A decade of antimicrobial resistance research in social science fields: a scientometric review

Antimicrobial Resistance (AMR) occurs when microorganisms develop resistance to antimicrobial medications such as antivirals and antibiotics. AMR is increasingly becoming a global threat, both economically as well as in terms of health. AMR can lead to the treatments becoming ineffective, making it more difficult to treat infections, and can then cause an increase in the risk of the disease spreading [1]. Particularly, diseases such as tuberculosis, malaria, HIV and Influenza are developing more drug resistant cases, leading to higher healthcare costs and a longer duration of illness. By 2050, the impact of AMR is estimated to lead to 10 million deaths per year, with a global economic cost of $100 trillion [2]. Within the European Union alone, the average yearly additional costs caused by AMR between 2015 and 2030 is estimated to be $252,215 USD per 100,000 persons per year [3].

Several factors have been found to cause AMR. Human use of antimicrobial drugs, particularly antibiotics, is recognized as one of the primary drivers of AMR. Antibiotics are among the most familiar of medicines by the public, and the global consumption of antibiotics has been estimated at more than 35 billion daily doses at 2015 [4]. Several studies have found a positive association between antimicrobial use and AMR [5‐7]. Although in the past AMR research has focused on direct public consumption of antimicrobial drugs, drugs use in livestock are an increasingly concerning issue. Antibiotics are an integral part of industrial agriculture to ensure healthy livestock and promotion of growth [8]. The Food and Agriculture Organization (FAO) expects that two-third of the future growth of antimicrobial use will be linked to animal production [9]. Human consumption of these animals can lead to resistant microorganism transmission between hosts. This can then in turn impact the wider environment through animal and human waste affecting soil and land.

Different approaches have been proposed to tackle AMR. One effective way is to provide interventions through antimicrobial stewardship programmes, which aims to promote the appropriate use of antimicrobials [10, 11]. On a larger scale, the idea of ‘one health’ has been increasing in popularity for combating AMR. One health is an approach for the improvement of public health through the promotion and collaboration of multidisciplinary areas [12]. Within the AMR space, this typically involves sectors from animal health, human health, food safety, and environment working together to prevent the spread of resistant microorganisms. The World Health Organisation (WHO) highlighted the one health approach as a guide for the global action plan and framework on AMR [13, 14].

Social sciences, encompassing a wide range of academic disciplines, such as sociology, psychology, communication, and economics, are expected to play a significant role in antimicrobial stewardship and the one health approach. For example, antimicrobial stewardship often requires theories, frameworks, and methods from behavioural and psychological sciences at a micro level. The one health approach also inspires social sciences to address AMR from the societal, historical, and economic perspectives. However, within the social science field, academic literature on AMR has historically been lacking [15]. It is only in recent years that the values of social sciences are increasingly recognised in AMR research [16‐18].

As the global efforts to reduce AMR rapidly increases, and social science is becoming more prominent in the field, existing gaps and future directions need to be identified. Therefore, this study investigates the social science literature surrounding this topic and how the landscape has changed and developed over the last decade. We aim to analyse the bibliometric data of the literature in the field via network analyses. Frid-Nielson et al. [15] have utilized similar techniques and revealed that the social science contribution to AMR is weak and peripheral compared to its contributions to other research areas such as climate change. However, our study bears significant differences and extensions from their study. First, our study focuses on literature within the last decade instead of looking at a wider scope of several decades. This is to ensure that the most recent development of social science AMR research are highlighted by avoiding the impact of early clinical and medical literature on network analyses. Second, we will perform more in-depth analyses than the previous study, including the presentation of co-authorship networks and keyword co-occurrence networks. We will also perform analyses based on the influential articles within the field in the last decade. These analyses are critical as they can reveal the most recent state of international contributions and emerging research topics in the field, allowing further identification of research gaps and future directions.

We followed the PRISMA guidelines [19] for conducting the review and reporting results. The Social Science Citation Index (SSCI) in the Web of Science (WoS) database was selected as the data source for our review. SSCI is a reputable scientific index that covers over 3000 social sciences journals across more than 50 disciplines. In addition, the hosting WoS database provides structured and consistent indexes of authorships and references and is well respected within academia. The index allows wide access of social science research on AMR and relatively accurate network analysis of authorships and co-citations.

A topic search in SSCI for “antimicrobial resistan*” and “antibiotic resistan*” published from 2010 to 2019 was conducted on Jan 14, 2020. Inclusion criteria included publications in English and limits were set to full-text peer-review journals to ensure the accuracy of follow-up co-citation analyses. The search produced 787 unique articles published in 294 journals with no duplicates. Bibliographic data including full reference lists of each article were extracted. Data of all articles were included in the analysis.

VOSviewer is a software tool that was used to analyze and visualize the bibliographic networks. VOSviewer can provide distance-based bibliometric maps in which the strength of the relation between two items can be visualized as the distance between the two items, allowing easy understanding and visualization of clustering [20]. Co-authorship analysis based on countries were conducted to showcase the current state of international contributions. The keyword co-occurrence network was examined to explore the research landscape and understand research trajectories. Top cited and co-cited articles were briefly reviewed to detail the existing landscape, complementing the keyword network analysis.

Finally, the citation and co-citation network of journals were examined. The citation network, where a connection is demonstrated when a journal was cited by another journal, is used to reveal the knowledge flow in AMR social science research. The co-citation network, where a connection is demonstrated when a journal was co-cited with another journal by a research document, often indicates the topic similarity between journals. In this study, we use the co-citation network to reveal the knowledge foundation of the research field, as it can present and cluster all journals that were cited by research documents in the current dataset.

AMR is a global threat to the sustainability of human health and development. Social sciences can contribute to tackling this public health issue by providing domain knowledge that other sectors may not be able to provide to the field. By analysing bibliometric data of SSCI literature on AMR, this study demonstrated the landscape of social science research on AMR in the last decade and identified research gaps and future directions of AMR social science research.

Though social science research within the area of AMR has rapidly increased in the past few years, the field remains relatively new and requires continued and intense international collaborations. Findings from our authorship analysis revealed that only 31.2% of social science studies on AMR involves non-high-income countries that house 84.3% of the global population. This phenomenon may be due to several important reasons besides the obvious economic factors. Countries at lower income levels may be less aware of the AMR issues than those at higher income levels. For example, AMR is largely a neglected problem in South-East Asia Region, as only 6 of the 11 regional countries carried out national AMR surveillance up to 2014 [43]. Also, lack of laboratory capacity for AMR surveillance in less developed countries may limit the development of social science research on AMR. Many countries in Africa and South-East Asia reported challenges surrounding AMR surveillance and research due to lack of standard surveillance frameworks, weakness in laboratory testing of drug resistance, and lack of trained staff [43, 62]. Therefore, besides coordination at the policy levels [63], international collaborations among countries at different income levels should be encouraged by the academic society, which will be essential to help the development of AMR surveillance and follow-up research in less developed countries.

Another important finding that warrants attention is that regional collaboration networks of AMR social science research remains undeveloped. While WHO has been promoting political coordination in data sharing and dissemination of AMR research [43], our co-authorship network analysis reveals significant gaps of academic collaborations within WHO regions in the last decade. This is consistent with the 5C Framework for assessment of minimum collaboration utilized by Ardal et al. [63] indicating that collaborative actions require higher levels of joint arrangements and decisions than coordination. However, unlike the political ones, academic collaborations can often bypass political boundaries because they can take various forms and be fulfilled through personal networks, academic networks sharing the same interests, and existing institutional connections [64]. Therefore, regional collaborations should be encouraged at the academic level as it may in turn promote future political collaborations.

Importantly, our analyses of keywords, articles, and journals showed cohesive evidence that one health was marginally linked to social science AMR research, though it has been largely connected with medical and microbial sciences. Given that the one health approach has only been recognized as a fundamental approach to combat AMR in the past five years, this disconnection is understandable. Nevertheless, as agriculture and food sectors are expected to account for two-thirds of the future antibiotic use in the coming decade [9], it is imperative to understand why and under what conditions antimicrobials are used so extensively in these sectors [65]. Social issues surrounding AMR campaigns targeting food industries are also in need of being resolved by social sciences [66].

In fact, under the umbrella of one health, all social science disciplines can play important roles in reducing AMR. For example, economic models can likely offer optimal solutions to balance profits and risks regarding antibiotic use, while sociology can utilize qualitative methods to understand and resolve tensions between different AMR stakeholders from the institutional perspective. The disciplines of communication and psychology can focus on a more micro-level to study attitudinal and behavioural mechanisms underlying debates of antibiotic use in food and animal sectors.

Consistent with Frid-Nielson et al.’s [15] findings, our results provided further evidence that AMR has not been incorporated into the core discourses of social science disciplines. This is surprising as some of the disciplines has been well engaged in the field of public health. For example, the disciplines of communication, and health communication particularly, have greatly contributed to academic discourses in various public health and medical topics including obesity and infectious diseases [67‐69]. WHO has called for programs to improve awareness of AMR through communication and education [43]. Given that these programs will inevitably involve communication media and techniques, health communication research can play an essential role in examining the impacts of media on AMR as well as revealing the public’s opinions and practices on antimicrobials via media (e.g., social media). Various health communication models such as the health belief model can be utilized to educate people on prudent antimicrobial use. These models can also help understand why the public inappropriately consume antibiotics for some common illness (e.g., influenza) when their effect is in fact very limited. Such research will inform better strategies for reducing antimicrobial use.

We only analysed bibliometric data from the SSCI WoS database, which may not reflect the complete set of research in the field. Nevertheless, as the SSCI database only includes journals that meet high research standards, the dataset ensures that our findings are based on articles that passed high standard peer-review processes and observe minimised bias.

AMR is a major health threat that will become more prominent in the coming decades. The battle with AMR requires global efforts from all academic fields. Though social science research on AMR was rapidly growing in the past five years, the field is still young and requires intense international academic contributions and collaborations. The one health approach should be addressed as it provides an overarching framework for contributions from different scientific disciplines. This study demonstrates the current landscape of AMR social science research and calls for international academic efforts from the one health perspective.