panel data were supplemented with (i) the density of paediatricians in each GP’s municipality of practice (DREES, Ministry of Health), and (ii) incidence rates for measles (by district) and seasonal influenza (by region) in 2009–13 in the general population (Santé Publique France and Réseau Sentinelles surveillance databases). Finally, we calculated, for each participant, a composite indicator of her/his degree of achievement of public health objectives other than vaccination, based on the P4P-ROSP programme as a proxy for their general adherence to guidelines (see Supplementary appendix S3); we used 2013 data about adherence to guidelines in prevention (excluding vaccination), chronic disease monitoring, and drug prescription. Statistical analysis All the samples were weighted for the four stratification variables to ensure their representativeness according to these variables. According to methods already published,15 we constructed a frequency score variable by summing each GPs’ responses on the Likert scales to the six items about their frequency of vaccine recommendations to patients (the higher the score, the higher the recommendation frequency, Supplementary appendix S4). We similarly constructed three other scores15 (Supplementary appendix S3): (i) ‘perceived vaccine risks’ summing their answers about their perceptions of the likelihood of links between various vaccines and potential severe adverse effects (six items, table 3); (ii) “doubts about vaccine utility”; and (iii) ‘trust in official and scientific sources’ (four items, table 3). Table 3 Description of the four scores constructed from the questionnaire responses (panel of French GPs, weighted data)  Rest of Francea (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) P-value (unweighted ANOVA) Mean SD Min Max Mean SD Min Max Mean SD Min Max Mean SD Min Max  Vaccine recommendation frequency scoreb 18.3 4.0 6.0 24.0 19.2 3.9 6.0 24.0 18.8 3.8 7.5 24.0 17.6 4.0 6.0 24.0 <0.0001 Score for perceptions of vaccine risksc 11.8 3.0 6.0 24.0 11.3 3.0 6.0 22.0 11.3 3.0 6.0 20.0 11.7 2.9 6.0 22.0 0.0032 Score for doubts of vaccine utilityd 3.73 1.6 2.0 8.0 3.7 1.5 2.0 8.0 3.3 1.4 2.0 8.0 4.0 1.7 2.0 8.0 <0.0001 Score for trust in official/scientific sourcese 12.8 2.1 4.0 16.0 12.8 1.9 5.0 16.0 13.2 2.1 4.0 16.0 12.5 2.0 5.0 16.0 <0.0001   Rest of Francea (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) P-value (unweighted ANOVA) Mean SD Min Max Mean SD Min Max Mean SD Min Max Mean SD Min Max  Vaccine recommendation frequency scoreb 18.3 4.0 6.0 24.0 19.2 3.9 6.0 24.0 18.8 3.8 7.5 24.0 17.6 4.0 6.0 24.0 <0.0001 Score for perceptions of vaccine risksc 11.8 3.0 6.0 24.0 11.3 3.0 6.0 22.0 11.3 3.0 6.0 20.0 11.7 2.9 6.0 22.0 0.0032 Score for doubts of vaccine utilityd 3.73 1.6 2.0 8.0 3.7 1.5 2.0 8.0 3.3 1.4 2.0 8.0 4.0 1.7 2.0 8.0 <0.0001 Score for trust in official/scientific sourcese 12.8 2.1 4.0 16.0 12.8 1.9 5.0 16.0 13.2 2.1 4.0 16.0 12.5 2.0 5.0 16.0 <0.0001  a Sample, not including any of the GPs in the three regional samples considered here. b The higher the score, the more frequently the GPs recommended vaccines to their patients. c The higher the score, the more likely the GPs were to consider vaccines causally associated with severe adverse effects. d The higher the score, the more doubtful the GPs about the utility of vaccines generally. e The higher the score, the more trust the GPs had in official/scientific sources to provide reliable information about vaccines. Table 3 Description of the four scores constructed from the questionnaire responses (panel of French GPs, weighted data)  Rest of Francea (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) P-value (unweighted ANOVA) Mean SD Min Max Mean SD Min Max Mean SD Min Max Mean SD Min Max  Vaccine recommendation frequency scoreb 18.3 4.0 6.0 24.0 19.2 3.9 6.0 24.0 18.8 3.8 7.5 24.0 17.6 4.0 6.0 24.0 <0.0001 Score for perceptions of vaccine risksc 11.8 3.0 6.0 24.0 11.3 3.0 6.0 22.0 11.3 3.0 6.0 20.0 11.7 2.9 6.0 22.0 0.0032 Score for doubts of vaccine utilityd 3.73 1.6 2.0 8.0 3.7 1.5 2.0 8.0 3.3 1.4 2.0 8.0 4.0 1.7 2.0 8.0 <0.0001 Score for trust in official/scientific sourcese 12.8 2.1 4.0 16.0 12.8 1.9 5.0 16.0 13.2 2.1 4.0 16.0 12.5 2.0 5.0 16.0 <0.0001   Rest of Francea (n = 1 289) Northwestern (n = 487) Central-western (n = 321) Southeastern (n = 489) P-value (unweighted ANOVA) Mean SD Min Max Mean SD Min Max Mean SD Min Max Mean SD Min Max  Vaccine recommendation frequency scoreb 18.3 4.0 6.0 24.0 19.2 3.9 6.0 24.0 18.8 3.8 7.5 24.0 17.6 4.0 6.0 24.0 <0.0001 Score for perceptions of vaccine risksc 11.8 3.0 6.0 24.0 11.3 3.0 6.0 22.0 11.3 3.0 6.0 20.0 11.7 2.9 6.0 22.0 0.0032 Score for doubts of vaccine utilityd 3.73 1.6 2.0 8.0 3.7 1.5 2.0 8.0 3.3 1.4 2.0 8.0 4.0 1.7 2.0 8.0 <0.0001 Score for trust in official/scientific sourcese 12.8 2.1 4.0 16.0 12.8 1.9 5.0 16.0 13.2 2.1 4.0 16.0 12.5 2.0 5.0 16.0 <0.0001  a Sample, not including any of the GPs in the three regional samples considered here. b The higher the score, the more frequently the GPs recommended vaccines to their patients. c The higher the score, the more likely the GPs were to consider vaccines causally associated with severe adverse effects. d The higher the score, the more doubtful the GPs about the utility of vaccines generally. e The higher the score, the more trust the GPs had in official/scientific sources to provide reliable information about vaccines. We used a five-step hierarchical linear regression to study the links between GPs’ self-reported vaccine recommendation frequency (dependent variable) and their practice location (variable of interest: inclusion in one of the three regional samples, or in the multiregional sample) and other individual and contextual factors (listed below). In step 1, the starting model (M1) included their practice location and was adjusted for the four stratification variables. Then, we successively added groups of variables to study the extent to which each group affected the association between the dependent variable and practice location: step 2 (M2), GPs’ professional characteristics and experience of VPD and adverse effects of vaccines (table 1); step 3 (M3), scores of perceived risks and doubts about vaccine utility and of trust in official and scientific sources; step 4 (M4), attitudes regarding official vaccine recommendations and indicator for guideline adherence; and step 5 (M5), contextual variables (paediatrician density, since these specialists share with GPs an important role in vaccine prescriptions in France, and incidence rate of seasonal influenza, MenC and measles). Finally, we further adjusted model M5 for the age structure (<16 and >60 years) of GPs’ patient lists (information available for 91% of participants). We used partial R-squares to quantify the partial contributions of each explanatory variable to the variation of the dependent variable.21 We computed the variance inflation factor (VIF) to test for multicollinearity and interpreted VIF values <5 as presenting no multicollinearity issues.22 All statistical analyses were performed with SAS version 9.4 (SAS Institute, Cary, NC) and based on two-sided P-values, with P ≤ 0.05 indicating statistical significance. Results Overall, 2823 GPs (41% of those contacted and eligible) agreed to participate in the panel, which entailed responding to five different cross-sectional surveys over 3 years. Among them, 2586 (92%) participated in the vaccination survey (details in table 1). Participants did not significantly differ from those who joined the panel but did not participate in the vaccination survey according to stratification variables, with the exception of workload in NW France (table 1). GPs in NW France practiced complementary medicine less often (9%) than those in the multiregional sample (13%, P < 0.05), and GPs in SE France reported fewer CME sessions (P < 0.001, table 1). The percentage of GPs reporting at least one patient with a VPD in the last 5 years was lowest in NW France (86 vs. 89% for the rest of France, P < 0.05, table 1). The percentage of GPs relying on their own judgment rather than on official vaccination recommendations was lower in NW (29%) and CW France (25%) than in the rest of France (34%, P < 0.05). GPs’ guideline adherence was lowest in SE France (table 2). GPs’ scores for vaccine recommendation frequency and trust in official sources were lowest in SE France, and those for perceived risks and doubts about vaccines utility were lowest in CW and NW France (table 3). Multiple linear regressions with the sample for the rest of France as the reference All variables included in the models were significantly associated with the recommendation score (dependent variable, table 4). Occasional practice of complementary medicine, the scores for perception of risks and doubts about vaccine utility, GPs’ perceived role in vaccination, and the indicator for guideline adherence contributed notably to the models, as indicated by their partial R-square (table 4) while none of the studied contextual factors contributed to the models. Table 4 Multiple linear regression coefficients and corresponding partial R-squares from the five-step modelling (dependent variable: score of vaccine recommendations, panel of French GPs and unweighted data)a  Coefficients (partial R²)  M1 M2 M3 M4 M5 Variable of interest     GPs’ practice location (ref. Rest of Franceb)              Northwestern 0.69** (0.02) 0.66** (0.02) 0.58** (0.02) 0.38* (0.02) 0.59* (0.02)         Central-western 0.16 (0.02) 0.04 (0.02) –0.30 (0.02) –0.34 (0.02) –0.21 (0.02)         Southeastern –0.55* (0.02) –0.49* (0.02) –0.33 (0.02) –0.29 (0.02) –0.30 (0.02) Professional characteristics     Practice (ref. Solo)              Group  0.48** (0.01) 0.34** (0.01) 0.27 (0.01) 0.28 (0.01)     Occasional practice of complementary medicinec (ref. No)              Yes  –2.76** (0.06) –1.46** (0.06) –1.26** (0.06) –1.24** (0.06)     CME completed last year > 1 session (ref. No)              Yes  0.74** (0.01) 0.43* (0.01) 0.46* (0.01) 0.46* (0.01)     Any patients with at least one VPD in the past 5 years (ref. No)              Yes  0.76** (0.00) 0.59** (0.00) 0.61** (0.01) 0.63** (0.01)     Any patients with a potential vaccine adverse effect (ref. No)              Yes  –0.96** (0.01) –0.22 (0.01) –0.07 (0.01) –0.08 (0.01) GPs’ perceptions of vaccines and information sources about vaccines     Score for perceptions of vaccine risks   –0.19** (0.07) –0.15** (0.03) –0.15** (0.04)     Score for doubts about vaccine utility   –0.76** (0.09) –0.66** (0.06) –0.65** (0.06)     Score for trust in official/scientific sources   0.22** (0.01) 0.16** (0.01) 0.16** (0.01) GPs’ attitudes regarding official vaccine recommendations and general adherence to guidelines     Considers that a GP's role is always to encourage vaccination (ref. No)              Yes    1.89** (0.06) 1.93** (0.06)     Relies on her/his own judgment rather than official vaccine recommendations (ref. No)              Yes    –0.49** (0.02) –0.49** (0.02)     Has consulted the new (2013) vaccination schedule (ref. No)              Yes    1.84** (0.01) 1.95** (0.01)     Degree of guidelines adherence (ref. Q1)              Q2    –0.26 (0.01) –0.23 (0.01)         Q3    0.14 (0.01) 0.17 (0.01)         Q4    0.67** (0.01) 0.74** (0.01) Environment     Density of paediatricians (ref. Q1)              Q2     0.20 (0.00)         Q3     0.26 (0.00)         Q4     0.61** (0.00)     Measles incidence rate, 2009–13, by districtd          Mean = 8.03 [0.71; 60.98]     –0.03** (0.00) N 2583 2583 2570 2097 2097 R² adjusted 0.08 0.17 0.34 0.35 0.35   Coefficients (partial R²)  M1 M2 M3 M4 M5 Variable of interest     GPs’ practice location (ref. Rest of Franceb)              Northwestern 0.69** (0.02) 0.66** (0.02) 0.58** (0.02) 0.38* (0.02) 0.59* (0.02)         Central-western 0.16 (0.02) 0.04 (0.02) –0.30 (0.02) –0.34 (0.02) –0.21 (0.02)         Southeastern –0.55* (0.02) –0.49* (0.02) –0.33 (0.02) –0.29 (0.02) –0.30 (0.02) Professional characteristics     Practice (ref. Solo)              Group  0.48** (0.01) 0.34** (0.01) 0.27 (0.01) 0.28 (0.01)     Occasional practice of complementary medicinec (ref. No)              Yes  –2.76** (0.06) –1.46** (0.06) –1.26** (0.06) –1.24** (0.06)     CME completed last year > 1 session (ref. No)              Yes  0.74** (0.01) 0.43* (0.01) 0.46* (0.01) 0.46* (0.01)     Any patients with at least one VPD in the past 5 years (ref. No)              Yes  0.76** (0.00) 0.59** (0.00) 0.61** (0.01) 0.63** (0.01)     Any patients with a potential vaccine adverse effect (ref. No)              Yes  –0.96** (0.01) –0.22 (0.01) –0.07 (0.01) –0.08 (0.01) GPs’ perceptions of vaccines and information sources about vaccines     Score for perceptions of vaccine risks   –0.19** (0.07) –0.15** (0.03) –0.15** (0.04)     Score for doubts about vaccine utility   –0.76** (0.09) –0.66** (0.06) –0.65** (0.06)     Score for trust in official/scientific sources   0.22** (0.01) 0.16** (0.01) 0.16** (0.01) GPs’ attitudes regarding official vaccine recommendations and general adherence to guidelines     Considers that a GP's role is always to encourage vaccination (ref. No)              Yes    1.89** (0.06) 1.93** (0.06)     Relies on her/his own judgment rather than official vaccine recommendations (ref. No)              Yes    –0.49** (0.02) –0.49** (0.02)     Has consulted the new (2013) vaccination schedule (ref. No)              Yes    1.84** (0.01) 1.95** (0.01)     Degree of guidelines adherence (ref. Q1)              Q2    –0.26 (0.01) –0.23 (0.01)         Q3    0.14 (0.01) 0.17 (0.01)         Q4    0.67** (0.01) 0.74** (0.01) Environment     Density of paediatricians (ref. Q1)              Q2     0.20 (0.00)         Q3     0.26 (0.00)         Q4     0.61** (0.00)     Measles incidence rate, 2009–13, by districtd          Mean = 8.03 [0.71; 60.98]     –0.03** (0.00) N 2583 2583 2570 2097 2097 R² adjusted 0.08 0.17 0.34 0.35 0.35  VPD, vaccine-preventable disease; CME, continuing medical education; Q, quartile. a Adjusted for age, gender, GP density in municipality of practice and 2012 workload (stratification variables). b Sample, not including any of the GPs in the three regional samples considered here. c Complementary medicine: e.g. homeopathy and/or acupuncture. d Other variable not shown (not significant): incidence rate of seasonal influenza. * P < 0.05. ** P < 0.01. Table 4 Multiple linear regression coefficients and corresponding partial R-squares from the five-step modelling (dependent variable: score of vaccine recommendations, panel of French GPs and unweighted data)a  Coefficients (partial R²)  M1 M2 M3 M4 M5 Variable of interest     GPs’ practice location (ref. Rest of Franceb)              Northwestern 0.69** (0.02) 0.66** (0.02) 0.58** (0.02) 0.38* (0.02) 0.59* (0.02)         Central-western 0.16 (0.02) 0.04 (0.02) –0.30 (0.02) –0.34 (0.02) –0.21 (0.02)         Southeastern –0.55* (0.02) –0.49* (0.02) –0.33 (0.02) –0.29 (0.02) –0.30 (0.02) Professional characteristics     Practice (ref. Solo)              Group  0.48** (0.01) 0.34** (0.01) 0.27 (0.01) 0.28 (0.01)     Occasional practice of complementary medicinec (ref. No)              Yes  –2.76** (0.06) –1.46** (0.06) –1.26** (0.06) –1.24** (0.06)     CME completed last year > 1 session (ref. No)              Yes  0.74** (0.01) 0.43* (0.01) 0.46* (0.01) 0.46* (0.01)     Any patients with at least one VPD in the past 5 years (ref. No)              Yes  0.76** (0.00) 0.59** (0.00) 0.61** (0.01) 0.63** (0.01)     Any patients with a potential vaccine adverse effect (ref. No)              Yes  –0.96** (0.01) –0.22 (0.01) –0.07 (0.01) –0.08 (0.01) GPs’ perceptions of vaccines and information sources about vaccines     Score for perceptions of vaccine risks   –0.19** (0.07) –0.15** (0.03) –0.15** (0.04)     Score for doubts about vaccine utility   –0.76** (0.09) –0.66** (0.06) –0.65** (0.06)     Score for trust in official/scientific sources   0.22** (0.01) 0.16** (0.01) 0.16** (0.01) GPs’ attitudes regarding official vaccine recommendations and general adherence to guidelines     Considers that a GP's role is always to encourage vaccination (ref. No)              Yes    1.89** (0.06) 1.93** (0.06)     Relies on her/his own judgment rather than official vaccine recommendations (ref. No)              Yes    –0.49** (0.02) –0.49** (0.02)     Has consulted the new (2013) vaccination schedule (ref. No)              Yes    1.84** (0.01) 1.95** (0.01)     Degree of guidelines adherence (ref. Q1)              Q2    –0.26 (0.01) –0.23 (0.01)         Q3    0.14 (0.01) 0.17 (0.01)         Q4    0.67** (0.01) 0.74** (0.01) Environment     Density of paediatricians (ref. Q1)              Q2     0.20 (0.00)         Q3     0.26 (0.00)         Q4     0.61** (0.00)     Measles incidence rate, 2009–13, by districtd          Mean = 8.03 [0.71; 60.98]     –0.03** (0.00) N 2583 2583 2570 2097 2097 R² adjusted 0.08 0.17 0.34 0.35 0.35   Coefficients (partial R²)  M1 M2 M3 M4 M5 Variable of interest     GPs’ practice location (ref. Rest of Franceb)              Northwestern 0.69** (0.02) 0.66** (0.02) 0.58** (0.02) 0.38* (0.02) 0.59* (0.02)         Central-western 0.16 (0.02) 0.04 (0.02) –0.30 (0.02) –0.34 (0.02) –0.21 (0.02)         Southeastern –0.55* (0.02) –0.49* (0.02) –0.33 (0.02) –0.29 (0.02) –0.30 (0.02) Professional characteristics     Practice (ref. Solo)              Group  0.48** (0.01) 0.34** (0.01) 0.27 (0.01) 0.28 (0.01)     Occasional practice of complementary medicinec (ref. No)              Yes  –2.76** (0.06) –1.46** (0.06) –1.26** (0.06) –1.24** (0.06)     CME completed last year > 1 session (ref. No)              Yes  0.74** (0.01) 0.43* (0.01) 0.46* (0.01) 0.46* (0.01)     Any patients with at least one VPD in the past 5 years (ref. No)              Yes  0.76** (0.00) 0.59** (0.00) 0.61** (0.01) 0.63** (0.01)     Any patients with a potential vaccine adverse effect (ref. No)              Yes  –0.96** (0.01) –0.22 (0.01) –0.07 (0.01) –0.08 (0.01) GPs’ perceptions of vaccines and information sources about vaccines     Score for perceptions of vaccine risks   –0.19** (0.07) –0.15** (0.03) –0.15** (0.04)     Score for doubts about vaccine utility   –0.76** (0.09) –0.66** (0.06) –0.65** (0.06)     Score for trust in official/scientific sources   0.22** (0.01) 0.16** (0.01) 0.16** (0.01) GPs’ attitudes regarding official vaccine recommendations and general adherence to guidelines     Considers that a GP's role is always to encourage vaccination (ref. No)              Yes    1.89** (0.06) 1.93** (0.06)     Relies on her/his own judgment rather than official vaccine recommendations (ref. No)              Yes    –0.49** (0.02) –0.49** (0.02)     Has consulted the new (2013) vaccination schedule (ref. No)              Yes    1.84** (0.01) 1.95** (0.01)     Degree of guidelines adherence (ref. Q1)              Q2    –0.26 (0.01) –0.23 (0.01)         Q3    0.14 (0.01) 0.17 (0.01)         Q4    0.67** (0.01) 0.74** (0.01) Environment     Density of paediatricians (ref. Q1)              Q2     0.20 (0.00)         Q3     0.26 (0.00)         Q4     0.61** (0.00)     Measles incidence rate, 2009–13, by districtd          Mean = 8.03 [0.71; 60.98]     –0.03** (0.00) N 2583 2583 2570 2097 2097 R² adjusted 0.08 0.17 0.34 0.35 0.35  VPD, vaccine-preventable disease; CME, continuing medical education; Q, quartile. a Adjusted for age, gender, GP density in municipality of practice and 2012 workload (stratification variables). b Sample, not including any of the GPs in the three regional samples considered here. c Complementary medicine: e.g. homeopathy and/or acupuncture. d Other variable not shown (not significant): incidence rate of seasonal influenza. * P < 0.05. ** P < 0.01. In M1 (table 4), GPs’ vaccine recommendation frequency was lowest in SE France and highest in NW France; M2 (adjusted for GPs’ professional characteristics) did not show a notable change in the coefficient associated with GPs’ practice location except for SE France. In M3, the coefficient for SE France was no longer significant after adjustment for the scores for perception of risks and doubts about vaccine utility. M4, adjusted for GPs’ attitudes regarding official vaccine recommendations and their adherence to guidelines, showed a reduction of the coefficient for NW France. In M5, further adjustment for environmental variables (table 4) did not notably change the results of M4 regarding GPs’ practice location. Multicollinearity presented no issues in the linear models. Results were similar after additional adjustment of M5 for the age structure of GPs’ patient lists. Discussion This study is the first to investigate and confirm geographical variations of self-reported vaccine recommendation practices by French GPs among large regional and multiregional samples. GPs’ vaccine recommendation frequency was highest in NW France and lowest in SE France. The distribution of GPs’ professional characteristics and of those of their practice environment did not completely explain these differences between regions. Differences in GPs’ perceptions of vaccine utility and risks and their trust in information sources explained the low vaccine recommendation frequency in SE France. GPs’ perceived role, attitudes regarding the official vaccination schedule and guideline adherence, together with their perceptions of vaccines, partially explained the high recommendation frequency score in NW France. The previously reported strong associations between GPs’ vaccine recommendation frequency and their perceptions of vaccine utility and risks15 suggest that VH exists among these professionals. Past and recent controversies in France concerning some vaccines considered in our study (especially hepatitis B and HPV vaccines) and insufficient or ineffective training of physicians in the area of vaccination may have fuelled GPs’ uncertainties about them.23 More generally, the more space for physicians’ personal beliefs, the more variations in medical practices.17,24 The regional variations in GPs’ vaccine recommendation frequency, perceptions of vaccine utility and risks and trust in sources of information about vaccines suggest that the prevalence of VH among GPs25 varies according to region in France. This important finding coincides with reports of similar gradients in both the prevalence of unfavourable attitudes towards vaccination in general in the French population and in VC for various vaccines among the general population and GPs.4,6,26 These observations should be taken into account in public health intervention strategies addressing VH, which should target not only parents but also health professionals and should be tailored according to regional characteristics. The French government’s 2017 decision27 to extend compulsory vaccination from 3 vaccines to 11 in children younger than 2 years makes an effort to restore trust in vaccines, more necessary now than ever.8 These regional variations in GPs’ vaccine recommendation frequency (table 3) may be related, at least in part, to regional practice styles and may result from variations in the content and modalities of initial training provided by regional schools of medicine, as observed, e.g in the USA28 and the UK.29 Despite the national curriculum for undergraduate medical training in France, medical schools have some flexibility in its implementation at the regional level.23 Moreover, the differences by region in GPs’ adherence to various public health guidelines (table 2) and the effects of this variable on the relation between practice location and vaccine recommendation frequency (M4) support the existence of regional medical practice styles and their potential influence on vaccination practices. The limited effect of CME on this relation (M2) may be linked to the diversification of CME availability in France, which no longer depends only on the regional medical schools.30 In any case, efforts at the level of these schools are necessary to improve GP training about vaccination. The absence of impact of contextual factors (paediatrician density and measles incidence) on regional variations in GP vaccine recommendation frequency is somewhat counterintuitive: local epidemiological and organizational settings have been shown to influence doctors’ daily practices.19 Our result is probably due to a lack of statistical power related to the low number of geographic units in our study. Nonetheless, the negative correlation between GPs’ vaccine recommendation frequency in 2014 and local measles incidence during the 2008–11 measles epidemic in France is interesting. Although this result must be interpreted cautiously because of the lack of a clear temporal correlation, it suggests that GPs tended to recommend the MMR vaccine less frequently in areas most affected by the measles epidemic, even after it ceased. This potential behavioural inertia might also reflect the persistence of pockets of vaccine reluctance among local populations despite repeated official warnings about this epidemic.5 These areas exist in SE France (in the Alps), where many well-educated and/or affluent people have settled, and numerous studies have reported that this population is more likely to delay or refuse vaccines than others.31 On the other hand, the fact that GPs recommended vaccines more frequently in areas with higher local density of paediatricians might result from a positive influence of informal collaborations between GPs and paediatricians caring for the same families. Some limitations of this study should be discussed. Given that GPs, by joining the panel, agreed to take part in five different surveys during a 30-month period, the commitment rate (41%) is fair, and higher than in other panels of primary-care physicians.32 This point does not, however, rule out the possibility of selection bias. In particular, panel participants and non-participants differed in workload in one of our samples. Nonetheless, weighting samples according to stratification variables should have limited any potential selection bias that might have arisen from those differences. Moreover, to limit potential selection bias that could have resulted from GPs’ particular attitudes about vaccination, this topic was not mentioned to GPs during the inclusion phase. Vaccine recommendation frequency was self-reported, which is a limitation that our study shares with previous publications on this subject: declaration or desirability biases cannot be excluded. However, questionnaire data appear to overestimate vaccination rates by <10%,33 and self-reported VC (e.g. for pandemic or seasonal influenza) in hospital healthcare workers has been shown to be a good proxy for recorded VC.34 Since this vaccination survey is cross-sectional and retrospective, no causal inferences can be drawn. Controlling for the proportion of GPs’ patients younger than 16 years and older than 60 years in their patient list did not affect our results. However, other patient characteristics, especially their hesitancy, might influence GPs’ vaccine recommendation behaviour.35 Because we had only three regional samples, we were unable to use multilevel analysis, which is usually used to analyse hierarchical data. Finally, we failed to explain all of the variability between the four samples (model M5). Unobserved characteristics of GPs and patients or contextual factors (socio-cultural, organizational, local policies etc.), alone or in combination, might explain these residual variations. In addition to the practice styles discussed earlier, cultural values might also explain differences in individuals’ (not only GPs’) propensity to adhere to official recommendations.36,37 Exploring these values is beyond the scope of this article, but more research is warranted to understand them better and to study their interactions with medical practices and especially VH.38 This remains one of the necessary steps to propose tailored interventions to help GPs cope with VH. Conclusion Our findings suggest that GPs’ vaccination practices may vary according to their practice region. Such unexplained variations probably exist in other countries, among physicians and other health care workers involved in vaccination activities. Further work is needed to improve our understanding of these regional differences in practices to help adapt intervention strategies to restore the confidence of GPs, patients and parents in vaccination in each regional context. This exemplifies the importance of collecting data on VC and VH at local and regional scales. Acknowledgements We thank Jo Ann Cahn for her help in editing the article. Funding Funding was received from DREES (convention n° 2101172809), INPES (convention no. 06/13 -DAS) and IReSP (convention ‘Prévention Primaire 2013’ no. PP-S1-14). We thank the National Agency for Research (ANR) and the National Institute of Health and Medical Research (Inserm) for their financial support. F.C. received a PhD grant from the (non-profit) Méditerranée Infection foundation (http://www.mediterranee-infection.com). Conflicts of interest: None declared. Key points The frequency of general practitioners (GPs’) vaccine recommendations varied across regions in France. These variations were partially explained by GPs’ perceptions of vaccine utility and risks. Regional differences in GPs’ general adherence to guidelines also played a role. References 1 Centers for Disease Control and Prevention . ChildVaxView | Interactive Child Vaccination Coverage.| CDC. Available at: https://www.cdc.gov/vaccines/imz-managers/coverage/childvaxview/data-reports/index.html (24 May 2018, date last accessed). 2 Centers for Disease Control and Prevention . 2015–16 Influenza season vaccination coverage report. https://www.cdc.gov/flu/fluvaxview/reportshtml/reporti1516/reporti/index.html (24 May 2018, date last accessed). 3 Screening & Immunisations Team NHS Digital. NHS Immunisation Statistics, England, 2015–16. 2016, https://files.digital.nhs.uk/publicationimport/pub21xxx/pub21651/nhs-imms-stat-eng-2015-16-rep.pdf (26 July 2018, date last accessed). 4 Santé Publique France . Couverture vaccinale-[Vaccine coverage] 2017. 2018;24. Available at: http://invs.santepubliquefrance.fr/Dossiers-thematiques/Maladies-infectieuses/Maladies-a-prevention-vaccinale/Couverture-vaccinale/Donnees (May, date last accessed). 5 Antona D , Lévy-Bruhl D , Baudon C , et al.  Measles elimination efforts and 2008-2011 outbreak, France . Emerg Infect Dis   2013 ; 19 : 357 – 64 . Google Scholar Crossref Search ADS PubMed   6 Pulcini C , Massin S , Launay O , Verger P . Factors associated with vaccination for hepatitis B, pertussis, seasonal and pandemic influenza among French general practitioners: a 2010 survey . Vaccine   2013 ; 31 : 3943 – 9 . Google Scholar Crossref Search ADS PubMed   7 The Strategic Advisory Group of Experts (SAGE) . Report of the SAGE working group on vaccine hesitancy  . 2014 , http://www.who.int/immunization/sage/meetings/2014/october/1_Report_WORKING_GROUP_vaccine_hesitancy_final.pdf?ua=1 (26 July 2018, date last accessed). 8 Rey D , Fressard L , Cortaredona S , et al.  Vaccine hesitancy in the French population in 2016, and its association with vaccine uptake and perceived vaccine risk–benefit balance . Eurosurveillance   2018 ; 23 :pii=17-00816. 9 Lieu TA , Ray GT , Klein NP , et al.  Geographic clusters in underimmunization and vaccine refusal . Pediatrics   2015 ; 135 : 280 – 9 . Google Scholar Crossref Search ADS PubMed   10 Carrel M , Bitterman P . Personal belief exemptions to vaccination in california: a spatial analysis . Pediatrics   2015 ; 136 : 80 – 8 . Google Scholar Crossref Search ADS PubMed   11 Henry KA , Stroup AM , Warner EL , Kepka D . Geographic factors and human papillomavirus (HPV) vaccination initiation among adolescent girls in the United States . Cancer Epidemiol Biomarkers Prev   2016 ; 25 : 309 – 17 . Google Scholar Crossref Search ADS PubMed   12 Slåttelid Schreiber SM , Juul KE , Dehlendorff C , Kjær SK . Socioeconomic predictors of human papillomavirus vaccination among girls in the danish childhood immunization program . J Adolesc Health   2015 ; 56 : 402 – 7 . Google Scholar Crossref Search ADS PubMed   13 Verger P , Flicoteaux R , Schwarzinger M , et al.  Pandemic influenza (A/H1N1) vaccine uptake among French private general practitioners: a cross sectional study in 2010 . PLoS One   2012 ; 7 : e41837 . Google Scholar Crossref Search ADS PubMed   14 Freed GL , Clark SJ , Cowan AE , Coleman MS . Primary care physician perspectives on providing adult vaccines . Vaccine   2011 ; 29 : 1850 – 4 . Google Scholar Crossref Search ADS PubMed   15 Verger P , Fressard L , Collange F , et al.  Vaccine hesitancy among general practitioners and its determinants during controversies: a national cross-sectional survey in France . EBioMedicine   2015 ; 2 : 891 – 7 . Google Scholar Crossref Search ADS PubMed   16 Paterson P , Meurice F , Stanberry LR , et al.  Vaccine hesitancy and healthcare providers . Vaccine   2016 ; 34 : 6700 – 6 Google Scholar Crossref Search ADS PubMed   17 de Jong JD . Explaining medical practice variation: social organization and institutional mechanisms. PhD diss., Utrecht University, 2008 . 18 Mousquès J , Renaud T , Scemama O . Is the “practice style” hypothesis relevant for general practitioners? An analysis of antibiotics prescription for acute rhinopharyngitis . Soc Sci Med   2010 ; 70 : 1176 – 84 . Google Scholar Crossref Search ADS PubMed   19 Westert GP , Groenewegen PP . Medical practice variations: changing the theoretical approach . Scand J Public Health   1999 ; 27 : 173 – 80 . Google Scholar Crossref Search ADS PubMed   20 Le Maréchal M , Collange F , Fressard L , et al.  Design of a national and regional survey among French general practitioners and method of the first wave of survey dedicated to vaccination . Med Mal Infect   2015 ; 45 : 403 . Google Scholar Crossref Search ADS PubMed   21 Collange F , Fressard L , Pulcini C , et al.  General practitioners’ attitudes and behaviors toward HPV vaccination: a French national survey . Vaccine   2016 ; 34 : 762 . Google Scholar Crossref Search ADS PubMed   22 Rogerson P . Statistical Methods for Geography  . London : SAGE , 2001 . 23 Kernéis S , Jacquet C , Bannay A , et al.  Vaccine Education of Medical Students: a Nationwide Cross-sectional Survey . Am J Prev Med   2017 ; 53 : e97 – e104 . Google Scholar Crossref Search ADS PubMed   24 Wennberg JE , Barnes BA , Zubkoff M . Professional uncertainty and the problem of supplier-induced demand . Soc Sci Med   1982 ; 16 : 811 – 24 . Google Scholar Crossref Search ADS PubMed   25 Verger P , Collange F , Fressard L , et al.  Prevalence and correlates of vaccine hesitancy among general practitioners: a cross-sectional telephone survey in France, April to July 2014 . Eurosurveillance   2016 ; 21 : 30406 . Google Scholar Crossref Search ADS PubMed   26 Gautier A , Chemlal K , Jestin C . GBS. Adhésion à la vaccination en France: résulats du Baromètre santé 2016 [Vaccination adherence in France: results of the 2016 “Baromètre santé”] . Bull Epidemiol Hebd. 2017; Hors série   2016 ; 21 – 7 . 27 Ministère de la santé LOI n°2017-1836 du 30 décembre 2017 - art. 49 (V) | Code de la santé publique. | Legifrance 2018. Available at: https://www.legifrance.gouv.fr/affichCode.do? idSectionTA=LEGISCTA000006171171&cidTexte=LEGITEXT000006072665 (24 May 2018, date last accessed). 28 Wennberg JE , Fisher E , Goodman D , Skinner J . Tracking the Care of Patients with Severe Chronic Illness: The Dartmouth Atlas of Health Care  . Hanover, NH : The Dartmouth Atlas Project , 2008 . 29 Mcmanus I , Elder AT , De Champlain A , et al.  Graduates of different UK medical schools show substantial differences in performance on MRCP(UK) Part 1, Part 2 and PACES examinations . BMC Med   2008 ; 6 : 5 . Google Scholar Crossref Search ADS PubMed   30 Garattini L , Gritti S , De Compadri P , Casadei G . Continuing Medical Education in six European countries: a comparative analysis . Health Policy   2010 ; 94 : 246 – 54 . Google Scholar Crossref Search ADS PubMed   31 Bocquier A , Ward J , Raude J , et al.  Socioeconomic differences in childhood vaccination in developed countries: a systematic review of quantitative studies . Expert Rev Vaccines   2017 ; 0 : 1 – 12 . 32 Joyce CM , Scott A , Jeon S-H , et al.  The “medicine in Australia: balancing employment and life (MABEL)” longitudinal survey–protocol and baseline data for a prospective cohort study of Australian doctors’ workforce participation . BMC Health Serv Res   2010 ; 10 : 50 . Google Scholar Crossref Search ADS PubMed   33 Brien S , Kwong JC , Buckeridge DL . The determinants of 2009 pandemic A/H1N1 influenza vaccination: a systematic review . Vaccine   2012 ; 30 : 1255 – 64 . Google Scholar Crossref Search ADS PubMed   34 Llupià A , García-Basteiro AL , Mena G , et al.  Vaccination behaviour influences self-report of influenza vaccination status: a cross-sectional study among health care workers . PLoS One   2012 ; 7 : e39496 . Google Scholar Crossref Search ADS PubMed   35 Larson H , Karafillakis E . European Centre for Disease Prevention and Control. In: Vaccine hesitancy amongst healthcare workers and their patients in Europe : a qualitative study  , 2015. Available at: https://ecdc.europa.eu/sites/portal/files/media/en/publications/Publications/vaccine-hesitancy-among-healthcare-workers.pdf (26 July 2018, date last accessed). 36 Blume S . Anti-vaccination movements and their interpretations . Soc Sci Med   2006 ; 62 : 628 – 42 . Google Scholar Crossref Search ADS PubMed   37 Streefland P , Chowdhury AM , Ramos-Jimenez P . Patterns of vaccination acceptance . Soc Sci Med   1999 ; 49 : 1705 – 16 . Google Scholar Crossref Search ADS PubMed   38 Ward J , Raude J . Understanding influenza vaccination behaviors: a comprehensive sociocultural framework . Expert Rev Vaccines   2014 ; 13 : 17 – 29 . Google Scholar Crossref Search ADS PubMed   © The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)


This is a preview of a remote PDF: https://academic.oup.com/eurpub/article-pdf/29/1/2/27596989/cky146.pdf

Collange, Fanny, Zaytseva, Anna, Pulcini, Céline, Bocquier, Aurélie, Verger, Pierre. Unexplained variations in general practitioners’ perceptions and practices regarding vaccination in France, European Journal of Public Health, 2019, 2-8, DOI: 10.1093/eurpub/cky146