Science Shops as key intermediary structures to respond to the current health research agenda bias: Evidence from the InSPIRES project

Abstract Introduction To increase the likelihood of research responding to societal needs, intermediary structures such as Science Shops are being created. Science Shops respond to research needs identified and prioritized through participatory processes involving civil society. However, these are not mainstream structures, and most research needs addressed by the scientific community are not defined by a diversity of stakeholders (including citizens) but are mostly prioritized by researchers and funders. Literature shows this often leads to bias between the research topics investigated and the research needs of other relevant stakeholders. This study analyses how 14 Science Shops contribute to decreasing bias in health research agenda setting. Methodology We compare the research priorities identified through participatory processes by the Science Shops, which participated in the European Union‐funded project InSPIRES (2017–2021), to the available research addressed in the literature (identified in Web of Science), which we use as a proxy for current research priorities. Results Science Shop projects contributed to decreasing the existing bias in health research agenda setting: (1) between drug and nondrug treatments and (2) between clinical trials of treatments for illnesses affecting high‐income versus middle‐ and low‐income countries, which leads to a lack of local strategies for high disease burdens in nonhigh‐income regions. Conclusion This study provides the first evidence of Science Shops' effectiveness in addressing current biases in health research agenda setting. We conclude they could play a key role in shaping local, national and international research policies.


| INTRODUCTION
To increase the likelihood that research will have an impact on society, intermediary structures are being created.One such structure is the Science Shop, which provides 'independent, participatory research support in response to concerns experienced by civil society'. 1,2Science Shops respond to research needs identified and prioritized through participatory processes involving civil society.
Science Shops emerged in The Netherlands in the 1970s, facilitated by a favourable political context that allowed university staff and students to start partnerships with civil society organizations (CSOs).Students, supervised by senior researchers, undertook research projects for free upon CSOs' requests. 3,4In the 1980s, the Dutch Science Shop movement inspired other European countries.The establishment of those Science Shops was enhanced by the environmental movement, which motivated collaborations with emerging university departments in environmental sciences.The Science Shops concept gained further traction in the 1990s, driven by the Information and Communication Technologies revolution, which promoted a shift towards a knowledge-based economy with emerging debates about the active role that society should have in science, moving away from the industrial economy in which society was just a receiver of scientific knowledge.The European Commission (EC) funded several projects on Science Shops at this time.In the late 1990s, the movement spread in eastern and central Europe 3 and outside Europe to countries such as South Africa. 5er the past decade, renewed attention to Science Shops has emerged with innovative approaches being used around the world.These efforts have been supported by international organizations such as the EC, which has funded new research and innovation (R&I) projects to*** promote more open and inclusive approaches influenced by different science policies.7][8][9] Science Shops differ in the activities they conduct, how they are organized, in which context and fields they work, and how they operate. 4However, although no dominant organizational structure defines a 'Science Shop', they all aim to respond to research topics identified and prioritized with a bottom-up approach. 1spite the growing interest, their role is currently relatively small as few higher education institutions have a (well-functioning) Science Shop.Although participatory methods are increasingly used in research, most research today still addresses topics identified and prioritized by researchers and funders (public and private) that are not necessarily aligned with the research needs of a broader community of stakeholders (including citizens). 10,11This may result in what Knottnerus and Tugwell 12 describe as a research agenda bias in clinical research.They point out that this bias is reflected in the 'striking disbalance' between efforts in: (1)  (3) proportion of research and burden of diseases, both within highincome countries and between high and nonhigh-income regions, such as common infectious diseases and neonatal disorders in South Asia and (4) drug research versus research on related healthcare policies such as the availability of essential medicines in low-income countries. 12bias in health research is also described by Crowe et al., 10  were asked to identify their priorities, they wanted more rigorous evaluation of the effects of physiotherapy, surgery, education, and coping strategies rather than more studies of drugs.
If we do not address the current research agenda bias, which is leading to key unmet needs regarding nondrug treatment research, we will not be able to achieve integrated care.Since integrated care seeks to better co-ordinate care around people's needs, it is important to also align research agendas with those needs.One solution to address the research agenda bias and to respond to 'the much too often hidden, implicit international and national research agendas' 12 is to involve different types of stakeholders in the research agenda-setting process.
Science Shops are one of the intermediary structures responding to research needs identified and prioritized through the participation of stakeholders. 2,3,9,13Structures that follow more open and inclusive approaches use different multistakeholder participatory methods to set research priorities after identifying the challenge.5][16][17][18][19] They also have different approaches for identifying, recruiting, retaining and ensuring inclusion of diverse and representative stakeholders with multiple interests and concerns in safe and cooperative arrangements and for integrating (transdisciplinary) knowledge and dealing with the complexity of challenges. 18,20Once research priorities are defined, they reformulate the research question and design and implement participatory research projects with methods such as citizen science or community-based participatory research 21 'while also continuing to rely on approaches wherein problems are solved by practice of combining and adapting existing knowledge from different sources without the "scientific research" dominating the process'. 9wever, little is known about the effectiveness of these approaches in addressing research agenda bias.
Our study analyses whether Science Shops contribute to decreasing bias in health research agenda setting.We compared the research priorities identified through participatory processes by 14 Science Shops that participated in the European Union (EU)funded project InSPIRES and the available research addressed in the scientific literature indexed in the Web of Science (WoS), which we use as a proxy for current research priorities.

| METHODOLOGY
The research was conducted within the framework of the EU-funded project InSPIRES (Ingenious Science Shops to promote Participatory Innovation, Research and Equity in Science, https://www.inspiresproject.com/)which ran from 2017 until 2021.The project aimed to bring together different stakeholders across and beyond Europe to co-design, jointly pilot, implement and roll out innovative models of Science Shops with participatory approaches.This study analyses data from projects conducted by the 14 Science Shops that participated in the InSPIRES project and from publications in WoS during the same period conducted worldwide and in specific countries where the Science Shops were located and/or performed their research.
The Science Shops were located in Benin, Bolivia, Ecuador, France, Greece, Hungary, Italy, The Netherlands, Romania, Spain, Tunisia, Turkey and Uganda (13 countries and 14 Science Shops with 2 Science Shops located in Spain), and some of them performed their research in other countries, such as Nepal and Ecuador (where no local Science Shops were involved).The Science Shops were committed to identifying and prioritizing research topics through participatory processes, as described in the framework for Science Shop processes developed within InSPIRES. 9Eight of the Science Shops were InSPIRES partners and they were located in: Europe (Spain, The Netherlands, France, Hungary, Italy), Africa (Tunisia) and South America (Bolivia).The other six were associated partners recruited through an open call, and were located in Europe (Greece, Romania, Turkey), Africa (Benin, Uganda) and South America (Ecuador).The next sections describe the methods for the comparative analysis of Science Shops projects with the publications indexed in WoS, applied within two different phases: (1) data collection and selection of Science Shop research projects and (2) analysis of research fields.

| Phase 1: Data collection and selection of science shop research projects
The data set included basic descriptive information from each Science Shop project.Table 1 shows the categories of data collected for each project: (1) basic descriptive information, (2) characteristics of the process for identifying and prioritizing research needs (e.g., number and diversity of stakeholders involved, participatory methods applied, level of participation achieved), (3) topic (e.g., human immunodeficiency virus [HIV], Chagas, Leprosy), (4) subtopics addressed (e.g., HIV treatment, HIV-related stigma) and ( 5) research field as defined by NordForsk, a funding agency under the Nordic Council of Ministers, which produced bibliometric analyses of the Nordic countries on a regular basis during the period 2009-2019 15 (e.g., Social sciences, natural sciences, humanities, engineering).Each Science Shop provided this data in a shared file delineating the data categories and a description of each of the categories.The level of participation was categorized according to Arnstein's Ladder of Citizen Participation 16 (Box 1) which includes participatory and nonparticipatory levels that were useful to be able to exclude the projects that applied nonparticipatory methods.Bilateral online meetings were held with Science Shop representatives to ensure that the instructions for collecting the data were clear.
To demonstrate that Science Shop research projects addressed research needs identified and prioritized through participatory processes, characteristics of the process to identify research needs (Category 2 in Table 1) were analyzed.A descriptive analysis was conducted defining the percentages of types of participatory methods used in the Science Shops, the diversity of stakeholders involved during the identification phase, and the level of stakeholder (non-)participation (cf.Arnstein's Ladder of Participation).
Finally, three data sets were defined with different inclusion criteria, as described in Table 2 and Figure 1.

| Phase 2: Analysis of research fields
In this phase, we conducted a bibliometric analysis comparing the research fields in the selected projects in Data Sets 2 and 3 with research fields in the publications in the WoS database.WoS is the world's largest and most authoritative database for studying publication output, 24 covering 34,000 journals across all fields of science, although its coverage of Arts, Humanities and Social sciences is less complete as its coverage of Science, Technology, Engineering and Mathematics and medical fields.Similarly, the database has an overrepresentation of English-language journals, 25,26 and thus an underrepresentation of, for example, Spanish-language journals, 27 although in recent years, many South and Latin American journals have been indexed in WoS.Despite these shortcomings, it is still the most used database for studying publication output in the Social sciences as well as other journal-oriented fields, as very few countries have national databases with complete coverage of all Social sciences output.
First, we identified all research publications related to the topics HIV/AIDS, Chagas and Leprosy in 2017-2020.We searched WoS article titles, abstracts and keywords using the search terms specified in Table 3.The search terms were considered unique for each specific disease: for example, when mapping papers about Chagas, the search terms used included 'Chagas' and the drug names 'benznidazole' and 'nifurtimox', the main treatments for this disease.These search terms were defined with the support of two clinical specialists.For subtopics, we used the same keywords as for the topics, adding the condition to also look for keywords related to the specific subtopic (e.g., 'stigma' and 'discrimination' in the case of HIV-related stigma).
Second, we identified all research publications related to 'Health': the search terms used were 'Health', 'Medicine' and 'Medical'.This analysis was only carried out at the global level.However, for the topics, it was also carried out for the countries where the selected projects were implemented and for the countries where the organization coordinating the project was located: HIV/AIDS in Spain,

Number and diversity of stakeholders involved
Because the degree of different stakeholders' involvement varies according to the type of social demand, 9 we collected data on the diversity of stakeholders involved in the identification of research needs (e.g., researchers, policy makers, education community or business and industry representatives).

Level of stakeholder participation
Description of the level of participation achieved during the process to define research needs.The level was described according to Arnstein's Ladder of Citizen Participation, 16 which ranges from manipulating the community to support a chosen research need to citizen control, where the community decides the priority research need.
3 Broad topic and topic Broad topics included 'Health', 'Environment', 'Health and Environment' and 'Other'.Topics could include, for example, HIV/AIDS, hepatitis, mental health or physical activity.
4 Subtopic of a topic Information about the topic that more specifically defines the research's main focus (e.g., for a project tackling the topic of HIV, the research subtopic could be 'HIV-related stigma').

5
Research field Information about research fields was extracted from the Web of Science database, which has 259 subject categories used for classifying journals (not individual papers).All journals are classified within one or more of these categories.NordForsk 22 proposed a reclassification of these fields into 16 broad categories, such as 'Biomedicine and molecular biosciences' (including subject categories such as virology, immunology, and pharmacology and pharmacy) or 'Social sciences' (including subject categories such as sociology, education and communication) (see Supporting Information S1: Appendix A).NordForsk's reclassification was based on a network analysis approach: that is, how journals' research fields correspond to publications from journals that cite each other.Because the NordForsk analysis is based on journals and not topics, it distinguishes between different types of research related to the same disease.This means, for example, that papers on HIV may be differently classified.A paper on HIV from Health Policy & Services will be classified as 'Health sciences', while an HIV-related paper in Pharmacology and Pharmacy will be classified as 'Biomedicine and molecular biosciences'.The projects were implemented in 30 countries (Figure 2).
In most of these projects, the research (sub-)topics were identified through participatory methods that ensured stakeholders' active participation in their identification and prioritization.The participatory methods used were highly varied.3).In 4.3% of the projects, these participatory methods were combined with nonparticipatory methods, such as observations, literature reviews and analysis of publications in the media.
As Figure 4 shows, consultation (46.2%) and partnership (35%) were the most frequent levels of stakeholder participation in identifying and prioritizing the research (sub-)topics.Although consultation is a low level of participation, the projects using it also applied methods with bidirectional exchange of ideas, such as workshops.A few projects (4.3%)only used therapy, and these were excluded from the analysis.
A wide diversity of stakeholders was involved in identifying and prioritizing research topics for the projects (see Figure 5).Of the 6035 participating stakeholders, the most common were nonorganized citizens (46%), who are citizens that do not participate as members or participants of an organization, and young people (students; 33%, as the Science Shop in Spain focused their projects on high schools).

| Analysis of research fields
Here, we present the results of our analysis of research fields for the broad topic 'Health' and for the topics grouped in the 16 research fields.Some results in the following tables have been grouped into an 'Others' category because they were not relevant for our analysis.
Thus, instead of 16 categories, the tables display 14 categories of research fields.4-Consultation citizens are allowed to be heard through surveys, interviews, neighbourhood meetings, and public hearings, but they lack the power to ensure that their views will be heeded by the powerful.

| Broad topic: Health
3-Informing citizens are informed but no mechanisms are put in place for feedback-one-way flow of information.

Levels of nonparticipation:
2-Therapy powerholders set up public participation to convince citizens that they are the problem and they need to be 'cured'.
1-Manipulationpowerholders invite the public to participate with the express purpose of 'educating' themparticipation as a public relations tool.
Source 23 projects mostly focused on Health sciences (56.2%) and the Social sciences (26%), with low involvement of Clinical medicine and Biomedicine and molecular biosciences (6.8% and 4.1%, respectively).

| Topics: HIV/AIDS, Chagas and Leprosy
Here we present our results regarding the percentages of WoS papers and Science Shop research projects on the topics of HIV/ AIDS, Chagas and Leprosy.
Globally, WoS publications about HIV/AIDS were mostly focused on Clinical medicine (31.6%) and Biomedicine and molecular biosciences (25.6%), followed by Health sciences (16.5%) (  F I G U R E 4 Level of stakeholders' participation during the identification and prioritization of research topics and subtopics (%), using Arnstein's levels. 16 I G U R E 5 Percentage of typologies of stakeholders participating during the research agenda setting processes (%).Comparisons between the countries reveal how Science Shops contributed to the development of research projects on specific diseases in low-and middle-income countries where the total number of publications on those topics was low, but the national burden of disease was substantial (according to 2019 measures of disabilityadjusted life years (DALYs) from The Global Health Observatory 21 ).
That was the case for HIV/AIDS in Bolivia, with 7 publications out of 63,021 worldwide, and the DALYs represented 0.3% of the total disease burden in the country, and in Tunisia, with 41 publications and the DALYs represented 0.4% of the total disease burden in the T A B L E 5 Distribution of HIV/AIDS WoS publications and SS projects (%) across research fields, globally, and in selected countries.

Note:
The color scaling reflects the magnitude of the percentages, from red (low) to green (high).
Abbreviations: AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus; SS, Science Shop; WoS, Web of Science.
country.For Chagas, there were 48 publications in Bolivia, out of 3610 worldwide, and the DALYs represented 0.6% of the country's total disease burden.For Leprosy, although the DALYs in Nepal and Ethiopia did not make up high percentages of these countries' national disease burdens (0.003% and 0.002%, respectively), these countries are amongst the 16 countries with the highest number of new cases of leprosy (2000+ each year).Just three countries report more than 10,000 new cases each year (India, Brazil and Indonesia).
Despite the high numbers of new cases in these two countries, there were only 47 (Nepal) and 93 publications (Ethiopia) publications on leprosy, out of 4968 worldwide., or the predominance of some professions in research (e.g., towards medical and surgical fields). 11,12These various barriers can only be addressed with methods that follow an iterative approach, where the different stakeholders can reflect together in action-learning spirals, and identify not only the research priorities but also the systemic barriers and possible strategies to address them.

| DISCUSSION
Regarding the diversity of stakeholders involved, the Science Shops participating in the InSPIRES project are not representative of Science Shops in general.Science Shops usually follow the more traditional model in which CSOs are the main stakeholder group. 1,2,9The Science Shops in our study followed the open and inclusive approach of the latest Science Shop models.
Involving a higher diversity of stakeholders in defining and prioritizing research projects may increase the likelihood that the research will be aligned with their needs and that systemic barriers will be addressed and thus that this approach will contribute (more) to decreasing existing bias.We therefore suggest that Science Shops evolve towards involving a larger diversity of stakeholders.
In our study, the diversity of stakeholders could have been improved with a better representation of business/industry and funding agencies, which were hardly involved.Therefore, further research is needed on the diversity of stakeholders types required, and on similar comparisons regarding approaches for 'recruiting, retaining, or ensuring the inclusion of diverse and/or representative stakeholders in governance activities' 18 and also for addressing the systemic barriers.
Another limitation is the use of the Arnstein Ladder of Participation.The model has been criticized for: neglecting power relations and differences, 24 ignoring different forms of knowledge and expertise, 30 reducing participation to 'a hierarchical set of social relations that are devoid of context', 31 considering 'consultation' as participatory process, even though it does not facilitate the necessary reflexive processes, and lacking 'insights into how participation might be progressed'. 31 Note: The color scaling reflects the magnitude of the percentages, from red (low) to green (high).
Abbreviations: SS, Science Shop; WoS, Web of Science.
Study and Conference on Improving Public Access to Science through Science Shops, Improving Science Shop Networking, Training and Mentoring of Science Shops, Public Engagement with Research and Research Engagement with Society, Enhancing Responsible Research and Innovation through Curricula in Higher Education, Enhancing the Responsible and Sustainable Expansion of the Science Shops Ecosystem in Europe), Ingenious Science shops to promote Participatory Innovation, Research and Equity in Science research fields within drug research, such as research on the effectiveness of starting or prescribing drugs versus research on drug cessation (withdrawal, deprescribing, stopping, discontinuation, or reduction of drug treatment), diagnosis and prognosis; (2) research fields in pharmaceutical versus nonpharmaceutical interventions, like psychotherapy, health education or rehabilitation which 'are severely underserved'; who compared the research topics addressed in registered trials with the research needs identified and prioritized by patients and clinicians over the same time period (2003-2012).The main priorities addressed in the internationally registered clinical trials during that time focused on drug evaluations.However, the priorities identified by patients and clinicians emphasized the importance of nondrug interventions, such as education and training, psychological therapy, or social care.Tallon et al. 11 in their exploration of research priorities on the management of osteoarthritis of the knee, found a similar bias towards the development of new drugs.When patients and clinicians

T A B L E 1 Basic descriptive information 1 . 1
Data collected for each Science Shop research project conducted within the InSPIRES period (2017-2021).Categories of data collected for each Science Shop research project 1 Name of the Science Shop research project 1.2 Research aim 1.3 InSPIRES partner who led the research 1.4 Country where the partner is located 1.5 Country where the project was implemented 1.6 Period of implementation 2 Characteristics of the process to identify research needs 2.1.Participatory methods Description of approaches and methods for engaging the different stakeholders in identifying the research needs.

3 | RESULTS 3 . 1 |
This section reports the results of each phase3.1.1 | Science Shop research projectsWe identified 117 projects implemented by 14 Science Shops located in 13 countries during the InSPIRES project lifespan (2017-2021).

BOX 1 8 - 7 - 6 - 5 -
Levels of participation according to Arnstein's Ladder of Citizen ParticipationLevels of citizen power with decreasing degrees of decision-making: Citizen control citizens obtain full managerial power.Delegated power powerholders hand over some degree of decision-making power to citizens.Partnership citizens can negotiate and engage in tradeoffs with traditional powerholders, ultimately sharing decision-making responsibilities.Levels of tokenism: Placation citizens are allowed to advise by having a (few) seat(s) on committees or boards, but they are easily outvoted by powerholders or powerholders retain the right to decide whether or not to follow their advice.

F I G U R E 1
Abbreviations: AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus.

Table 4
shows percentages of WoS papers and Science Shop research projects, distributed by research fields.Within the broad topic 'health', WoS publications focused mostly on Clinical medicine (30.8%), followed by Health sciences (17.4%) and Biomedicine and molecular biosciences (16.7%).All other research fields have a much lower presence (5% or less).In contrast, Science Shops research

Table 5
Description of each data set and the inclusion and exclusion criteria used.
(13%) and none on Biomedicine and molecular biosciences.Country distribution of WoS publications in HIV/AIDS followed a similar pattern to the global distribution.In all countries, Clinical medicine was the main research field and Biomedicine and molecular biosciences were second (except in Ecuador, where Social Sciences was second).Low-and middle-income countries were substantially less present in Biomedical and molecular bioscience studies, as shown in comparing the countries outside Europe with The Netherlands and Spain (except for Bolivia, although it might not be relevant as results were based on only seven papers).Health sciences publications on HIV/AIDS differed were mostly biomedical (almost 50%), and about 27% of the papers were in Clinical medicine.In Bolivia, the publications were more evenly distributed amongst these two research fields (35.4% in each).In all cases, studies in the Social sciences and Health sciences were practically nonexistent.Amongst the Science Shop research projects, there were no biomedical projects at the global level nor in Spain and T A B L E 2 Abbreviations: AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus.
[10][11][12]as shown that the Science Shop research projects mostly identified and prioritized research topics through participatory processes with a wide diversity of stakeholders, mainly nonorganized citizens and young people (high school students).Furthermore, the Science Shop projects made much more use of Social sciences and Health sciences, mainly focusing on nondrug treatments, compared to the global publications in WoS, which focused on Clinical medicine and Biomedicine and biomolecular sciences, largely aiming at the development of drug treatments.This difference in research focus is very much in line with findings from previous studies on research agenda bias.[10][11][12]We,therefore, argue that the Science Shop Distribution of Chagas WoS publications and SS research projects (%) across research fields, globally and in selected countries.
The WoS results (Tables5-7) also indicate a global 'division of labour' in the distribution of research fields in publications.Low-and middle-income countries are more involved in projects 'on-site', while more basic research and development are carried out at laboratories cultural contexts, as, for example, cultural norms on who should participate in agenda setting and the role of research may be very different in Africa, South America and in Europe.This research is key if the Science Shop model is to be scaled up to different cultural contexts.T A B L E 6 Distribution of Leprosy WoS publications and SS projects (%) across research fields, at the global level and in selected countries.
Despite these shortcomings, we particularly used Arnstein's model to label the methods being used in Science Shops, including those that were nonparticipatory (which enabled us to exclude projects), and those that may not T A B L E 7