Communicating effectively with patients about vaccination: A systematic review of randomized controlled trials

Background Good communication between healthcare professionals and their patients is essential to enlighten the benefits and risks of vaccination. Despite the availability of effective vaccines, reluctance prevails, sometimes fuelled by sub-optimal communication leading to a lack of trust. An evaluation of the effectiveness of a communication strategy for which healthcare professionals are trained has yet to be carried out. Objective Systematic review of studies with a randomized controlled trial (RCT) to define and evaluate the impact of healthcare professionals’ communication on patients’ vaccine adherence. Methods We performed a structured search on Medline, Embase, CENTRAL, PsycINFO and CINAHL. The studies selected include those involving healthcare professionals authorized to administer vaccines according to Canadian guidelines. Primary outcomes include vaccination rate or vaccine hesitancy rate. Results Nine articles were included. Five studies (n=5) reported intervention effectiveness according to vaccine adherence. The results are largely represented by parental vaccine hesitancy for human papillomavirus (HPV) or childhood vaccination, while three studies (n=3) target the general population. The risk of bias relative to the studies is either low (n=7) or of some concern (n=2). Conclusion The effectiveness of communication varies according to the studies and knowledge acquired through training. Future studies will need to examine communication with healthcare professionals in order to establish a consensus on optimal and appropriate training.


Introduction
Vaccination is effective in preventing many diseases and their serious forms. However, some patients are reluctant to be vaccinated, despite the potentially harmful consequences for their health and that of the population as a whole. This hesitancy stems from multiple, complex and sometimes interconnected factors (1)(2)(3)(4)(5)(6)(7). Possible reasons include a lack of trust in healthcare professionals and institutions, healthcare professionals' lack of patient communication skills (4,5,7), or difficulties in navigating the sometimes contradictory information available (1)(2)(3)5).
Physician-patient communication is defined in the literature as a key component of the therapeutic relationship, enabling the development of a bond of trust that leads to optimal care (5,(7)(8)(9). The bond of trust is important when discussing vaccination, since the decision-making process has an impact on individual and community safety (1). Given the importance of communication in healthcare decision-making, it is possible that a communication intervention with healthcare professionals could influence vaccine adherence. Given the coronavirus disease 2019 (COVID- 19) pandemic and its repercussions, including the lack of educational resources in patient communication skills, a communication intervention is all the more important to address the limitations of healthcare institutions and mistrust of the COVID-19 vaccine. In the absence of intervention, current limitations may lead to mistrust of future vaccines in times of health crisis. The effectiveness of intervention has yet to be systematically evaluated.

Objectives
We conducted a systematic review of randomized controlled trials (RCTs) to define and evaluate the impact of healthcare professionals' communication on patients' vaccine adherence.

Protocol and registration
This systematic review was conducted in accordance with AMSTAR 2 (A Measurement Tool to Assess Systematic Reviews) standards (10) and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (11). The protocol has been registered with the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42022330645).

Eligibility criteria
All RCTs in which participants were healthcare professionals authorized to administer vaccines (doctors, nurses, pharmacists and resident physicians) were eligible. We included studies in which communication on vaccine adherence was the main intervention. We excluded studies in which the healthcare professionals were medical, nursing or pharmaceutical students (not authorized to administer vaccines according to Canadian guidelines). We also excluded studies where the intervention was aimed at patients rather than healthcare professionals. Non-peerreviewed articles, conference abstracts, letters, editorials and commentaries were not eligible.

Information sources
Two electronic search reviews (12)

Search
The search strategy (Supplemental material A) was developed by an information specialist with the research team and revised by a second information specialist as suggested in the Peer Review of Electronic Search Strategies (PRESS) guide (12). Eligibility criteria (Box 1) included no language or publication date limits. A filter for published RCTs was applied (13). The search strategy was developed in Medline and then translated into the other databases. Key search concepts included MeSH terms related to vaccine adherence, healthcare professionals and communication. Only studies published and available in French or English were considered. The list of references cited in the included studies was also searched. The final list of included studies was reviewed by content experts to confirm their relevance.

Selection of studies
Studies were uploaded to a web-based software program, Covidence (version 2.0, Veritas Health Innovation, Melbourne, Australia) (14), and duplicates were removed. A pilot assessment tool, developed by the research team and tested on 30 randomly selected articles (Supplemental material B), was refined until subjectively acceptable agreement was established among the judges. Evaluation of each level of inclusion was carried out by pairs of independent reviewers, and conflicts were resolved by a third party.
Population: healthcare professionals authorized to administer vaccines (physicians, nurses, pharmacists and resident physicians) Intervention: communication training for healthcare professionals to be used during vaccination consultations only Comparison: a control group of healthcare professionals who received no communication intervention Outcome: vaccine adherence, defined as receiving, intending to receive or being less reluctant to receive the series of disease-preventing vaccines according to the schedule suggested by the national immunization authority Study date: no limit Method: randomized controlled trial Publication language: no initial limit Publication date: no limit Box 1: Search strategy eligibility criteria

Data extraction
A data extraction grid (Supplemental material C), developed by the research team, was tested by the same reviewers. Extraction was performed in duplicate by pairs of reviewers and consensus by a third party. Extracted data include publication characteristics (name of lead author, year of publication, data collection sites), study characteristics (objective, study design and context, number of healthcare professionals, outcomes), type of healthcare professional, intervention details and results.

Risk of bias inherent in each study
Pairs of reviewers assessed included studies for risk of bias according to the Risk of Bias Tool 2 for Randomized Controlled Trials (RoB 2) (15). The tool assesses the risk of bias attributed to study design, conduct and data reporting. For each area, a questionnaire is used to establish the level of risk as "low," "some concern" or "high." All areas must be predominantly low risk for the study to be considered reliable (15).

Data summary
A description of all included studies is presented in tables containing information on demographic, clinical and methodological quality. The results are summarized qualitatively, given the heterogeneity of the included studies.

Selection of studies
The search identified 6,484 studies. After eliminating duplicates, 4,014 studies were assessed for eligibility, including 57 full-text articles, 48 excluded studies and 9 included studies ( Figure 1).

Characteristics of selected studies
The included studies (n=9) employed communication training in a variety of formats targeting different knowledge areas, including understanding the virus, how the vaccine works, assertive communication, effective recommendations and the patient perspective. The vaccination context was childhood diseases (n=2), pneumonia/influenza (n=3), or human papillomavirus (HPV) (n=4). Six studies (16-21) focused on parental vaccine hesitancy, and three on adult vaccine hesitancy (22)(23)(24). General characteristics are shown in

Summary of results
Among the studies (n=9) included, the effectiveness of the interventions varied greatly according to the training format (5 effective (17,18,21,22,24); 3 no significant difference (16,19,20); 1 ineffective (23)). A descriptive analysis of the communication adopted and its results are presented below. The measurement tools, primary outcomes and results with statistical significance are summarized in Table 2.

Effectiveness of communication training
Effective training First, we note some training courses that proved effective in the HPV context. These included educational resources and patient-adapted recommendations. Following a self-guided webinar and two group sessions (18), the application of motivational interviewing during physician-patient interactions improved HPV vaccine adherence in adolescents. Similar training consisting of a webinar with three interactive modules and weekly encouragement to reveal common patient questions also improved vaccine adherence (21).  We also observed that good physician-patient communication includes a good understanding of the virus, the vaccine and the reasons for vaccine hesitancy. The study by Muñoz-Miralles et al. (24) shows a positive effect in patients aged 60 and over following a brief standardized intervention in the context of influenza. Although this communication depended on a directive guide, doctors and nurses were encouraged to adapt their communication by using empirical evidence to address the reasons for vaccine hesitancy, gathered beforehand.
This example can be enriched by the intervention proposed by Abdel-Qader et al. (22), who integrated the patientpartner perspective into the training material. The training, organized in 16 virtual sessions in a private Facebook group, invited pharmacists to be trained by eight doctors and eight pharmacists. However, the training sessions particularly included testimonials from patients discussing their experiences with the health crisis and vaccination. The patient-partner perspective justified the importance of patient-adapted communication. This study shows a significant reduction in vaccine hesitancy and an increase in vaccination rates. It should be noted, however, that the self-reported results of this study may be biased.
Training courses based on assertive communication cannot be overlooked. Brewer et al. 's study (17) demonstrated improved HPV vaccine adherence using an announcement, i.e., a vaccine recommendation given on the day of the consultation. The same study also evaluated the effectiveness of a conversation with the patient to present the vaccine for shared decision-making, but this intervention noted no significant difference.

Risk of relative and cross-study bias
Seven studies (16)(17)(18)(19)21,23,24) have low risk and two studies (20,22) are of some concern (see Table 3). A follow-up bias is present, as the healthcare professionals would have been aware of the result of randomizing to an intervention or control group. We consider this risk unavoidable, based on ethical considerations of informed consent, despite the fact that it may have had an impact on study results. The second bias (20,22) (measurement bias) is taken into account, since self-reported surveys were used, which can influence the validity of the results.

First author, year
Results measurement tool(s) Name of primary outcome(s) Conclusion of primary results Henrikson, 2015 (20) Mother's score on the "Parental attitudes to childhood vaccines" test.
Childhood vaccines by PACV percentage of mothers reluctant to vaccinate.
Six single-item self-efficacy questions on communicating with parents about childhood vaccines (email survey).
The intervention had no effect on the mother's vaccine hesitancy (p=0.78). Adjustment for baseline PACV score and race yielded similar results Proportion of adolescents receiving HPV vaccination.
The rate of missed opportunities decreased in intervention vs. control practices by 6.8% (95% CI: 3.9-9.7) for HPV vaccination initiation. No significant difference was observed for subsequent vaccination.
The rate of missed opportunities decreased between the start of the study and the intervention period by 2.4% (95% CI: 1.2-3.5) in intervention vs. control practices.
For adolescents with at least one office visit during the intervention period, HPV vaccine initiation was 3.4% (95% CI: 0.6-6.2) higher in intervention vs. control practices. No significant difference was observed for subsequent vaccination.

Interpretations
The effectiveness of interventions does not seem to depend simply on the presence of communication that adopts epidemiological and medical knowledge, but also on communication that is adapted to the patient, understanding the factors that influence the vaccination decision. The most effective interventions (24,25) focused on HPV and targeted parents of minor patients. These studies have potentially been built on a better understanding of parental vaccine hesitancy, since the reasons for vaccine hesitancy and HPV have previously been addressed through research, improved communication and the development of quality recommendations (25). An adapted intervention, such as motivational interviewing (18), is consequently viewed favourably in the literature and by healthcare professionals (6,(26)(27)(28) (29)(30)(31), several socio-political factors have prompted a change in the public's approach to vaccination (32). Social and medical perception seems to depend on multiple variables including ideology, customs, understanding of health, collective responsibility, trust and accessibility to healthcare (33).
Given the complexity of vaccine hesitancy, we would like to hypothesize that effective communication must take into account the above variables. The literature points to the inefficiency of a universal algorithm. In 2015, a systematic review on vaccine hesitancy demonstrated the need for a call for strategies tailored to the target population, the reasons for hesitancy and their context (34). We note in particular that effective studies tended to form recommendations with subjectivity according to the patient's concerns, but the integration of all these variables remains to be applied to establish a bond of trust with patients. Further socio-culturally adapted communication interventions would be needed to study this topic.

Limitations
There are several limitations to note. Other diversified studies would have enabled a better scope of conclusions, as well as a meta-analysis to understand the relationship between different groups of healthcare professionals, different diseases and vaccines, and then different communication training. Studies may be missing given the broad scope of the search strategy, the exclusion of articles published neither in English nor French, and the fact that only studies involving healthcare professionals authorized to administer the vaccine in Canada were included. Some studies also included different clinical locations and determining variables that may have been ignored or absent, such as regional infection rates, the context of the intervention (e.g. a national or regional vaccination program) and the demographics of specific patient groups. RCTs only were included in the study because of their rigorous methodology. It would also have been possible to include cohort studies with the same type of intervention.

Conclusion
The effectiveness of vaccination-related communication varies according to the studies and knowledge acquired through training. This systematic review confirms the need for studies that focus on communication with healthcare professionals to build consensus around optimal, tailored training that increases trust in healthcare institutions. There is thus a need for studies that take into account initiatives that include the patient perspective in communication with healthcare professionals.
Author's statement CD -Participation in study design, writing-original draft, data acquisition and evaluation, data analysis, writing-revision and editing, final approval MD-L -Participation in study design, writing-revision and editing, final approval CPC -Data acquisition and evaluation, writing-revision and editing, final approval MR -Data acquisition and evaluation, writing-revision and editing, final approval TAG -Data acquisition and evaluation, writing-revision and editing, final approval P-MD -Data acquisition and evaluation, writing-revision and editing, final approval JLH -Participation in study design, writing-revision and editing, final approval SF -Participation in study design, writing-revision and editing, final approval MC -Participation in study design, writing-revision and editing, final approval NL -Search strategy development, writing-revision and editing, final approval SB -Participation in study design, development of search strategy, data analysis, writing-revision and editing, final approval

Competing interests
No conflicts of interest have been declared.

Funding
This systematic review was funded by the Public Health Agency of Canada.

Supplemental material
These documents are available in the Supplemental material file.
Supplemental material A: Search strategy Supplemental material B: Effective communication strategies Supplemental material C: Data extraction grid