Simulation in clinical learning.

Abstract

Simulation in clinical learning is becoming increasingly commonplace. The use of simulation spans both the undergraduate and postgraduate domains and has come about as a result of a number of driving forces. Medical education reform has been widespread throughout the world.1, 2 These reforms recognise the need to prepare undergraduate learners for their roles in the ever changing world of medical practice. Moreover, widespread adoption of the concept of clinical governance has implications for learning in both undergraduate and postgraduate arenas.3 The purpose of clinical governance is to ensure that the highest quality of care is provided to patients. Areas within the remit of clinical governance upon which simulation may have an impact include risk management, lifelong learning, education, training and continuing personal and professional development, staffing and staff management, continuous quality improvement and the management of poor performance. Technological advances in recent years have seen the development of a variety of models and manikins ranging from simple replications of body parts for task-based learning of some examination skills, to high fidelity patient simulators driven by complex pathophysiological computer models developed to provide a highly valid replication of clinical environments. Anaesthesia has been one speciality where the use of high fidelity simulation has been seen to be of great value in promoting patient safety and in developing competence and on-going training,4 drawing from the world of aviation and its experience in flight crew training and teamwork.5 Simulation has extended into other fields, including emergency medicine,6 and surgical training7 and assessment,8 and has crossed into the undergraduate educational arena. In this latter field, opportunities can be created to practise the management of clinical conditions which would otherwise not be possible, safe in knowledge of them taking place in a protected environment.9 At the same time, there has been criticism of medical education from within the profession of the relative paucity of sound educational research that underpins much of medical education innovation.10 Medical simulation offers tremendous opportunities for the advancement of our understanding of learning because it is consistent with very different ways of conceptualising learning, and because research in very different paradigms11 can be accommodated. A more explicit discussion of theories of learning, and of paradigms of research, might therefore helpfully inform future practice and future research on simulation. We begin here to explore something of what this might mean. Early writing on behaviourism focused on the acquisition, retention and decay of new behaviours, and emphasised the effects of such factors as the role of reward and ‘punishment’, the most effective proportions of each in a learning situation, and the consistency with which rewards and punishments are given after a particular action has been performed. For a summary, see Gross, 199212 (p.166 et seq). More recent work in this area has considered the importance of feedback on how closely the student's response matched the ideal response (feedback known as ‘knowledge of results’), the frequency of such feedback and the way in which learners are encouraged to interact with that feedback (e.g. comparing it with their prior estimate of the error involved in their action).13, 14 Since behaviourism ignores the ‘black box’ of the mind, it lends itself to ‘low road’ models15, 16 for the transfer of learning (making the behaviour automatic so that it can be widely applied). This can be achieved through such tactics as over-learning (see Magill 1989,17 p.395–6). In this supplement, Glavin & Maran make explicit reference to this area of learning theory.18 Constructivism has its basis in Piaget's notion of learners constructing their understanding of the world through their interactions with it. Learning can occur when new experiences of the world fit into a learner's existing cognitive structures (assimilation), and when new experiences cannot be made to fit, and therefore challenge those existing cognitive structures which then have to be changed (accommodation). This theory of learning has had a longstanding influence on thinking about science education.19 The notion of assimilation suggests that a learner will have a richer understanding of a concept when they have experienced a sufficiently broad range of examples such that they appreciate the full range of application of that concept. The notion of accommodation suggests that learners' conceptual structures will be expanded, modified and made more sophisticated when they encounter unexpected experiences (cognitive conflicts) that reveal the inadequacy of existing structures. These processes of assimilation and accommodation happen in individuals' informal learning through their experiences of the world and can also inform formal processes of learning. It is of course possible to imagine another reaction to a cognitive conflict: disengagement, dismissal of the unexpected experience as flawed in some way, and re-entrenchment into the old position. However, in the classroom situation the teacher has an important role in encouraging the positive, constructive response and avoiding disengagement. Important things that the teacher will do include exploring learners' preconceptions in advance so that the learner is explicitly aware of their initial thinking; providing the cognitive conflict; drawing attention to the mismatch of expectation and experience in ways that emphasise that the unexpected experience is valid (e.g. making clear that this was not a mistake, or trick, or fault in the equipment); asking questions about the mismatch to prepare the learner to be receptive to new ideas (‘What does this mean? How can we make sense of this?’); teaching those new ideas; explicitly pointing out that the new ideas work where the learner's preconceptions worked and that they work where the learner's preconceptions did not.20 The constructivist model of learning also stresses the importance of a learning environment that is perceived to be safe – one where preconceptions can be exposed without risk of ridicule21 (and perhaps, in the simulation context, without the risk of injury to the patient). Social constructivism22 puts emphasis on the importance of the social context of learning. Rather than emphasising the previous constructivist focus on the learner's individual interaction with the environment, social constructivism stresses the role of social interaction in helping learners to construct their new understanding. One consequence of this is a concern with how the teacher can ‘scaffold’ the student's learning, providing support enabling the student to advance the boundaries of their knowledge, and then withdrawing that support in an appropriate manner so as to encourage independence. There are many ideas about the ways in which scaffolding can be achieved.23 We would see the elements of teaching listed in the previous paragraph as ‘scaffolding’ of a sort, but other possibilities include the provision, at first, of simplified experiences that highlight the key issues, modelling of the appropriate problem-solving behaviour by the teacher with explanations of what is being done and why, discussion to focus the learner's attention on the most salient parts of an experience, and discussion about the processes of problem solving that have taken place (to enhance meta-cognition). The links with the theoretical underpinning of problem-based learning (PBL) are quite evident.24 A teacher planning ‘scaffolding’ would also plan ways in which the support of this scaffolding can gradually be withdrawn, and through explicit discussion of how this learning might apply to other situations, would emphasise ‘high road’ transfer of learning involving thoughtful consideration of the range of application of the learning.15, 16 These theories of learning clearly have application to professional learning in the simulation environment, but are also relevant in many other situations where learning is intended (e.g. the school classroom or the medical school physiology class). The remaining theories apply particularly (though not exclusively) to situations of professional learning. Reflective practice25, 26 places emphasis on supporting the learner in planning, acting (including reflection-in-action), evaluating (including reflection-on-action) and re-conceptualising (thinking differently about the situation as a result of the experience that has been reflected on). A key idea is that theories, contextual information and values each have a role to play in each of these stages of reflection. The interaction amongst these factors can be complex; for example different theories may point to different plans for action (at the planning stage) or to different understandings of actions (at the evaluation stage). The choice between theories may be strongly influenced by the details of the particular context in which the action is being taken, or by the participant's values, or both. The theories that should be considered in each stage of reflective practice include formal theories that are relevant to the action. These might come from very different fields (e.g. theories relevant to a particular action in a simulator might come both from pharmacology and from the social psychology of group leadership.) However, theories also include personal theories (preconceptions, craft knowledge, intuitive knowledge). New theories (especially new personal theories) can be created in the re-conceptualisation phase. The contextual issues that should be considered include the resources available, as well as the personal histories of the people involved and the recent history of the institution in which the action is taking place, the expectations of the stakeholders, the other characteristics of the people involved (e.g. their motivation and abilities), and the network of local and national policies relevant to the actions. Re-conceptualisation may lead to decisions to change some aspect of the context (e.g. to buy a different resource) or it may lead to a determination to see a fixed contextual factor in a different way. Values include the values held by the learner as well as those of the institution and the national setting. Re-conceptualisation may lead to changes in personal values or a decision to lobby for broader change. Issues for teachers who are working within a reflective practice model include how to help students to make explicit their personal theories, their craft knowledge and their values, how to encourage critical engagement with a range of formal theories, how to look in depth at the physical and social elements of a given context, how to understand their actions in light of these factors, and how to ask not only how they might act differently in the future, but also how they might think differently about the particular scenario in which they have been engaged. A further issue for the teacher is how to make their own thinking about such issues explicit to the student. Lave & Wenger27, describing situated learning, emphasise that learning takes place as students enjoy ‘legitimate peripheral participation’ in a community of practice. For teachers, it can be quite challenging to seek guidance from this theory, as Lave & Wenger 199127 (p.40) explicitly argue that it is not ‘an educational form, much less a pedagogical strategy…but a way of understanding learning’. Nevertheless, it might be valuable to regard simulator experience as legitimate peripheral participation in the ‘real’ world of medical provision for human patients. Seen in this way, some advantages of the simulator are immediately apparent. For example, Lave & Wenger argue (p.96) that in most situations of legitimate peripheral participation, the component parts of a complex practice have to be learnt in an order that differs from that in which those parts are deployed in practice. This is necessary so that ‘less intense, less complex, less vital tasks are learned before more central aspects of practice’. Because issues of patient safety do not arise in the simulator, this is less of an imperative and the learning experiences offered to students can have greater authenticity. If we choose to see simulation as legitimate peripheral participation, some interesting issues arise. We will consider just three of these here. First, since peripheral participation might be empowering (as the student gradually moves to more and more central forms of participation), or disempowering (as the student is constrained to remain on the edge) how can an empowering transition from the simulator to the ‘real world’ best be managed? Secondly, since communities of practice do not necessarily imply communities with agreed positions (but rather communities in which people are engaged in critical discourse), how might potentially vulnerable students engage with this critical discourse – how might they pilot a path through the different opinions and practices of the range of people who will support (and assess) them (e.g. those responsible for simulation and those who take a less positive view of it)? Thirdly, since learning may not be as closely structured by our formal pedagogic strategies as we might traditionally expect, but rather is shaped by informal opportunities – so that ‘apprentices learn mostly in relation with other apprentices’ (Lave & Wenger p.93) – how can appropriate opportunities be provided for students to learn in this way? Situated learning might lead us to view the simulator as a part of the community of practice (or activity system) that is, for the most part, concerned with real patients. Engeström's activity theory28 emphasises the learning that occurs within, and across the boundaries between, the different activity systems in which a learner operates. In this model, we might think of simulation as one activity system and ‘real world’ application of the same understandings and skills as another. We might then be tempted to extend the model and see the problem-based learning class, and the biochemistry seminar as further activity systems in which the student is engaged. For Engeström, each activity system consists of individual and group actions directed towards something (the object). Each activity system should be considered as a whole, key questions being about the interplay of the different voices, each with its different history, role, dispositions and concerns. The object-oriented actions of the activity system are ‘always, explicitly or implicitly, characterized by ambiguity, surprise, interpretation, sense making, and potential for change’ (Engeström,28 p.134). They are influenced and mediated by the social setting of the activity system. As in the case of situated learning, this raises questions about how students (in the simulator setting, PBL class or ‘real world’ operating theatre) can be helped to be clear about the object of the system, how they can be ‘tuned in’ to the different influences within the system and their interconnections, how they can respond positively to differences of view, and how they can be creative about ambiguity and surprise (rather than reacting negatively or defensively to such circumstances). In addition, activity theory emphasises that related activity systems jointly construct meanings about objects on their boundary and come to jointly shared understandings about these boundary objects. It is through these boundary objects that meanings can be shared across the systems and enhanced learning can result. This implies that, as well as trying to structure each separate system to support students' learning, we should consider how meanings will pass from one to the other, and what reconstructions might be required of students in that process in order to be responsive to the different characteristics of the systems. Simple notions of learning theory in the lecture theatre and applying it on the ward, or developing a skill in the simulator and using it in the operating theatre are problematised by this approach. One might summarise the scientific and interpretive paradigms (after Lincoln & Guba29) as follows. There is a single tangible ‘reality’, parts of which can be studied independently; The whole is the sum of the parts; It is possible to separate the observer from the observed, the knower from the known; What is true at one time and place may well be true at another; Causality is linear – causes lead to effects; and Any enquiry can be value free. Scientific paradigm research is concerned with validity and reliability, and seeks generalisability, such that results obtained from a sample of a population can be taken as applying to that population as a whole. Realities are multiple, and are individually (or socially) constructed; The knower cannot be separated from the known; We can only make statements that are time and context bound; All entities are continually shaping each other; and Inquiry is inevitably value-bound. Such research is concerned with issues such as authenticity, completeness, coherence, trustworthiness, and plausibility and seeks transferability (through the thoughtful application of ideas from one context to a different context). It is a feature of much current simulator research that it strives to meet the demands of the scientific paradigm and is often framed in terms of experimental or quasi-experimental studies. However, it tends to rely on expert judgement of learners' performance (e.g. experienced colleagues' judgement of performance as recorded on video-tape). As such it struggles with the problems of the validity and reliability of the data. Small sample sizes are also common, with the consequent risk of overlooking real effects that, because of sample size, fail to reach statistical significance. Perhaps one way forward for simulator research is to begin to encompass the interpretive paradigm; another might be to be more targeted about the use of the scientific paradigm so that studies that lie within this paradigm are designed to capitalise upon its particular ‘range of convenience’. Another issue is that the learning theories on which the teaching interventions in the simulator are based are often implicit. As a result, actions may not always be consistent with all the demands of the relevant theory. Whatever the paradigm of the research, increased theoretical clarity with respect to learning theory might be worth striving for. Together, these ideas could lead to rather different kinds of enquiry. We give some examples of this below. Since behaviourist learning theories do not concern themselves with the ways in which the learners make sense of the stimuli and responses that they receive and make, perhaps data from the simulator itself (rather than data from interviews with the participants) is particularly relevant in instrumenting research of this kind. Objective data are available in high fidelity simulations, e.g. the values taken by the variables in the underlying mathematical models, and the data on the actions of the learner (drugs administered, timing in relation to changes in the model parameters). These could be presented to the learner in ways that are consistent with the guidance that behaviourist learning theory can provide, e.g. on the timing and frequency of feedback. This kind of research could be conducted with confidence within the scientific paradigm. Issues of the validity and reliability of the data would be technical issues related to the adequacy of the simulator models. Sample size may remain a problem. Perhaps the sharing of data across simulator sites (possibly via the internet) could enable researchers to assemble data on a sufficient number of learners to begin to address this. Perhaps an increased use of simulators in the undergraduate medical curriculum will also help – as long as the data available are used as research data (and appropriate ethics procedures are followed). In a constructivist approach one might encourage a learner to reveal their thinking about a set of symptoms (perhaps by drawing a concept map), then arrange the simulator to respond to these symptoms so that the match between the predictions based on the learner's thinking and the outcomes in terms of the simulation response can be explicitly seen. Discussion of the accepted theories related to the symptoms can be followed by the clarification that these theories lead to predictions that match the simulation response. There are good opportunities for real clarity here as the theories inform the mathematical models, so the simulator response should (!) match those theories very well. Within activity theory, if the simulator suite is an activity system, what is the object of that system? Perhaps it is unequivocally the learning of the learner (whereas in the ‘real’ situation, the object may be less clearly defined, slipping between the learner's learning and the patient's needs for diagnosis and treatment within an ethically sound framework). A question then arises: what are the boundary objects that allow discourse across the boundaries between the simulation activity system and the activity system of the ward, or A&E department? There is a suggestion (K. Seddon, personal communication) that theories may serve as boundary objects in teachers' learning, hinting that maybe there are links to be made between activity theory and reflective practice. In a constructivist, or activity theory approach, it would be more appropriate to conceptualise the research within the interpretive paradigm: taking seriously the meanings that are constructed by the people involved in a learning situation; recognising that these meanings will have implications for what those people think and do. The interpretive paradigm leads to concerns with what the learners feel about simulation, what they think they have learnt, and what enabled or disrupted that learning. It is not concerned with the hard data of the scientific instrument, nor does it seek to generate generalisable insights based on representative samples. It is an approach that leads to the understanding of a situation in depth from the perspective of the participants (amongst whom figure the researchers who ultimately, put their own interpretation on the data). One particular problem with simulator research is the difficulty of controlling for increasing simulator familiarity during a scientific study of skills development through simulator teaching. This is much less a problem for research in the interpretive paradigm as one can interrogate interview data to see if issues of familiarity arise, or can actively prompt for them during the collection of data. Learning logs kept by learners, or observations followed up by interview in a ‘stimulated recall’ approach could also shed light on the issue of familiarity. If the interpretive paradigm is to find a bigger place in simulator research, it will be essential to pay attention to the criteria for quality: as noted above, the concepts of validity, reliability and generalisability have to be rethought in terms such as completeness, trustworthiness and transferability. Analysis of interview data has to be carried out with the same precision (and the same opportunities for creativity) as is the case with statistical analysis of data in the scientific paradigm. There must be clear protocols for how interview transcripts are to be analysed: for example through immersion in the data, and the identification of emerging themes30 or through detailed coding of units of meaning in the transcript and a subsequent search for the relationships between the codes.31 There must be methods for checking this analysis: do other researchers identify the same themes?; do others code the transcript in the same way?; once codes are agreed, do others assign the comments to the same codes? There must be systematic ways of reporting the interviews – so that the bad journalist's preoccupation with the extreme or the dramatic is replaced by the good researcher's preoccupation with integrity. The interpretation of the analysed data should also be explicit, clarifying the basis from which the researcher chooses to bring out the particular meaning that they do bring out, and reconsidering the data to see if there is anything that can disconfirm the emerging interpretation. In this stage of the work the opportunities for participant validation of the interpretation should be created wherever possible. The papers presented in this supplement, individually and collectively, make a significant contribution to our understanding of the place and impact of simulators in medical education of various kinds. Perhaps the issues raised above could broaden still further the scope of simulator use and of simulator research. If so, several future versions of this supplement may well be justified.