Cognitions and behaviours of general practitioners in France regarding HPV vaccination: A theory-based systematic review.

Human papillomavirus (HPV) vaccination is safe and efficacious to prevent persistent HPV infection, precancerous anogenital lesions and cervical cancer. However, in countries where vaccination programmes are implemented outside of schools, such as France, reaching high HPV vaccination coverage of the target population is challenging. Many studies have been performed in France to assess cognitions of general practitioners (GPs) regarding HPV vaccination. However, the evidence is not consistent about which cognitions are key. To provide a comprehensive overview, we performed a systematic review of studies conducted in France on GPs' cognitions regarding HPV vaccination and used the reasoned action approach to extract and synthesize data. The systematic search was performed up to July 2020 in Medline via PubMed, PsycINFO, PsycARTICLES, Embase, CINAHL Plus, Web of Science, Pascal and Francis databases. Grey literature was searched for in the French Public Health Database, Cairn. Info, yahoo.fr, and Google Scholar. Twenty-five scientific publications were selected based on eligibility criteria and assessed for quality. Our qualitative synthesis highlights that although 73% of GPs report recommending HPV vaccination, up to 50% would not recommend it because of concerns, including changes in patients' health behaviours and doubts about safety and/or efficacy. GPs' injunctive norms, i.e. trust in institutional information, were shown to be associated positively with GPs' willingness to recommend HPV vaccination. Parents' fears, girls' age, and potential connection with sexuality do not seem to affect GPs' recommendations. These results will inform the development of a professional educational intervention targeting GPs in France.

Introduction

Human papillomavirus (HPV) vaccination has been shown to be highly effective to reduce the prevalence of persistent infections with targeted HPV types and the incidence of high-grade precancerous cervical lesions. Thus, it is currently being implemented in many countries, though with variable coverage rates (Bruni et al., 2016). In some high-income settings, such as Canada and the United Kingdom, HPV vaccination coverage can reach more than 80%. It is generally much lower in countries where almost all vaccinations are delivered outside of schools such as the United States and France (Bruni et al., 2016). In France, where general practitioners (GPs) are responsible for counselling and vaccinating adolescent girls, HPV vaccination coverage of the target population, i.e. girls aged 11–14 year since 2012, has not exceeded 30% since its implementation (Santé publique France, 2019). Barriers and facilitators to HPV vaccination have been intensely studied in France (Verrier et al., 2019; Blondel et al., 2019; Institut National du Cancer INCa, 2019). For French parents and young girls, the main barriers regarding HPV vaccination are concerns about serious side effects (Verrier et al., 2019; Blondel et al., 2019; Institut National du Cancer INCa, 2019). Socioeconomic determinants have been shown to influence HPV vaccination uptake with a lower uptake observed in more deprived groups (Blondel et al., 2019). Similarly to others primary care professionals, the HPV vaccination decision of French parents is driven by the recommendations of their GPs (Institut National du Cancer INCa, 2019; Karafillakis et al., 2019; Newman et al., 2018). GPs' barriers mainly concern the efficacy and safety of HPV vaccination, and also relate to the national recommendations regarding HPV vaccination age and groups (Institut National du Cancer INCa, 2019; Karafillakis et al., 2019). The French National Cancer Institute and/or other national agencies have implemented interventions to address GPs' barriers (Campana et al., 2019; Fonteneau et al., 2019). Changes in the national recommendations regarding the age of girls did not lead to an increase in HPV vaccination uptake and it is too early to report on the efficacy of extending HPV vaccination to boys (Fonteneau et al., 2019; Haute Autorité de Santé HAS, 2019). The provision of information about HPV vaccination to GPs has been reinforced since 2017 (Santé publique France, 2020). Although they have not been formally evaluated, these actions are speculated to contribute to the upward trend in coverage observed in France since 2015 (Fonteneau et al., 2019). The gaps and the apparent contradiction in the evidence suggest that some important determinants regarding French GPs' cognitions might have been missed.

A systematic review focusing on French GPs' cognitions and going beyond only knowledge and attitude, which are usually reported, will fil this research gap (Rosen et al., 2018; Herzog et al., 2013). Using a behaviour theory as a framework to extract evidence can provide a more comprehensive overview of these cognitions. We chose to conduct a reasoned action approach (RAA) theory-based systematic review of studies performed in France in order to provide a comprehensive overview of French GPs' cognitions related to HPV vaccination. The RAA is the most recent derivative of the theory of planned behaviour, which has been the most used theoretical framework to explain and predict health practitioners' intentions and behaviours regarding HPV vaccination (Ferrer et al., 2015; Ajzen, 1991; Fishbein and Ajzen, 2010). In addition, the RAA provides a more detailed framework with each of the cognitions of the theory of planned behaviour represented by pairs of distinct, but related, subcomponents, which have been shown to be important in predicting and explaining behaviour (Fishbein and Ajzen, 2010; McEachan et al., 2016). The RAA differentiates the following cognitions (Fig. 1):

Fig. 1

Theoretical framework used in the review based on the reasoned action approach (RAA) (Fishbein and Ajzen, 2010). In this theory, behaviour is driven by intention, which in turn is driven by instrumental and experiential attitudes towards the behaviour, injunctive and descriptive norms, and PBC's capacity and autonomy. Intention regarding HPV vaccination was not assessed in the review, because no data were available.

GP = general practitioner; PBC = perceived behavioural control.

Instrumental (i.e. perceived outcomes of behaviour adoption) or experiential (feelings associated with behaviour) attitudes;

Injunctive (i.e. perceived expectations of important referent individuals or social groups) and descriptive (i.e. perceptions of what important referent individuals or social groups do) norms;

Perceived behavioural control's capacity (i.e. individuals' confidence, beliefs about the necessary skills and abilities they have to perform the behaviour) and autonomy (i.e. individuals' beliefs that they have control over the behaviour).

We used this framework to (i) report GPs' cognitions and beliefs, and (ii) examine the impacts of these cognitions on GPs' behaviours. The impact of background variables related to GPs, i.e. self-related, contextual, or sociodemographic, was also investigated, because they could be influential to French GPs' cognitions and/or behaviours regarding HPV vaccination (Fig. 1) (Rosen et al., 2018). The results of this review will inform the design of an educational intervention on HPV vaccination targeting GPs in France.

Methods

We used the PRISMA statement and the checklist to report explicitly and comprehensively all the recommended components (Moher et al., 2009).

Study search and selection

The systematic search was performed by PV up to December 2018 in the following databases: Medline via PubMed, PsycINFO, PsycARTICLES, Embase, CINAHL plus, Web of Science, Pascal and Francis. The search for grey literature was performed by both GB and PV in the French Public Health Database (Banque de données en santé publique), Cairn.Info, yahoo.fr and Google Scholar. Search updates from December 2018 to July 2020 were conducted by PV. The full research query in Medline via PubMed is available in Appendix, Table A.1.

To be considered for inclusion, titles/abstracts and then full texts were required to meet the following criteria: (a) published in any language, (b) the study population was GPs or included GPs, (c) used quantitative (self-reported questionnaire) and/or qualitative (semi-structured interview, focus group) method, (c) reported outcomes related to any information regarding GPs' cognitions and/or behaviours related to HPV vaccination, (d) based in France or included France among other countries, and (e) was an original study and either a peer-reviewed article or a medical dissertation. Studies were excluded when (a) focused on psychometric validation of questionnaires, or (b) were review, conference abstract, opinion and/or a published survey that did not provide a full description of the methodology and results or (c) was a medical dissertation whose related results were published in a peer-reviewed article, which was part of the review.

Data extraction

A data extraction form and coding method was developed and tested first with some of the included studies. The variables of interest were those presented in the tables in the Result section or the Appendix. The variables were descriptive and included studies characteristics, GPs' characteristics, GPs' cognitions as reported in the RAA model (Fig. 1) and background variables that have been shown to influence primary care professionals' behaviour regarding HPV vaccination (Rosen et al., 2018) (Fig. 1).

We also extracted all variables to explore a possible relationship or an association between a) GPs' attitudes, norms, perceived behavioural control and either cognitions or behaviour regarding HPV vaccination b) background variables and either GPs' attitude or behaviour regarding HPV vaccination.

Data from all included full-text publications were identified and extracted by two investigators (GEB and PV). Any disagreement in data extraction was resolved by consensus.

Assessment of study quality

Two of the authors independently assessed the methodological quality of the included articles using the Newcastle-Ottawa quality assessment scale (NOS) for cohort studies adapted for cross sectional studies and the Critical Appraisal Skills Programme (CASP) framework for qualitative articles (Herzog et al., 2013; Stang, 2010; Singh, 2013). An additional column was added to the NOS tool to highlight articles. Any disagreement in quality grading was resolved by consensus.

Synthesis

Because of heterogeneity in data measurements and the outcomes, qualitative synthesis was applied to synthetize the data (Seers, 2015). No study was excluded based on its quality. However when synthetized the findings and when appropriate, the research outcomes were evaluated taking into account the quality of the studies involved.

Results

We identified 2361 abstracts through our systematic search and 73 additional through a non-systematic search (Fig. 2). We screened 1512 abstract for eligibility. We assessed 78 records based on full text evaluation. Twenty-five studies were included in the synthesis (Fig. 2).

Fig. 2

Studies included in the review based on the PRISMA flow diagramme (Moher et al., 2009).

Description of the included articles

Among the 25 studies identified for this review (Agrinier et al., 2017; Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Casimont, 2015; Chauvet, 2016; Collange et al., 2016; Degoue, 2019; Erpeldinger et al., 2012; Gougenheim-Fretin, 2014; Houdjal, 2017; Killian et al., 2016; Lamirand, 2015; Lasset et al., 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Martinez et al., 2016; Pelissier and Bastides, 2008; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Thierry et al., 2016; Tutala, 2019; Verger et al., 2015), 18 reported quantitative and seven qualitative designs (Table 1). The majority of the studies were conducted either before 2011 (Barjhoux and Glatigny, 2009; Erpeldinger et al., 2012; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Pelissier and Bastides, 2008; Piana et al., 2009; Plessis et al., 2012; Thierry et al., 2016) with some around 2007–2008 i.e. the period where HPV vaccination was implemented in France (Barjhoux and Glatigny, 2009; Erpeldinger et al., 2012; Lutringer-Magnin et al., 2011; Pelissier and Bastides, 2008; Piana et al., 2009), or over the 2011–2015 period (Agrinier et al., 2017; Bouvret et al., 2016; Casimont, 2015; Collange et al., 2016; Gougenheim-Fretin, 2014; Killian et al., 2016; Lamirand, 2015; Manolitsi, 2012; Martinez et al., 2016; Raude et al., 2016; Verger et al., 2015). The most recent studies were conducted in 2016 (Chauvet, 2016; Houdjal, 2017; Leicht, 2016; Sadki, 2016) or in 2019 (Degoue, 2019; Tutala, 2019). The sample size ranged from ten to 31 GPs in the qualitative studies and from 96 to 1598 in the quantitative studies. Except for five articles in which the participants represented a national panel (Agrinier et al., 2017; Collange et al., 2016; Leicht, 2016; Thierry et al., 2016; Verger et al., 2015), the studies included participants from local panels. Only one study used random sampling to select participants (Leicht, 2016).

Table 1

Characteristics of studies included in the review.

Author (publication year)Year(s) of data collection	Documenta	Study design	Measurement	Participants /Invitedb(Type of recruitment)	Aim	Risk of biasc
Agrinier et al., 20172013–2014	Article	Quantitative	Self-reported questionnaire	1038/1712(National panel)	To measure discrepancies between vaccination recommendations by GPs for their patients and practices for their children	Moderate
Barjhoux and Glatigny, 20092008	Medical dissertation	Quantitative	Self-reported questionnaire	278/1200(Local panel)	To evaluate the modalities and difficulties encountered by GPs when offering the vaccine to patients	High
Bouvret et al., 20162014–2015	Article	Quantitative	Self-reported questionnaire	96/140(Local panel)	To assess opinions, practices, and difficulties of GPs regarding HPV vaccination	Low
Casimont, 20152014	Medical dissertation	Qualitative	Interview	10(Local panel)	To describe opinion of GPs about changes in strategies to prevent cervical cancer	High
Chauvet, 20162016	Medical dissertation	Quantitative	Self-reported questionnaire	143/495(Local panel)	To evaluate tools that aim to help GPs to deliver information about HPV vaccination	High
Collange et al., 20162014	Article	Quantitative	Self-reported questionnaire	1598/1712(National panel)	To study GPs' perceptions of HPV vaccination risks and efficacy and their recommendation behaviour; and the relative importance of factors associated with the frequency of their recommendations	Moderate
Degoue, 20192018–2019	Medical dissertation	Quantitative	Self-reported questionnaire	337(Local panel)	To describe GPs' practice regarding HPV vaccination	High
Erpeldinger et al., 20122009	Article	Quantitative	Self-reported questionnaire	518/1193(Local panel)	To describe the knowledge of GPs on infection with HPV and Gardasil®, and to determine the impact of training and information on knowledge and the attitude towards this vaccine	High
Gougenheim-Fretin, 20142013	Medical dissertation	Qualitative	Interview/focus group	10(Local panel)	To highlight the reluctance of GPs towards HPV vaccination	Low
Houdjal, 20172016	Medical dissertation	Qualitative	Interview	12(Local panel)	To highlight GP representations of HPV vaccination	High
Killian et al., 20162013–2014	Article	Quantitative	Self-reported questionnaire	693/2839(Local panel)	Comparison of GPs' HPV immunization practices for their patients and their children	Low
Lamirand, 20152015	Medical dissertation	Qualitative	Interview	11(Local panel)	To describe impacts of the media and new recommendations about HPV vaccination on GPs' practice	Low
Lasset et al., 20142010	Article	Quantitative	Self-reported questionnaire and interview	271/290(Local panel)	To investigate the evolution of practices and opinions regarding HPV vaccination among GPs	Low
Leicht, 20162016	Medical dissertation	Quantitative	Self-reported questionnaire	171/350(National panel)	To identify obstacles of GPs in France for HPV vaccination	High
Lutringer-Magnin et al., 20112007–2008	Article	Quantitative	Self-reported questionnaire and interview	279/5973(Local panel)	To examine the perceptions, attitudes, and practices of GPs in relation to HPV vaccination	Moderate
Manolitsi, 20122012	Medical dissertation	Quantitative	Self-reported questionnaire	145/932(Local panel)	To examine and to understand the obstacles of GPs regarding HPV vaccination	Moderate
Martinez et al., 20162013	Article	Qualitative	Interview/focus group	36/622(Local panel)	To conduct an exploratory qualitative study with GPs to identify determinants of their commitment to vaccination	Low
Pelissier and Bastides, 20082007	Article	Quantitative	Self-reported questionnaire	252/545(Local panel)	To describe prevention behaviours and to examine perceptions of HPV vaccination	Moderate
Piana et al., 20092008	Article	Quantitative	Self-reported questionnaire	359/1000(Local panel)	To assess the standpoint of GPs regarding HPV vaccination and to evaluate the factors associated with a favourable standpoint	Low
Plessis et al., 20122009–2010	Article	Qualitative	Interview/focus group	16/19(Local panel)	To better understand GPs' decisions about HPV vaccination and their role in cervical cancer	High
Raude et al., 20162013–2014	Article	Quantitative	Self-reported questionnaire	1582/1712(Local panel)	To improve the understanding of the role of institutional trust in practices related to vaccination	Low
Sadki, 20162016	Medical dissertation	Qualitative	Interview	17/28(Local panel)	To highlight GPs' opinions and practices about HPV vaccination and their attitude towards reluctance to vaccination	Low
Thierry et al., 20162010	Article	Quantitative	Self-reported questionnaire	363/706(National panel)	To evaluate the vaccine coverage, according to eligibility for vaccination in a sample of girls who were seen in general practices in France	Moderate
Tutala, 20192018–2019	Medical dissertation	Quantitative	Self-reported questionnaire	216/4148(Local panel)	To evaluate the role of health practitioners regarding HPV vaccination in Occitanie region	High
Verger et al., 20152013–2014	Article	Quantitative	Self-reported questionnaire	1582/1712(National panel)	To assess the prevalence of vaccine hesitancy among GPs through the frequency of their vaccine recommendations, and the determinants of these recommendations	Low

GP = general practitioner; HPV = human papillomavirus.

Peer-reviewed article or medical dissertation.

Invited to participate in the study, where available.

Appendix, Tables A.2 and A.3.

The quality of the included articles varied widely, with ten articles (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Killian et al., 2016; Lamirand, 2015; Lasset et al., 2014; Martinez et al., 2016; Piana et al., 2009; Raude et al., 2016; Sadki, 2016; Verger et al., 2015), considered at low risk (i.e., high-quality studies), six (Agrinier et al., 2017; Collange et al., 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Pelissier and Bastides, 2008; Thierry et al., 2016) at moderate risk (i.e., moderate-quality studies), and nine at high risk of bias (Barjhoux and Glatigny, 2009; Casimont, 2015; Chauvet, 2016; Degoue, 2019; Erpeldinger et al., 2012; Houdjal, 2017; Leicht, 2016; Plessis et al., 2012; Tutala, 2019) (i.e., low-quality studies) (Appendix, Tables A.2 and A.3).

Although the characteristics of the participants were not systematically reported in the articles, those that did report the characteristics described the GP study groups as mainly men (more than 50% in 20 of the 25 articles), middle-aged (mean age range: 42.3–54 years) (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Chauvet, 2016; Degoue, 2019; Gougenheim-Fretin, 2014; Houdjal, 2017; Lamirand, 2015; Lasset et al., 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Pelissier and Bastides, 2008; Piana et al., 2009; Sadki, 2016; Thierry et al., 2016; Tutala, 2019), with more than 10 years of experience in general practice (Bouvret et al., 2016; Houdjal, 2017; Manolitsi, 2012; Plessis et al., 2012; Sadki, 2016), working mainly in urban and suburban areas (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Casimont, 2015; Chauvet, 2016; Degoue, 2019; Gougenheim-Fretin, 2014; Killian et al., 2016; Lamirand, 2015; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Martinez et al., 2016; Pelissier and Bastides, 2008; Plessis et al., 2012; Sadki, 2016), and with a workload of between 3000 and 6000 consultations per year for more than 50% of GPs (Agrinier et al., 2017; Collange et al., 2016; Killian et al., 2016; Lamirand, 2015; Raude et al., 2016) (Appendix, Table A.4).

GPs' attitudes, norms, perceived behavioural control, and underlying beliefs regarding HPV vaccination

GPs' cognitions regarding HPV vaccination are summarized in Table 2. Attitudes (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Casimont, 2015; Collange et al., 2016; Degoue, 2019; Erpeldinger et al., 2012; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Martinez et al., 2016; Pelissier and Bastides, 2008; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Thierry et al., 2016; Tutala, 2019; Verger et al., 2015) and perceived behavioural control towards HPV vaccination (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Casimont, 2015; Chauvet, 2016; Degoue, 2019; Gougenheim-Fretin, 2014; Houdjal, 2017; Lamirand, 2015; Lasset et al., 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Martinez et al., 2016; Pelissier and Bastides, 2008; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Thierry et al., 2016; Tutala, 2019; Verger et al., 2015) were evaluated in a total of 21 studies each, whereas norms were examined in only 11 articles (Bouvret et al., 2016; Casimont, 2015; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Martinez et al., 2016; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Verger et al., 2015). None of the articles distinguished between the subcomponents of each cognition when reporting the results.

Table 2

General practitioners' cognitions regarding HPV vaccination.

Type	Subcomponents a
Attitude	Experiential attitudes (Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Erpeldinger et al., 2012; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Sadki, 2016; Tutala, 2019)➢ In favour or approve HPV vaccination (Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Sadki, 2016; Tutala, 2019): > 74% ➢ Prudent, enthusiast, doubt, worried (Erpeldinger et al., 2012; Gougenheim-Fretin, 2014) Instrumental attitudes (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Collange et al., 2016; Degoue, 2019; Erpeldinger et al., 2012; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Martinez et al., 2016; Pelissier and Bastides, 2008; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Thierry et al., 2016; Tutala, 2019; Verger et al., 2015)➢ Trust in efficacy and safety (Tutala, 2019): 81% ➢ Useful and necessary (Collange et al., 2016; Degoue, 2019; Houdjal, 2017; Raude et al., 2016; Sadki, 2016; Thierry et al., 2016; Verger et al., 2015): 75% ➢ Doubt about clinical benefits or efficacy (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Collange et al., 2016; Manolitsi, 2012): 21% -35% ➢ Worried about potential side effects and risks (Bouvret et al., 2016; Collange et al., 2016; Degoue, 2019; Gougenheim-Fretin, 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Thierry et al., 2016): 30% -50% ➢ HPV vaccine is dangerous (Barjhoux and Glatigny, 2009; Degoue, 2019; Houdjal, 2017): < 10% ➢ Other consequences (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Leicht, 2016; Manolitsi, 2012; Piana et al., 2009; Plessis et al., 2012)o Decrease of cervical cancer screening (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Houdjal, 2017; Manolitsi, 2012; Piana et al., 2009): 16% -29% o Decrease of condom use (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Manolitsi, 2012; Piana et al., 2009; Plessis et al., 2012): 16%–24% o Encourage riskier sexual behaviour or premature sexual relation (Gougenheim-Fretin, 2014; Lasset et al., 2014; Leicht, 2016; Piana et al., 2009): 5%–10%
Norms	Injunctive norms (Bouvret et al., 2016; Casimont, 2015; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Martinez et al., 2016; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Verger et al., 2015)➢ GPs role regarding public health policies (Martinez et al., 2016; Plessis et al., 2012) ➢ Need to follow health authorities' recommendations (Casimont, 2015; Houdjal, 2017; Piana et al., 2009; Plessis et al., 2012; Sadki, 2016) ➢ Trust in institutional information provided by official sources (Bouvret et al., 2016; Casimont, 2015; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Martinez et al., 2016; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Verger et al., 2015): 81–94.5% ➢ Public health (Lasset et al., 2014; Lutringer-Magnin et al., 2011; Martinez et al., 2016; Plessis et al., 2012): 60% Descriptive norms➢ Rely on pairs opinion and practice to decide to vaccinate (Houdjal, 2017; Martinez et al., 2016)
Perceived Behavioural control	Capacity (Bouvret et al., 2016; Chauvet, 2016; Gougenheim-Fretin, 2014; Leicht, 2016; Pelissier and Bastides, 2008; Piana et al., 2009; Raude et al., 2016; Thierry et al., 2016; Tutala, 2019; Verger et al., 2015)➢ Have enough information about HPV vaccination and related subjects (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Leicht, 2016; Pelissier and Bastides, 2008; Piana et al., 2009; Tutala, 2019): > 68% ➢ Feel confident to inform about HPV vaccination (Chauvet, 2016; Pelissier and Bastides, 2008; Raude et al., 2016; Tutala, 2019; Verger et al., 2015): 19% to 88% ➢ Feel able to implement recommendations (Pelissier and Bastides, 2008): 70% ➢ Feel comfortable with adolescents care (Thierry et al., 2016): 95% Autonomy (essentially barriers) (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Gougenheim-Fretin, 2014; Houdjal, 2017; Lamirand, 2015; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Martinez et al., 2016; Plessis et al., 2012; Sadki, 2016; Tutala, 2019)➢ Parents (Degoue, 2019): 56% ➢ Parents' fear of side effects (Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Gougenheim-Fretin, 2014; Houdjal, 2017; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Sadki, 2016): > 60% ➢ Age of girls (Barjhoux and Glatigny, 2009; Casimont, 2015; Gougenheim-Fretin, 2014; Houdjal, 2017; Lamirand, 2015; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Sadki, 2016) ➢ Scarcity of pre-adolescents or adolescents' consultations (Gougenheim-Fretin, 2014; Lamirand, 2015; Lasset et al., 2014; Sadki, 2016) ➢ Socio-cultural characteristics of patients as barriers (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Gougenheim-Fretin, 2014; Lamirand, 2015; Manolitsi, 2012; Plessis et al., 2012): 17.5% to 25% ➢ Lack of time (Barjhoux and Glatigny, 2009; Casimont, 2015; Gougenheim-Fretin, 2014; Martinez et al., 2016; Sadki, 2016; Tutala, 2019)

HPV = human papillomavirus.

Percentage of general practitioners was reported when available.

Other consequences (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Leicht, 2016; Manolitsi, 2012; Piana et al., 2009; Plessis et al., 2012)

Attitudes and underlying beliefs

Among the 21 studies reporting GPs' attitudes, we identified that one reported GPs' experiential attitudes only (Casimont, 2015), nine reported instrumental attitudes only, (Barjhoux and Glatigny, 2009; Collange et al., 2016; Leicht, 2016; Martinez et al., 2016; Pelissier and Bastides, 2008; Plessis et al., 2012; Raude et al., 2016; Thierry et al., 2016; Verger et al., 2015) and 11 reported both (Bouvret et al., 2016; Degoue, 2019; Erpeldinger et al., 2012; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Sadki, 2016; Tutala, 2019) (Table 2). Experiential attitudes were reported mostly in terms of favourable versus unfavourable opinions towards HPV vaccination, confidence versus concern, worry or doubt, and/or enthusiasm (Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Erpeldinger et al., 2012; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Sadki, 2016; Tutala, 2019). Instrumental attitudes were reported in terms of trust regarding efficacy and security, perceptions of the efficacy of HPV vaccination, its benefits, its usefulness and/or its risks represented by fear of side effects, especially in relation to autoimmune diseases (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Collange et al., 2016; Degoue, 2019; Erpeldinger et al., 2012; Gougenheim-Fretin, 2014; Houdjal, 2017; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Martinez et al., 2016; Piana et al., 2009; Raude et al., 2016; Sadki, 2016; Thierry et al., 2016; Tutala, 2019; Verger et al., 2015). Instrumental attitudes were also reported by assessing beliefs related to the impact of HPV vaccination on other health behaviours, i.e., cervical cancer screening, condom use, and/or sexual behaviours (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Houdjal, 2017; Lasset et al., 2014; Leicht, 2016; Manolitsi, 2012; Piana et al., 2009; Plessis et al., 2012).

The results showed that GPs found HPV vaccination to be useful and necessary (Collange et al., 2016; Degoue, 2019; Houdjal, 2017; Raude et al., 2016; Sadki, 2016; Thierry et al., 2016; Verger et al., 2015). When assessed, and independently of study quality, more than 74% of GPs were in favour of or approved of HPV vaccination (Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Sadki, 2016; Tutala, 2019). When study quality was taken into account, i.e. reported percentage adjusted to reflect overall quality of the studies, 21%–35% of GPs reported doubts about the clinical benefits or efficacy of HPV vaccination (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Collange et al., 2016; Manolitsi, 2012) and about 30–50% reported worries or concerns about side effects (Bouvret et al., 2016; Collange et al., 2016; Degoue, 2019; Gougenheim-Fretin, 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Thierry et al., 2016). More specifically, rates of GPs with worries and concerns were 19–50% in high-quality studies (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Piana et al., 2009), 23–60% in moderate-quality studies (Collange et al., 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Thierry et al., 2016), and 23–34% in low–quality studies (Degoue, 2019; Leicht, 2016). The date of study did not seem to explain these results. Three low-quality studies reported that up to 10% of GPs even considered HPV vaccination to be dangerous (Barjhoux and Glatigny, 2009; Degoue, 2019; Houdjal, 2017). In studies performed before 2015, for up to 61% of GPs these doubts about efficacy and benefits/risks were linked to the “novelty” of HPV vaccination and the relatively short period since the introduction of the vaccine (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Collange et al., 2016; Manolitsi, 2012; Pelissier and Bastides, 2008; Plessis et al., 2012). Irrespective of study quality and date, 16–29% of GPs believed that HPV vaccination would decrease patients' participation in cervical cancer screening (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Houdjal, 2017; Manolitsi, 2012; Piana et al., 2009), 16–24% of GPs reported that HPV vaccination would decrease the use of condoms (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Manolitsi, 2012; Piana et al., 2009; Plessis et al., 2012) and 5–10% that it would lead to taking sexual risks and/or encourage premature sexuality (Gougenheim-Fretin, 2014; Lasset et al., 2014; Leicht, 2016; Piana et al., 2009). Furthermore, in five articles mostly with high (Barjhoux and Glatigny, 2009; Casimont, 2015; Erpeldinger et al., 2012; Plessis et al., 2012) to moderate (Pelissier and Bastides, 2008) risk of bias, GPs reported that screening and/or condom use are more effective than HPV vaccination (Barjhoux and Glatigny, 2009; Casimont, 2015; Erpeldinger et al., 2012; Pelissier and Bastides, 2008; Plessis et al., 2012).

Norms and underlying beliefs

Regarding norms, influence of pairs was approached in two studies (Houdjal, 2017; Martinez et al., 2016) and injunctive norms were mainly assessed (Bouvret et al., 2016; Casimont, 2015; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Martinez et al., 2016; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Verger et al., 2015) (Table 2) These norms included GPs' perceptions of their role in public health policies, the need to follow health authorities' recommendations, and trust in institutional information. Overall, 81–94.5% of GPs trusted that institutional information (i.e., from the ministry of health, health agencies, scientific sources, and the opinions of scientists and specialist physician colleagues) about the benefits and risks of immunization was reliable (Bouvret et al., 2016; Casimont, 2015; Houdjal, 2017; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Martinez et al., 2016; Piana et al., 2009; Plessis et al., 2012; Raude et al., 2016; Sadki, 2016; Verger et al., 2015). HPV vaccination was seen as a requirement for preventive medicine and a benefit for public health (Lasset et al., 2014; Lutringer-Magnin et al., 2011; Martinez et al., 2016; Plessis et al., 2012). For instance, the beneficial effects of vaccination on primary prevention of HPV and public health were highlighted by 60% of GPs (Lasset et al., 2014; Lutringer-Magnin et al., 2011). GPs emphasized their role in public health, which consisted of informing patients about recommended vaccines, explaining, convincing, and justifying the vaccination recommendations, addressing discussions about sexuality, and vaccinating their patients (Martinez et al., 2016; Plessis et al., 2012). GPs also stated that they vaccinated their patients to be in line with recommendations (Casimont, 2015; Houdjal, 2017; Piana et al., 2009; Plessis et al., 2012; Sadki, 2016).

Perceived Behavioural control and underlying beliefs

Capacity

Except for one study, which did not indicate rates (Gougenheim-Fretin, 2014), more than 68% of GPs felt that they were well-informed about HPV and HPV vaccination (Bouvret et al., 2016; Gougenheim-Fretin, 2014; Leicht, 2016; Pelissier and Bastides, 2008; Piana et al., 2009; Tutala, 2019). GPs felt confident in justifying vaccine recommendations and in explaining the utility of the vaccine (up to 88%) and its safety (up to 77%) (Chauvet, 2016; Pelissier and Bastides, 2008; Raude et al., 2016; Tutala, 2019; Verger et al., 2015). However, this rate was lower (up to 58%), when confidence in explaining the role of vaccine adjuvants was investigated (Chauvet, 2016; Raude et al., 2016; Tutala, 2019; Verger et al., 2015). GPs reported being comfortable with having adolescents as patients (Thierry et al., 2016).

Autonomy

Parents' fear of side effects of HPV vaccination is one of the most important barriers, reported by more than 60% of GPs (Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Gougenheim-Fretin, 2014; Houdjal, 2017; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Sadki, 2016).

GPs explained these fears by the relatively short period since the introduction of the vaccine (Bouvret et al., 2016; Sadki, 2016) and controversies about vaccination in general or about specific vaccines, such as those against hepatitis B virus, H1N1 influenza virus, and HPV (Bouvret et al., 2016; Casimont, 2015; Gougenheim-Fretin, 2014; Houdjal, 2017; Pelissier and Bastides, 2008; Plessis et al., 2012). Independently of their quality, studies reported high variability in the rate of GPs who considered the socio-cultural characteristics of patients and their parents (e.g., religion and personal conviction) to be important barriers (17.5–25%) (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Gougenheim-Fretin, 2014; Lamirand, 2015; Manolitsi, 2012; Plessis et al., 2012).

The main practice-related barriers reported were the age of girls (Barjhoux and Glatigny, 2009; Casimont, 2015; Gougenheim-Fretin, 2014; Houdjal, 2017; Lamirand, 2015; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Sadki, 2016) and the scarcity of consultations with adolescents (Gougenheim-Fretin, 2014; Lamirand, 2015; Lasset et al., 2014; Sadki, 2016). Before 2012, when vaccination was recommended for girls aged 14 years, 28–72% of GPs would have preferred to vaccinate against HPV at as young an age as possible (i.e., 11–14 years) to avoid having to discuss sexually transmitted infections and sexuality (Barjhoux and Glatigny, 2009; Lutringer-Magnin et al., 2011). Indeed, addressing sexually transmitted infections was reported as an issue for 11–31% of GPs (Barjhoux and Glatigny, 2009; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012). After 2012, when HPV vaccination became recommended for girls aged 11–14 years, GPs no longer felt an obligation to discuss sexuality, and therefore they perceived this change as beneficial (Lamirand, 2015; Sadki, 2016). Nevertheless, some GPs continued to address sexuality as part of the information on HPV vaccination or to satisfy parents' expectations and requests (Houdjal, 2017; Lamirand, 2015; Sadki, 2016). This did not seem to be a barrier anymore (Tutala, 2019). Meanwhile, new issues were reported by GPs regarding younger ages of patients: girls being too young to be concerned (Casimont, 2015; Houdjal, 2017; Lamirand, 2015), having to rely on parents negative decision as reported by 56% of GPs (Degoue, 2019), and parents being insecure to talk about sexuality at such young age and/or frightened of encouraging premature sexuality (Casimont, 2015; Houdjal, 2017; Lamirand, 2015; Sadki, 2016). These issues affected GPs' practice as both parents and GPs had the tendency to delay HPV vaccination to older ages (i.e. 14) (Houdjal, 2017; Sadki, 2016).

The scarcity of consultations with adolescents was perceived as a serious obstacle (Gougenheim-Fretin, 2014; Lamirand, 2015; Lasset et al., 2014; Sadki, 2016). Therefore, any reason for consultation (i.e., medical consultation other than vaccination, certificate of fitness for sport) was reported to be used to propose the vaccine (Casimont, 2015; Gougenheim-Fretin, 2014; Sadki, 2016).

GPs' behaviours regarding HPV vaccination

Main results are presented in Fig. 3 and in details in Appendix (Table A.5). In the more recent studies of moderate to high quality and with larger number of participants, 73% of GPs reported either always (46%) or often (27%) recommending HPV vaccination to the target population (Collange et al., 2016; Verger et al., 2015), and 72.9% of GPs reported proposing to vaccinate these girls against HPV (Bouvret et al., 2016).

Fig. 3

Association between general practitioners' cognitions and behaviour variables assessed in the review. Sens of arrow reflects which variable was used as dependent vs independent in the studies. Association was either not significant (X) or significant and either positive (+) or negative (−). Descriptive norms were not assessed. PBC = perceived behavioural control; HPV = human papillomavirus.1Bouvret et al., 2016; 2Collange et al., 2016; 3Lasset et al., 2014; 4Lutringer-Magnin et al., 2011; 5Manolitsi, 2012; 6Piana et al., 2009; 7Raude et al., 2016; 8Verger et al., 2015.

Impact of cognitions

Five articles used quantitative methods to examine the effects of specific cognitions on the behaviour of interest, essentially to recommend or propose (Bouvret et al., 2016; Chauvet, 2016; Collange et al., 2016; Raude et al., 2016; Verger et al., 2015). Five articles examined the links between cognitions and dependent variables, i.e., experiential attitudes (Lasset et al., 2014; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009), vaccine hesitancy (Raude et al., 2016), or perceived behavioural control (Raude et al., 2016).

Importance of instrumental attitudes

Among attitudes tested, GPs' perceptions of risks, including change in patients' health behaviour and doubts about vaccine utility/efficacy, were shown to be negatively associated with proposing and recommending HPV vaccination (Bouvret et al., 2016; Collange et al., 2016; Raude et al., 2016; Verger et al., 2015). GPs' concerns about side effects of HPV vaccination were shown to be negatively associated with favourable (experiential) attitudes (Manolitsi, 2012; Piana et al., 2009).

Importance of injunctive norms

Results from three studies of either moderate quality (Collange et al., 2016) or high quality (Raude et al., 2016; Verger et al., 2015) were in favour of a positive and significant association between trust and recommending HPV vaccination. GPs' trust in institutional information showed a positive and significant association with HPV vaccination recommendation, either directly (Verger et al., 2015) or indirectly and mediated through a decrease in vaccine hesitancy (Raude et al., 2016). Trust was also positively correlated with ability to explain the utility, safety, and adverse events of vaccination (Raude et al., 2016). Beneficial effects of vaccination in public health and its role in primary prevention were positively correlated with favourable (experiential) attitudes (Lasset et al., 2014).

Perceived behavioural control: inconsistent results

The impact of abilities on behaviour was contradictory from two high-quality studies. Confidence in GPs' ability to explain the utility of vaccines, the safety of vaccines, and the role of adjuvants in general (including HPV) was reported to be either associated (positively) or not associated with recommendation of the vaccines (Raude et al., 2016; Verger et al., 2015). Ability to talk about sexuality was positively associated with favourable opinions about HPV vaccination (Piana et al., 2009), but HPV vaccine knowledge, by itself, seemed not to have effect on experiential attitudes (Manolitsi, 2012).

When autonomy, i.e., barriers related to either patients or practice (i.e., reason for consultation, age for vaccination, questions asked by patients, parents' presence and reluctance to address sexuality, and necessity to address the issue of sexually transmitted infections) were examined, they were not associated with GPs' recommendation (Collange et al., 2016) and it is not clear if it is associated with GPs' attitudes (Lutringer-Magnin et al., 2011; Manolitsi, 2012).

Impact of background variables on behaviours and experiential attitudes regarding HPV vaccination

Ten articles reported effects of background variables on behaviours and/or experiential attitudes (i.e., favourable opinion), with age and gender being the most studied variables (Agrinier et al., 2017; Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Chauvet, 2016; Collange et al., 2016; Erpeldinger et al., 2012; Killian et al., 2016; Lasset et al., 2014; Lutringer-Magnin et al., 2011; Piana et al., 2009) (Appendix, Table A.6).

When study quality was taken into account, results from studies assessing age of GPs (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Chauvet, 2016; Collange et al., 2016; Erpeldinger et al., 2012; Lutringer-Magnin et al., 2011; Piana et al., 2009) showed that younger GPs were more favourable to proposing vaccination and to vaccinating against HPV (Bouvret et al., 2016; Lutringer-Magnin et al., 2011; Piana et al., 2009). The results showed no significant association between gender and behaviours or attitudes (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Chauvet, 2016; Collange et al., 2016; Erpeldinger et al., 2012; Lutringer-Magnin et al., 2011), except in one high-quality article in which men were more prone to have favourable opinions (Piana et al., 2009). Practice-related variables (workload reported as consultations or visits >100/week or > 3000/year, consultation duration <20 min, seeing more women or children/teenagers) were positively associated with favourable opinions and recommendation/vaccination (Collange et al., 2016; Lutringer-Magnin et al., 2011; Piana et al., 2009). HPV vaccination behaviour was not affected by practice of alternative medicine (e.g., homeopathy, acupuncture), practice of Pap smear or gynaecology, or recommendation of other vaccines, although these variables were shown to affect favourable opinions towards HPV vaccination (Bouvret et al., 2016; Chauvet, 2016; Erpeldinger et al., 2012; Lutringer-Magnin et al., 2011; Piana et al., 2009). Variables that were not related to behaviours or attitudes were type of practice (i.e., solo/group), number of years of practice, cost of vaccine, and number of injections (Bouvret et al., 2016; Chauvet, 2016; Killian et al., 2016; Lutringer-Magnin et al., 2011).

Discussion

The use of the RAA theoretical framework to extract and analyse data offers new insights into French GPs' cognitions and behaviours regarding HPV vaccination. Our review shows that, in terms of attitude, these are fears of risks (i.e., concerns about safety and change of behaviour) and doubts about utility/efficacy regarding HPV vaccination, which drive GPs' decisions about whether to recommend HPV vaccination (Bouvret et al., 2016; Collange et al., 2016; Raude et al., 2016; Verger et al., 2015). Taking into account that very strong ethical norms are in place in GPs as a professional group, including the precautionary principle, which is very well embedded in norms in France regarding public health (David, 2009), it is more likely that GPs with concerns and doubts, i.e., 30–50% of GPs (Barjhoux and Glatigny, 2009; Bouvret et al., 2016; Collange et al., 2016; Degoue, 2019; Gougenheim-Fretin, 2014; Leicht, 2016; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Piana et al., 2009; Thierry et al., 2016) will neither propose nor recommend HPV vaccination to girls. This figure may even be underestimated, because GPs' responses are potentially biased due to social desirability (GPs respond what they think they should say) and/or wishful thinking (GPs respond what they would like to be true). It is worth noting that the percentage of GPs with concerns and doubts is very high and in line with the rate of unvaccinated adolescent girls in France, i.e. about 70% in 2019 (Santé publique France, 2019).

Although up to 88% of GPs report confidence in explaining the safety and efficacy of HPV vaccination (Chauvet, 2016; Pelissier and Bastides, 2008; Raude et al., 2016; Tutala, 2019; Verger et al., 2015), the percentage of GPs who will be able to answer related specific questions is probably much lower, as reported in France and in other settings (Tutala, 2019; McSherry et al., 2018). This result together with the fact that up to 50% of GPs reports doubts and concerns regarding HPV vaccination is intriguing. Indeed, reliable information related to HPV vaccination safety and efficacy is easily accessible to GPs through authoritative health websites (Santé publique France, 2020). It is possible that French GPs, who have been reported to work at least 50 h a week, may not have time to both access HPV vaccination-related information and attend continuing education (Chaput et al., 2019). Another possibility, which might be also related to their heavy workload, is that GPs use heuristics to process HPV-related information. These are mental shortcuts that enable them to make decisions and process information more rapidly and based on incomplete, uncertain, or/and peripheral information (Stroebe, 2011). Heuristics have been shown to be part of health practitioners' practice and driven by constraints such as time (Marewski and Gigerenzer, 2012). Moreover, it has been shown that heuristics-based information is less resistant to counter-argument and less predictive of behaviour than systematic processing (Stroebe, 2011). It is thus possible that GPs when facing parents would not be able to address parents' arguments and even recommend HPV vaccination, which is a driver for parents to vaccinate (Institut National du Cancer INCa, 2019; Karafillakis et al., 2019; Newman et al., 2018). This may explain the discrepancy between the GP-reported recommendation/vaccination rate (i.e. 70%) (Bouvret et al., 2016; Collange et al., 2016; Verger et al., 2015) and actual HPV vaccination rates (of less than 30%), which is also reported in other settings (Rosen et al., 2018; Vollrath et al., 2018). The positive effect of higher (>3000 consultations/year) workload on GPs behaviour (Collange et al., 2016) seems to contradict our hypothesis on time constraints stated above. This threshold may be underestimated according to the average number of consultations per GP per year i.e. 5100–5800 (Chaput et al., 2019; Jakoubovitch et al., 2012). Thus, instead of considering a positive effect of high workload, we may assume that GPs with a workload lower than average (e.g. semi-retired, working part-time or as substitutes), recommend less HPV vaccination, what may be explained by seeing less patients including girls.

Our results suggest that the ways in which HPV vaccination-related information is currently provided to GPs in France are not adequate. Vaccination teaching in the French medical curriculum has shown some limitations (Kernéis et al., 2017). Continuing medical education is not mandatory, eventhough completing at least one training is part of the GPs' annual performance scheme. Information tools, including continuing education, should be developed to favour less heuristic processing of information and take into account GPs' constraints.

Our review highlights the central place of injunctive norms, i.e., trust in institutional information, in positively affecting GPs' beliefs and behaviours regarding HPV vaccination (Raude et al., 2016; Verger et al., 2015). This is understandable because GPs are recognized as a professional group with shared strong professional norms. Regarding descriptive norms, to the best of our knowledge, no studies have yet explored in details the role of these norms in GPs' behaviours regarding HPV vaccination, although they have been shown to be a driver in physicians' practice (Keating, 2017). Future work is needed in this field. Both the importance of HPV vaccination and the role of GPs in public health should be highlighted when developing information tools and education modules on HPV vaccination.

Parents' fear of side effects of HPV vaccination is reported by more than 60% of GPs as one of the most important barriers to recommend and vaccinate (Bouvret et al., 2016; Casimont, 2015; Degoue, 2019; Gougenheim-Fretin, 2014; Houdjal, 2017; Lutringer-Magnin et al., 2011; Manolitsi, 2012; Sadki, 2016). Despite the change of the target age group from 14 years to 11–14 years, the age of girls remains an issue for GPs, raising even more concerns. The main reason is that HPV vaccination remains strongly connected to sexuality or sexual behaviour, as reported in other settings, such as the United States and Canada (Rosen et al., 2018; Vollrath et al., 2018). However, age/sexuality/parents fears does not seem to impact GPs' recommendation decision regarding HPV vaccination (in our review). Vaccination might be an issue due to girls' parents but GPs have a strategy in place i.e. vaccinate girls later, at around age 14 years, which is also highlighted in other reviews (Rosen et al., 2018; Vollrath et al., 2018). This is interesting as research has been developed in recent years focusing on patients as a barrier and developing interventions to improve the abilities of health providers to recommend HPV vaccination. Interventions using this approach have been shown to be effective (Brewer et al., 2017; Reno et al., 2018); however neither sustainability of these interventions nor their effects on GPs' cognitions have been shown so far (Malo et al., 2018). Moreover if motivational interviewing seems to be promising, time required to do it properly and based on ethical principles render it difficult to be integrated in GPs' busy schedule (Passmore, 2013).

Our results suggest that these are actually GPs themselves that could be considered as a barrier to HPV vaccination and that research should centre on how HPV vaccination information is effectively conveyed to them. While this study included only GP-related interventions, the findings will be applicable to other primary care professionals (e.g. nurses), in healthcare systems where the patient consultations are not limited to GPs (Malo et al., 2016). Moreover, taking into account that HPV vaccination of girls aged 11–14 years often seems to be delayed to older ages and that sexuality always seems to be connected to this vaccination whatever the age of the girl, one option to facilitate HPV vaccination would be to increase the age of the target group to 15–18 years. This is supported by recent scientific evidence obtained by our group showing efficacy of HPV vaccination for this age group with a two-dose schedule (Basu et al., 2019), but should be balanced against the possibility of missing opportunities to vaccinate before sexual debut among early initiators.

Our study has the usual limitations of synthesising evidence from retrospective surveys based on self-reporting, including recall bias, social desirability bias, and wishful thinking bias. It should be highlighted that taking into account the nature of the population surveyed, i.e., GPs, social desirability and wishful thinking may be particularly important. The included studies covered a twelve year period (Table 1), over which GP's cognitions regarding HPV vaccination might have changed. This was not formally assessed in our analysis.

The quality of this review can be acknowledged regarding the accuracy of the literature research and the use of recommended review guidelines and tools to assess bias. Compared with other literature reviews that explore cognitions related to HPV vaccination and focus on attitudes and knowledge, our review goes further by (a) using the RAA theoretical framework (b) addressing methodological differences between studies, (c) providing an overview of the impacts of all these cognitions on vaccination behaviour, and (d) focusing on only one type of health providers, i.e., GPs. The results of this review could easily be used and adapted in countries that have similar health policies and similar HPV vaccination issues.

Conclusion

To our knowledge, no GP-level interventions have been successful in improving HPV uptake rates in France. Our results will inform the development of a professional educational intervention on HPV vaccination targeting GPs in France. HPV vaccination is one of the main pillars of the recently launched WHO initiative to eliminate cervical cancer around the world and removing barriers to vaccination in different settings and from different perspective including health professionals is of paramount importance for success.

Funding information

P. Villain/IARC obtained funding from the Institut National du Cancer, France (INCa, grant nr. 2016-009) and Métropole de Lyon / Cancéropôle Lyon Auvergne-Rhône-Alpes, France (CLARA). The sponsors had no involvement in the study design, collection, analysis and interpretation of data, the writing of the manuscript, or the decision to submit the manuscript for publication.

Disclaimer

Where authors are identified as personnel of the International Agency for Research on Cancer / World Health Organization, the authors alone are responsible for the views expressed in this article and they do not necessarily represent the decisions, policy or views of the International Agency for Research on Cancer / World Health Organization.

Declaration of Competing Interest

None to declare.