Refinement: promoting the three Rs in practice

Abstract

Refinement of scientific procedures carried out on protected animals is an iterative process, which begins with a critical evaluation of practice. The process continues with objective assessment of the impact of the procedures, identification of areas for improvement, selection and implementation of an improvement strategy and evaluation of the results to determine whether there has been the desired effect, completing the refinement loop and resulting in the perpetuation of good practice. Refinements may be science-driven (those which facilitate getting high-quality results) or welfare-driven or may encompass both groups, but whatever the driver, refinements almost always result in benefits to both welfare and science. Refinements can be implemented in all aspects of animal use: improved methodology in invasive techniques, housing and husbandry, and even statistical analyses can all benefit animal welfare and scientific quality. If refinement is not actively sought, outdated and unnecessarily invasive techniques may not be replaced by better methods as they become available, and thus outdated information is passed down to the next generation, causing perpetuation of old-fashioned methods. This leads to a spiral of ignorance, leading ultimately to poor practice, poor animal welfare and poor-quality scientific data. Refinement is a legal and ethical requirement, yet refinements may not always be implemented. There are numerous obstacles to the implementation of refinement, which may be real or perceived. Either way, in order to take refinement forward, it is important to coordinate the approach to refinement, validate the science behind refinement, ensure there is adequate education and training in new techniques, improve liaison between users and make sure there is feedback on suitability of refinements for use. Overall, refinement requires a coordinated ongoing process of critical appraisal of practice and active scrutiny of resources for likely improvements. In the busy world of biomedical research, this process needs help. In order to develop these themes further, a workshop was held at the LASA Winter Meeting 2006, UK, to assist in identifying potential obstacles to refinement, and then to explore and develop strategies for overcoming these obstacles in key areas. A range of strategies appropriate to different circumstances was identified, which should facilitate the implementation of refinements.

Refinement of scientific procedures carried out on animals to minimize adverse effects and to maximize the scientific benefit gained is a legal and ethical requirement, under the Animals (Scientific Procedures) Act 1986, the Animal Welfare Act 2006, and the Animal Health and Welfare Strategy for Great Britain, and also in the wider European and worldwide context (http://scienceandresearch.homeoffice.gov.uk/animal-research/; http://www.defra.gov.uk/animalh/ahws). However, refinements are not always implemented and the reasons for this are numerous (Hartley et al. 2004, Karas 2006).

In attempting to define or describe animal welfare, there is broad consensus that the ‘five freedoms’, as laid down by the British Farm Animal Welfare Council in 1979, can be taken as general indicators of animal wellbeing (FAWC 1979). These are freedom from hunger and thirst, freedom from discomfort, freedom from pain, injury and disease, freedom to express normal behaviours and freedom from fear and distress. Scientific research has the ability to impact on all these. For example, food and water restriction paradigms are commonly used in behaviour studies, metal cages and grid floors can produce discomfort, experimental surgery can cause pain, single housing and restricted space can impact on behaviour and handling methods or procedures can cause fear or distress. There is inevitably tension caused by the potential of science to impact on welfare and the need to maintain the welfare of animals used (Smith 2006). Researchers have an obligation to minimize the impact of the procedures they do on these five freedoms through the active implementation of refinements.

Fundamentally, the drivers for refinement are always the same, improving scientific quality or improving animal welfare, but on a practical level, cultural and organizational factors will impact on how change (including refinement) occurs in different settings.

Culture

The concept of academic freedom is highly prized, sometimes at the expense of more effective management. Within an academic environment the inadequacy of implementation of refinements may be recognized, but there is often a failure to take any action to address this because of the lack of management criteria such as key performance indicators and use of standard operating procedures (SOPs). Conversely, there is much within the culture of pharmaceutical research and development that encourages innovation including refinement. Within the drug hunting process, highly organized and disciplined teams work towards common goals. There are good communication channels and a focus on quality assurance. Investment in facilities, people and equipment is strong, with an inherent focus on technology and innovation, but there is not an ‘open cheque-book’ and a business case must be made for all expenditure to be effectively prioritized.

Drivers for refinement

Academic research tends to be very focused into narrow areas. Scientists are strongly driven to do high-quality research in their field, but may lack similar drive when evaluating the influence of methodology on the scientific quality of their results or on the welfare of animals used. The implementation of the 3Rs is seen almost as a separate subject area, whereas it would be better if seen as a normal part of the research package and integrated into it. Within the pharmaceutical sector, strong corporate drivers for consistency and harmonization within global organizations can be used to bring about refinement (e.g. housing improvements) and in an increasingly tough external environment, a strong focus on corporate reputation can also bring useful leverage to welfare initiatives. The project focus in drug discovery can provide impetus to refinement, such that refinements driven by project need happen ‘automatically’ but conversely refinement cannot be allowed to interrupt project progression. It can be more difficult to influence practice (e.g. change an anaesthetic regimen or animal model) with an established screening cascade: timing and opportunity are critical.

Scale

Academic research groups tend to be small and work independently, so refinements implemented at the individual project level have little impact in a wider institutional setting, whereas in industry, change can take time but when it happens it can typically be on a large scale across a whole department or site.

The refinement loop

Refinement is an iterative process, which begins with a critical evaluation of practice, leading to recognition and assessment of poor or suboptimal welfare, identification of the causes of this, selection of improvement strategies and implementation of these strategies. Any changes then have to be evaluated for their efficacy, thus the process begins again, forming the refinement loop.

Critical evaluation of practice

Researchers have to look critically at what they do and identify if there is a better way of conducting the research that will minimize welfare implications and maximize the scientific output, thus improving the harm: benefit balance and the justification for doing the work. It is particularly important that this is seen as an ongoing requirement: new developments in refinement may turn what was yesterday's best practice into today's outdated methodology.

Objective assessment of animal welfare and scientific quality

This requires first and foremost an understanding of wellbeing and what particular animals need. Only then is it possible to identify appropriate parameters to assess wellbeing and scientific quality in a meaningful manner. Sufficient resources have to be provided to perform the assessments effectively. This requires the researcher to schedule observations appropriately to maximize the likely detection of problems, to recognize where there is deviation from normal, taking into account species, strain and inter-animal differences and to be able to identify when there is room for improvement.

Selection of improvement strategies

Having identified that there is a problem and a likely cause for this, the researcher has to be able to select an appropriate improvement strategy. This requires active and effective utilization of an ever-increasing variety of sources of information and advice. This may include the veterinarian and animal care technicians, other researchers, journals or internet-based resources.

Implementation of improvement strategies

The implementation of any new methodology is accompanied by a period of inexperience, during which time it may not be possible to discern any improvement, and indeed may lead to a period of increased losses while researchers develop the necessary skills for the new technique. This may cause researchers to question the new method. Systems should be in place at institutions to deal appropriately with such a reaction, and make sure that the refinements are implemented so that the long-term benefits are realized. Researchers must also recognize that this will be the case and ensure that they have help, advice and support available while they are learning the new technique.

Evaluation of improvement strategies

Having implemented an improvement strategy, this must be evaluated to determine whether there has been an improvement. It may be that the new method is found to be better, or that the current method is confirmed as being the best available method: either way the information is valuable to others in the field. This information should then be disseminated to others working in the same area, ideally by publication in mainstream journals, but as a minimum it should be used for internal reference.

This process forms the refinement loop (Hartley et al. 2004) (Figure 1).

Figure 1

The refinement loop

Obstacles to refinement

Obstacles to refinement may be real or perceived, each requiring different strategies to overcome them. Real obstacles such as lack of money, facilities, personnel or equipment will require a patient, tactical approach using all organizational levers available to overcome them. Perceived obstacles arise from misunderstanding or misinformation and need a different approach: providing evidence-based strategies to tackle key issues, educating and informing to allay unfounded fears and influencing and persuading to change attitudes.

Numerous factors have been identified, which can impede the progress of refinement (Karas 2006). Particular issues include:

Lack of ownership: There are many people involved and it is not always clear who has the responsibility to implement refinement. Often no one wants to take responsibility and it is passed between the investigator, the project leader, the veterinarian, the animal care technicians, the ethical review committee, the regulatory official and even the funding agency. In addition, there may be ethologists, other scientists or professional ‘committee goers’ involved, who simply cause overload and confusion. Although the implementation of refinements requires team working and coordinated networks to develop and exchange information, there still needs to be one named individual with the responsibility to make it happen. Scientists may not feel responsible for the implementation of refinements, yet they are ultimately accountable for what happens to the animals, and should be encouraged to accept this personal responsibility. There is a necessity to define who is responsible for the implementation, and for them to take ownership of the issue. This person must understand what the animals need, what the science needs and be supported by others at the institution and empowered to actually make it happen.

Lack of support for the development of a culture of care: Veterinarians and animal care technicians need institutional support for their advisory role. A ‘flawed’ ethical review process (ERP) and inappropriate attitudes from investigators can undermine these key people, since they may not always have the authority and support from the institution to deliver change when it is needed.

Lack of academic credibility for refinement research: There may be a problem with the academic credibility of refinement research: subject areas such as laboratory animal science, animal welfare and ethology are not perceived as real science and are therefore not seen as relevant. Coupled with concerns about the effects of refinements on the science, and the inability to perform objective assessment of the effect of refinements on animals and on the science, this leaves the scientist with a conflict of interests, which can make interpretation of indices of welfare more difficult. The professional integrity of those who can perform objective assessments should be recognized.

Lack of resources: Many refinements do need extra resources. This needs to be recognized and understood by the senior management team who will usually not be part of the animal management team. These two groups need to communicate effectively.

Lack of knowledge on the part of the investigator about techniques available, and inability to utilize sources of information appropriately. The information is complicated, not concise, and not seen as relevant. Regulation is veiled in a raft of advice and guidance, which often confuses. It is an ever-moving picture and the scientist needs support and guidance, which is tailored to their individual needs.

Failure to actively seek refinement leads to a situation where techniques are handed down unchanged through generations of practitioners. This is particularly the case in biomedical research since although new researchers undergo a very basic training course the majority of their learning is ‘on the job’. This leads to a spiral of ignorance, in which techniques used become further and further removed from ‘best practice’ (Figure 2).

Figure 2

The spiral of ignorance

Making refinement work

Experience from both industry and academia has shown that it is possible for refinement to be taken forward. Important factors for success include determining what type of obstacle is involved, then determining the appropriate strategy to tackle it, including which people are the best ones to help. A lack of resources requires a different approach to a problem of attitude or knowledge. Strategies for success include capitalizing on opportunities, influencing and persuading, engaging with stakeholders to tackle concerns and good team working. Using a tactical team approach and all organizational levers is critical for success.

In a workshop held at the LASA Winter Meeting 2006, UK, potential obstacles were explored in more depth. Participants were divided into four groups of between eight and 10 participants, and each group was tasked with identifying strategies for overcoming obstacles to refinement in an area where the main issues were attitudinal.

The groups considered four key areas:

(1) How can we provide education and support for new researchers to stop perpetuation of poor techniques?

The key issues identified were inexperience and lack of confidence. They simply may not have the knowledge or confidence to try new techniques or challenge their supervisors. Also, new researchers are at the point of unconscious incompetence (Chapman 2007) (Table 1) – they may not recognize that they have a problem. This may also apply to more senior researchers, who have always used a particular technique and do not recognize that times have moved on.

Table 1

Conscious competence-learning matrix (adapted from Chapman 2007)

	Incompetence	Competence
Unconscious	(1) Unconscious incompetence the person is not aware of the existence or relevance of the skill area the person is not aware that they have a particular deficiency in the area concerned the person might deny the relevance or usefulness of the new skill the person must become conscious of their incompetence before development of the new skill or learning can begin the aim of the trainer or supervisor is to move the person into the ‘conscious competence’ stage, by demonstrating the skill or ability and the benefit that it will bring to the person's effectiveness	(4) Unconscious competence the skill becomes so practised that it enters the unconscious parts of the brain – it becomes ‘second nature’ common examples are driving, sports activities, typing, animal handling, listening and communicating it becomes possible for certain skills to be performed while doing something else, for example, knitting while reading a book the person might now be able to supervise and teach others, although after some time of being unconsciously competent the person might actually have difficulty in explaining exactly how they do it – the skill has become largely instinctual this may require longstanding unconscious competence to be checked periodically against new standards
Conscious	(2) Conscious incompetence the person becomes aware of the existence and relevance of the skill the person is therefore also aware of their deficiency in this area, by attempting or trying to use the skill the person realizes that by improving their skill or ability in this area their effectiveness will improve the person ideally makes a commitment to learn and practise the new skill, and to move to the ‘conscious competence’ stage the person has a measure of the extent of their deficiency, and a measure of what level of skill is required for their own competence	(3) Conscious competence the person achieves ‘conscious competence’ in a skill when they can perform it reliably at will the person needs to concentrate and think in order to perform the skill the person can perform the skill without assistance the person will not reliably perform the skill unless thinking about it – the skill is not yet ‘second nature’ or ‘automatic’ the person should be able to demonstrate the skill to another, but is unlikely to be able to teach it well to another person the person should continue to practise the new skill; practice is the most effective way to move from stage 3 to 4

These issues can be addressed by considering training, mentoring and providing a culture of support, which is the responsibility of the institutional official (in the UK this would be the certificate holder). He or she should ensure that there is both induction and refresher training available in the use of new techniques. There should be a clear chain of command and support for all new researchers, which should involve veterinarians, animal care technicians and scientists (including the principal investigator/study director). Best practice should be described and captured in SOPs – these should be widely available so everyone can see and use them. SOPs should not be viewed as ‘tablets of stone’ nor act as an obstacle to refinement. They need to be reviewed frequently and updated as necessary to ensure they remain best practice. A support network of experienced colleagues for new researchers, such as a user group, can provide a valuable source of advice. Care must be taken to ensure that the advice is up-to-date. There should be a system for confidential reporting of problems, so that appropriate experienced persons can observe procedures in progress where problems have been identified to give advice and identify/provide any training needs. A system should be set up for dealing with the consequences of identification of incompetence. Rather than being ignored, incompetence should be tackled by a system of support and retraining.

New researchers could benefit from having a person who is not their supervisor or manager appointed to provide support and advice. This person could act as a mentor, being available to listen to problems and offer help in a confidential way. Supervision is the responsibility of the principal investigator, so they have a key part to play in this. However, they may actually be part of the problem, and may also need mentoring and support. Supervisors need to be encouraged to do better by informal or formal mentoring themselves. This should be encouraged by the institutional official.

The culture at an establishment very much depends on the attitude of its leader. It is necessary to promote a culture of advice-seeking, so that new researchers feel confident to ask for help in areas where they are uncertain. New researchers must recognize that they are not competent just because they have completed training. The veterinarian and animal care technicians need to be available to give advice, and their expertise should be respected. A culture of support, not a culture of blame, needs to be fostered by the institutional official. It needs to be accepted that it is alright to need support.

(2) How can we improve the attitudes of senior researchers and supervisors and encourage them to move forward?

The main issue identified was that it is difficult to get people to accept that what they are doing and have always done may no longer be best practice, and that things have progressed. If challenged, the usual reaction is to defend the position held, ‘I've always done it this way and never had any problems’. This is a difficult issue to tackle and may take time to change. Change may be incremental and it may be necessary to go backwards before moving forwards while scientists are learning a new up-to-date technique. Different strategies may be needed for different cases. The terminology used is important, and this may need plenty of gentle persuasion and diplomacy. It is necessary to get researchers on side, so they will come back and ask for help again.

Here, it is necessary to explore why there is reluctance to change, explain why change is beneficial, and demonstrate that this occurs in practice.

Exploring why there is reluctance to change involves psychology – asking ‘Why do you do it this way?’ comes across as a challenge, and is likely to result in the person defending their position. Rather than attacking the methodology being used, it is more effective to use open, non-threatening questioning, such as ‘Have you considered x?’ Change requires time and effort, additional resources, and acceptance that what has always been done may no longer be appropriate.

Legal and ethical requirements encourage and require refinement, and there may be some leverage in explaining the potential benefits of the proposals. Both the local ERP and funders may look more favourably on proposals using more refined methods. It may also be possible to take advantage of academic competitiveness by citing academic competitors using new techniques or the possibility of attracting new talent (and possibly funding) to the laboratory.

It may be possible to demonstrate that refined methods work better or are more likely to attract funding. This may need some kind of proof, such as good-quality papers or the opinion of other experts to back it up.

Many people can be involved in this process, including the veterinarian, institutional official and scientific colleagues in the team who the researcher might respect or listen to. Sometimes it may be appropriate to involve the regulatory body, or draw the attention of the researcher to the guidelines used by funders or journals. Lay members from the ERP can be useful as they can ask naive questions and still be seen as non-threatening.

In the event that there is still reluctance to change and there may be implications for animal welfare, it may not be acceptable to wait to implement refinements. Sometimes the establishment has to be up front and say no. There is an ongoing legal and ethical requirement for refinement to be implemented, and the institutional official is responsible for ensuring compliance at their establishment. There should be a forum into which concerned individuals can feed their concerns, knowing that appropriate action will be taken.

(3) How do we improve the perceived credibility of animal welfare science?

Some scientists may perceive research into refinements as less important or peripheral, and such research is rarely published in mainstream journals where scientists get to see it. Welfare research is perceived as poor-quality science and is not considered relevant to many scientists. Welfare scientists should be encouraged to publish welfare research in more mainstream journals, where there is peer review by scientists, to improve the actual credibility of welfare science. There may be some difficulties getting mainstream journals to publish this type of research, e.g. there may be a need for control groups with less refined procedures/housing, increasing animal use and suffering to provide a comparison with more refined method, which may be a barrier to evidenced-based research.

A stock of ‘best-practice’ papers to be used as sources of information can be useful. These should be research evidence based, relevant and multidisciplinary with good statistical analyses and must be critically evaluated before being given to scientists. In particular websites such as the National Centre for the 3Rs (NC3Rs) and Centre for Alternatives to Animal Testing act as useful portals to up-to-date 3Rs information and ‘best practice’. It must be noted that often experiments are condition-dependent, and the conditions at the establishment may be different from those used in the study, which can render the results less useful. The institutional official should ensure that there is a team-based approach to this problem, facilitated through the ERP.

Scientists should be encouraged to include details of refinements – including those that did not have the desired effects – in their research papers. Awards for papers demonstrating commitment to the 3Rs could encourage scientists to seek and publish refinements. The editors of journals have a significant part to play in this, as do the funding organizations, who are in a position to evaluate research grant applications for best practice in animal welfare.

At present, scientists tend to require proof that refinements will not adversely affect the research that they are doing before they will consider implementing them. However, ideally it should be the other way round: scientists should find proof that the refinement does impact on their research negatively before rejecting a refinement. Also, refinements often improve the scientific quality of the results (Ritskes-Hoitinga et al. 2006): scientists should make it clear that they have evaluated the possible positive effects of any refinement on their research before rejecting it. The animals should always get the benefit of the doubt.

(4) How do we enhance the dissemination of information about welfare-based refinements, to get refinement onto the main agenda among the scientific community?

Information about refinements is often passed on by the veterinarian, animal care technician, or regulatory official, and sometimes researchers, but scientists often work very much in isolation. Refinements may be published in specialized laboratory animal science, ethology or animal welfare journals, which the scientists do not have time to read, and there is simply information overload. Animal welfare scientists may also work in isolation: they should be open to new developments in research techniques, in order to address the 3Rs' value of these developments and encourage their use in other disciplines.

A strategy is needed for finding relevant information on refinements simply. This may be achieved by encouraging the involvement of 3Rs experts at the study design stage so that all opportunities for refinement are recognized, encouraging interaction between scientists using intranets or newsletters and setting up fora for the sharing of ideas. Relevant information can come from new staff, laboratory animal science meetings, journals or discussions with experienced colleagues. Most mainstream science groups have 3Rs groups and run workshops, but to maximize the use of appropriate sources of information it may be necessary to employ the services of an information scientist. Such an individual could be a resource that is shared between establishments. All facilities should have access to the services of a 3Rs specialist adviser. Scientists should be encouraged to present refinement data in peer-reviewed manuscripts and at mainstream scientific meetings, to reach a wider audience and to demonstrate that there is benefit to them in terms of finance, scientific quality and time. Case studies can demonstrate the benefits of refinement, such as changing from tail to ear clips or saliva sampling for genotyping (Hawkins et al. 2006). Table 2 shows some key points and key personnel for implementation of refinement.

Table 2

Key points and key personnel for implementation of refinement

Obstacle	Solution	Key personnel
Provision of education and support for new researchers	Training	Institutional official
	Mentoring	Principle investigator
	Providing a culture of support	Veterinarian
		Animal care technicians
		ERP
Improving attitudes of senior researchers	Explore why there is reluctance to change	Institutional official
	Explain why change is beneficial	Veterinarian
	Demonstrate benefits in practice	Scientific colleagues
		Regulatory body
		ERP
Improving the perceived credibility of animal welfare science	Publish welfare papers in mainstream journalsUse journals and online resources as sources of best practice papersPublish refinements in scientific papersEvaluate research proposals for implementation of best practice in animal welfare	Welfare scientistsTeam approach led byinstitutional officialScientistsJournal editorsFunding bodies
Enhancing dissemination of information about welfare-based refinements	Involve 3Rs experts at study design stage	Veterinarian
	Encourage interaction between scientists	Animal care technicians
	Use information scientists to identify relevant informationPresent refinement data in mainstream publications or meetingsEvaluate new methods for applicability in other situations	Regulatory officialScientistsJournal editorsAnimal welfare scientists

ERP = ethical review process

Conclusions

• European legislation (the Animals (Scientific Procedures) Act 1986 in the UK) requires implementation of the three Rs.

• Refinement is an iterative process starting with critical evaluation of practice.

• Failure to actively seek refinement leads to a spiral of ignorance and perpetuation of poor practice.

• Numerous obstacles to refinement have been identified, which can prevent refinements from being implemented.

• To overcome these, it is essential to coordinate the approach to refinement within an institution or organization. It may be beneficial to identify someone with overall responsibility for this: a ‘refinement champion’ who is responsible for ensuring that refinements are actively implemented at the establishment.

• The institution needs to ensure that support is available for key people in developing a culture of care, and that the professional integrity of the experts in welfare assessment is recognized. The animal management team needs to communicate effectively with senior management, to make sure that adequate resources are provided to implement refinements appropriately. There should be a clear process for the implementation and reporting of problems, and the development of a culture of advice-seeking and support, not blame and isolation.

• Scientists need to recognize that there is a learning curve to all new methods, and accept that while inexperience may lead to new or increased problems in the short term, there will be longer-term benefits from perseverance.

• There is a need to enhance information dissemination, to support and guide scientists through the overabundance of advice and regulations to help them identify what is best practice.

Refinements may be developed to enhance animal welfare, to improve scientific quality or sometimes simply as a defensive reaction to prevent criticism of old-fashioned methodologies. The underlying reason of why a procedure is done matters less than what is done, when, how and by whom, and the validity of the results obtained. For the animal it is not important what we think, what we monitor or how we score it, but what actions are taken that directly affect it (Wolfensohn & Honess 2007): for the animal it is actual quality of life that matters.