Knowledge, attitudes, beliefs and behaviors of general practitioners/family physicians toward their own vaccination: A systematic review

ABSTRACT

Context: General practitioners and family physicians (GP/FPs) play a key role in the vaccination of the public in many countries and serve as role models for their patients through their own health behaviors. Objectives and Methods: a) To search for and document recommended/mandated vaccines for GP/FPs in high-income countries; b) To systematically search and review the literature on these physicians' knowledge, attitudes, beliefs, and behaviors (KABB) toward their own vaccination with the recommended/mandated vaccines and the factors determining it. Results: a) The 14 countries included recommended or mandated as many as 12 vaccines; b) The systematic review identified 11 studies published in the last 10 y. All considered seasonal influenza vaccination but differed in the variables investigated. Discussion/Conclusions: This review highlights the need for further studies on this topic, including qualitative and interventional studies (based on behavior change theories). These should cover occupational vaccines and determinants known to be associated with vaccine hesitancy.

Introduction

General practitioners and family physicians (GP/FPs) – medical practitioners who provide personal, primary, and continuing care to individuals, families, and a practice population, irrespective of age, sex, and illness ¹ – are, like other healthcare professionals, required or advised to have certain immunizations to protect themselves and their patients from some vaccine-preventable diseases.^2–5 Two recent reviews showed that recommendations about occupational vaccines for health care workers (HCWs) in acute care facilities vary considerably from one country to another.^2,3 Unfortunately nothing similar to the tool developed by the European Centre for Disease Prevention and Control (ECDC) to present and compare European vaccines schedule for the general populations exists for occupational vaccines for HCWs.⁶

In high-income countries, these physicians are usually patients' first contact in the healthcare system (either in private practices or in local primary health centers). Although their role may differ from one country to another,⁷ they are key actors in the implementation of vaccination policies in most of these countries.^8–11 In France, for example, they administer nearly 90% of vaccines,¹² and around 65% in New South Wales, Australia.¹³ GP/FPs not only deliver vaccines by prescribing and administering them, they also must recommend the appropriate vaccines to their patients according to national guidelines, explain vaccination and vaccines to them, and answer their questions on the subject.

Their recommendations play an influential role in their patients' vaccination behavior. Their patients perceive them as a reliable source of information about vaccines,^14–16 and their consultation can also be a key factor in patients' decision to be vaccinated or not.^17–21 Moreover, their recommendations to patients about vaccinations are associated with their vaccination behaviors and attitudes for themselves.^22–26 Their own vaccination might be important as a proxy for their potential “vaccine hesitancy,” – a term that describes vaccination reluctance and was first studied in the general population^27,28 and, more recently, also among GP/FPs.^28,29 For all these reasons, identifying factors associated with their own vaccinations is important for improving vaccination coverage rates in the general population.

No review has yet been published specifically on GP/FPs' knowledge, attitudes, beliefs and behaviors (KABB) toward their own vaccination and and the factors determining it.

Our review focused on high-income countries (where the economic, organizational, and health environments of physicians are relatively comparable).³⁰ Our objectives were: a) to search for and document recommended/mandatory vaccines for GP/FPs; and b) to search for and review articles that address GP/FPs' KABB toward their own vaccination and the factors associated with their use of these vaccines for themselves.

Review results

Recommended/mandatory vaccines for GP/FPs in 14 countries

Table 1 presents the recommended occupational vaccines for GP/FPs in 14 high-income countries, based on our consultation of national guidelines. As many as 12 vaccines were recommended in some countries, always including influenza and hepatitis B vaccines.

Table 1.

National policies for occupational vaccination of GP/FPs (not working in acute care facilities) by vaccine and by country.

Country	Influenza	Mumps	Measles	Rubella	Varicella	Hep A	Hep B	Diphteria	Tetanus	Pertussis	Polio	BCG
Austria 59	R	R	R	R		R	R	R	R	R	R
Belgium 60,61	R	R	R	R			M	R	R			spRa
France 62	R		R		R		R*	*	*	R	*	*
Germany 63	R	R	R				R			R
Italy 64	R	R	R	R	R		R					spRb
Ireland 65	R	spRc	spRc	spRc	R	spRd	R			R		spRb
The Netherlands 66	R				R	R	R	R	R	R	R
Spain 67	R	R	R	R	R		R	R	R	R
Sweden 68	R						R					R
Switzerland 69	R	R	R	R	R	spRb	R	R	R	R	R
United Kingdom 70	R	R	R	R	R		R	R	R		R	R
Australia 71	R	R	R	R	R	spRe	R			R		spRf
Canada 72	R	R	R	R	R		R	R	R	R	R	spRb
USA 73	R	R	R	R	R		R	R	R	R

R: recommended

spR: special recommendations

^arecommended vaccination if frequent contacts with at-risk populations

^bonly in case of high-risk exposure

^cfor GP/FPs born in Ireland since 1978 or born outside Ireland and at risk of exposure

^din case of local outbreak

^efor GP/FPs who work in remote Indigenous communities or with Indigenous children

^ffor GP/FPs who may be at high risk of exposure to multidrug-resistant cases of tuberculosis (depends on state or territory guidelines)

^*mandatory for medical students in France.

Characteristics of the studies included in the systematic literature review

The systematic literature search initially identified 956 articles. A screening of titles and abstracts selected 21 studies for full text review (Fig. 1). Of these, 10 were excluded because they met at least one of our exclusion criteria. Table 2 presents the characteristics of the 11 studies finally included in the review. Screening the reference lists of these articles identified no additional study.

Figure 1.

Flow diagram of study selection procedure for the systematic literature review of knowledge, attitudes, beliefs, and behaviors of GP/FPs toward their own vaccination with recommended/mandatory vaccines.

Table 2.

Characteristics of the 11 studies included in the systematic literature review.

Year of publication and authors	Study location	Study time period	Study design	GP/FP population (n)	Questionnaire survey participation rate (%)	Vaccine(s)a
2015, Massin et al.37	France	2010 and 2012	Cross-sectional	1136	79.4	seasonal influenza
2014, Joseph et al.20	Gironde (France)	April-May 2012	Cross-sectional	225	48.3	seasonal influenza
2013, Castilla et al.38	7 Spanish regions	March-April 2012	Cross-sectional	923	36.2b	seasonal influenza
2013, Pulcini et al.39	France	June-December 2010	Cross-sectional	1431	36.8	hepatitis B, pertussis, seasonal influenza
2013, Paya et al.31	Loire (France)	Spring 2010	Cross-sectional	288	64.0	DT-Polio, pertussis, seasonal influenza, hepatitis B, BCG
2013, Socan et al.34	Slovenia	February 2010	Cross-sectional	289	24.0c	seasonal influenza
2011, Ward et al.32	New South Wales (Australia)	February-April 2009	Cross-sectional	79	30.0	seasonal influenza
2010, Opstelten et al.35	Netherlands	February 2010	Cross-sectional	670	83.0	seasonal influenza
2008, Opstelten et al.36	Netherlands	February 2008	Cross-sectional	698	96.0	seasonal influenza
2006, Cowan et al.40	United States	May 2004	Cross-sectional	195	39.4	seasonal influenza
2006, Semaille et al.33	the Walloon & Brussels-Capital regions of Belgium	2004-2005	Cross-sectional	730	60.8	seasonal influenza, diphtheria, tetanus, poliomyelitis, pertussis, hepatitis B, rubella, BCG, hepatitis A

Abbreviation: GP/FPs – General practitioners and family physicians.

^aOther than the pandemic 2009 A/H₁N₁ influenza vaccine

^bOverall participation rate for healthcare workers including GP/FPs, pediatricians, and nurses

^cOverall participation rate for healthcare workers including GP/FPs and six other medical specialties.

Seven countries were represented. Four of the 11 studies were conducted in France and 2 in non-European countries (Australia and the United States of America).

All studies identified by our literature review were cross-sectional studies that used questionnaires to assess the GP/FPs' self-reported knowledge, attitudes, beliefs and/or behaviors toward their own vaccination. All studies included seasonal influenza vaccination. Three of them also considered other vaccines.

Table 3 reports the critical appraisal of the risk of bias of these studies. Scores ranged from 2 to 7 (for a maximum score of 10, corresponding to the lowest risk of bias).

Table 3.

Results of the critical appraisal of the risk of bias of the studies^a included in this review.

	Selection (maximum 3 asterisks)	Comparability (maximum 2 asterisks)	Outcome (maximum 3 asterisks)	Scoreb
	Representativeness of the sample	Sample size	Non-respondents	Ascertainment of exposure	Based on design and analysis	Assessment of outcome	Statistical test
Study	(Ø or *)	(Ø or *)	(Ø or *)	( Ø, * or **)	(* or **)	( Ø,* or **)	( Ø or *)
2015, Massin et al. 37	*			*	**	*	*	6
2014, Joseph et al.20	*			*		*		3
2013, Castilla et al.38	*			*	**	*	*	6
2013, Pulcini et al.39	*			*	**	*	*	6
2013, Paya et al.31	*	*	*		**	*	*	7
2013, Socan et al.34	*			*	*	*		4
2011, Ward et al.32	*			*		*		3
2010, Opstelten et al.35	*			*	*	*		4
2008, Opstelten et al.36	*		*	*	**	*	*	7
2006, Cowan et al.40	*			*	**	*	*	6
2006, Semaille et al.33	*					*		2

^aThe studies included were all cross-sectional studies

^bThe maximum score is 10, and corresponds to the lowest risk of bias.

Legend: allocation of asterisks as stated by the rating scale (see additional file 4).

Knowledge, attitudes and beliefs of GP/FPs toward their own vaccination

All the studies reported GP/FPs' vaccination behaviors, but less than half (3/11) also investigated their knowledge, attitudes, and beliefs (Table 4). Studies that provided results about knowledge showed that most of these doctors (65 to 74%) reported that they were aware of occupational recommendations.^31,32 Some, however, had misconceptions about vaccines: for example, in an Australian study,³² 24% incorrectly believed that the injectable seasonal influenza vaccine contains live viruses and may cause the disease. A Belgian study ³³ showed a substantial fraction of GP/FPs to be unfamiliar with the duration of protection of occupational vaccines: 29% believed that the protection from the vaccine against pertussis was lifelong.

Table 4.

General practitioners' and family physicians' knowledge, attitudes, and beliefs toward their own vaccination (n = 3 studies).

Year of publication and authors	Location	Vaccine(s)a	Knowledge	Attitudes/beliefs
				-Attitude toward booster vaccines: 90% reported they took care of their own booster vaccines, with memory aids (74% with a vaccination record, 9% with a computer alert and 6% because they were reminded by their monitoring of their children's vaccination )
2013, Paya et al. 31	France	Diphtheria-tetanus-poliomyelitis, pertussis, 2009-10 seasonal influenza, hepatitis B, BCG	-Awareness of the French High Council of Public Health recommendations regarding occupational vaccines: 76% said they were aware of them	-Adherence to occupational vaccines (recommended by the French High Council of Public Health): 59% reported they followed them, 25% that they did not follow them and 16% did not know if they follow them
			-Knowledge of influenza vaccine characteristics: 24% thought that injectable influenza vaccine contains live viruses that can cause influenza	-Opinions about risk-benefit balance of the vaccine: 96% agreed that the influenza vaccine was safe, 92% agreed that yearly vaccination was effective, 92% reported that being vaccinated was important to protect their patients
2011, Ward et al. 32	Australia	2007 to 2009 seasonal influenza	-Awareness of recommendations for GP/FPs to receive flu immunization: 9% were not aware of them and 10% were not sure if they were aware of any recommendation	-Potential barriers to vaccination: 10% reported that this vaccination would not benefit them because they do not get the flu, having to pay for the vaccine was a barrier for 30% of them
2006, Semaille et al. 33	Belgium	2004-05 seasonal influenza, diphtheria, tetanus, poliomyelitis, pertussis, hepatitis B, rubella, BCG, hepatitis A	-Knowledge of duration of protection: 29%, 37%, 50% and 60% reported that protection from, respectively, pertussis, poliomyelitis vaccine, hepatitis B and rubella vaccine was lifelong; 32% reported that tetanus vaccine protected for more than 10 y	—

Reasons for (not) being vaccinated

Figure 2 depicts the reasons GP/FPs report that they are (not) vaccinated against seasonal influenza (5 studies); ^20,33–36 one study of other vaccines has also examined these reasons.³³

Figure 2.

Percentages of GPs/FPs self-reported reasons for being or not being vaccinated against seasonal influenza (n = 5 studies).*One side effect was specified: influenza-like illness post-vaccination.

The two most commonly reported motives for GP/FPs' own vaccination were to protect themselves (55 to 88.5% according to the study ^20,35,36) and to protect their patients (36 to 84.7% ^20,34,36). The most common reasons for non-vaccination were the belief that frequent exposure already protected them (28 to 39% ^33,35,36), concerns about side effects (6 to 25% ^20,33,35,36), forgetfulness (9 to 28.6% ^33,35,36), and doubts about the vaccine's efficacy (2.5 to 16% ^20,35,36).

Factors associated with GP/FPs' own vaccination

Table 5 presents the factors found to be significantly and independently associated with GP/FPs' own vaccination in 7 studies.^31,35,40 Lower self-reported rates of influenza vaccination (Fig. 3) were consistently associated with female gender and the practice of a complementary or alternative medicine (CAM), such as homeopathy or acupuncture.^36–39 Conversely, increasing age (except in Paya et al.) and workload were associated with higher vaccination rates.^36,38 Castilla et al. asked physicians about their concerns about acquiring influenza from patients and about transmitting it to them: both factors were associated with higher vaccination rates.³⁸ Two French studies showed that preventive behaviors (besides vaccination) for oneself (having one's own physician rather than mainly treating oneself, testing one's own lipid profile) or being risk-averse (evaluated by 3 questions on GP/FPs' individual risks attitudes in daily life, personal finances, and their own health) were also associated with increased vaccination rates.^37,39 Finally, an American study found that access to vaccine and the presence of an incentive policy in the workplace were positively associated with vaccination rates.⁴⁰ For the other vaccines that were studied, increasing age was associated with decreased self-reported vaccination rates for hepatitis B and pertussis vaccines,³⁹ as well as for DT-Polio, seasonal influenza, measles, and hepatitis B vaccines (these four vaccines were grouped as one variable in the study).³¹

Table 5.

Factors independently and significantly associated with GP/FPs' own vaccination for occupational vaccines in multivariate analyses (n = 7 studies).

						Factors significantly associated with higher GP/FP self-reported own vaccination rates
Study	Location	Vaccinea	VCb (%)	Dependent variable	Association measures	Negative associations	Positive associations
2015, Massin et al. 37	France	2009-10 seasonal influenza	77	GP/FP self-reported vaccination	Marginal effects	-Female gender −7.9% PP < 0.01	-High workload 15.4% P P P < 0.05
2013, Castilla et al.38	Spain	2008-09 to 2011-12 seasonal influenza	63-63-58-54	GP/FP self- reported vaccination for each season	OR	-Female gender 0.72 [0.54–0.97] -Concerns about transmitting the disease to family 0.60 [0.41-0.88]	-Increasing age 2.79 [1.65–4.70] -Having a major chronic condition 2.12 [1.39-3.25] or a household member with one 2.57 [1.84-3.59] -Confidence in vaccine effectiveness 3.03 [1.94-4.73] -Concerns about becoming ill 4.35 [3.19-5.93]
2013, Pulcini et al.39	France	Hepatitis B	73	GP/FP self-reported vaccination	RR	-Increasing age 0.89 [0.82-0.98] -Using Internet at the office once a week or less 0.92 [0.85–.0.99]	-Other preventive behaviors: having one's own physician rather than mainly treating oneself 1.19 [1.04–1.36] and lipid profile 1.20 [1.08–1.33]-Vaccination against pertussis 1.32 [1.20–1.44]or seasonal influenza 1.20 [1.07–1.33]
		Pertussis	64			- Increasing age 0.76 [0.68–0.86]	-High workload 1.19 [1.05–1.35] -Vaccination against hepatitis B 1.49 [1.30–1.70] or A/H1N1 pandemic influenza 1.12 [1.01–1.23]
		2008-09 seasonal influenza	78			-Female gender 0.89, [0.83–0.96] -Occasional practice of CAM 0.86, [0.76–0.96]	-High workload 1.11 [1.01–1.22] -Participation in CME sessions 1.08 [1.02–1.15] -Vaccination against hepatitis B 1.14 [1.05–1.23] or pertussis 1.08 [1.00–1.15] or A/H1N1 pandemic influenza 1.63 [1.49–1.79]
2013, Paya et al.31	France	DT-Polio	81	GP/FP self-reported vaccination for all 4 vaccinations	OR		-Increasing age 0.95 [0.92–0.98]
		2009-10 seasonal influenza	73
		Measles	64c
		Hepatitis B	87
2006, Cowan et al.40	United-States	2003-04 seasonal influenza	87	GP/FP self-reported vaccination	OR		-Access to vaccination 2.57 [1.45–4.57] -Workplace's incentive policy 2.24 [1.30–3.85] -Positive perception of the benefit-risk balance of vaccination 1.94 [1.02–3.68] -Feeling professionally responsible for being vaccinated 3.97 [1.86–8.46]
Study	Location	Vaccinea	VCb (%)	Dependent variable	Association measures	Factors significantly associated with lower GP/FP self-reported vaccination rates
						Negative associations	Positive associations
2010, Opstelten et al.35	Netherlands	2009-10 seasonal influenza	36	GP/FP self-reported vaccination	RR	0	0
2008, Opstelten et al.36	Netherlands	2007–08 seasonal influenza	84	GP/FP self-reported vaccination	RR	-Female gender 1.20 [1.07–1.34]	-Young age 1.20 [1.05–1.37]

Abbreviations: GP/FPs: General practitioners and family physicians; VC: vaccine coverage; CAM: Complementary and alternative medicine (e.g., homeopathy or acupuncture); OR: Odds Ratio; RR: Risk Ratio; CME: Continuing medical education; DT-Polio – Diphtheria-Tetanus-Poliomyelitis.

^aOther than pandemic 2009 A/H1N1 influenza vaccine.

^bPercentage of GP/FPs who reported being vaccinated.

^cPercentage of GP/FPs who reported being vaccinated or who had measles.

Figure 3.

Factors significantly and independently associated with GPs' own vaccination against seasonal influenza in multivariate analyses (n=6 studies).

Discussion

This systematic review summarizes the last decade's findings on GP/FPs' own vaccinations. We identified very few studies on this topic (n = 11), which we found surprising since GP/FPs are key actors in the vaccination policies in most high-income countries. Moreover, nearly all the studies (9/11) were conducted in Europe, and very few (3/11) assessed occupational vaccines other than the seasonal influenza vaccine (which was included in all studies), even though up to 12 vaccines are recommended for these professionals (Table 1). Finally, all studies were quantitative and used questionnaires and self-reported vaccination coverage rates.

We note that we had decided from the start to exclude studies that presented data for GP/FPs combined with those of other HCWs, because findings that their own vaccination rates are usually higher than those of other HCWs ^2,39,41 suggest that they may behave differently.^25,42 Our literature search identified and excluded two such studies: one qualitative study (focus group) ⁴³ and one interventional.⁴⁴

The studies we included differed widely in the variables that they investigated and in the association measures they used. Therefore, their results cannot be combined and must be evaluated individually.⁴⁵

This review has several strengths: 1) it documents recommended/mandatory vaccines for GP/FPs; 2) it specifically addresses their KABB toward their own vaccination and the factors determining it according to a standardized and validated process (PRISMA guidelines) and assessed the quality of selected studies; 3) it focused on high-income countries, thereby limiting heterogeneity in organizational and health contexts. Nevertheless, the results should be interpreted in the light of several limitations: 1) our inventory of recommended/mandatory vaccines for GP/FPs is not exhaustive and is limited by availability and comprehension of this type of information; 2) we cannot exclude a publication bias as we only used the Medline database for our literature search.

Assessment of the risk of bias and the quality of the studies (with the Newcastle-Ottawa scale) highlighted several methodological deficiencies. Almost all the studies used an acceptable sampling method, but most did not report a satisfactory response rate and did not test comparability between respondents and non-respondents. This may lead to potential selection bias. Most studies also failed to use a validated instrument to measure these physicians' KABB toward their own vaccination.

Previous reviews and studies of influenza vaccination among HCWs have shown that the main categories of vaccination determinants are: personal (e.g., age, gender), perceived benefits and risks of the disease/vaccination for oneself and patients/family, and factors related to convenience (e.g., sufficient free time, accessibility of vaccine); ^46–48 some but not all of these variables were also studied among GP/FPs in the studies we included in our review. Increasing age and male gender were consistently associated with higher coverage rates in all reviews published in the literature,^2,47,48 consistent with our findings. It is difficult to compare our results for the other factors to the published data targeting all HCWs, due to the limited number of studies included in our review. Reasons for (not) being vaccinated did not differ in any major way between the GP/FPs in our review and physician populations mixing several specialties working in primary care or even other HCWs in general.^2,46,48–50 The 2 main reasons for being vaccinated were to protect oneself and to protect one's patients, with self-protection the leading factor. Feeling already protected and fear of side effects were the main reasons for not being vaccinated.

Our review identified several factors that have been associated with negative opinions toward vaccination or less inclination to be vaccinated and that could be considered to contribute to vaccine hesitancy according to the literature: practicing alternative or complementary medicine and a negative perception of the vaccine benefit-risk balance.^28,51 Moreover, one of the studies showed that GP/FPs' own vaccination depended on the context,³⁸ and such context-dependence is – according to the SAGE group ^51,52 – an important aspect of vaccine hesitancy. Castilla et al. also reported that the pandemic 2009 A/H₁N₁ influenza outbreak might have caused GP/FPs' own vaccination rates for seasonal influenza vaccination to decline.³⁸ However, some factors related to vaccine hesitancy in the literature, such as personal experience with illness or confidence toward health authorities, were not studied among GP/FPs.²⁹

Improving understanding of these physicians' KABB toward their own vaccination and the factors determining it is essential because they not only act as role models for their patients, but because their behaviors for themselves can be determinant in their recommendations to their patients. This has been shown in the case of vaccination against influenza (positive association between their own and their patients' vaccination against seasonal influenza,²⁰ personal acceptance of vaccination against pandemic influenza predictive of GP/FPs' recommendation of this vaccine to their patients²²), but also, for example, for prescription of psychotherapy (GP/FPs who have undergone psychotherapy are more likely to recommend it⁵³) or help in quitting smoking (GP/FPs who are current smokers are less likely to provide advice on quitting⁵⁴).

Conclusions

Studies assessing the determinants of GP/FPs' own vaccination are urgently needed, especially to the extent that their own non-vaccination may be considered a marker of vaccine hesitancy and is associated with the vaccination recommendations they make to patients. All types of studies, including qualitative and interventional studies (based on behavior change theories), are necessary, and all occupational vaccines should be studied, since behaviors can vary substantially from one vaccine to another.³⁹ Determinants known to be associated with vaccine hesitancy and those that may be modified by interventions should also be included,²⁹ because dealing with GP/FPs' vaccine hesitancy is essential for improving vaccine coverage among patients.

Methods

Search for and document GP/FPs recommended/mandatory vaccines in high-income countries

The high-income countries we selected were those that posted the information on vaccines mandatory or recommended for GP/FPs not working in acute care facilities on the websites of the national health authorities.

Search for and review articles that address GP/FPs' KABB toward their own vaccination with the recommended/mandated vaccines and its determinants

Data sources

We performed a systematic review focusing on GP/FPs' knowledge, attitudes, beliefs, and behaviors toward occupational vaccines and on the determinants of these vaccines' coverage rates among them. We used the recommendations for Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA)⁵⁵ as a guide to conduct the review and report the results (additional file 1). According to the PICOS approach included in PRISMA, the review objectives included:

1. Population: GP/FPs, as defined in the introduction.

2. Intervention/comparator: not applicable.

3. Outcomes: KABB of GP/FPs toward their own vaccination and determinants of the coverage rates occupational vaccines.

4. Study designs: All types of study design (including quantitative and qualitative surveys).

We searched the Medline database on April 14th, 2015, for articles published in the past 10 y, in English or in French. We used a combination of terms and synonyms referring to vaccination, coverage, and general practitioners/family physicians (the full research query is available in the additional file 2).

Study selection

We included all studies assessing any type of recommended or mandatory occupational vaccine for GP/FPs, except the vaccine against the pandemic 2009 A/H₁N₁ influenza, since a number of literature reviews have already studied it, among HCWs in general as well as GP/FPs in particular.^23,46,56,57 We excluded studies: that covered only the pandemic 2009 A/H₁N₁ influenza vaccine; with data collected in a country not listed among the high-income countries (those with a gross national income per capita less than $12,736); ³⁰ which were not original research studies or reviews; for which no abstract was available; which did not present results separately for GP/FPs (i.e. presented results mixing data for GP/FPs and other HCWs); for which no multivariate analysis was performed when studying determinants of GP/FPs' own vaccination; when vaccination coverage data were presented alone (without other data on attitudes to vaccination or determinants of vaccination); and when the sample size was too small (all exclusion criteria are listed in the additional file 3). Two authors independently screened all titles and abstracts to assess whether the studies met the inclusion criteria. Discrepancies were resolved by consensus. The reference lists of the articles selected were also screened to identify additional relevant references.

Data extraction and critical appraisal

Information on study design and population, vaccines, KABB of GP/FPs toward their own vaccination, (self-reported) vaccine coverage rates, and the results of associations tested between different factors and GP/FPs' own vaccination were extracted from all included studies. The risk of bias was evaluated independently by 2 authors using an appropriate and previously used tool – the Newcastle-Ottawa Scale adapted for cross-sectional studies (additional file 4).⁵⁸ Disagreements were resolved by consensus.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We are grateful to Jo-Ann Cahn for her help with language editing.

Authors' Contributions

FC and CP designed the review, performed the literature search, and wrote the first draft. PV and OL critically reviewed the outline of the paper and contributed significantly to the writing of the final version of the manuscript.

Funding

This research has received funding from the French National Research Agency (call for proposals issued in 2015) and joint funding from the French National Health Insurance Fund for Employees (CNAM-TS), the French Directorate General of Health (DGS), the Arc Foundation for Cancer Research, the French National Cancer Institute (INCa), the INPES, the French National Institute of Health and Medical Research (INSERM), the French Interdepartmental Agency for the Fight against Drugs and Addictive Behaviors (Mildeca) and the French Social Security Plan for the Self-Employed (RSI) as part of the “Primary Prevention” call for proposals issued by the French Institute for Public Health Research (IReSP) and the INCa in 2013. F.C. received a PhD grant from the (non-profit) Méditerranée Infection foundation (http://www.mediterranee-infection.com/).