Doctors learn to become competent and to make good judgements through the experience of treating patients with appropriate levels of guidance and supervision. They also learn by consulting and studying textbooks, reference books and journals. Then thirdly they participate in formally organised teaching and discussion. All three can be found in clinical settings; studying and teaching also occur in other settings -- library, seminar, lecture room, and a range of workshops or laboratories. This chapter focuses only on clinical settings, recognising that there is a continuum from treating patients under the watchful eye of more senior doctors, through engaging in discussions about those patients, to participating in discussions about patients one has not treated or being taught about the diagnosis and management of a patient in a mini-seminar or lecture soon after visiting that patient. The use of literature informs both the treatment itself and preparations for presentations and further discussions about that treatment.

A well-conceived training programme will take all three modes into account. This involves the selection of clinical settings, the selection of patients for attention by the trainee, the trainees' level of responsibility for each type of patient at a given stage in their training, and the nature and amount of supervision or guidance or teaching given. Another critical factor is continuity of care: over what period of time does the trainee have contact with and/or responsibility for various types of patient; and is this sufficient for the proper development of competence? Lowdermilk and McGaghie (1991) report improvements in medical care as residents gain longer periods of contact with patients.

Although more formal attention is given to opportunities for clinical experience when selecting clinical settings, consideration may, indeed should be given to the wider clinical environment. The ethos and values of the workplace and the prominence of good role-models can have a major impact on the professional development of doctors. The role-model aspect is more formalised in the North American residency programmes through the evaluation of clinical teachers (see pp

). A related issue concerning individual teachers as well as training programmes concerns the quality of supervision. While most research has focussed on the supervision of house officers (Section 4.1) and trainee surgeons (Section 4.2), its disappointing findings may apply equally to registrars and trainee physicians. Supervision is not only central to the role of a clinical educator, but also an important extension of patient-centred values to conduct beyond a supervisor's direct personal contact with patients.

Research into the outcomes of training programmes is sparse, although it can serve a range of useful purposes. One of these is establishing realistic standards of competence, so that none of the parties affected -- the doctors, their patients or those who employ the doctors -- misunderstands the level and range of their competence. The difficulties of achieving even this level of information are discussed in the section on assessment; but it is nevertheless useful to consider what use might be made of evidence already gathered by the formal assessment processes embodied in the Royal College examinations and the award of the Certificate of Completion of Specialist Training (CCST). Such evidence might contribute to the formative evaluation of training programmes, the planning of CME for newly appointed consultants and their duties when appointed, and help to validate (if possible) that most flexible of research instruments -- the self-report. This is important, because the small numbers involved, in regional training programmes, and the diversity of experience within those programmes make it unlikely that controlled comparative research will yield significant results. Even if marked differences in outcome were to be observed, it would be difficult to attribute those differences unambiguously to particular programme features.

4.1 Evaluations of Postgraduate Basic Training in the UK

Pre-Registration House Officers (PRHOs) are technically postgraduates, although not yet registered as doctors, so they have been included. Studies of PRHO training emerging throughout the 1990s, used different methods and foci but nevertheless presented a consistent picture. Two major improvements have been a substantial reduction in PRHO hours of work and clear programmes of educational seminars; but progress in other areas has been slow. Four conclusions from Calman and Donaldson's (1991) critical incident study still apply today: the need for all PRHOs to receive proper feedback on their performance, for access to someone senior to talk to, for training in working with dying patients and for greater use of the learning opportunities provided by their clinical experience.

A questionnaire survey by Gillard et al. (1993) of PRHO training in eight English regions combined questions about hours, conditions of work and inappropriate duties with questions about key aspects of competence and quality of supervision. While 91% and 61% reported receiving adequate guidance in basic and advanced cardiopulmonary resuscitation, only 65% (basic) and 10% (advanced) felt confident in performing it unsupervised. Only 41% reported adequate guidance in breaking bad news and 44% in pain control. Ratings on three aspects of supervision were 2.7 (briefing), 2.9 (problems) and 2.6 (career), all on the ``dissatisfied'' side of neutral on a 5 point scale.

Wilson (1993) studied the consultants' perspective on PRHO training, concluding that while they welcomed the concept of an educational supervisor ``theoretically'', many felt unable to take on this added responsibility. Teaching ward rounds were the standard method of instruction but most consultants estimated their total teaching time to be less than 30 minutes a week. Of the 33 respondents, only 8 had one to one teaching for their house officers; while 24 said they delegated some of that responsibility to other junior medical staff and to ward sisters. 25 said they would welcome feedback on their teaching and 18 would have liked training in educational methods and principles; and they were all concerned about not downgrading the apprenticeship system. Their problem did not appear to be their attitude towards teaching, but the pressure they felt from their lack of time to give to PRHOs when overwhelmed by service demands.

Two in-depth studies in particular localities by Dowling and Barrett (1991) and Eraut et al. (1997) focused respectively on the management and learning of PRHOs. Dowling and Barrett highlighted the range of expectations of PRHOs, the disparate and distributed nature of their supervision, their lack of clear goals and systematic training, and the absence of any coherent management of their work and training. Eraut et al. found that only a minority of 33 PRHOs in three hospitals received formal feedback from a consultant; and that even informal feedback was totally absent in important areas such as communication with patients. (Roche et al., 1997, found a similar neglect of interactional skills during the intern year in New South Wales). They would have liked a wider range of clinical experience, especially in clinics and surgery (c.f. Toogood et al., 1996), and more bedside teaching. Teaching and learning in clinical areas was always on a case by case basis, not one respondent could recall reviewing a group of patients with a similar condition or engaging in what the Americans call a ``chart review''. Some received quite a lot of teaching on ward rounds, some virtually none. There was good practice in every hospital but also areas where teaching was neglected. There was no guarantee that official recommendations would be implemented and no quality assurance of training (c.f. Goldacre et al., 1997). Organisational factors such as frequent changes in composition and distributed working locations prevent the formation of the kind of stable working teams with which the ideal form of apprenticeship is frequently associated.

An earlier survey by Grant et al. (1989) of the training of Senior House Officers (SHOs) reported similar levels of dissatisfaction. On a scale of 0-3, statements receiving agreement with means of over 2.0 included:

SHOs provide service more than receive training	2.6
SHO hours are not conducive to learning	2.6
SHOs do not think they are receiving postgraduate training	2.2
No one sees teaching as a defined part of their job	2.2

Only about 53% reported receiving most teaching from their consultants; and their views of the learning methods on which SHOs relied most differed from their consultant supervisors (see Table 4.1 below).

Table 4.1: Priority given to 12 methods of SHO learning by consultants and SHOs

Learning Method	% included in top 5
(24 less used methods not listed)
	Consultants	SHOs
Ward round with consultant **	84	55
Ward teaching with registrar/senior registrar	41	47
Formal teaching with consultant	23	19
Supervised practice	17	12
Unsupervised clinical practice **	19	44
Discussion with peers *	34	44
Library work or reference to journals *	25	37
Reading about cases seen *	33	50
Following up outcomes if cases seen	20	28
Presenting patients at small meetings *	24	12
Teaching in outpatients *	20	10
Teaching for Membership exams	20	13
(* difference of 10-20%, ** difference of over 20%)

The authors note that the ``apprenticeship'' model, if unanalysed, ``perpetuates the unhelpful confusion between training and service, to the extent that providing the service may become identified with receiving training''. If consultants are to be given responsibilities as teachers or supervisors, the necessary time must be built into their routine schedule of work. Whereas SHOs seem to be regarded both as full-time doctors in the hospital service and as students when off-duty, with the usual obligation to study in what little time is left to them.

Barker et al.'s (1994) study of SHOs and registrars in four types of hospital concluded that variation in the content of training was caused as much by lack of structure as by type of hospital. Most consultant teaching occurred during business rounds, though the teaching element was rarely formally defined. There was great variety in the amount of ward-based teaching, depending on the propensity of individual consultants to teach. Similar ad hoc access to learning opportunities was found in outpatient clinics and theatres. The authors also noted that there was little attempt to cater for the differing educational needs of different grades of junior doctor. However Paice et al. (1997) were able to show significant improvements in SHOs' rating of their post between 1992-3 and 1994-5: the proportion recommending the post to a friend as `good' or `excellent' rose from 45% to 57%, and ratings of their consultant supervisor rose from 48% `good' and `excellent' to 61%. There was also a large reduction in the frequency with which they felt forced to cope with a problem beyond their experience or competence: only 2% felt this at least once a day (previously 23%), 12% once or twice a week (17%) and 46% once or twice a month (28%). Nevertheless Paice and Leaver (1999) reported that the new GMC guidance document The Early Years did not address the ``real difficulty'' of balancing service with teaching; and Calman (1999) made the same point about specialist registrars:

Bunch et al.'s (1998) review of Basic Surgical Training in Yorkshire showed good consultant support for trainees, but deficient clinical experience was found in 32% of general surgical and 70% of orthopaedic posts. This was due mainly to poor organisation of their clinical activity (for example 70% of orthopaedic SHOs never attended outpatient clinics), though it was not helped by the lack of PRHOs in orthopaedics. Only 8% of trainees had undergone an appraisal and ``the concept of bleep-free periods for educational purposes was non-existent.''

4.2 Learning Procedures

Wigton and his colleagues have published a series of papers on the learning of procedures on residency programmes in Internal Medicine in the U.S. (Wigton, 1981b; Wigton et al., 1989,1990; Wigton, 1992). Since these programmes are organised from (though not confined to) a single site he was able to survey the Programme Directors. His 1989 paper had two main purposes: (1) to find out from each Director which procedures should be mastered by all residents and which procedures are mastered by all their residents, and (2) to find out how many of each type of procedure were needed to attain and to maintain mastery. For several procedures the gap between should and are was 40% of all respondents; and many procedures commonly done in practice were mastered by all residents in less than half the programmes. The medium number of procedures required for mastery ranged from 5 for thoracentesis to 50 for chest roentgenogram interpretation. Once more a gap is revealed between performance on completion of training and the expected competence of a qualified doctor.

This raises questions about how the teaching of procedures is organised. One successful approach to this problem is described by Gabryel et al. (1991). It involves (1) a requirement that procedures can only be conducted by a resident who has been formally credentialed for that particular procedure; (2) a phased timetable for acquiring the agreed list of necessary procedures; (3) regular feedback to residents on their current status and one-to-one consultation sessions for those who are falling behind. The proposal that ``residents should be credentialed to perform, supervise and teach certain clinical tasks, particularly procedures, by performing those tasks under direct supervision'' was approved by the Educational Policy Committee of the American College of Physicians (American College of Physicians, 1989); and Gabryel et al.'s system made it more workable. The advantage of a credentialing system is that it requires competence to be demonstrated to those properly qualified to assess it. Not all aspects of medicine are suited to such a system, but it could be extended beyond procedures to include areas such as the administration of drugs (Bain et al., 1990). Credentials for cardiopulmonary resuscitation skills (Morris et al., 1991; Wheatly and Redmond, 1993) and others requiring regular practice should presumably be datestamped.

A further question relates to the context(s) where certain procedures can be most effectively taught and most validly assessed. Giving thought to this issue and arranging short term placements could be helpful. At a more advanced level, Burke et al. (1996) report that learning nasendoscopy on awake patients in an ENT clinic is an excellent introduction to fibreoptic skills and transfers to nasotrachial intubation in anaesthetised patients.

4.3 Learning in the Operating Theatre

A Scottish study by Steele et al. (1989) on the technical training of general surgical trainees covered experience, supervision and indirectly competence. 47% of those in teaching hospitals reported that their operative workload was too little (23% in district general hospitals); 35% said they did too little operating alone and 58% too little supervised operating. More worrying was the response from 32% that they were ``sometimes out of depth'' when doing emergency operations alone in DGH (15% in teaching hospitals, 9% for elective surgery). As in the Gillard study, trainees' concerns relate both to the amount of experience (resuscitation or operating) and the degree of supervision (in ward or theatre), the two most central features of learning in clinical settings. Steele later participated in an audit of colorectal cancer surgery in three regions for the period 1990-4 (Aitken et al., 1999) which revealed both low levels of supervision of registrars and SHOs and a much lower level of operating experience than that offered by an US fellowship programme for residents. This concern was backed up by a Scottish mortality study of avoidable deaths after colorectal surgery (Macarthur et al., 1998), a finding which prompted Collins (1999) to suggest that surgical training deserved the priority attention given to waiting lists.

Concern about supervision in the operating theatre was raised again by a survey of unsupervised `first time' procedures by Wilson (1997), who also raised the problem of excessive `first time' surgery by young consultants and arrangements for the introduction of innovative procedures. In response O'Riordan and Shaper (1997) noted a cultural change towards defining a well trained surgeon in terms of the number of operations performed under supervision rather than the number of operations performed solo; while Galasko and MacKay (1997) noted improvements in training subsequent to Wilson's survey and the introduction of an assessment form to be sent directly from trainees to the supervisory body. Earlier Reed (1993) had urged that careful evaluation would be needed as surgical training was changed from ``a protracted apprenticeship, where endurance is eventually rewarded, to a shorter, more intense, residency style programme'' (page 198s).

In the United States, Fallon et al.'s (1993) audit of over 4000 cases showed a correlation between the absence of a supervisor and increased complication and mortality rates, but was not able to take case severity or complexity into account. Similar concerns about supervision have been raised about emergency departments; though Sox et al.'s (1998) study of 3600 US patients revealed better compliance with guidelines when supervisors were present but no significant changes in patient satisfaction or reported problems with care. Forde (1994) reviews how the urgent training problem associated with the introduction of Minimal Access Surgery was approached.

4.4 Vocational Training of General Practitioners

Hindmarsh et al.'s (1998) review of evidence from several countries of the effectiveness of vocational training, concluded that its introduction had improved the preparation of GPs, but comparisons between different models of vocational training were beyond the researchers' brief. The most searching study of programme outcomes was conducted in Holland by Grol et al. (1989). This measured changes in the consultation skills and medical performance of General Practitioners during a one year programme of vocational training with a similar structure to that used in the UK, i.e. 4 days a week in the practice of a trainer and 1 day a week on a formal course. Both the pre-test and the post-test were based on 20 audio-taped consultations, rated on 26 variables by a panel of 4 assessors. Two groups of trainees were compared, a cohort of 32 from Nijmegen and a cohort of 31 from Gronigen. Significantly greater gains in medical performance resulted from the Nijmegen programme based on systematic skills development than on the Gronigen programme based on problem-based learning. Increases in consulting skills were similar for both groups in four areas but for ``psychosocial performance'', the Gronigen group maintained a high starting level without further improvement while the Nijmegen group improved from a lower starting level. Performance on knowledge tests also increased and was strongly correlated with both consultation skills and medical performance. Moreover, those with above average knowledge scores at the beginning made the greatest improvement in medical performance. Such improvements resulted from both an increase in obligatory performance and a decrease in superfluous performance, judged against protocols for the most common conditions. Neither programme achieved a desired level of performance across the full range of common conditions; and the vocational training programme in Holland was extended to 2 years.

The British system of vocational training became mandatory in 1982, but had been used in some regions for several years. However, mandation made it necessary to establish criteria for the approval and reapproval of trainers or to tighten up the existing local criteria. The benefit is documented by a before-after study in the West of Scotland (Kelly and Murray, 1992), which showed an increase to 84% of trainees receiving a tutorial weekly (or most weeks), and a decrease to 4% of those rating their training as less than ``fairly good''. Most British research into the training of General Practitioners has focused on consultations in GP surgeries and the weekly tutorials given by trainers. The main aim of these tutorials is to improve the quality of consultations both from a medical perspective and in terms of communication skill. Since formal summative assessment at the end of the practice-training year and for membership examination of the Royal College of General Practitioners requires the submission of a video of several consultations, this aim has high priority for all concerned. The content of the tutorials typically comprises some combination of cases and issues raised by the trainee, a randomly chosen set of cases seen by the trainee in the previous week, a previously agreed medical topic, and discussion of the consultation process itself, often prompted by notes of a consultation at which both had been present or a video of a trainee consultation (Gray, 1998). Urgent cases are usually discussed at the time but may be reviewed again at a tutorial. Trainers have all received training for their role; and will almost certainly make use of the construct of a patient-centred consultation.

Shapiro (1990) and Shapiro and Talbot (1991) have explored the parallel between trainer or trainee-centred tutorials and doctor or patient-centred consultations. However, both Marvel (1991) and Pitts and Coles (1996) have shown that many experienced doctors are patient-centred in their consultations but trainer-centred in their tutorials. Gray (1998) used both telephone interviews and videos of tutorials, accompanied by separate confidential comments from each participant, to study the nature and effectiveness of tutorials. This raised issues of style, focus, continuity and effective use of tutorial time, all highly pertinent to the training of trainers and meetings of local trainer networks. The system can be described as the regular, flexible and reflective use of two principles established as important for effective social skills training; regular systematic feedback and facilitation of self-assessment of both verbal and non-verbal communication by video recording. While patient-centredness is usually advocated by the trainer and perceived as a desirable goal by the trainee, no particular approach to consultation is formally modelled and taught. General practice trainees are learning by doing as well as from tutoring, so it would be impossible to isolate the contribution of tutoring; nor would a control group study without training or without video be acceptable.

Not surprisingly, such a short training period cannot cover all conditions, so there have been many articles suggesting areas where GPs need more training. One of the best arguments has been in Palliative Care, where Lloyd-Williams and Lloyd-Williams's (1996) survey found that only 15% of a sample of GPs had received a tutorial, and less than a third felt they had received adequate training in pain and symptom control; while 85% would have liked to have a placement in a hospice as part of their training. Another was in the area of rheumatic disorders where a large survey by Lanyon et al. (1995) showed that less than a half of GPs had received any special training, although such disorders account for 10% of an average general practice workload.

Two Australian articles are also relevant. A critical incident study of GPs six month into training and 12 to 18 months later (Sim et al., 1996) demonstrated several major areas of positive change as well as some new or continuing areas of difficulty. The latter include gynaecology, pharmacotherapy and dermatology, the diagnosis of common complaints with uncommon presentations, the skill of managing difficult or angry patients, the organisation for the follow-up of patients with potentially severe disorders, and managing feelings of guilt over missed diagnoses or poor management. The second is Shah et al.'s (1997) report on a remediation programme for trainees placed in the borderline zone at their final assessment. This provides additional intensive training and supervision; and having such a ``middle pathway'' clearly contributes to the maintenance of standards.

Snadden and Thomas (1998) report the use of portfolios by General Practice Registrars (GPRs) for a number of purposes: ongoing planning of a ``curriculum'' to meet the needs of the individual trainee, as a stimulus to reflection, as a mechanism to capture examples of significant learning in general practice, to reach the more difficult areas of a GPR's experience and to help feedback. They found that a portfolio was useful in establishing confidence during the first half of the year, provided that the relationship between trainer and trainee was reasonably good. However, it ceased to be useful when preparation for the MRCGP examination became the dominant feature of their planned learning.

4.5 Ambulatory Care Settings

This North American term includes both hospital outpatient clinics and private offices (general practice surgeries). These settings are now increasingly used in US residency training, especially in Family Medicine and Internal Medicine, because they ``allow learners to:

We are especially indebted to Irby's comprehensive 1995 review of research into teaching and learning in ambulatory care settings ``where 95% of doctor-patient encounters occur''.

Several researchers (Osborn et al., 1993; Napodano et al., 1984; Krackov, 1982) report better case mix, more continuity of care, more trainee responsibility for patients and more effective trainer-trainee interaction in private offices than clinics; and Osborn et al. found that private office experience led to better performance on an oral examination of clinical competence. In general, however, there is great variation across ambulatory settings which affects both what and how much is learned. Lowdermilk (1986) lists the most critical factors as: the organisation of the clinic, interference from inpatient responsibilities; limited time, space and supervision; access to attending physicians for advice and feedback.

Whereas there is some discretion, at least in theory, over the pacing of visits to inpatients, clinics and office consultations are timetabled in a manner that determines the overall time spent with patients. A major concern, therefore, is the time available for interaction between trainer and trainee and how that time was used. There is a significant amount of evidence from American research but little from Europe. One indicator is the proportion of cases discussed, another is the duration of the teaching component of that discussion. The proportion of cases discussed varies greatly with location and trainee's experience. Thus Williamson et al. (1988) reported consultation rates in a family medicine clinic of 48% for year 1 residents, 28% for year 2 and 26% for year 3; faculty saw 20% of the patients of first year residents and 12-13% of patients of year 2/3 residents. The average duration for discussion was 8 minutes for year 1 residents and 6 minutes for year 3. Consultations were more likely to be resident initiated as they became more experienced (McGlynn et al., 1978). (Gennis and Gennis, 1993) found a 7 minute average for interactions involving presentation and discussion but a 15 minute average when the interaction included a visit to the patient. Knudson et al. (1989) found that 50% of interaction time was taken up by case presentation, 25% by clarifying questions and 25% by discussion. These discussions were largely focused on patient management, differential diagnosis and/or patient compliance (Hekelman et al., 1993; Lowdermilk and McGaghie, 1991; Williamson et al., 1988); even in family medicine clinics, family and psychosocial issues were less frequently discussed (Marvel and Morphew, 1993).

Three recent American papers discuss approaches to using the limited time for interaction between trainer and resident as effectively as possible. Ferenchick et al. (1997) focus on strategies during clinic/office hours, such as (1) scheduling patients so that some are seen first by the resident then jointly, while others are seen by the trainer alone, (2) case presentations in the presence of the patient followed by a short burst of oral questioning by the trainer, and (3) using reflection on experience to develop short teaching scripts. McGee and Irby (1997) discuss the selection and use of teaching points, role modelling and preparation for and follow-up of doctor-patient encounters. DaRosa et al. (1997) also discuss strategies which do not impinge on clinic time, such as learner preparation for seeing patients, various forms of case conference, and linked independent study.

Adaptations of morning reports for ambulatory contexts are discussed by Paccione et al. (1989) and Malone and Jackson (1993). Residents pass 1 page summary reports of all patients seen each day to their trainer, who then selects cases for discussion at a teaching session the following morning. They found that a wider range of medical conditions was discussed in the ambulatory setting and that greater attention was given to the medical interview and to social issues, increasing residents' reflection on their outpatient experiences.

Feedback is also a critical issue. Irby's (1995) review reported that a majority of students found feedback being given during 3-6% of case discussions, the majority of it being positive. However, Gennis and Gennis (1993) found that supervisors changed their views about diagnosis and management quite frequently if they visited a patient after a resident's presentation and discussion, often judging them to be more seriously ill and becoming less confident about the resident's performance. Moreover, Cope et al. (1986) showed that counselling combined with detailed feedback on patient satisfaction scores significantly improved the performance of residents with below average scores, when compared with a matched control group. Ende et al. (1995) conducted an ethnographic study of trainer's corrections of interns' errors, which found four different patterns of response, all of which they describe as soft ``face-saving'' strategies, e.g. asking subsequent questions that contain hints to lead the intern to the correct answer. They suggest that, while successful in reducing social distance between teacher and learner, these indirect responses may fail to promote accurate self-assessment by the interns who may not grasp the central message about what was wrong. The observed behaviour was more compatible with Mizrahi's (1984) report of the social denial of mistakes by house staff by reference to medicine as an art with a ``gray area'', repression, or redefinition of mistakes as non-mistakes than with a learning climate which facilitates learning from mistakes (see also Wu et al., 1991).

Failure to use ambulatory settings was one major criticism of Nicol et al.'s (1995) survey of urological training in Australia. Hospital-based training alone left trainees ill prepared for outpatient or ``office'' urology, and without experience of non-acute conditions such as urinary incontinence or infertility. Communication skills and critical appraisal of the literature were not well developed and feedback on progress was inadequate. The majority of respondents felt that mentors required specific training to facilitate feedback to trainees. Positive changes reported in the US include the introduction of a clinical-hospice/palliative medicine rotation for fellows training in haematology/oncology (Von Gunten et al., 1995), and the development of a Housestaff Ambulatory Block for internal medicine residents. This block covered areas prioritised by an analysis of unmet training needs -- dermatology, rheumatology, health promotion, women's health, behavioural medicine, geriatrics and sexually transmitted diseases (Lucas et al., 1997).

Another area of research concerns the role and performance of the clinical teacher. A useful starting point is the policy adopted at McMaster University for the evaluation of the teaching performance of physicians on their residency training programme in internal medicine (Guyatt et al., 1993). 14 domains of performance were identified, and trainees' rating forms provided evidence to the Director of the Clinical Teaching Unit in each hospital. These domains were:

The inclusion of biophysiology reflects the determination in this particular programme to emphasise basic science as well as evidence-based medicine for which it is renowned. The other categories map well onto Ullian et al.'s (1994) four roles of effective clinical teachers in family medicine: physician role model, effective supervisor, dynamic teacher and supportive person. Irby (1995) used these roles as a framework for his summary (below) of research on teachers in ambulatory settings, concluding that all 14 studies reviewed were congruent both with each other and with studies of inpatient teaching.

Effective Teachers in Ambulatory Settings

Physician Role Model: Excellent clinical teachers served as positive physician role models. Such physicians were characterised as being knowledgeable and clinically competent, having good rapport with patients, and being perceived by residents as good role models.

Effective Supervisor: In the role of supervisor, effective clinical teachers gave residents responsibilities for patient care, provided opportunities to do procedures, reviewed patients with residents, and involved the residents in patient care. Excellent supervisors and teachers provided direction and constructive feedback as well. Greater delegation of responsibility for patient care was significantly associated with higher overall ratings of teachers by medical students.

Dynamic Teacher: Excellent clinical teachers were interested in teaching, were enthusiastic, made an effort to teach, spent time with individual residents, were available, engaged in dialogue with residents, provided explanations, and answered questions. Characteristics identified in other studies include availability and accessibility, organisation, clarity and enthusiasm, asking questions and giving directions, and directing residents' learning.

Supportive Person: Excellent clinical teachers demonstrated support for the resident, were easy and fun to work with, were friendly, and were helpful and caring. Characteristics mentioned by other authors include establishing rapport with learners, demonstrating a positive attitude toward teaching and motivating learning, and creating an educational environment that facilitates learning.
Irby (1995)

Ullian et al. also found that first year residents preferred more didactic teaching and one-to-one discussion, whereas third year residents took more responsibility for their learning and wanted more information and explanation.

To conclude this section, we draw attention to the report by Bordage et al. (1998) on a policy conference convened to discuss research needs for redesigning education in ambulatory settings. This reveals how little is known about this topic, even in North America. The members of the conference gave the highest research priority to the outcomes of ambulatory education, followed by teaching strategies, faculty development and the influence of organizational culture.

4.6 Inpatient Settings

We were unable to find any more recent survey of SHO learning patterns in hospital settings than Grant et al. (1989), which we reported in Section 4.1, nor any comparable study of registrars. Nevertheless that survey reminds us that unsupervised clinical practice and discussion with peers are important sources of learning as well as planned (or unplanned) teaching by senior doctors. This needs to be kept in mind when reviewing the research which is (a) dominated by studies of teaching and teachers and (b) largely from North America. In Britain patient-based teaching takes place mainly on the wards, but also at weekly conferences/seminars where case presentations are made to a wider group of doctors than the resident's own firm (there are called `grand rounds' in North America). Ward rounds are described as `teaching rounds' or `business rounds', but the distinction between them is not precise and most ward rounds are described as a mixture of the two. Some of the time on teaching rounds may be spent in spaces close to the ward but away from the patient, if they are available.

In North America there are four types of activity which correspond to ward rounds: attending rounds, patient management rounds, teaching rounds and morning reports. Attending rounds involve the residents and an ``attending physician'', the American term for a fully trained doctor with overall responsibility for the patient (the nearest American equivalent to a consultant). Patient management rounds are defined as patient care rounds without an attending physician. Teaching rounds are rounds formally dedicated to teaching, but in practice often combined with attending rounds. Shankel and Mazzaferri's (1986) survey of 123 residency programmes in medicine showed that both the content and the location of teaching and teaching/attending rounds were identical. According to the chief residents, 15% of teaching rounds were only at the bedside, 22% only in the conference room, and 63% used both locations. The trend towards greater use of conference rooms was viewed negatively (though in Britain where the balance is more ward-based, access to other space might be envied). Morning report is a distinctly separate activity in North America, whereas in Britain its function is assumed either by informal conversations between consultants and registrars or by a scheduled ward round. The former would be a purely ``business'' transaction, the latter might also be used for teaching. Shankel and Mazzaferri's (1986) review indicated that in about half or more teaching hospitals, about half a day 5 days a week was occupied by rounds and morning reports. Attending rounds were conducted 5 days a week and usually lasted an hour or more. Teaching rounds averaged 3 days a week and a length of 50-90 minutes. Most respondents considered the optimal length to be 2 to 3 hours. Residents assigned greater educational value to (1) morning reports (2) teaching rounds, programme directors valued (1) grand rounds and (2) teaching rounds.

Differences of perspective were also found on a smaller scale study by White and Anderson (1995) using Critical Incident Interviews to investigate factors facilitating or restraining learning by residents. Both residents and faculty emphasised the importance of interaction and reflection in discussion. Faculty also emphasised concrete experience (with some support from residents), technical rationality and faculty expertise; whereas residents stressed faculty involvement and commitment, consideration of multiple perspectives and personal relevance. Two major restraining factors for residents (with some faculty support) were insufficient time and opportunity to learn and low faculty involvement and commitment.

Reilly and Lemon's (1997) paper on Evidence-Based morning reports describes an innovative development of the teaching round to incorporate searches for research evidence and feedback on outcomes. Each morning report session lasts an hour and is divided into four phases: reporting search results from the previous day, reporting on new admissions, case presentations of three new patients (selected by the residents because they provided the group with clinical challenges whose resolution required research evidence), and formulating questions for a literature search and report back next day. On Fridays the third phase is devoted to follow-up reports on all the cases presented earlier in the week. Evidence of the impact on patient outcomes is currently being collected, but meanwhile a major change in departmental climate is reported, particularly in the emphasis given to evidence-based medicine and patient preferences and values.

Another adaptation of traditional morning report practice is reported by Wartman (1995) who also incorporated literature reviews commissioned the previous day to encourage the paradigm of evidence-based medicine. His prime innovation, however, is the presentation of a ``discharge case'' selected by the attending residents to get a good case-mix each month. The case review is expected to cover the reasons for admission, decisions and progress made each day in hospital, discharge and follow-up plans and costing for the hospital visit. The rationale was the lack of attention given by traditional morning reports to patient outcomes, psychosocial issues, humanness of the care rendered, rehabilitation and cost. Both the innovation and the new balance it gave to the programme were positively evaluated by the residents. Wartman was partly responding to Wenger and Shpiner's (1993) follow-up study of morning reports. They found that 24% of medical cases discussed at morning reports did not have a final agreed diagnosis when discharged. A review 6 month later revealed that 39% were still uncertain, while 36% now had a confirmed diagnosis which differed from the best ``morning report diagnosis''. Moreover 24% of those cases that had reached a ``final diagnosis'' now had a ``late final'' diagnosis from that first proposal (17% changed while in hospital and 7% subsequently). Their conclusion was that ``misconceptions of disease presentation and appropriate diagnostic evaluation are likely to occur if cases are never revisited.''

Wyte et al. (1996) conducted a controlled trial with First Year residents on an Emergency Medicine rotation comparing bedside teaching alone with bedside teaching and written course materials. This showed learning gains on a written test but no greater gain for the group given the materials; and the satisfaction rate was higher for those without materials! However, the provision of on-line materials in emergency departments, designed for rapid consummation rather than in-depth study proved popular with House Officers in several British hospitals. In this case the main success criterion was HOs using the facility and finding it helpful (Grant and Marsden, 1992).

Griffith et al. (1997) reviewed research studies on the relationship between internal medicine housestaff training and patients' outcomes, examining in particular whether trainees' inexperience or their their workload affected affected patient care, the effects of the structure of the teaching service, and possible benefits of having an internal medicine training programme. They were unable to reach any valid conclusions and were rather wary of studies more than 10 years old.

The introduction of cognitive science perspectives (see Chapter 3) has led to a more sophisticated analysis of interactions on teaching rounds and case discussions. Thus Irby's detailed analysis of instructional thinking and decision-making by six highly rated but very different physicians conducting teaching rounds in a conference room following a resident-led ward round incorporates the concept of ``teaching scripts''. Though not claiming a representative sample, his framework of analysis is well matched to teaching in clinical settings and suggests many areas for further research.

Clinical Teaching Decisions by Six Physicians

Planning: informed by periodic learning needs assessments with their team, they set priorities in advance for the allocation of time. Half of them contacted the resident ``on call'' the night before for an up-to-date review of their patients. Decisions were made about which cases to highlight (for difficulty or for educational value) and which content to teach (team requested, teacher-determined, filling gaps in the team's knowledge). Five of the six read some literature and four prepared simple learning material (handouts, reprints, exercises, slides, etc.)

Diagnosing the Patient's Condition: the main difference from the diagnostic context described in Section 1 is the greater reliance on second hand information from other members of the team.

Diagnosing the Learner's Understanding: inferred from the presentation and a few key questions. One physicians commented that ``the effectiveness of the presentation is directly proportional to how much I can remember of it'', another that if one cannot follow the script, the thinking is probably disjointed.

Interactive Thinking: this involves continually monitoring learners' responses and adjusting to their needs, making mid-action decisions about educational opportunities, selecting from their repertoire of teaching scripts, etc.

Teaching: four of the six trainers organised their teaching around cases, but appeared to use cases to trigger ``fixed scripts'' for the topic, which they then improvised around using a ``Socratic style of interchange''. The key teaching points for a given case (the only protocol case used in the study) varied greatly between physicians.

Reflecting: all trainers reported later reflection on their teaching rounds, some about the patients, some about the house staff, some about both.
Irby (1992)

Hewson's (1991) report of an out-of-phase interaction between a resident and a trainer was based on their reflected comments, and suggests how material for constructive reflection on clinical teaching might easily be generated.

The eight principles of learning advocated by Irby's six physicians were: ``actively involve learners and ask lots of questions, capture attention and have fun, connect the case to broader concepts, go to the bedside, meet individual needs, be practical and relevant, be selective and realistic, and provide feedback and evaluation'' (Irby, 1994b, page 336) to which he later added ``model professional thinking and action'' and ``create a collaborative learning environment'' (Irby, 1994a). Their practice demonstrated consistent use of these principles. Moreover, they often used their knowledge in an encapsulated mode, making a precise teaching point but elaborating on the deeper structure of that knowledge, when needed or asked. Irby (1994b) also advises that, while faculty development workshops are effective in developing general teaching skills and improving understanding of learners, ``knowledge of case-based teaching is best learned in the context of departmental teaching, improvement and mentoring programmes where content-specific, case-based teaching scripts and strategies can be shared'' (page 340). Irby's expert teachers said it took them six or seven years to discover a teaching style they felt comfortable with. With more local sharing of expertise that time might become a little shorter.

4.7 Summary

There have been several evaluations of postgraduate basic training programmes in the UK. Though there have been a few improvements, the overall impression is still negative. Many features of the educational policy seem to be appropriate, but they are not being implemented in many hospitals. There is insufficient supervision and feedback. Educational goals are subordinated to service demands. While many house officers receive good clinical teaching, a minority do not and assurance of educational quality is weak. Learning goals are only specified at a very general level, so there is little clarity about priorities, especially at the PRHO stage. Management of the educational process at local level is given little time and little authority.

Continuing concern has been expressed about the survey evidence on basic surgical training. Sometimes the problem is too little supervision of operations by house officers, sometimes the house officers get insufficient clinical experience. Operating under supervision is seen as the most critical feature of learning to be a surgeon and there is not enough of it. Some authors recommend the greater use of 'skills labs' and simulators. (See also Chapters 3 and 5).

The learning of procedures in medical posts has been criticised for being too haphazard: there is often little continuity of experience and guidance is often provided by doctors who are themselves not very experienced. Systematic attention like that found in many parts of North America (when it sometimes includes a credentialing system for procedures) has not been reported in the UK.

Research in Holland into the outcomes of their 1-year training programme for General Practitioner registrars led to the extension of the programme to 2 years. Not surprisingly research investigating particular areas of expertise has resulted in long lists of needs for GPs continuing medical education (see Chapter 6. In areas such as palliative care and psychiatry the argument seems particularly strong, in other areas one might look to other ways of distributing more specialist expertise within primary care organisations. GP training contrasts favourably with basic hospital training in its ability to provide tutorial support on a regular basis; and the commitment to quality improvement appears to be much greater.

Outside general practice, there is much more in-depth research in North America than in Britain. This has given particular attention in the last decade to learning in ambulatory care settings, a term which covers both family medicine and hospital clinics. More use is now being made of such settings in order to give doctors a broader experience of medicine, especially when significant aspects of care are being moved out of hospitals.
continued over page

Summary -- continued

The key issues emerging from this research are:

The time allocated for trainer-trainee interaction and making the best use of that time;
Methods of providing feedback;
Qualities of teachers rated as 'good' by trainees and their colleagues.

Methods for finding time for trainer-trainee interaction cover both time created within clinics by patient scheduling and the use of clinic experience for later case discussion and chart review. The papers in this area include many useful practical suggestions as well as evaluations of practice. In general, learning in ambulatory settings was found to give rise to discussion of a wider range of medical conditions, and greater attention to the medical interview and to social issues.

Issues relating to feedback included in the findings that:

Supervisors changed their views quite often when they visited a patient after having heard the resident's report;
Counselling combined with detailed feedback on patient satisfaction scores improved the performance of residents with below average scores;
The effectiveness of feedback tended to be reduced by ``face-saving'' strategies which ameliorated or even disguised the central message.

Qualities of good teachers inferred from rating studies can be grouped under the headings of Physician Role Model, Effective Supervisor, Dynamic Teacher and Supportive Person.

Learning in inpatient settings is also researched in greater depth in North America. One gets the impression that American residents receive considerably more clinical teaching than their British counterparts, but there is no British data to enable a proper comparison. The variation in the amount of training received by British trainees is reported as considerable, raising issues of quality assurance and trainee entitlement. The two major constraining factors on learning by US residents were insufficient time and opportunity to learn, and low faculty involvement and commitment. Innovations receiving strong positive evaluations included adaptations of the Morning Report system to incorporate the teaching of Evidence-Based Medicine; and case reviews of patients whose diagnosis had changed while in hospital or within 6 months of leaving hospital.

Research is reported on instructional thinking and decision making by highly rated clinical teachers. This covers planning, diagnosing the patient's condition, diagnosing the learner's understanding, teaching and interactive thinking during discussions with learners, and reflecting.

4. Learning in Clinical Settings