Breaking barriers to ethical research: An analysis of the effectiveness of nonhuman animal research approval in Canada

ABSTRACT

In Canada, all institutions that conduct publicly funded, animal-based research are expected to comply with the standards of the Canadian Council on Animal Care (CCAC). The CCAC promotes the use of animal alternatives, and uses the “3Rs” principles of Replacement, Reduction, and Refinement as a guiding ethical framework. To ensure these standards are strictly enforced, internal ethics committees at each institution are tasked with creating “Animal Use Protocol” (AUP) forms to be filled out by researchers and evaluated by the committees.

In this paper, we assess AUP forms from Canada’s top research universities to identify the extent to which they conform to, or advance, the 3Rs framework. Our results show various deficiencies that call into question the quality of information elicited by these forms. To remedy this, we recommend that the CCAC assume responsibility for creating a standardized 3Rs section to be used on all AUP forms. In addition, proposal forms and experimental results for all research at CCAC-certified institutions should be digitized and uploaded into a national database. We argue that this would offer higher quality information for researchers at the experimental design stage, while strengthening the CCAC’s mandate to be accountable to the Canadian public.

KEYWORDS:

• 3Rs framework
• Canadian Council on Animal Care
• research ethics
• nonhuman animal subjects
• public policy

Introduction

In the summer of 2018, we collected Animal Use Protocol (AUP) forms from Canada’s top research universities, identified as the U15 Group of Canadian Research Universities. Drawing on the definitions of the 3Rs provided by the Canadian Council on Animal Care (CCAC) as well as their guide for applying them, the Animal Use Protocol [AUP] Worksheet, we developed a set of questions to assess the extent to which the AUP forms conform to the CCAC’s standards. This paper describes our findings, summarized in five tables, and offers some proposals in light of them.

After first explaining the 3Rs, detailing the research approval process for research conducted at publicly funded institutions in Canada, and explaining our methodology, we analyze and compare AUP forms from the Canadian U15 Group. Our analysis uncovers deficiencies within individual forms and inconsistencies across forms. Our findings suggest that AUP forms often fail to elicit information that Animal Care Committees (ACCs) need to make informed decisions that are consistent with 3Rs standards. This in turn undermines confidence in the ethics review process. We propose a solution: the standardization and digitization of a 3Rs section on all AUP forms. We recommend that the CCAC is tasked with creating this section, thereby avoiding the inconsistencies and deficiencies that arise when ethics committees develop their own unique forms. Furthermore, we suggest that information from proposed and ongoing research, and from ethics committee decisions, be aggregated and archived into a searchable database. With access to this database, researchers and ethics committees would be better equipped to comply with the 3Rs – i.e., to find adequate inanimate or “less sentient” experimental subjects, reduce the number of animals in research to that which is scientifically necessary, and refine protocols and husbandry to reduce suffering and improve the wellbeing of animal subjects.

We conclude by describing how our proposal can be used to strengthen the CCAC’s commitment to communicate with the Canadian public as well as reflect Canadian values in the decision-making process. With information about research and ACC decisions compiled and made accessible, the CCAC could increase public transparency in a controlled way. We argue that an increase in transparency, combined with the opening of new communication channels with the public, would make the research approval process effectively more accountable to Canadians.

Policy and principles—Positioning the 3Rs

In Canada, laws governing the use and treatment of nonhuman animals (hereafter, animals), including those in research, fall primarily under provincial jurisdiction (Griffin and Locke 2016, 272). Apart from the general protection against cruelty to animals (Criminal Code 1985, ss. 444–447), the Government of Canada has yet to use its authority to directly regulate the use of animals in research,¹ with one notable exception.² Instead, the regulation of research using animals is entrusted to the CCAC, which is directed by its mandate to “act in the interest of the people of Canada” (CCAC 2019f) and is, at least in principle, accountable to them. Whether the current structure and practice of research approval meets this goal of accountability to the Canadian public is one of the issues that makes our project urgent. Interestingly, there is no direct ministerial oversight. Instead, the CCAC is supported by and reports to Canada’s federal research funding agencies (the Tri-Council agencies).

In the first step of the research approval process, proposals involving animals are approved for funding primarily on the basis of having valuable research goals. While federal funding bodies reserve the right to decline funding on the basis of ethical concerns, they give the CCAC the responsibility to set and maintain standards for animal care in research, and position themselves as supporting the CCAC in fulfilling this role (Canada 2018; NSERC 2013). After being approved for public funding, proposed research is then evaluated by institutional ACCs, which are regulated by the CCAC. Projects should only be approved once ACCs are confident that the research design causes no harm beyond that which is necessary to meet research goals, and that the research goals are worth pursuing despite the cost to animals.

The “3Rs” principles of Replacement, Reduction and Refinement were first introduced by William Russell and Rex Burch in 1959 and have since become recognized internationally as the standard for animal-based research. The CCAC adopted the 3Rs framework early in their mandate (1989) and enforces compliance with these standards through a process of certification. All Canadian institutions receiving federal funding for animal-based research must adhere to policies that promote the 3Rs or risk losing federal funding.

The CCAC’s commitment to the 3Rs is explicit. They offer definitions of the 3Rs on their microsite, capturing the central aim of each of Russell and Burch’s principles. Replacement “refers to methods which avoid or replace the use of animals in an area where animals would otherwise have been used” (2019a). There are two categories of Replacement alternatives: Absolute Replacement, which involves replacing “animate systems” with “inanimate systems,” such as mathematical or computer models, and Relative Replacement, which involves replacing more sentient organisms with less sentient or non-sentient organisms or tissues (2019e).³ Reduction “refers to any strategy that will result in fewer [sentient] animals being used” (2019a). Reduction alternatives include any alternative strategy that will maximize information obtained while using the fewest animals and can include using animals more than once, sharing tissues or data, and gathering more data. Choosing an appropriate animal model, an appropriate experimental design, and consulting with a statistician can lead to Reduction alternatives. Refinement “refers to the modification of husbandry or experimental procedures to minimize pain and distress” (2019a) and generally addresses the welfare of the animals used. Refinement alternatives can include the use of medication for pain, distress, and/or illness, environmental enrichment, and finding less invasive measuring procedures. We assume these definitions for the remainder of the paper.

It is not only important to be clear on the definitions of each of the 3Rs, their order is also significant. Questions of Replacement must always be addressed first, as the necessity of animal use must be proven prior to consideration of the other two Rs. As Russell and Burch note, “absolute replacement may be regarded as the absolute ideal” (1959, 70 [see also, 66]). After all, the best way to ensure that animals do not experience pain and distress in research or associated husbandry practices is not to use them at all. If an animal model can be absolutely replaced questions of Reduction and Refinement simply will not arise. Reduction to the minimum number of animals for scientifically valid results or maximizing the information obtained per animal is the second step, with questions of Refinement coming third. In short, this means researchers should make every effort to ensure that where possible they do their research without using animals, only addressing the other two Rs when they have determined that using animals is scientifically necessary.

Certainly, the CCAC gives every impression of being ardently committed to the 3Rs and sincere in their obligation to the Canadian people. However, they are not, for the most part, charged with making decisions about what research is done. The work of applying these ethical standards is mostly performed by institutional ACCs and researchers. Ultimately, the CCAC’s success in ensuring that research using animals conforms to the 3Rs framework can only be instantiated indirectly through the robust application of these principles by ACCs. It follows that we should expect the CCAC to create policies that facilitate an efficient, rigorous, and good faith application of 3Rs principles by ACCs.

Research approval process

As noted above, research using animals in Canada usually goes through two stages of approval. The first stage typically involves approval for funding by private or charitable organizations (such as the Heart Foundation) or, alternatively, approval by one of the two agencies responsible for funding public scientific research through grants and awards—the Canadian Institutes of Health Research (CIHR) and the Natural Sciences and Engineering Research Council of Canada (NSERC). Each of these organizations has an internal peer-review process that allocates funds to scientifically promising research. It is only in the second stage that the ethical justification and the social value of research is considered for approval by ACCs. To be eligible for funding from CIHR or NSERC, institutions are required to maintain a certification from the CCAC. This certification depends, in part, on having properly composed ACCs and AUP forms. Moreover, certified institutions are subject to regular and interim visits from members of the CCAC to comprehensively evaluate the institution’s animal research program (Griffin and Locke 2016, 274–6). Included in this visit is the submission of the institution’s AUP form template, which is judged according to general guidelines.

Several policy documents outline what the CCAC requires ACCs to put in their AUP forms. The CCAC Guidelines on Animal Use Protocol Review (1997) (1996) and Terms of Reference for Animal Care Committees (2006b) both highlight the importance of incorporating the 3Rs principles of humane treatment into the AUP forms. The latter provides examples of a number of Replacement alternatives that are worth consideration, and a number of possible refinements that might be employed in order to “protect and enhance animal health and welfare” (CCAC 2006b, 6). These documents also note that each ACC is responsible for ensuring that all members and animal users are familiar with CCAC guidelines and policies (CCAC 2006b, 9). While the CCAC requires AUP forms to make explicit reference to the 3Rs, there is no universal set of questions that must be included. Instead, ethics committees are expected to be familiar enough with the 3Rs (and with the CCAC policies and guidelines) to adequately address scientific and ethical concerns.

Differences in legal jurisdictions, as well as differences in university expertise, research priorities, and facilities, may necessitate some differences in AUP forms. However, the justification for giving ACCs the freedom to create their own AUP forms is broader in scope: the CCAC wants to allow AUP forms and ACC decision-making to reflect the “nature and culture of the institution” (CCAC 2010, 7). Given that ACCs are expected to have a solid understanding of both the 3Rs and their own role in approving research on scientific, social, and ethical grounds, one would expect to see only trivial differences in how key 3Rs concepts are approached. Indeed, it is only on this assumption that giving ACCs the freedom to create their own AUP forms can be justified.

However, two studies raise concerns about giving ACCs the freedom to create their own AUP forms. The first, a survey which was conducted by past and present members of the CCAC, casts doubt on whether the 3Rs have been well understood by both researchers and members on ethics committees (Fenwick and Griffin 2011). Participants were asked to answer a series of questions intended to evaluate their attitudes toward the 3Rs. Out of 414 participants, 72% were principal investigators (PIs) and the remaining were “graduate students, post-doctoral researchers, research associates, and animal care staff” (319). The survey showed that while participants had an adequate understanding of both Replacement and Reduction, their understanding of Refinement fell short with only 49% mentioning “minimizing pain and distress” in their definition (320). This is troubling because, of those surveyed, 37% identified as past or present members of animal ethics committees (319). Given just this information, there is a worry that some members who serve on ethics committees have an inadequate understanding of the 3Rs portion of the AUP forms that they are tasked with creating and evaluating.

The second, an ethnographic study, asked twenty-eight animal ethics committee members how they approached their role in evaluating research proposals (Schuppli 2011). The study found that twenty-three participants (82%) thought that “approval by peer review [the first stage of approval] was an assurance not only of scientific merit but also of social value and that the research was justified to proceed” (415). In other words, the majority of committee members took for granted that benefits of research, outweighing the cost to animal subjects, had already been established by funding bodies. It is not surprising, then, that the study also found that only half of the participants specified that they used a “harm-benefit assessment” when evaluating research proposals, despite this type of analysis being one of the main functions of ACCs (413–14).

While some time has passed since these studies were conducted, there is no indication that these issues have been resolved. While they call into question the 3Rs training and relevant ethics competency of ACC members and researchers, one might expect to have more confidence in the AUP forms that aid their work. After all, AUP forms are subject to regular evaluation by the CCAC (CCAC 2019b). Presumably, it is the responsibility of the CCAC to correct any deficiencies that may exist in these forms during the process of certification. We should therefore expect these forms to meet the ethical standards set out in the 3Rs framework. In the next section, we articulate what, given the CCAC’s avowed commitments and 3Rs resources, one could reasonably expect to find in AUP forms across all CCAC-certified institutions. These expectations inform the questions that we use to interrogate the forms in the subsequent section and the analysis and recommendations that follow.

What is implied by the CCAC’s commitments

As noted above, ACCs and AUP forms only come into play during the second stage of research approval. Although these forms differ across institutions, they are all expected to conform with the standards set by the CCAC. One of the most important resources for the development of AUP forms is the CCAC’s AUP Worksheet (2009), which is an informative guide for researchers to use (at their discretion) to ensure they have adequately investigated 3Rs alternatives in their proposals. The AUP Worksheet contains a series of questions that help researchers to comprehensively break down their search for alternatives into actionable steps. As such, it is reasonable to expect that a considerable portion of the information relevant to a thorough search for 3Rs alternatives would be elicited by all AUP forms, ensuring that all proposed research using animals in Canada be subjected to consistent ethical standards.

What’s more, as the CCAC requires that AUP forms make explicit reference to the 3Rs it is reasonable to expect definitions and examples provided for each of the Rs (CCAC 2006b, 5–6). The 3Rs should also be listed in the order in which they should be considered by researchers: Replacement, Reduction, and Refinement, respectively. In the section dedicated to the 3Rs framework, we should also expect to find questions that prompt the researcher to explain how each R was considered at the experimental design stage.⁴ Without this information, ACCs cannot determine the extent to which these ethical considerations have been incorporated into the overall plan for the proposal.

CCAC guidelines require that AUP forms prompt researchers to provide “a description of why sentient animals must be used for the project, [and] of how the applicant arrived at this conclusion (e.g., searches of databases on alternatives)” (2006b, 6). This requirement ensures that ACCs can follow the reasoning and access the sources used by the researcher to determine which alternatives were worth consideration, incorporation, and rejection. Additionally, researchers must provide “a description … of possible replacement alternatives … and justification if these are not to be employed” (2006b, 6). We should therefore expect that AUP forms elicit information about researchers’ search for Replacement alternatives, including where and how they searched for relevant information, and what the results were. AUP forms should then ask which potential alternatives have been found and considered and why they were rejected. These questions allow ethics committees to assess the sources and reasoning employed by researchers, thereby allowing them to determine how well research proposals meet the standard of Replacement. It is difficult to see how ACCs can identify flaws in research proposals or recommend improvements without this information. Of course, the same reasoning follows for questions about Reduction and Refinement alternatives. Unless AUP forms elicit specific information about the experimental design process it is unclear how ACCs can ensure that researchers (and ACCs) are up-to-date on best practices in animal research and that no valuable information is overlooked.

Forms should also ask researchers to give a scientific justification of their animal model. Any experiment that uses animals when they are poor models for answering the research question under investigation cannot be ethical because it has no reasonable expectation of producing scientific knowledge that might outweigh the harms to the animals in question. Justifying the animal model is a necessary step for scientifically valid, and thus ethically acceptable, animal research. A scientifically robust and clearly articulated justification of the animal model allows ACCs to determine whether the research question can be answered by the proposed research program (an increasingly pertinent issue, given recent questions about the translatability of animal research [Garner et al. 2017]), prior to asking whether there are any viable Replacement alternatives.⁵

Researchers should also be prompted to provide a calculation showing that the minimum number of animals are used under the proposed experimental design. Asking the researcher to provide this calculation ensures that the fewest number of animals required for significant results are being requested. This question also makes transparent those factors that affect the minimum required sample size, which could help with considerations of alternative experimental designs. This calculation, however, does not contain information about whether there may be Reduction alternatives available to the researcher to further minimize the number of animals used (e.g., choosing to take more or different data).

To determine that no obvious Reduction alternatives have been overlooked, we should expect that AUP forms include a question prompting researchers to compare the number of animals requested to the number used in similar experiments. This comparison would allow researchers to uncover new possibilities for 3Rs alternatives that may have otherwise been overlooked and would set a conservative upper bound to the number of animals used in similar experiments. It would also enable ACCs to review how 3Rs alternatives have been employed in relevantly similar projects, helping to ensure that best practices are being followed to minimize the number of animals used.

Forms should ask researchers whether there are any opportunities to share data, to share tissues, or to re-use animals. Again, researchers would be expected to conduct a search for these opportunities. Each of these are recognized by the CCAC as important Reduction alternatives (2009). While the implementation of each of these alternatives will be more or less manageable depending on the context of the proposed research, they must at least be considered for research to be consistent with 3Rs principles.

With regard to Refinement, forms should explicitly mention pain and distress, or equivalent terms, and the researcher should be asked how they plan to mitigate them. This question is important to ensure that the researcher has thought about each component of the research proposal with respect to its impact on the animal subjects’ wellbeing and that the researcher has planned to resolve pain and distress (or, indeed, other harms) where possible. We note that Refinement problems may result from efforts to address Reduction, particularly when animals are used for multiple experiments, so AUP forms should be designed to prompt sufficient information so that ACCs can evaluate this risk.

Researchers must also consider Refinement alternatives in light of their institution’s Standard Operating Procedures (SOPs). Researchers are required to adhere to SOPs which contain information about animal handling and care. There can be upwards of 150 SOPs for any single institution (McGill 2019) and as such comprehensive analysis of SOPs is beyond the scope of our project. As a consequence, we did not expect to see specific information about the housing and handling of animals in research facilities on AUP forms. Our focus for the Refinement section is, thus, primarily concerned with highlighting exceptions to SOPs. Moreover, we think AUP forms should still address general harm reduction strategies to ensure that the researcher has considered Refinement in their experimental design.

Finally, we should expect forms to ask about the “Category of Invasiveness” (CI) and the “Purpose of Animal Use” (PAU) of the proposed research, as required by the CCAC (2006b, 5). The CI is a marker of how much pain and distress animals are subjected to in an experimental protocol. CI categories range from A-E, where “A” represents no pain and distress and where “E” represents the highest degree of pain and distress (CCAC 1991). The principle of Refinement clearly requires that the CI be kept to a minimum where possible. In addition, “the intensity of protocol review should depend on the level of invasiveness of the procedures … with the most invasive procedures subject to the most intense level of review” (CCAC 2006a, 21). The CI therefore indicates how much scrutiny the goals and methods of the project should be accorded in the ACC’s evaluation.

The PAU value, on the other hand, represents the kind of social and/or scientific benefit of the research under consideration. The PAU value includes the following six distinct categories of use (represented as PAU0 to PAU5 respectively): breeding, general scientific investigation, medical science, testing to meet environmental regulations, product or appliance testing, and education (CCAC 2018, 10). While the CI is an important marker that indicates a greater ethical cost of research, the PAU indicates the kind of benefits that are being pursued. As such, both play an important role in any harm-benefit analysis, which researchers and ACCs are expected to use in evaluating research proposals.

The expectations outlined above help to foreground the kinds of questions that should be included on AUP forms. Forms that do not contain these questions may fail to elicit all of the information necessary to properly evaluate a research proposal. It is surely reasonable to expect that all AUP forms will be able to prompt all information relevant for research assessment following 3Rs principles, despite their differences.

In the next section, we describe our methods, using these expectations to develop a set of items by which we evaluate AUP forms. These items are intended to capture one or more of our expectations in a way that allows us to measure whether they have or have not been met. Our method allows us to evaluate how well these forms meet the 3Rs principles both absolutely and relative to one another. This, in turn, gives us the information needed to evaluate the extent to which the CCAC’s commitment to the 3Rs principles is reflected in the practices and policies of the people applying them.

Materials and methods

We focus our analysis on the U15 Group of Canadian Research Universities, as they produce the majority of research funded by competitive peer-reviewed grants (79%) as well as the majority of contracted private-sector research (83%) in Canada (U15 2019).⁶ As leaders, the U15 are expected to maintain high standards in research, including ethical standards. Not only do these institutions have the resources and institutional infrastructure to create and maintain first-rate policies but they are highly incentivized to do so given the potential reputational costs and financial risks of failing to meet acceptable standards in animal care.

To collect AUP forms, we accessed those available on university websites and requested access to those forms that were not readily available. Because we could not access some of the forms directly our data depends on the good faith efforts of those research administrators who sent us the forms; we cannot discount the possibility that some information was lost in this process. Forms were then reviewed using a set of thirty items, developed from our expectations outlined above. (To avoid confusion, we refer to the questions that we pose to AUP forms for the purposes of comparison as “items”, and questions on AUP forms are referred to as “questions”.) The items we pose are split into four categories: a set of general items and a set pertaining to each of the 3Rs. These items were used to examine and compare forms with respect to the differences and similarities in how they elicit the appropriate information from researchers. After first reviewing the forms on the basis of our initial set of items, strengths and weaknesses that had been revealed were then used to incorporate new items. This reflective process was used to generate the list of items used in the final analysis, which comprise the items listed in each of the five tables in the next section. For each item we determined whether any given form failed or met these standards. Results were then tabulated and discussed.

Most items have four possible valuations with respect to an AUP form: “Yes (3Rs section)”, “Yes (Other section)”, “No” and “Other”. The distinction between “Yes (3Rs section)” and “Yes (Other section)” is that, while both valuations indicate that the form elicits the required information (or close enough), the former indicates that it appears in the dedicated 3Rs section, while the latter indicates it appears elsewhere. This distinction helps to determine where different information is located on an AUP form. Items satisfied outside of the 3Rs section are often not placed in the same ethical context as questions located within this section. When a form failed to elicit the required information it received the valuation “No”. There is a fourth category of valuation, “Other”. Here we include any form where it is debatable whether an item is addressed or which frames the subject matter in some noteworthy way, such as misleadingly or with a restricted scope. In effect, we applied “Other” to indicate those forms that were neither clearly “Yes” nor clearly “No” for a given item.

Results

We have developed five tables that show the results of our analysis. Tables 1, 3, 4, and 5 display results by item category (General, Replacement, Reduction, and Refinement, respectively). Table 2 displays the results of all of the database-related items taken from the original four categories. Noteworthy results that are not necessarily visible in the tables are discussed in the body of the text. For the sake of providing a transparent analysis, we have also provided discussion for some of the more difficult interpretations.

Table 1. This table displays the results of general items. They include the items on the left side, the valuations above, and a summation of the values for the 14 AUP forms we analyzed on the right. General items posed address the logic of the 3Rs, such as ordering (1.5) and their clear articulation (1.1, 1.2, 1.3, 1.4, 1.6), as well as issues beyond the 3Rs that are nonetheless crucial for ethics evaluation, such as the social value of research (1.7), Purpose of Animal Use (1.8) and Category of Invasiveness (1.9). For the second item (1.2), asking if there is dedicated 3Rs section, forms that have a section with a concentration of 3Rs related questions and resources was considered sufficient for a “Yes.” Likewise for the items concerning subsections (1.4) and the order of subsections (1.5).

General items		Yes (in 3Rs section)	Yes (other section)	No	Other
1.1	Does the form use a heading to distinguish the 3Rs section?	8		1	5
1.2	Is there a 3Rs section?	14
1.3	Does the form use headings to distinguish between each of the 3Rs?	8		6
1.4	Are there Replacement, Reduction and Refinement subsections?	11		3
1.5	Are the 3Rs correctly ordered in the 3Rs section (Replacement, Reduction, Refinement)?	10		4
1.6	Links to 3Rs search tools and literature	8		3	3
1.7	Asks about the social value of research (benefits to humans or animals, or scientific knowledge)	1	7	3	3
1.8	Asks researcher to provide Purpose of Animal Use (PAU)	2	11	1
1.9	Asks researcher to provide Category of Invasiveness (CI)		14

Table 2. This table addresses items related to database searches for 3Rs alternatives, including keywords used for the search and the results. Items are listed on the left side, and the 3Rs category that we applied the items to appear above. Four numbers displayed in the represent the number of “Yes (in 3Rs section),” “Yes (other section),” “No” and “Other” values respectively.

Alternatives Search items		Replacement	Reduction	Refinement
2.1	Asks if alternatives search was performed	1, 0, 12, 1	0, 0, 14, 0	0, 0, 14, 0
2.2	Asks which databases were searched	1, 0, 13, 0	0, 0, 14, 0	0, 0, 14, 0
2.3	Asks which keywords were searched	1, 0, 13, 0	0, 0, 14, 0	0, 0, 14, 0
2.4	Asks about relevant search results	1, 0, 13, 0	0, 0, 14, 0	0, 0, 14, 0

Table 3. This table displays the results of Replacement-specific items. Items are listed on the left side, and the 14 forms are classified on the basis of whether they addressed the item in a specific 3Rs section, addressed the item in a different section, did not address the item, or had questions that might on some readings be thought (albeit ambiguously) to address the item, which are counted as “Other.” Items address Replacement definitions (3.1) and examples (3.2) as well as Replacement-specific issues such as the justification of the animal model (3.3), possible alternatives (3.4), and the reasons for their rejection (3.5). Any form that asks the researcher generally to justify their choice of animal or species in relation to their goal and/or in the context of Replacement, satisfies the justification item (3.3). For the fourth Replacement item (3.4), identifying rejected Replacement alternatives, any form that asks researchers whether they “considered” alternatives fails to satisfy this item but is counted as “Other.” Questions asking “how” researchers considered alternatives, while better than those asking “whether” they have, are still considered too vague and count as “Other.” The reasoning behind this choice is described in more detail in the body of the text.

Replacement items		Yes (in 3Rs section)	Yes (other section)	No	Other
3.1	Defines Replacement	6		3	5
3.2	Provides examples of Replacement	4	1	3	6
3.3	Asks to justify animal model	10	3		1
3.4	Asks researcher to identify existing Replacement alternatives that have not been incorporated into the project, or to describe (with sources) how it was determined that no such alternatives exist	6		1	7
3.5	If Replacement alternatives that were not incorporated into the project have been identified, asks why they are not suitable	6			8

Table 4. displays the results of Reduction-specific items. Items are listed on the left side, and the 14 forms are classified on the basis of whether they addressed the item in a specific 3Rs section, addressed the item in a different section, did not address the item, or had questions that might on some readings be thought (albeit ambiguously) to address the item, which are counted as “Other.” Items address reduction definitions (4.1) and examples (4.2) as well as items that clearly address reduction specific requirements, such as the search for opportunities to share data (4.5) and tissue samples (4.7) and re-use animals (4.6). For the third item (4.3), which addresses the calculation showing that the fewest animals required for statistically significant results has been proposed, any form that did not ask for specific calculations to this end did not satisfy this item. With respect to the items addressing opportunities to re-use animals (4.6), any form that mentions animal re-use in the context of either sourcing animals or what happens after research is satisfies this.

Reduction items		Yes (in 3Rs section)	Yes (other section)	No	Other
4.1	Defines Reduction	6		3	5
4.2	Provides examples of Reduction	2		5	7
4.3	Asks researcher to show calculation that number of animals used is the fewest required for statistically significant results	4	5	3	2
4.4	Asks researcher to compare the number of animals used in similar experiments (if applicable)	2		11	1
4.5	Asks if search found opportunities for data sharing across institutions	2		12
4.6	Asks about opportunities for animal re-use/transfer	2	2	10
4.7	Asks about opportunities for tissue sharing	2		12

Table 5. This table displays the results of Refinement-specific items. Items are listed on the left side, and the 14 forms are classified on the basis of whether they addressed the item in a specific 3Rs section, addressed the item in a different section, did not address the item, or had questions that might on some readings be thought (albeit ambiguously) to address the item, which are counted as “Other.” Items address Refinement definitions (5.1) and examples (5.2) as well as items that clearly address Refinement specific requirements, such as plans for mitigating pain and distress (5.4) and deviations from SOPs (5.5). Regarding the fourth item, which asks how the researcher plans to mitigate pain and distress (5.4), any form that asks which refinements are going to be included or that asks anything substantially similar to this will satisfy this item.

Refinement items		Yes (in 3Rs section)	Yes (other section)	No	Other
5.1	Defines Refinement	6		3	5
5.2	Provides examples of Refinement	1		5	8
5.3	Mentions “pain and distress” or equivalent	8	4	2
5.4	Asks researcher how they plan to mitigate pain and distress	8	1	3	2
5.5	Asks researcher to justify deviations from standard environment conditions (usually determined by SOP policies)	2	7	3	2

As shown in Table 1, while all AUP forms analyzed had a dedicated 3Rs section, many did not explicitly refer to the 3Rs. Instead, these forms titled this section “Justification of Animals Use” and/or referred to “Alternatives” or the “Consideration of Alternatives.” Three of the eleven AUP forms that had Replacement, Reduction, and Refinement subsections did not use subsections headings. Of the three forms that did not have 3Rs subsections, one asked a single question addressing all 3Rs, one did not contain a Refinement question and one had neither a Refinement nor a Reduction question. Of the forms that had subsections, one ordered Reduction first and one ordered Reduction last.

Recall that CCAC policies require that PAU and CI are listed on AUP forms (CCAC 2006b, 5). While all forms asked researchers to list CI, and all but one asked to list PAU, no forms asked this in the 3Rs section. Additionally, only one form of eight that asked the researcher about the social value of the proposed research did so in the 3Rs section. Three of the forms that were counted as “Other” for this item did not specifically ask the researcher about benefits to humans, animals or scientific knowledge. Instead, these forms simply asked the researcher about the “benefits,” “potential benefits,” and “expected benefits” of the proposed research.

Of the eight forms that provided links to 3Rs search tools and literature, six only linked to the CCAC’s 3Rs microsite. One of the remaining two linked to their own internal 3Rs website, while the other linked to the defunct Go3R tool. Forms counted as “Other” either gave dead or generic links with no substantive 3Rs information or referred to search tools without linking them.

As we explain above, it is reasonable to expect forms to elicit information about how 3Rs alternatives have been sought and considered as well as why prima facie plausible alternatives were rejected. Table 2 reports our findings on these items. Only one university addressed the search for alternatives for Replacement, and one that was counted as “Other” for two of the four items instructs the researcher to perform a search for Replacement alternatives and asks whether alternatives were found (see Table 2). No university asked specifically about alternatives searches done for either Reduction or Refinement. Notably, many universities asked very general questions about how the researcher “considered” or “addressed” the 3Rs. Though it is possible that some researchers may provide information about their search for alternatives in response to these general questions, this information is not clearly elicited.

No university provided definitions of just one of the 3Rs but not the others (see Tables 3–5). However, they varied more with regard to giving examples of alternatives. Five forms did not contain any 3Rs definitions but instead linked or referred researchers to the CCAC’s 3Rs microsite where definitions are provided. Many forms also relied on their links or references to search tools and literature for examples of alternatives. Since these items track how well 3Rs concepts are clearly articulated and reinforced on AUP forms, all forms that indirectly linked or referred to outside sources that contain 3Rs definitions and examples were counted as “Other.” It is, however, noteworthy that some of these forms did this more explicitly for the purpose of providing definitions and examples of alternatives compared to other forms that provided links and references as a more general resource.

Only one form failed to ask the researcher to justify their animal model (see Table 3). Three forms, however, asked this outside of the 3Rs section of the form (such as, for example, in a dedicated “justification” section). Forms were much less consistent with the remaining Replacement items.

The first of these, which asks the researcher to identify existing Replacement alternatives that have not been incorporated into the project, or to describe how it was determined that no such alternatives exist, attempts to capture how well the forms prompt the researcher to describe what the next best alternatives are or why there are no viable alternatives, if none exist. We counted six forms that did this sufficiently, while one form entirely failed to approach the problem identified by this item (by only asking why animals must be used). The remaining seven forms were counted as “Other” because, while they approached this problem, they asked researchers questions that may fail to elicit the information intended by this item as described.

Of those seven, most asked whether or how the researcher considered Replacement alternatives or the tenet of Replacement, or they asked whether there are alternatives, often accompanied with a space where the researcher is prompted to respond in some depth. These were counted as “Other” because they cannot be expected to consistently elicit information in a way that ensures that researchers have made sincere and well-informed efforts to find Replacement alternatives and have good reasons for rejecting them (though, the openness of these questions means that some answers, perhaps many, might supply the appropriate information). Notably, three forms, all of which were evaluated on this item differently, framed the problem exclusively in terms of Absolute Replacement, despite the fact that Relative Replacement alternatives must also be considered.

Similar problems were found with the final Replacement item which asks researchers to explain why alternatives were found to be unsuitable. Particularly noteworthy are three forms that satisfied the previous item but counted as “Other” for this item. The first of these takes for granted that any viable alternatives should have been incorporated into the project, and so this form only asks the researcher to describe how it was determined that there were no remaining alternatives, after first asking which alternatives have been incorporated. The other two ask the researcher to describe alternatives that have been considered, but do not ask specifically about what makes these alternatives unsuitable for pursuing the goals of the proposed research. Additionally, two forms that asked the researcher whether there were any Replacement alternatives, which were evaluated as “Other” for the previous item, satisfied the final Replacement item.

While the results of the items addressing Replacement questions were mixed, Reduction questions (see Table 4) were even more concerning. Five out of nine forms that asked the researcher to provide the sample size calculation showing that the fewest possible animals were being used did so outside of the 3Rs section. One form that was counted as “Other” did not make it clear that statistical significance was an important aspect of justifying animal numbers, and the other referred the researcher to justify animal numbers in an appendix form that was not made available to the authors for analysis. This latter form was also counted as “Other” for the item dealing with comparing requested animal numbers to numbers used in other, similar research. Just two forms asked for such a comparison or suggested to the researcher that such a comparison was a possible basis for justifying requested animal numbers.

Some forms prompted researchers to explain the general Reduction strategies to be employed. However, the majority did not explicitly request information that the CCAC has identified as being important Reduction considerations for researchers (2019d): prompts about opportunities for inter-institutional data sharing, animal transfer/sharing, and tissue sharing were absent from almost all forms.

Refinement items also provided mixed results (Table 5). While all but two forms mentioned pain and distress (or equivalent), several forms did so outside of the Refinement subsection. For a related item, we evaluated forms to see whether or not researchers were asked to explain how they plan to mitigate pain and distress and found that nearly a third of the forms did not. AUP forms that fail to mention steps taken to alleviate discomfort of animal subjects may rely on their institutional SOPs to address the necessary refinements which would account for the absence of this question. However, we found that three forms did not ask researchers to justify deviations from SOPs, which is information ACCs must take into consideration during their deliberations.

As shown in the tables above, we found significant variations in AUP forms across fourteen members of the U15 Group. In the discussion section, we note the implications of the results above for the CCAC given the potential barriers these variations pose to ethical animal-based research.

Discussion

Limitations

Before discussing the results of our analysis, some limitations should be noted. First, the AUP forms analyzed were made available to the authors as read-only documents. As such, web-based protocol forms were reformatted by the institutions prior to being released for our review. This leaves open the possibility that information was lost when the forms were transferred into this read-only format. While no obvious gaps in the forms were identified by the authors, this possibility cannot be ruled out. The results above take for granted that the forms received were complete and accurate.

Second, while some universities make their AUP forms available to the public through their website, others have restricted access. Because the majority of forms were received as classified documents, it was necessary to anonymize the results of our findings. As a consequence, we were not able to provide an extended table of our results in an appendix. It was determined that this information could be used to match AUP forms to their corresponding universities through a process of elimination (i.e., by comparing our extended results table with those forms made available online to the public). Despite their mandate to communicate with the Canadian public, the CCAC does not require that ACCs make AUP forms public. As such, the level of opacity is left to the discretion of each institution. This limitation of our analysis was, as a result of this policy, unavoidable.

Finally, a potential limitation exists with the sample size. Although our response rate was high, 93%, with only one of the U15 Group failing to respond, the data does not include institutions outside of the U15. This, however, should not be seen as impacting the quality of our results for one important reason: the variations found in the forms we have analyzed show deficiencies that, even if other forms happened to elicit the appropriate information from researchers, would still indicate a failing on the part of the CCAC to ensure proper adherence to the 3Rs. Moreover, because, as mentioned above, U15 universities produce approximately 80% of competitively funded and privately contracted research in Canada, we speculate that these forms (or their precursors) likely were used to approve the use of the majority of sentient animals currently in research in Canada.

Problems: Ambiguity, inconsistency, and insufficient information

When considered together, these results make it difficult to be confident that all animal-based research conducted in federally-funded institutions in Canada has gone through a consistent, rigorous, and ethically robust review process. Our analysis found that the majority of forms did not ask researchers to provide the databases used to inform their considerations of Replacement, Reduction, and Refinement alternatives. If AUP forms do not request sources for alternatives searches (including both the databases searched and the keywords used), then ACCs must deliberate without being able to retrace the reasoning used by researchers. Without access to this information, ACCs may have difficulty determining that a thorough investigation of alternatives has been conducted. This could be remedied by requiring that all forms elicit this information.

AUP forms were also found to use imprecise or ambiguously worded questions. Many forms, for example, asked researchers whether they had considered alternatives, which is technically a yes/no question. Others asked researchers how they had considered alternatives which, though still open-ended, at least prompts the researcher to elaborate. These questions could be strengthened by explicitly asking the researcher to provide the sources and reasoning used to determine that alternatives have been sought out and incorporated into the project when suitable.

Finally, it was found that forms often failed to provide definitions and examples for each of the 3Rs. Instead, a number of forms linked to the CCAC’s microsite which houses this information. This leaves open the possibility that at least some researchers will not review the link, assuming (wrongly, in some cases) that they already possess a sufficient understanding of 3Rs principles. Given that nearly half of experts working in animal-based research surveyed failed to refer to “minimizing pain and distress” when asked to provide a definition of Refinement (Fenwick and Griffin 2011), we cannot take for granted that researchers have a sufficient understanding of the 3Rs or possible alternatives when they are submitting research proposals.⁷ Even experienced researchers may submit proposals without adequate consideration of alternatives because they assume that they do not need to revisit the 3Rs concepts or they assume that they are up to date on best practices.

Just as forms often relied on the 3Rs microsite to provide definitions and examples of alternatives, the majority relied on it to act as the sole resource for 3Rs literature and search tools (Table 1 [1.6]). However, at the time of writing, the page dedicated to providing resources for a literature search contained a number of dead links: out of twelve links, six were dead (CCAC 2019g). Access to search tools and reliable resources is crucial for ensuring that 3Rs alternatives are thoroughly sought out and implemented. Universities rely on the CCAC’s microsite to provide this information, and so the CCAC should ensure that there are current and active high-quality resources available.

The CCAC asks institutional ACCs to create their own AUP forms so that questions can be tailored to suit the “nature and culture of their institution” (CCAC 2010, 7) as well as any legal requirements specific to their province. While this strategy may be preferable for some parts of AUP forms, the ethical considerations contained in the dedicated 3Rs section are based on a set of pre-established guidelines that should not change from institution to institution. Given the morally conservative character of the 3Rs, it is difficult to see how a dedicated 3Rs section could fall afoul of provincial laws, which, at worst, promise to add more restrictions and requirements in addition to those imposed by the CCAC. The results of our analysis suggest that when AUP forms are created by institutional ACCs, these universal principles are not enforced adequately and consistently. As such, we suggest that the CCAC take responsibility for creating and standardizing the 3Rs portion of AUP forms to ensure that these concerns are remedied.

Solution: Standardization and digitization

Our findings reveal a number of problems across AUP forms and it is the obligation of the CCAC to resolve them. On this basis, we recommend that a dedicated 3Rs section be created by the CCAC as boilerplate language that must be used on AUP forms at all certified institutions in Canada. This would allow the CCAC to incorporate their policies and guidelines into a comprehensive set of questions that elicit all necessary information from researchers. Standardizing this portion of AUP forms would also give the CCAC direct control over how questions are worded, eliminating the problem with imprecision and ambiguity identified in our analysis. With standardized 3Rs sections on AUP forms, ACCs would be better positioned to fulfill their role in evaluating the merits of research proposals using cost-benefit analyses, as they would have access to all of the information necessary for this purpose.

Another significant advantage of standardization would be the ability to digitize and archive research proposal documents. This would allow for the creation of a national database of past and current research proposals, as well as their evaluations by ACCs. In this way, AUP forms would function as a seamless method of registering animal-based research similar to the registration of clinical trials using human subjects. Indeed, such a database would be difficult to create without standardization. A standardized form that is filled out and submitted electronically could be formatted and archived in a machine-readable mark-up language. This archive could then be opened to ACCs and researchers, who would then be able to make nuanced search queries. Such a database could yield valuable information about similar research proposals, experimental models, hypotheses, results, and the ways in which the ethical standards of the 3Rs were considered by all parties in the approval process.⁸ Without a publicly supported national database (as the situation currently stands), Canadian researchers are dependent on private organizations or other countries for providing access to information that is necessary for fully evaluating the ethical merits of Canadian research. Additionally, a shared national database would expand the possibility for more international collaboration in animal research ethics.

The inter-institutional sharing of information that is facilitated by our proposal would act as a safeguard against unnecessary harm done to animals in research by providing more resources to ACCs and researchers to search for 3Rs alternatives (see Strech and Dirnagl 2019 for a similar proposal). This sharing could also improve the scientific value of research being done, as it would give ACCs and researchers the ability to do the following: (i) avoid unnecessary and harmful duplications; (ii) track the failures and successes of similar models; (iii) develop better hypotheses and models; and (iv) understand what ethical decisions were made in similar circumstances. Standardization, digitization, and the distribution of information across institutions would allow for better and more consistent ethical outcomes in the research approval process.

The benefits of our proposal become particularly salient when we consider the problem of publication bias in research. Publication bias occurs because research that fails to disprove the null hypothesis (i.e., research that fails to support its own hypothesis) is less likely to be published, either due to authors’ self-selection or editors’ lack of interest in null results. The problem of publication bias is prevalent in many areas of scientific research, but it has an even greater negative impact in areas of research where there is an ethical cost to performing experiments, such as in animal research. Publication bias wastes time, resources and increases the likelihood that published results are false positive findings. If researchers and ACCs only have access to published results, they are needlessly working with a smaller pool of information. This leads to fewer experimental alternatives being identified in the research approval process.

Since the CCAC is committed to the 3Rs, they should be similarly committed to eliminating the negative effects of publication bias in animal research, which necessitates that they make all animal research results available to researchers and ACCs across Canada. Where information is considered proprietary, relevant parts could be embargoed for a limited amount of time so long as enough information is available to avoid pointless duplication and provide clear evidence of 3Rs compliance. Our proposal to standardize and digitize AUP forms, and to make all AUP submissions and evaluations available to researchers and ACCs across Canada, could be readily and easily amended to include making experimental results available. This tool would also be able to document research proposals, indicating whether experiments were underway and also would be a place for researchers to indicate when studies had been terminated for unforeseen reasons, or had produced results that were not published. Eliminating the negative effects of publication bias in the research approval process is consistent with the call for “[f]unding agencies [to] encourage investigators to publish both positive and negative results of the primary hypotheses tested in the research projects they have funded” (Joober et al. 2012, 152).⁹ Moreover, because the research registration process would be effectively folded into the research approval process it would not create an additional burden to researchers.

In addition to addressing these concerns, the database we propose serves another important function: to gather and sort quality information about 3Rs alternatives, making this information readily available to researchers. For researchers to determine whether or not their scientific objective can be achieved using non-sentient models (or methods which use fewer animals or that re-use animals without producing unacceptable Refinement issues) it is imperative that researchers have at hand a comprehensive set of search results.

As this database grows in Canada, and as other nations follow suit, a new possibility will arise: information from each database could be integrated into a larger database (a “superbase”) which also has the ability to pull sources from already established databases (PubMed, TOXNET, etc.). More advanced search tools, such as semantic search engines, could then retrieve information from the superbase thereby increasing the resources at the researchers’ disposal. Perhaps the most important addition here would be integration with a database of new Replacement alternatives – such as new computer simulations or tissue chip technologies. Because successful Absolute Replacement strategies by researchers will lead to experiments that do not have to go through ACC approval, the database that we describe here will never be sufficient as the sole search tool used be researchers and ACCs to ensure 3Rs compliance.

The idea that this would be useful for ending the unnecessary use of and harm to animals in research is not new. The Go3R semantic search engine, which is now defunct, once served this purpose. This tool was well-received by those concerned with the humane treatment of animals in research (Sauer et al. 2009; Liebsch et al. 2011) and by those care committees in the U15 Group who, as we saw, recommended this tool to their institutions’ researchers. Go3R was also well-received by the CCAC, which is evidenced by their endorsement of the tool (CCAC 2019g). This endorsement serves as a testament to the usefulness of well-developed search tools for proper adherence to the 3Rs.¹⁰

Despite the enthusiasm with which Go3R was received, and despite the promise it offered to the scientific community, the project fell short on funding and the website hosting it was removed due to lack of funding (2017). In the future, through international collaboration, the database we have proposed could be integrated into a larger project that resembles the Go3R search tool and superbase. In the meantime, creating a national database would solve some of the immediate problems uncovered through our analysis by offering a useful resource for ACCs and researchers. Needless to say, adequate funding from NSERC and CIHR would be essential to creating and maintaining such a database.

In brief, the standardization of the 3Rs section of AUP forms promises not only to correct some of the inconsistencies among forms revealed by our analysis, but also provides important opportunities for improving 3Rs compliance. Having research proposals, ACC approval decisions, and experimental results digitized and archived in a national database – something that requires standardization if it is to be efficiently implemented and effectively searchable – would allow researchers and ACCs across the country to use this information to effectively investigate and assess how scientific aims can be met through the use of 3Rs alternatives.

In the next section we consider how, in addition to producing better ethical outcomes, implementing our recommendation would place the CCAC in a better position to fulfill its mandate of being accountable to the Canadian public. Accountability requires that the interests of the public be accurately tracked and taken into consideration by the CCAC, which requires a degree of transparency that does not currently exist. As such, we outline how our proposal can facilitate accountability by increasing transparency in the research approval process.

Accountability and transparency¹¹

To some extent, the CCAC has shown its commitment to communicating with the Canadian public by making their policies and guidelines readily available through their website and welcoming public comment on these guidelines when they are developed. It is through these documents that concerned members of the public can stay informed of the ethical standards, the stated mandates, and the responsibilities of the CCAC. In addition, data compiled on the number of animals used in research at CCAC-certified institutions is released to the public in the CCAC Annual Data Report (CCAC 2019c). This report provides the public with the total number of animals used, divided by species, and assigns a Category of Invasiveness (CI) and a Purpose of Animal Use (PAU) value for each group of research subjects (recall that the CI value indicates the degree of pain and distress experienced, while the PAU value represents the kinds of benefits being pursued by research).

Through the annual report, the Canadian public should be able to evaluate the research that takes place at CCAC-certified institutions (CCAC 2006a, 6). Together, the CI and PAU offer the public a rough estimation of the ethical costs and the scientific benefits of research, respectively. However, these broad categories are insufficient to conduct an accurate ethics assessment. To conduct a harm-benefit analysis, it must be established that the harms specific to a research proposal be justified by the benefits. In-depth explanations of how the 3Rs were considered is a requirement for all submissions reviewed for approval, not simply CIs and PAUs alone. If ACCs cannot be expected to evaluate proposals with these broad categories, then the Canadian public should not be expected to do so either.

While the public does not have access to the specifics of those proposals that have been approved, it is the role of ACCs to act as proxies for the Canadian public. These committees are composed not only of experts in the field (scientists, veterinarians, and researchers) but also include laypersons who act as community representatives. The role of lay members is to “help identify and address the public’s opinions and concerns,” and in so doing reflect the interests and values of Canadians (CCAC 2006, 12). Like other committee members, community representatives are expected to be familiar with both the 3Rs framework and how to conduct harm-benefit analyses. Community representatives also have access to experts in the field, removing the worry that decisions will be considered without the appropriate level of scientific literacy.

To facilitate transparency and communication with various publics, we recommend that the information contained within our proposed database be accessible to all Canadians. The potential risks – for instance, to researchers conducting controversial, though approved, experiments – can be mitigated in two ways: First, the public should have a different level of access from that of researchers. What this means is that, unlike researchers, the public will not be able to access sensitive or identifying information. Details about where the research is taking place and about who is involved in the project need not be released, as this information is not required to evaluate the merits of the research being done. Second, we recommend that a line of communication be opened up with the public. This could be done by providing a forum where Canadians can discuss their concerns with the CCAC. This way, concerns that reflect the values of the public can be addressed and investigated without disrupting the research process.

Currently, the interest and values of the general public are represented by those who serve on ethics committees. Annual reports are released with the intention of keeping Canadians informed of the research is being conducted. However, we have shown that the information contained within these reports is insufficient for a thorough evaluation. Without the ability to review the specifics of research being done, the Canadian public has no way of confirming that their various interests are being upheld and, as such, it would be reasonable for them to lack confidence in the process. Our proposal to increase transparency with the general public and to provide a forum for feedback will, if implemented, ensure that the values of Canadians are better integrated into the research approval process. If the CCAC implements the recommendations outlined in this section, their mandate to communicate with the public – and, as a result, to be accountable to the public – will be significantly strengthened.

Concluding remarks

Our analysis has uncovered deficiencies across AUP forms that challenge the integrity of the research approval process. The CCAC has the responsibility to reform policies and practices that act as barriers to informed ethical research. As such, we have proposed that the CCAC standardize and digitize AUP forms and archive research proposal documents. We have also recommended that the CCAC use this process to develop more effective search tools for researchers, and that they reconsider the degree of transparency and responsiveness to Canadian values present in the current system. Our proposal will help to avoid the ethical costs of causing unnecessary harm to animals in research by providing the CCAC with actionable policy reform measures.

The CCAC recognizes that “each ACC has an ethical, scientific, and social responsibility to apply protocol review and approval criteria in a fair, equitable and consistent manner” (CCAC 1996, 1). While the CCAC requires consistency within each institutional ACC, it is our suggestion that consistency across institutions is of equal importance. Having the CCAC create a standardized 3Rs section will ensure that all institutions are subject to the same ethical standards. Standardization could then be used as a stepping stone to create a database which would help Canadian researchers to better adhere to ethical research practices.

The ethical costs of failing to appropriately respond to the deficiencies we have uncovered are high. Our aim in analyzing and revealing these deficiencies is constructive: we want to support the CCAC in meeting their obligations by reforming policy to better align with their commitment to the 3Rs standards. The CCAC claims that monetary cost should not be a deciding factor in the consideration of 3Rs alternatives given the ethical weight of what is at stake (CCAC 1996, 3). The same principle should apply to the evaluation of the research approval process itself.

Acknowledgments

The authors would like to thank Vaughan Black, Andrew Fenton, Paul Locke, an anonymous reviewer, and an audience at the Canadian Society for the History and Philosophy of Science meeting in 2019 for their helpful feedback. We also gratefully acknowledge support from the Social Sciences and Humanities Research Council of Canada, the Canada Research Chair in Bioethics and Philosophy and the Canada Research Chair in Environment, Culture and Values.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Canada Research Chair in Bioethics and Philosophy, the Canada Research Chair in Environment, Culture and Values; and the Social Sciences and Humanities Research Council of Canada [430-2016-00828].

Notes

1. See Hughes (2003) for criticism.

2. This one exception came into effect in June 2019 with the passage of the Ending the Captivity of Whales and Dolphins Act. Now in order to conduct research on cetaceans in captivity a license must be sought from the province (Ending the Captivity of Whales and Dolphins Act, 2019).

3. We note that it is not at all clear on the face of it what might be meant by “less sentient.” The CCAC defines it as follows: “animals that current expert peer advice and interpretation of scientific evidence indicate have a significantly lower potential for pain perception, such as some invertebrates” (CCAC 2019e). Unfortunately, the care they show on this point is not universally exercised. When it comes to the use of the term “animal” they equivocate between a broad meaning, the biological taxon, and a narrower one, sentient animals. This is problematic as it muddies some issues, for instance, questions around what counts as Absolute versus Relative Replacement. Russell and Burch (1959) clearly limit their discussion to sentient animals, which at the time they took to include all vertebrates (6).

4. This would have the salutary side effect of pressing researchers to consider the 3Rs during the first step of the research approval process when they are applying for funding.

5. Our point here is, in effect, about the logical relationship between research with scientific merit and research that is ethical. It is possible that research that has considerable scientific merit might fail to be ethical, because the costs to the research subjects outweigh the epistemic benefits. It is not possible that research without scientific merit could be ethical as there is nothing to weigh against the costs (barring, of course, external considerations about things like the benefit to research assistants of having work). The question of whether scientific merit alone, often thought of as “knowledge for knowledge’s sake,” should outweigh most ethical considerations is actually a controversial one. Although we acknowledge the difficulty of weighing epistemic goods against other considerations, we do note that it is uncontroversial in human-based research that costs to research subjects are given primacy of place when evaluating research proposals.

6. The U15 comprises the University of Alberta, University of British Columbia, University of Calgary, Dalhousie University, Université Laval, University of Manitoba, McGill University, McMaster University, Université de Montréal, University of Ottawa, Queen’s University, University of Saskatchewan, University of Toronto, University of Waterloo, and Western University.

7. Given that the survey was conducted nearly a decade ago, it may be the case that efforts to promote awareness of 3Rs principles has reduced the number of researchers who are unable to provide accurate descriptions for each of the Rs. However, when AUP forms are created in a way that presupposes a level of competence that may not always exist there is a danger that the 3Rs will not be properly addressed.

8. The National Institutes of Health (U.S.) RePORT database is similar to the tool that we recommend creating (RePORT 2019).

9. The database PROSPERO (2019) uses a similar method to remedy the problem of publication bias across a wide range of disciplines. While it has been well-received, registration is not mandatory (Stewart, Moher, and Shekelle 2012).

10. The Netherlands is developing The Interspecies Database (2019), which serves a related purpose.

11. Frasch (2016) expresses similar concerns about the transparency and accountability for animal-based research in the US.