Introduction

In 2020, an estimated 8 million animals (mostly vertebrates) were used for research and testing purposes in the EU [1] with more than 100 million vertebrates used globally [2]. Animal use in research results in both ethical and political challenges. Although there are those who oppose all animal experimentation, overall public acceptance is contingent on adherence to good welfare standards and the restriction of animal use to designated purposes [3].

One way to uphold humane use of animals for experimentation is via the implementation of the 3R principles (Replacement, Reduction, Refinement) proposed by Russel and Burch in 1959 [4]. This widely adopted framework aims to promote animal welfare by: replacing animals wherever feasible with another means of experimentation; reducing the number of animals used to the minimal needed to obtain scientifically valid results; and to refine experiments to avoid unnecessary harm to animals [5]. However, this framework does not provide guidance on which experiments should take place. In this regard, Beauchamp and DeGrazia [6] proposed six principles of morally justified research, which need to be satisfied for an animal experiment to be conducted. These are: 1) no alternative method to animal testing exists, 2) an expected net benefit to humans, 3) the experiment provides sufficient value to justify the harm caused to animals, 4) animals are not subjected to unnecessary harm, 5) basic needs of animals are met, and 6) there are upper limits to the harm allowable in an experiment. A significant part of these principles is the weighing of expected harms against potential research benefits, a process often referred to as a harm-benefit analysis (HBA). Even if these conditions are met, scientifically unsound research amounts to unnecessary harm [7] since the design features of an experiment affect the validity and reproducibility of its results [8]. For this reason, some researchers argued that an assessment of the scientific validity of animal research protocols should be an essential part of an HBA [8].

Different decision-making entities are tasked to carry out evaluations of animal research protocols, such as Animal Ethics Committees (AEC) or Institutional Animal Care and Use Committees (IACUC). These committees can be made up of veterinarians, practicing scientific researchers, lawyers, ethicists, and sometimes, lay members of the public. The composition of the committees varies depending on context and regulation. In the US, for example, IACUCs must include at least one veterinarian, one scientist, one non-scientist (ethicist, lawyer, or member of the clergy), and one member unaffiliated with the institution [9]. In contrast, EU legislation requires the presence of technical experts, animal welfare experts, and scientific researchers [10]. The framework for reviewing animal experiments also varies from country to country. Some countries, such as Canada, do not legislate for animal experiments directly but nonetheless have a self-regulating oversight system in the form of IACUCs, which incorporate frameworks such as HBAs and 3R [11]. In Japan, the 1973 Act on Animal Welfare and Management of Animals [12] regulates how experiments are carried out while deferring to self-regulated IACUCs at an institutional level. In the US, the Animal Welfare Act 1966 requires ‘review’ of animal experiments by an IACUC. Within this act, scientific evaluation of experiments is not required unless related to animal welfare. With regard to the EU, countries are subject to the 2010/63/EU directive [13]. In particular, article 36 requires an assessment of the 3Rs, a HBA, and an evaluation of experiment design by a competent authority designated by the state. A similar approach as in the EU is adopted by New Zealand [14] and Australia [15] whose legislation also requires an assessment of the 3Rs, HBA, and experimental design for scientific validity. Given these legal differences, it is understandable that the decision-making process differs between countries. Even between EU countries, there are significant differences in how EU legislation is being implemented [16,17].

Beyond national and regional differences, variations in decision-making have been noted between different committees in the same country. Poor reliability between committees was identified in final decisions made [18] and in assessments of costs to animals, expected benefits, and recommended modifications [19]. Given these inconsistencies, it is hardly surprising that public trust in regulatory systems around animal testing is low [3].

In 2012, a literature review charted some of the issues impacting AECs’ decision-making processes [20]. Key issues noted were: membership composition influencing the effectiveness of committees; narrow ethical views within committees; high rates of approval of protocols; and poor reliability for the justification of animal use. The introduction of European directive 2010/63/EU shortly preceded this 2012 review. This directive was followed by further updates to legislation in other countries. For example, the revision of the Australian Code of Care for the Use of Animals for Scientific Purposes (2013), and the US revision of the Policy on Humane Care and Use of Laboratory animals (2015). These updates require a re-evaluation of the decision-making process. Another recent systematic review focused on methods that have been used to evaluate IACUCs in the US [21]. There is some overlap in references used between this review and our current review although the research question is different. Our current review focusses on the decision making process of animal ethics committees and is not limited to US studies.

The overall objective of the scoping review is to explore the decision-making process for the evaluation of animal research proposals within AECs and IACUCs especially in light of legislative changes over the last decade. More specifically, it has the following aims: (1) to gain a deeper understanding of the decision-making process of AECs when evaluating animal research proposals, and (2) to critically summarize the available empirical literature on the different factors influencing, or likely to influence, decision-making by AECs when ethically evaluating animal research proposals.

Method

This scoping review adheres to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for scoping reviews (PRISMA-ScR) [22]. Its protocol was registered on the Open Science Framework (OSF) with the following DOI: https://doi.org/10.17605/OSF.IO/GZJMB.

The scoping review approach was chosen to: (1) identify the types of empirical evidence available on decision-making processes in AECs/IACUCs, (2) to catalogue the key characteristics or factors related to decision-making, and (3) to identify and analyse knowledge gaps [23]. Following JBI methodology for scoping reviews, the Population-Concept-Context (PCC) framework was used to guide the focus of the review [24]. To keep the search as broad as possible, context (animal experimentation) was omitted as an exclusion criterion to prevent overlap with the population (animal ethics committees). The search strategy was first developed for PubMed and then adapted for Scopus and Web of Science (core collection). The Medical Subject Headings (MeSH) and search terms were identified based on extensive background reading on decision-making and animal ethics committees. Only search terms in English were included. The search strategy was checked and validated by an information specialist of the University of Basel library. The search was performed first on 11.05.2023 and then updated on 03.06.2024 (dd.mm.yyyy).

As part of the inclusion criteria, the date range for publication of sources was 01.12.2012 to 03.06.2024 to include all publications since the time of the last review on the topic [25]. Only empirical studies published in peer reviewed journals were included in the languages known to the authors (English, French, German and Italian), there was no limitation on the type of empirical data to be collected. Theoretical papers, reviews, books, editorials, and book chapters were excluded as were articles that did not separate data on animal ethics committees from human ethics committees or institutional review boards. Papers were also excluded if they did not include data about animal ethics committees making decisions about experiments in biomedical research, i.e., experiments for education were not included.

The results were collated using EndNote X9 software and then transferred to Covidence, an online software used to remove duplicates and facilitate the screening process [26]. Title-and-abstracts and full texts screening were performed independently by two review authors (AM and ELM) with a third author (LDG) acting as arbitrator in the case of disagreement. Once the full texts were screened, a citation and reference search was carried out on all included reports. Independently extracted data were collated using a data charting form including general study characteristics: author names, year of publication, country of origin, journal name, study design, study participants and study aim. To meet the objectives of this review, we also collected information on the factors influencing the decision-making process or details related to the modus operandi of animal ethics committees. Data extraction was done by AM and ELM and reviewed by LDG and EDC to ensure accuracy.

Basic qualitative content analysis was used to map and identify key characteristics in the diverse data [27]. This process identifies the following procedure as appropriate for scoping reviews. To begin authors AM and ELM immersed in the data with several readings of the included reports. During the charting and analysis process, an inductive framework was developed. Initial thoughts and notes were recorded for each source to create a coding framework. Extracted information was then organized within this framework. The framework was then reviewed to develop overarching categories within the data.

Results

Nine hundred and sixty publications were obtained by database searches, in which 298 duplicates were found and removed. The remaining 662 title-and-abstracts were then independently screened by AM and ELM. This resulted in the inclusion of 95 studies for full-text screening by AM and ELM. Eighty three studies were excluded for the following reasons; irrelevant outcomes-they did not address the topic of decision making in animal ethics committees (n =  8); irrelevant language-it was published in a language unknown to the authors (n =  1); irrelevant population- the data collected in the study was not collected from animal ethics committees or animal ethics committee members specifically looking at animal research protocols either through self-report method or through analysis of documents related to committees (n =  3); or irrelevant study design- the papers did not include empirical evidence (n =  71). This resulted in 12 eligible studies available for data charting. Citation and reference searching for the eligible studies resulted in no additional sources. The process of study selection is outlined below (Figure 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA diagram systematic search and selection process.

https://doi.org/10.1371/journal.pone.0318570.g001

To present our findings, we will start with providing some general study characteristics (details of included studies are outlined in Table 1). Of the 12 included papers, 4 were quantitative, 5 qualitative and 3 mixed methods studies. Seven relied on self-reported data, while the other five consisted of documents generated by committees or transcripts of committee discussions. Six of the retrieved literature originated in the US and of those, four had the same first author. Of the studies involving a European cohort, three of the five were from Sweden. Most papers (n = 9) were published in animal specific journals, only three were published in general bioethics or research ethics journals. In a next step, we will discuss two main themes and various subthemes: (1) principles guiding the review process and (2) factors influencing the review process. Distribution of codes and the studies in which they were present is shown below (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of included publications (N = 12).

https://doi.org/10.1371/journal.pone.0318570.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Distribution of themes, subthemes and codes.

https://doi.org/10.1371/journal.pone.0318570.t002

Discussion

This scoping review provides an overview of the empirical literature on the decision-making process in animal ethics committees. The 3R principles, as they were first presented by Russel and Burch, have a hierarchy, in which replacement is given the top priority followed by reduction and refinement [4,5]. In contrast, our results show that refinement is given the most prominent place in discussions and replacement the least [35]. Given that scientists often form the majority of the committee members [29] and that scientists tend to rank refinement as higher than replacement [41], this is not a surprising result. Refinement is also considered by scientists to be the most achievable [41,42]. Additionally, our results show that committees typically have at least one veterinarian who is trained to address refinement techniques and animal welfare needs, and that veterinarians – alongside chairpersons – are contributing the most comments per person [36]. Therefore, discussions of animal research protocols could potentially be stirred towards refinement methods by the latter, at the detriment of replacement or reduction techniques. For other committee members, implementing refinement methods such as improved housing or enrichment may be more straightforward to understand and adopt. It is thus possible, that the availability and familiarity of knowledge surrounding each ‘R’ influence their weight in the evaluation process.

Reduction on the other hand is more challenging. A common reduction strategy is the use of statistics to ensure the minimum number of animals are used to obtain valid results. This requires knowledge of statistics, which our results show, is not always available to the majority of AEC members [33,40]. Other important means of reduction such as data and tissue sharing through databases such as the SEARCH framework in the US [43] or the 3R blackboard in the EU [44] were not discussed in all our retrieved literature. Similarly, other authors have demonstrated that knowledge of replacement methods is lacking and there are some concerns about the validity of replacement methods [45] or fears about publishers preferring animal studies [46], which may lead scientists to neglect replacement in favor of refinement or reduction. This is especially important given that the majority of AEC members are biomedical researchers [29]. Nonetheless, it is possible that animal methods bias (a preference for animal methods when alternatives are available) and scientific inertia or tradition are reasons why replacement methods are less considered [47,48].

HBAs are the linchpin for the justification of animal research, yet it is evident that there are a range of challenges with the implementation of an HBA. Two such challenges were found in our results. First, assessing benefits in comparison to harms is a subjective task, which is considered difficult and frustrating to carry out by committee members [37,49]. One cited reason in the retrieved literature was a lack of clear guidance on how to carry out a HBA [30]. The EU directive 2010/63 provides detailed schema to assess harms, whereas the legislation does not provide any schema to assess the benefits of research beyond confining research to certain purposes [17,50]. Similarly, US regulations provide only minimal guidance to ensure research has social or scientific value [39]. It is, therefore, not a surprise that our results which were predominantly from the US or Europe, demonstrate that harms were discussed more frequently and in more detail than benefits by committee members, and that the weighing of harms against benefits was even more sparse [30,35,36,39]. Second, our results show that the benefits of research are often optimistically assumed and not analyzed in detail [39]. Taking into account how few projects are considered to satisfy Bateson’s criteria for conducting an HBA retrospectively, this optimism and confidence in research is hard to justify [51]. One proposed solution is to diversify the committees’ composition [52]. At present only technical experts; animal welfare experts (including veterinarians); and scientific researchers are required for committees within the EU [49]. Scientists, however, are more likely to be positively biased towards research than other groups, such as animal activists [10]. Another potential challenge with HBAs is that they represent a narrow consequentialist approach to ethics [49]. Members of the public and key animal ethicists, on the other hand, have a range of ethical positions [53–56]. It is arguable that this may present a source of conflict within committees as to how to approach and conduct a HBA. Where these analyses are conducted using an ethical framework not supported by either ethicists or the public, it is possible that the review process may not be trusted. This may be one reason why evidence points to a lack of public trust in the review process [3].

Our review demonstrated that although the validity of using certain animal models is considered important [28,33,35], the scientific validity of an experiment was less likely to be discussed by committee members [39]. This holds great significance given the emerging evidence of poor reproducibility or translatability of completed animal research in recent years due to poor experimental design [57,58]. The strength of internal validity, reported in protocols submitted for ethical review, correlates with the strength of internal validity reported in publications [59]. This demonstrates that the quality of a research proposal cannot be assumed at ethical review stage and highlights the importance of assessing the scientific validity prior to commencing a project to prevent poorly designed studies going forward. Not only is there minimal benefit to invalid results, but invalid results also represent a waste of resources as well as unnecessary suffering of experimental animals. Despite this, several countries (including the US and Canada) do not require an assessment of scientific validity as part of their review process. For this reason, it is recommended that the HBA should only take place after an in depth assessment of the validity of a research protocol [8].

Our results demonstrated that the review process can be influenced by factors related to the nature of the review process, the applicants, or the members of the committee. The composition of AECs or IACUCs can be imbalanced in favor of researchers and veterinarians which can affect: (i) the expertise available and focus of the review; (ii) the independence of the review; (iii) the empowerment of all members. Although expertise in animal research can be useful in harm reduction, the lack of knowledge in animal research ethics, replacement, or reduction methods can result in a limited ethical review. This also appears to skew ethical review towards harm reduction rather than ethical evaluation. This concern has been echoed by other authors [52,60]. Questions were raised in our results about the independence of reviewers [31,33,34]. Another author explained that this may be a result of the difficulty in recruiting truly independent committee members and cautioned that committee members can be extremely reluctant to raise concerns on ethical grounds with their colleagues [60]. Our review also found that scientists and veterinarians contributed most and tended to lead discussions but that other types of members felt less influential and reported frustration in the process [36,37]. This may be a result of deficient processes for managing moral dilemmas [60]. Ethical review occurs after successful funding applications and our results demonstrate that this can influence reviewers to approve projects [39]. It has been speculated that institutional pressure to approve projects that receive funding may be at play here [60].

In comparison to a previous review on the decision-making process of AECs [25], two issues have remain unchanged. The committee composition still seems to affect its effectiveness and although committees may be more reliable in their analysis of refinement, the justification of animal use remains unclear. There were some other issues that this scoping review had hoped to address. For instance, the ethical codes, regulations and research standards used in practice by AEC members, cultural norms, or the ethical issues deliberated on. These topics however, did not arise in the retrieved literature. Literature on the causes of variations between different AECs or cultural norms was also missing as was any data about how committee members are selected.

Conclusions

Ethical review of animal protocols is a complex and contentious process on which animals depend for their protection. Through this review we have created a clearer picture of the decision making process and charted some of the deep seated issues surrounding it. At each stage of the evaluation, deficits in the decision-making process were found. The scientific validity of a protocol was often neglected, the 3R assessment tends to focus on refinement rather than reduction or replacement, and the harm benefit analysis often focusses on harms rather than an in –depth weighing of interests. The committees themselves influenced these processes significantly. Committee composition affects the knowledge and expertise available to address key issues, but also the focus of the review. Problems with group dynamics and independence of committees were also discussed. Remedies to address these issues were suggested and although the recommendations given above may be costly, they reflect the deep responsibility for those involved in animal experimentation to protect the animals they work with and ensure responsible research. Further research should delve into the ethical issues debated and the ethical theories informing decision making along with testing measures to address the aforementioned issues.

Supporting information

Acknowledgments

The authors thank Christian Appenzeller-Herzog from the University Medical Library of the University of Basel for his assistance in developing the search strategy for this project.