Chapter 45 Teaching clinical reasoning to medical students
Clinical reasoning is the process by which health practitioners evaluate and make decisions on the diagnosis and management of a patient. It is of particular importance when a patient presents with what has been described as an ill-structured problem (Barrows & Feltovich 1987). The development of clinical competence is dependent on increasingly refined and elaborated medical knowledge (Schmidt et al 1990) and judgement (Round 2001). The critical endpoint of the reasoning process will result in decisions, often based on exploration of a range of possibilities that may include further history, physical examination or investigation.
Both the nature of the learning (whether in traditional or in problem based programmes) and the timing of clinical experience need to be considered in helping medical students learn clinical reasoning. In addition to experiencing a particular curriculum, students are also developing within a broader framework of professionalism (Mann et al 2005). Medical students initially try to understand patients who may present with a bewildering, unsorted array of complex information (clinical, personal, social, emotional) of uncertain relevance. In the early stages, students have only limited knowledge on which to build their reasoning. They are often anxious about the appropriateness and effectiveness of their communication skills in seeking and clarifying relevant information. The challenge for teachers is to encourage the ordering and prioritizing of information based on the most cogent elements; the generation, testing and refining of hypotheses; and the formulation of clear, specific, answerable diagnostic or therapeutic questions. What are the most appropriate strategies to achieve these aims and how are they best made explicit to students?
The idea of a ‘generic’ form of clinical reasoning is appealing (Schuwirth 2002). However, it has been noted that clinical reasoning is both domain-specific and idiosyncratic. The challenge for medical educators is not only to make explicit the processes of reasoning (Kassirer 1989, 1995), but also to help students identify the relevant and necessary content information and efficient ways of retrieving these data.
Reasoning alone is inadequate for clinical decision making; knowledge and understanding of basic mechanisms of human function (both normal and abnormal) are essential. Basic biomedical sciences must be linked to clinical and epidemiological information. Data from several sources must be organized into coherent representations of disease processes (Boshuizen & Schmidt 1992, Schmidt et al 1990). Recently, debate has centred around the effectiveness of different methods to tie the acquisition of knowledge more securely to the development of clinical reasoning. As Round (2001) has noted, relevant perspectives include psychology, clinical psychology, clinical practice and clinical education. She questions whether teachable cognitive skills can exist independently of their context. Nendaz & Bordage (2002) also ask whether teaching reasoning separately from content can be successful.
One research approach has been to identify the components of expert diagnostic and management skills. A diagnosis often represents an explanation of an illness (Elstein 1995), implicitly emphasizing the need for mechanisms of health and disease to be understood. Although hypothetico-deductive reasoning was for years seen as the commonest form of diagnostic reasoning, it now may be considered a relatively weak conceptual approach (Coderre et al 2003). Neufeld et al (1981) suggested that skills of practitioners and students are attributable to experience rather than to superior reasoning. Associations of symptoms and signs generate patterns that experts recognize quickly; for students, the patterns have little meaning. Increasing medical knowledge must be ‘chunked’ for manageability (Schuwirth 2002) and integrated into logically organized and elaborated ‘pattern recognition’ structures or ‘illness scripts’ (Schmidt et al 1990) in order to aid rapid, accurate and relevant retrieval.
Effective clinical reasoning is based on iterative information gathering, a process in which hypotheses are framed, tested, modified or discarded (Kassirer 1995). This process requires skills in communication and physical examination, as well as the selective ordering and interpretation of investigations, using the best evidence available (Sackett et al 1997).
Much of the research on clinical reasoning has focused on diagnosis, while management, comprising by far the greater part of patient care, has been relatively neglected. Management requires more than understanding the processes of disease, mechanisms for repair and means of alleviating symptoms. Technical expertise must be combined with a deep and empathic understanding of the patient’s perspectives and needs that derives from face to face encounters requiring effective communication. Students therefore need longitudinal experiences if they are to see how the clinical reasoning process contributes to patient care over time.
Teaching needs to include a number of common elements: observations are made, and information – often disorganized and not expressed in medical terms – is collected; the data are ordered into more or less formal hypotheses based on existing medical knowledge and experience; further inquiry seeks clarification; diagnostic possibilities are identified that can be eliminated; a plan is developed for further investigation and/or immediate management. An experienced clinician often undertakes some of those processes in parallel, rather than sequentially. Given the reports of significant errors in over- or under-estimating probabilities (see Round 2001), explicit discussion of those biases with students as they gain experience could contribute to some improvement in the transfer of clinical reasoning skills.
In teaching clinical reasoning, medical teachers must determine overall curricular goals, identify essential content and design the processes for learning which will best support the development of an effective medical professional. Not all teachers acknowledge the need to make educational strategies explicit to students, and some have little insight into inconsistencies in their own performance. However, we argue that it is essential to focus explicitly both on the processes of clinical reasoning itself and on the educational methods that support its development. Students need to be engaged actively as informed partners.
For the novice with limited knowledge, the parallel processing and chunking of related information is restricted; some structure to information gathering is essential. Templates with rigorous steps are initially useful to ensure that essential information is not missed, but rigid adherence is inefficient in the long term. When templates or algorithms are used uncritically, students may fail to recognize priority issues and to develop appropriately structured knowledge for rapid responsiveness (Schmidt et al 1990). A recent approach to diagnostic problem representation (Nendaz & Bordage 2002) has been shown to enhance students’ capacity to describe and recall problems more effectively, although interpretation was not improved. Strategies of this kind seem to be effective for students early in their clinical studies. Round (2001) notes that using algorithms – although they can be effective – is seen as too time consuming.
Grant (1989) has encouraged students to share experiences and articulate the processes they use to work through diagnostic problems. In a supportive and safe atmosphere, her students express themselves honestly and receive specific, sensitive feedback; they also observe and model the strategies of others. Schuwirth (2002) has also stressed the essential importance of feedback. Perhaps the most significant benefit is the development of metacognitive skills so that self-aware learners identify their thinking processes and monitor their progress. The strategy may well appeal to those who are convinced of the individuality of mental processes or who question the notion of imposing a single best reasoning process.
Most medical schools define a certain number of preclinical years focusing on basic sciences, and a number of clinical years where students interact with patients, illness is emphasized and clinical reasoning introduced. The transition can be difficult for a number of reasons. Students learn science by hypothetico-deductive reasoning from first principles (Niaz 1993), processes that are appropriate to the biomedical or physical sciences (Patel and Kaufman 1995). In contrast, such strategies are used by skilled medical practitioners only when problems are particularly difficult or obscure (Norman et al 1994). Students find it hard to reason backwards when confronted with patients with ambiguous symptoms and signs (Barrows & Feltovich 1987, Patel et al 1991).
When individual subjects are taught in isolation, little information is transferred between them. Thus the conceptual linkages necessary for effective clinical problem solving (Schmidt et al 1990) are not readily established. The recent information knowledge explosion within existing disciplines and the inclusion of new topics (e.g. molecular biology, intracellular signalling) has increased the overload of medical curricula, militating against the thoughtful reflection required for deep understanding. Pushed to master an increasing volume of facts, students resort to surface learning at the expense of critical analysis and thinking. When assessments value recall, students are discouraged from reasoning at all (Ramsden 2003). The resulting deficiencies hamper the later development of effective clinical reasoning when basic and clinical subjects must be interrelated.
Attempts to overcome these difficulties generally rely on the importation of basic biomedical sciences into the clinical teaching arena and the importation of clinical cases into the biomedical sciences. Some medical sciences lend themselves to presentation from the perspective of the abnormal (e.g. endocrine excess and deficiency help in understanding normal balances and controls; the function of neuroanatomical structures is illustrated by lesions). Such examples can introduce students to aspects of clinical reasoning. Since fewer medically qualified staff now teach in the early years, conceptual links between basic and clinical sciences are less accessible to the teachers and thus to the students. In traditional programmes, specific approaches to integrating clinical experiences and basic sciences include those of Coles (1990), in which basic science examinations were delayed until after the first clinical attachments, and Patel & Dauphinee (1984), in which students learned some basic science during clinical years. In both examples, students had elaborated their knowledge and were better able to retrieve and use basic information in clinical settings.
Hospital and community settings can be used to provide clinical examples of important concepts for students in early years of traditional programmes. At their best, such experiences provide not only a sense of relevance to the basic studies but also opportunities for students to see and model aspects of clinical reasoning. To be effective rather than tokenistic, however, the exposure must be well planned and students must be actively engaged rather than passive onlookers. The aims must be explicit and directly related to other concurrent learning.