Gender Barriers to Knowledge Transfer and Exchange Among Vaccine Researchers in Low-, Middle- and High-Income Countries – An International Cross-Sectional Study in 44 Countries

Background: Globally, women constitute 30% of researchers. Despite an increasing proportion of women in research, they are still less likely to have international collaborations. Literature on barriers to knowledge transfer and exchange (KTE) between men and women, remain limited. This study aimed to elucidate perceived gender barriers to KTE activities in vaccination related research in low, middle- and high-income countries. Methods: This was a cross-sectional data from a self-administered questionnaire distributed to researchers in the eld of vaccination research. Structural factors to KTE were assessed using 12 statements measured with 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). An index ranging from 12 to 60 points was created to assess structural factors to KTE, with higher score indicating higher perceived barriers. Linear regression modelling was applied to examine the association between KTE barriers and gender. Results: Regardless of gender, researchers experienced challenges with respect to KTE activities, particularly factors related to the availability of human and nancial resources, and level of technical expertise among their target audience. We were also able to identify perceived facilitators among men and women, such as the presence of structures that link researchers and target audiences, the investment of target audiences in KTE efforts and the presence of stable contacts among target audience. Our linear regression analysis showed that women have a reduction of 1.069 in their scores in comparison with men (R 2 =0.014; B=-1.069; 95%CI -4.035;1.897). Conclusions: Men and women shared common perspectives on barriers to KTE. KTE activities could be strengthened by improving structural efforts to reduce gender differences and increase collaborations between researchers and their target audience.


Background
As one of the most cost-effective interventions in global health (1), vaccination has prevented approximately three million deaths annually (2).
Despite the bene ts in decreasing mortality and reducing costs to healthcare providers and families, inadequate vaccination remains a challenge in low-and middle-income countries (LMIC) (3).
Even though an estimated number of 1.5 million deaths could be prevented through overall improvement of global vaccination coverage, the global burden of vaccine-preventable diseases (VPD) remains high (2). To address these challenges, an advancement in evidence-informed decision by bridging research, policy and practice is warranted (4).
Knowledge Transfer and Exchange (KTE) can be used to link the three pillars of research, policy and practice in which knowledge can be converted into policy through an iterative process (5). Additionally, the process is an exchange of knowledge between research producers and research users (6). KTE related activities can be classi ed into four models, namely push efforts (7,8), pull efforts (9,10), exchange efforts and integrated efforts (7). Push efforts is de ned as the efforts of knowledge dissemination where researchers seek policy-relevant research questions (7,8), while pull efforts is de ned as seeking information in order to support decision makers in developing informed choices (9,10).
With respect to exchange and integrated efforts, these are de ned as collaboration between various actors and cooperation of different stakeholders to conduct KTE-related activities (7). Context plays a vital role in the effectiveness of KTE processes. For instance, in LMIC, evidence generated at a local level is often ignored due to power imbalances or inadequate data quality (11). Therefore, translation of extensive knowledge into relevant policies remains underutilized (12). The documentation of KTE activities in the eld of vaccination research remains limited as well (13)(14)(15). This could be due to having integrated KTE activities on the national policy and international initiatives levels to the point that the concept is not visible per se (16), or alternatively that KTE activities have not been carried out due to inadequate structures to support KTE (11).
Structural barriers to KTE activities among health researchers, such as limited access to databases and research ndings, nancial limitations, limited administrative and infrastructural capacity, and emergence of other priorities within the health system have been identi ed in previous studies (17)(18)(19)(20)(21). However, little has been documented on structural barriers to vaccination related research.
While there are some studies on structural barriers that health researchers face in KTE, literature on structural barriers to KTE between men and women, remain limited as well (22). Women constitute 30% of the world's researchers (23). Despite an increasing proportion of women researchers globally, they are still less likely to collaborate internationally (24). Men publish more research papers on average than women (24).
Men are also more represented when it comes to rst authorships. For every 1 article with a woman as rst author, there are around 1.93 articles rst-authored by a man (25).

Methods
This study aimed to elucidate perceived gender barriers to KTE activities in vaccination related research in low, middle-and high-income countries.
Our speci c objectives were to (1) compare perceived structural barriers and facilitators to KTE activities among men and women in vaccinerelated research; and (2) investigate the association between gender and structural factors to KTE activities in low, middle-and high-income countries.
This study was based on a cross-sectional data from a self-administered questionnaire distributed to researchers between 28 March and 22 April 2018. The questionnaire was developed and validated by the World Health Organization (WHO) and McMaster University, Canada (26).

Recruitment of participants
Participants were recruited based on identi cation of vaccination-related articles obtained from PubMed using the search terms "(vaccinate* [MeSH Terms]) OR (immunize* [MeSHTerms])". The articles were screened for publication between 1 January and 31 December 2017, availability of abstract and unique e-mail addresses, included human subjects, and were written in English language.
Based on these criteria, articles were included if the study population included children (< 18 years) or those in the proximity (e.g. parents, paediatrics, policy/programs targeting children); conducted quantitative or qualitative analysis; and systematic reviews. Additionally, articles were excluded if articles were based on opinions, comments or a case report; did not discuss VPD; did not include human subjects; was not written in English; and if e-mail of corresponding author was not provided.
Authors were invited to participate in the study via the e-mail addresses obtained from the articles identi ed as relevant to the topic. In order to increase the response rate, reminders were sent on several occasions during a one-month period (once per week during the rst two weeks; twice a week during the third week; daily during the fourth week).
During the recruitment process, a total of 717 researchers were identi ed and invited to participate in the survey. Of these, we included authors who had valid e-mail addresses, provided consent, and conducted research in vaccination-related eld. This resulted in a total number of 158 participants ( Fig. 1).

Data collection and measures
Structural factors to KTE were assessed using 12 statements measured with 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). The statements included: 1.
The cost for translating research on the health topic into action was very low 2.
KTE activities could be paid for through research grants for which I was eligible to apply 3.
Structures and processes existed to link researchers and your target audiences

4.
Personal and organizational contacts among your target audiences were quite stable over time (e.g., low turnover among representatives and/or members of your target audiences)

5.
Perceived crises in the health system drew attention away from research on the health topic 6.
Target audiences lacked the expertise for translating research on the health topic into action 7.
Target audiences had access to technical support for translating research on the health topic into action 8.
Target audiences created opportunities to develop joint research initiatives with them 9.
Target audiences did not make decisions about the health topic on the basis of research 10. Target audiences invested nancial and/or human resources in joint research initiatives 11.
Target audiences created events for knowledge transfer and exchange related to the health topic (e.g., forums that bring researchers and target audiences together for discussion)

12.
Target audiences invested nancial and/or human resources in knowledge transfer and exchange activities (e.g., hired staff to identify and make available relevant research).
The survey included gender (men; women), year of birth, country of primary a liation, education (medical doctor; bachelor's degree; master's degree; doctoral degree), and area of specialization (biomedical research; population and public health; clinical research; other). Based on the participants' country of primary a liation, the countries were divided into two income levels based on the country's Gross National Income (GNI) per capita, according to the de nition of the World Bank in 2018. Countries were categorized as LMIC if their GNI per capita was below or equal to 12,235 USD, while countries were categorized as high income countries (HIC) if their GNI per capita was above 12,235 USD (27). In addition, age was calculated as the difference between '2018' and 'year of birth'.

Statistical analysis
We tested the 12 statements for internal reliability using Cronbach's alpha. The results showed some inconsistencies between the statements (Cronbach's alpha = 0.504). In order to have consistent measure of the items, the following items were reverse coded: 'Target audience lacked the expertise for translating research on the health topic into action'; 'Target audience did not make decisions about the health topic on the basis of research', increasing the internal consistency of the 12 statements (Cronbach's alpha = 0.71). Further, we created an index with the 12 items, ranging from 12 to 60 points for KTE barriers, in which a lower score indicated more frequent experiences related to structural barriers regarding KTE activities.
To describe the study population, we computed descriptive statistics using Fisher's exact test for the variables 'country of primary a liation', 'research specialization' and 'educational attainment'. For the variable 'structural factors to KTE', Mann-Whitney test was carried to compare differences among men and women.
Linear regression analysis was applied to test the association between perceived KTE barriers and gender. In our model, the outcome (dependent variable) was continuous variable on KTE score and our independent variable was gender. We also included age and country of primary a liation as covariates. The variables gender and country of primary a liation were treated as binary variables. Dummy variables were created and coded as follows male = 0, female = 1, HIC = 0 and LMIC = 1. Age was included as a continuous variable. Results are presented using beta coe cients and 95% con dence intervals (95% CI). We considered alpha p < 0.05 statistically signi cant. All statistical analyses were computed using SPSS no statistically signi cant differences were observed (Table 1). In our study we found structural factors perceived as barriers to KTE among men and women. More than half of men (n = 14; 53.9%) and women (n = 15; 53.6%) did not perceive their target audience to invest human and nancial resources in KTE activities or in joint research initiatives (n = 11; 42.3% and n = 17; 60.7%). Another perceived barrier among men (n = 13; 50%) and women (n = 16; 57.2%) was lack of expertise among their target audience to translate research into action. In addition, 34.6% of men (n = 9) and 42.8% of women (n = 12) thought that their target audience lacked access to technical support to translate research into action. High costs for translating research into action was also a common perceived barrier among men (n = 12; 46.2%) and women (n = 14; 50%). Some structural factors were perceived differently according to men and women.
While less than a quarter of men (n = 11; 42.3%) perceived that research grants for KTE activities were available, about half of the women disagreed (n = 14; 50%). Less than half of men (n = 11; 42.3%) perceived crises in the health system as a barrier to KTE whereas, 42.9% of women (n = 12) disagreed (Table 2a). We were also able to highlight some structural factors that men and women perceived as facilitators to KTE. About half of the men (n = 13; 50%) and women (n = 18; 64.2%) perceived the presence of stable contacts among their target audience as a facilitator to KTE. The presence of structures linking researchers to target audiences was also perceived as a facilitator among men (n = 12; 46.1%) and women (n = 12; 42.9%). The engagement of target audiences in KTE activities was perceived by some (n = 12; 46.1%) and women (n = 11; 39.3%) as a facilitator (Table 2b.).

Index on KTE structural Factors
The mean score of the variable structural factors to KTE was highest among men a liated with HIC (33.94; min:23-max:42), and lowest among women a liated with LMIC (33.0; min:30-max:35). In our sample, women a liated with HIC (32.88; min:21-max:42) scored higher than women a liated with LMIC but lower than men a liated with LMIC (33.78; min:27-max:40). There were no statistically signi cant differences among the mean scores calculated (p = 0.09) (Fig. 2).

Association between perceived structural barriers for KTE activities and gender
A multivariable linear regression analysis was done to test if gender signi cantly predicted scoring. The results indicated that women have a reduction of 1.069 in their scores as compared to men (R 2 = 0.014; B=-1.069; 95%CI= (-4.03;1.89) but the difference was not statistically signi cant (Table 3).

Discussion
This study aimed to compare and investigate the association between perceptions of men and women to structural barriers and facilitators to KTE activities. We were not able to identify signi cant differences in the distribution of respondent characteristics.
The mean scoring of men and women shows that researchers regardless of gender and country of a liation experience some challenges when it comes to KTE activities. Men and women perceived factors related to availability of human and nancial resources as well as level of technical expertise among target audience as barriers. We were able to identify perceived facilitators among men and women, such as the presence of structures linking researchers and target audiences, the investment of target audiences in KTE efforts and the presence of stable contacts among target audience. To some extent these ndings are in-line with studies conducted in other settings (28)(29)(30).
When we ran our linear regression model, we could see a relation, despite not being signi cant, between gender and a reduction in score. While we were not able to identify speci c studies tackling gender barriers and KTE, it may be explained through the literature exploring challenges that women face in academia (23,24,31). A previous study indicated that men faculty members were able to positively engage in research and obtain funding while women found it harder to balance between undergraduate education and research efforts (32). Women not only publish less than men, they also still face challenges to get to decision making positions. Despite having more women in mid-management positions, men still dominate executive positions and full professor positions globally (24). This can be clearly seen in the eld of health, where women still represent lower cadres of health workers despite being about 75% of the global health workforce (24).
In our study, papers on barriers to KTE, that have been conducted in different settings, were available (28-30, 33, 34). However, we were not able to nd studies tackling barriers to KTE in vaccination and gender. To our knowledge, our study presents a novel idea in the eld of KTE and vaccination. It investigates KTE from the perspective of the research community, drawing comparisons between genders. It also contributes to the growing research on KTE in vaccination, which may still be considered as limited. Our study may also serve as a foundation to future research as it encompasses a global scope.
We can identify some limitations in our study as well. Respondent withdrawal and missing data were observed in many sections across the questionnaire. Despite having 158 respondents in total, 65.8% did not respond to questions related to structural factors. We reason the lengthy questionnaire to be a contributing factor to this. Another limitation in our study may be attributed to our selection process. The study excluded researchers who did not speak English and/or published in grey literature due to our questionnaire administration criteria. The questionnaire also included only 2 options for the variable gender. While we did not have missing data for this variable, it may have been more convenient to not limit gender identities to man and woman only.

Conclusion
This study was not able to highlight statistically signi cant differences between men and women when it comes to vaccine related KTE. Men and women shared common perspectives on barriers to KTE. The ndings of this study show that more efforts on a structural level need to be carried out to strengthen KTE activities. Based on the results, it is important to invest in nancial and human resources in KTE activities. These efforts should not be the sole responsibility of researchers. The target audience and decision makers need to be more engaged in strengthening the implementation of KTE activities. Future research may examine the barriers and facilitators to KTE at an organization level. For this study we received approval from the Stockholm regional ethics committee (Etikprövningsnämnden), Sweden (ID#: 2018/219-31).

Consent for publication
Informed consent was sought from all participants before they could ll in the survey. Respondents who did not ll in the informed consent were not included in the analysis.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Competing interests Figure 1 Schematic overview of recruitment of participants Score on KTE structural factors among men and women in HIC and LMIC