We sought for literature on the following structured question:

New teaching and learning activities and methods can be evaluated against a contemporaneous control educational strategy, or baseline before the new educational activity, using measures of patients' health gains (clinical outcome) or participants' learning achievements (educational outcome) as endpoints for evaluation of effectiveness and impact 91011 (Table 1). Learning achievement can be assessed separately for knowledge, skills, attitudes, and behaviour 10. Knowledge relates to issues such as remembering materials as well as grasping the meaning, for example, defining and understanding the meaning of Numbers Needed to Treat (NNT). If this knowledge can then be applied accurately to given problems, then this will be regarded as a gain in critical appraisal skills; for example, the ability to generate NNTs when baseline risks and relative risks are provided. Spontaneously acknowledging a need for the use of a certain piece of knowledge or skill in practice will be regarded as a change in attitude, for example, recognising without prompting the need for different NNTs for different clinical scenarios and intending to calculate the respective NNTs for different risk levels. Finally, a change in behaviour occurs when one seeks the necessary information and applies the knowledge and skills to solve the issue in practice, for example, searching the literature, finding relevant baseline risks and relative risks and calculating necessary NNTs to guide clinical practice. Ultimately by consistently applying these findings in practice, patients' outcomes can be expected to improve. Insistence that these competencies must translate into improved clinical outcomes or outputs for our patients, or that the proof can only be considered valid if generated through RCT is not the subject of our commentary. The empirical evidence is summarised below looking at various elements of interactive and integrated teaching and learning. Some readers may be concerned that strength of the empirical evidence may be overstated, and if they need assistance in interpreting the results of systematic reviews they may wish to consult existing texts on appraising such reviews 121314.

What is the best way to learn EBM? Can learning of EBM improve not just basic educational outcomes such as knowledge and appraisal skills, but also substantial outcomes such as attitudes and clinician's behaviour? Can EBM learning result in improved patient outcomes? There are several theoretical reasons why clinically-integrated interactive teaching of EBM may produce better results in comparison to standalone didactic teaching 25. These are explored in Figure 1. If learners are fully immersed in educational experience, feel relaxed alertness thereby eliminating fear while maintaining challenge, and actively participate in processing information, they will consolidate and internalise the learnt materials effectively 262728. Changes in their attitudes are likely to be important in bringing about sustained changes in behaviour, which is what will ultimately benefit patient care.

In a rapidly changing world, health care practitioners need to approach their profession with a view to lifelong learning 29. They need to identify their educational needs and develop a strategy to meet those needs in realistic and effective manner. This is likely to be achieved best with an a priori outline of a personal learning plan, which can serve as a vehicle for guiding educational activities over a specified time period when the progress can be assessed and a follow up learning plan formulated. In generating a learning plan health care practitioners should identify their own needs and set learning objectives accordingly. Their personal learning plan is based on the principles of adult learning 30313233, whereby they take responsibility for their own learning through a systematic programme of acquisition, renewal, upgrading and completion of knowledge and skills outlined in their professional development objectives. The objectives should be based on what they and their organisation need to learn. If there is cognateness between individual and organisational learning needs, the learning plan is more likely to be supported 34. There should be a practical, flexible and achievable strategy for learning which has a realistic hope of meeting the learning objectives within the limits of the resources available to the individual and their organisation. The choices should be driven by elements of effective CME outlined above.

Ideally, the plan for learning should be entirely voluntary and it should be driven by personal motivation of learner 35. Individuals should make the choice of the subject matter they wish to learn, the manner in which they will go about learning it, and they should learn independently without the need for close supervision. The ability to fulfil learning objectives will depend on critical reflection on experiences throughout the planned learning period 35. Therefore the plan should delineate learning outputs and outcomes to assess whether learning objectives have been met. The experience should help them critically analyse their performance against standard criteria using both formative and summative style assessments 36. In this manner a self-directed individual can be nurtured in the course of a CME cycle. Such an approach when applied diligently can help with personal development, revalidation etc., but above all it has a real chance of making a difference to our patients' outcomes.

Below we list elements of CME that we consider would enhance the value for the learner, and ultimately for their patients. Although some of the elements are based on empirical evidence, the others are based on educational theory and practice.

Learning using an interactive approach 16.

Learning incorporated into clinical practice 23.

Sequenced events that aid reinforcement, rather than single or episodic CME events.

Courses that identify and take into account learner needs.

Courses that take place in the context of CPD 2 and KT 3, as well as local and national service objectives.

Multi-faceted strategies in teaching and learning 3738.

Course that give individual feedback and the opportunity for self-assessment.

Various other effective approaches exist, such as educational outreach for prescribing, reminders, opinion leaders, as well as audit and feedback 3738. Moreover, not surprisingly, multifaceted interventions targeting different barriers have been shown to be more likely to be effective than single interventions 3738. These, when combined with effective teaching and learning, are likely to bring about desired outcomes for EBM.

Practitioners should view CME concerning teaching and learning of EBM in a holistic manner. CPD encompasses traditional CME as well the acquisition of other skills such as administration, management, teaching, and communication, and embodies elements such as self-directed, patient-centred and individualised learning as well as continuing appraisal 2. Work based learning and KT are elements of in-house CPD that focus on knowledge and skill application in the workplace 39. KT is the process whereby information is transferred to clinicians and applied in practice, a process that requires understanding of complex interactions between various individuals (medical and non-medical) and organizations 3. It focuses on health outcomes and changing behaviour 3. EBM, which can change physician's practice 23 and thus holds the potential to change health outcomes, can be a critical tool in KT. Whilst CME or CPD is primarily located in teaching settings, KT activities are centred in practice settings, and place a greater focus on group or team learning 3. KT is an element of CPD that is receiving wider attention in the field of CME.

Based on educational evidence, theory and principles we propose a hierarchy of teaching and learning methods for EBM (Table 2).

The concepts behind the above hierarchy are gaining in popularity. However, it needs to be recognised that many caveats need consideration. We believe that we have taken into account both empirical evidence and theory judiciously in generating this ranking of teaching and learning activities. However, there may be concern that empirical evidence has been weighted more than theoretical considerations. On balance, we think that the weight to be attached to theoretical evidence should be considered very carefully, as theory so often does not materialise as expected when put into practice. Thus, we are confident that our emphasis on empirical evidence, particularly on evidence from a systematic synthesis of literature, is justified.

Decision-making about choice of teaching and learning methods should be guided by comparison of effects and costs associated with various educational strategies. Our aim was indeed to generate numerical measures of the effect size(s), however, due to poverty of reporting and substantial heterogeneity (in populations, teaching methods, outcome definitions, assessment tools and study quality, amongst other features), we were unable to provide any statistical summaries. Two very important issues where there were no or scant data relate to long term effects of the teaching methods and the effect on clinical outcomes. Pragmatic decisions have to be made based on the information available and in the absence of meta-analytic summaries, qualitative syntheses within broad subgroups of teaching methods and educational outcomes stratified by study methodology minimise biased inferences 24. When dominance can be so obviously demonstrated, a basic economic evaluation does not require statistically sophisticated analyses. For example, in a recent study 42, the average cost of providing a critical appraisal workshop was approximately £250 per person but no improvement was demonstrated in knowledge or attitudes. These findings challenge the policy of funding 'one-off' educational interventions aimed at enhancing EBM. Cost-effectiveness analyses are often simple and straightforward when the issue of dominance is not finely balanced 43. Is it reasonable to propose that standalone teaching should be abandoned, when an effort without desirable benefits does incur substantial costs? We think yes.

There are many barriers to the feasibility of an interactive and integrated approach and to its acceptability to both teachers and learners. Those who endeavour to embark on higher level of EBM teaching methods in our hierarchy will need to study these carefully as part of an implementation plan 44454647. These methods may be considered similar to an innovation in many settings with the need for various phases for embedding them in practice 48, For example within problem- based learning, there has been discussion about the utility of introducing of interactive approaches to learners who have yet to acquire a baseline level of knowledge. Interactive approaches are also seen as challenging across different cultures that have a tradition of didactic teaching. In these situations, methods lower down in our hierarchy may provide the pre-requisite knowledge or encouragement to learners before they engage in interactive and clinically-integrated teaching. It is important to remember that education concerning EBM can take one through the key initial stages of change before one is prepared to adopt EBM in practice. EBM is a strategy for just-in-time learning 49 that increasingly is possible in a clinical environment. Teachers and learners should carefully examine their learning environment and circumstances when developing an implementation strategy 50 for their chosen EBM related educational activities.