Giving Effective Feedback in Medical Education: A Guide for Doctors

• Effective feedback should help someone improve, not simply describe their performance. “Good job” and “you need more confidence” may be well intentioned, but neither tells the learner what they should continue doing or change.
• Learners need to know what good looks like. Correcting individual faults is much less useful if the learner does not understand the overall standard they are trying to achieve. Make expectations visible rather than leaving them inside the teacher's head.
• Before giving corrective feedback, try to understand why the problem occurred. The same apparent weakness can result from lack of knowledge, difficulty applying knowledge, limited practice, cognitive overload, misunderstanding the task or problems within the learning environment. Different causes require different solutions.
• Feedback should usually be a conversation rather than a verdict. Asking learners how they thought something went gives you information about their insight and may reveal difficulties you could not see from observation alone.
• Be specific about behaviour and its impact. Feedback becomes much more useful when the learner can identify exactly what happened, why it mattered and what they could do differently next time.
• Psychological safety does not mean avoiding difficult feedback. Learners need to feel able to make mistakes and expose uncertainty, but doctors also have responsibilities to patients, colleagues and professional standards. Sometimes feedback needs to be clear and uncomfortable.
• Feedback models can provide useful structure, but no framework substitutes for judgement. A two-minute conversation after a clinical encounter needs a different approach from a formal discussion about repeated poor performance.
• The purpose of feedback is ultimately change. The important question is not whether you successfully delivered your feedback, but whether the learner understood it and is better able to perform next time.

Feedback is sometimes treated as the final stage of teaching: deliver the session, let the learner practise and then provide some feedback. In reality, it is woven throughout effective teaching. If you are supervising somebody taking a history, you may intervene with a brief suggestion and then let them continue. When teaching a procedure, you might reinforce one part they performed well before correcting another. During a small-group discussion, you may challenge someone's reasoning and ask them to reconsider an assumption. These are all forms of feedback.

The underlying purpose is the same. Learners need information about the gap between what they are currently doing and what they are trying to achieve, together with some idea of how that gap might be closed.

Simply identifying the gap is not always enough. Imagine telling a trainee after a presentation that they need to “be more confident”. They may agree completely. They may have known they lacked confidence before they started. What they still do not know is what you observed or what they should do differently. They don’t know if they were they speaking too quietly, reading from the slides, avoiding eye contact, apologising repeatedly, losing their structure when questioned, or speaking so quickly that the audience could not follow them.

Until you identify the behaviour, “be more confident” is a description of the impression they created rather than useful developmental feedback.

Feedback sometimes fails because we concentrate on telling learners what they did wrong without ever making clear what they were trying to achieve.

Imagine you are cooking a dish for the first time. You have never tasted it and have only watched somebody make it once. As you cook, the person teaching you keeps tasting it and telling you to add more salt. You follow their instruction, but you have no idea how the finished dish is supposed to taste, why it needs more salt or how you would recognise when you had added enough. You are responding to the feedback, but you are not really learning how to make the dish for yourself.

This happens surprisingly often in medical education. We ask somebody to take a history, present a patient, lead a ward round, give a presentation or have a difficult conversation, but the standard we are judging them against exists largely inside the experienced clinician's head.

The learner then performs the task and receives feedback afterwards: ‘you need to be more structured’; ‘that was too detailed’; ‘you didn't prioritise enough’;’ you need to sound more confident’;’ your explanation wasn't clear enough’.

They may accept every word of it and try hard to respond. The problem is that they still do not necessarily know what good looks like. Just like the person adding more salt, they are making adjustments in response to somebody else's judgement without yet understanding how to judge the result for themselves.

This can create a frustrating cycle. Someone told that their patient presentation contains too much detail may strip so much out next time that important information disappears. They are then told that they have left out too much, so they put some of it back. Someone told to be more confident may concentrate on sounding assertive without understanding that what was really needed was a clearer clinical recommendation. The learner moves from one correction to another without developing a clear picture of the performance they are trying to achieve.

Feedback works much better when learners have some conception of the destination. They need to understand not only what was wrong with their latest attempt, but what good performance looks like and why. Over time, the aim should be for them to become less dependent on somebody else saying “more salt” and increasingly able to recognise for themselves when they have got it right.

---

Define the standard before you correct the gap

Where possible, establish expectations before the learner performs the task. That does not require an elaborate marking scheme for every clinical encounter. Sometimes a brief explanation is enough.

If you ask a junior doctor to present a patient, explain what you are looking for: a concise account of the relevant information, evidence that they have interpreted it rather than simply collected it, and a clear view about what they think should happen next.

If somebody is preparing a teaching session, they need to know that success is not simply getting through all their slides. You might expect them to identify what the learners need, structure the material clearly, involve the group appropriately and establish whether people have understood.

For practical skills, demonstration can be particularly powerful. Showing a learner a good example gives them something concrete against which to compare their own performance. Worked examples, observation of experienced colleagues, exemplars, rubrics and clearly described criteria can serve the same purpose in different settings.

The important principle is that standards should not remain hidden inside the teacher's head.

---

Make some of your expert thinking visible

This becomes particularly important as doctors become more experienced. Experts often recognise good performance almost instinctively but find it surprisingly difficult to explain exactly what they are recognising.

A consultant may immediately know that a trainee's presentation is unfocused, that their clinical reasoning is immature or that a difficult conversation has been handled particularly well. Years of experience allow them to recognise patterns that the learner cannot yet see. The danger is assuming that the learner can somehow infer those standards simply by being told when they have fallen short.

Part of teaching is making that tacit expertise visible. Explain what you are noticing. Describe the decisions you are making. Show the learner how an experienced practitioner distinguishes important information from distracting detail or how they adapt when a situation changes.

The aim is not to turn expert practice into an inflexible checklist. It is to give learners enough access to your thinking that they can begin developing judgement of their own.

---

Link feedback to the standard

Once the learner understands what good performance looks like, feedback becomes much easier to interpret. Instead of saying, “That presentation was too detailed,” you can explain the gap between what happened and what was required: “The aim is for the listener to understand the main clinical problem and what you think should happen next. You gave me a very complete history, but the important information was buried within it and I still wasn't sure what you thought the main problem was.”

The learner now knows not only what went wrong, but why it was a problem and what they are trying to achieve instead.

There is something inherently inefficient about allowing somebody to perform badly, telling them what was wrong, letting them try again, identifying another problem and repeating the process until they eventually infer what you wanted in the first place. Yet this is effectively how some workplace learning happens. The learner receives a succession of corrective comments but never gets a coherent picture of good performance. Feedback becomes a process of avoiding mistakes rather than developing towards expertise.

A better approach is to make the goal visible, observe where the learner currently sits in relation to it and then use feedback to help close the gap. Effective feedback therefore involves three questions:

• Where am I trying to get to?
• Where am I now?
• What should I do next?

Recognising that something went wrong is not necessarily the same as understanding why it went wrong. Before deciding what feedback to give, it is worth trying to understand what produced the performance you observed.

---

Start by understanding what actually happened

One of the most important distinctions in feedback is between recognising a problem and diagnosing it. Experienced clinicians understand this instinctively in medicine. Breathlessness is not a diagnosis. Neither is abdominal pain. We gather information because different underlying causes require different interventions.

Educational problems are similar. Suppose a learner repeatedly misses important information when taking a clinical history. It would be easy to tell them to “take a more thorough history”. But why are they missing things? Perhaps they do not know which questions matter. Perhaps they know perfectly well but are concentrating so intensely on maintaining rapport that they forget them. Perhaps their history-taking is rigid and they become lost when the patient tells their story in an unexpected order. Perhaps they are rushing because they believe they have only five minutes. Perhaps they are anxious because they know they are being observed.

Those are different educational problems. More explanation might help the first learner. A structured aide-memoire might temporarily help another. Practice may be more useful for somebody who knows the theory but cannot yet perform the skill fluently. The anxious learner may need repeated lower-stakes opportunities rather than another lecture on history-taking.

Before deciding what somebody needs to do differently, therefore, try to understand why they performed as they did.

---

Ask the learner what they thought

One of the simplest ways to understand performance is to ask the learner. “How did you think that went?” is useful not because learners must always be allowed to give themselves feedback first, but because their answer gives you information.

Sometimes they identify exactly the same issue you noticed. In that situation, there may be little value in spending several minutes carefully revealing something they already know.

This can be particularly helpful when somebody has performed badly. As a teacher, you may be sitting there worrying about how you are going to tell them without upsetting them, carefully choosing your words and wondering how they will react. Then you ask how they thought it went and they say, “That was awful. I completely lost my structure and I knew halfway through that I wasn't answering the question.”

The difficult revelation you were preparing to make is no longer necessary. The learner already knows that something went wrong. The conversation can move much more quickly from whether there was a problem to why it happened and what they can do differently next time.

In my experience, this is one of the most useful reasons for hearing the learner's view before launching into corrective feedback. We sometimes become so concerned about the possibility of upsetting somebody that we assume they are unaware of their poor performance. Quite often, they are already painfully aware of it. What they need from us is not a carefully softened announcement that they performed badly, but help understanding what went wrong and how to improve.

At other times, their interpretation is completely different. You may think a consultation lacked structure while the learner is worried that they forgot one relatively minor clinical question. That tells you something about where their attention was directed and may suggest that they have not recognised the more important problem.

Occasionally, the learner identifies something you had not appreciated at all. They may tell you that they lost concentration because they did not understand an instruction, that they were trying to apply feedback somebody else had previously given them, or that something happened before the encounter that affected their performance.

Self-assessment is not infallible. Learners can underestimate themselves, overestimate themselves or simply fail to recognise an important problem. You still have a responsibility to provide your own judgement. But hearing their perspective tells you where the conversation needs to start. Sometimes it confirms the problem, sometimes it reveals a different one, and sometimes it saves you from delicately explaining something the learner already knows perfectly well.

---

Consider whether the problem is really the learner

When somebody struggles, our attention naturally goes to the individual. Sometimes that is where the problem lies. Sometimes it is not.

A trainee who repeatedly fails to perform a procedure independently may simply not have had enough opportunities to practise. Someone who seems disengaged may be working in an environment where teaching is repeatedly cancelled because of clinical pressure. A learner who performs poorly with one supervisor but well elsewhere may be responding to something within that particular relationship or setting.

There may also be personal circumstances affecting performance that you know nothing about. This does not mean explaining away poor performance. It means remaining curious about its cause. Educational supervision involves asking not only what the learner need to change but occasionally what needs to change around this learner.

Once you understand what you have observed and why it may have happened, the next task is to turn that judgement into something the learner can actually use.

---

Make feedback specific

It is surprisingly easy to give feedback that sounds meaningful but gives the learner very little information they can actually use.

As experienced clinicians and educators, we often form an overall impression of somebody's performance very quickly. A presentation feels disorganised. A consultation lacks confidence. An explanation is difficult to follow. A learner seems hesitant or insufficiently patient-centred. Those impressions may be perfectly valid, but they are the beginning of feedback rather than the finished product.

The learner needs to know what you observed that led you to that conclusion.

Otherwise, we end up giving feedback such as:

• “You did really well.”
• “You need to be more structured.”
• “Try to communicate better.”
• “Be more confident.”
• “Read around the subject.”

The problem is not necessarily that these statements are wrong. It is that they leave most of the work to the learner. What does “more structured” actually mean? Which part of their communication was ineffective? What did they do that made them appear unconfident? What exactly should they read, and what gap in their performance is that reading intended to address?

This is particularly problematic with feedback about qualities such as confidence, communication, leadership or professionalism. These can easily become labels rather than descriptions of performance. Telling somebody that they “didn't seem confident” may accurately describe the impression they created, but confidence itself is difficult to practise. Speaking more slowly, making a clear recommendation, avoiding unnecessary apologies or maintaining eye contact are behaviours they can actually work on.

Specific feedback therefore moves from impression to observation. Instead of saying that a presentation was poorly structured, you might explain that the audience did not discover the main argument until five minutes into a ten-minute presentation. Instead of telling someone to communicate better, you might point out that they used several technical terms without checking whether the patient understood them. Instead of simply saying that their clinical reasoning was weak, you might explain that they produced a reasonable differential diagnosis but did not use the information from the history to distinguish which diagnoses were more or less likely.

That gives the learner something they can examine and change. It also makes the feedback easier to discuss. They may agree with your interpretation, explain why they behaved that way or offer context you had not appreciated. You are discussing something that happened rather than simply passing judgement on them.

Specificity matters when performance is good as well as when something needs to change. “Excellent” or “You did really well” may be encouraging, but they tell the learner very little about what you actually observed. Positive feedback becomes more useful when it identifies what worked and why it worked. We will return to this shortly, because successful performance deserves rather more attention than simply saying “well done”.

Specificity does not mean producing an exhaustive commentary on everything you observed. In fact, that can make feedback less useful. The skill is to identify the observations that matter most, explain why they matter and give the learner something sufficiently concrete to work with. Often one or two well-chosen, specific points will change performance far more than a long list of general criticisms.

---

What if you are the learner receiving vague feedback?

Of course, you will not always be the person giving the feedback. As a doctor in training, you may find yourself on the receiving end of exactly the sort of vague comments we have described.

If somebody tells you that you need to “be more confident”, “communicate better” or “be more structured”, it is reasonable to ask them what they mean. Do not assume that you should somehow be able to translate a general impression into a development plan by yourself. Ask for an example. You might say, “What did I do that made me come across as lacking confidence?” or “Which part of the presentation did you find difficult to follow?” If somebody tells you that your answer needs more structure, ask what a better structure would have looked like.

You can take this a step further by asking about the standard you are trying to reach: “What would you expect someone doing this well to have done differently?”

That can be a particularly useful question because it moves the conversation beyond identifying faults. If you are repeatedly being told what not to do but still do not have a clear picture of what good performance looks like, asking for an example, demonstration or description of the expected standard may be much more useful than collecting further criticism.

You can also check your understanding before the conversation finishes. “So the main thing you want me to work on is making my clinical recommendation earlier and then explaining the reasoning behind it. Is that right?” This forces vague feedback to become something sufficiently concrete that you can actually act on it.

Not every supervisor will be equally skilled at giving feedback. Sometimes you may need to help them give you something useful. That does not mean challenging every comment or demanding that somebody justify every criticism. It means taking some responsibility for your own learning. If you leave a feedback conversation knowing that you need to “improve your communication” but having no idea what you should do differently tomorrow, the conversation has not yet given you much to work with.

As a learner, therefore, do not be afraid to ask three simple questions:

• What specifically did you notice?
• What would good have looked like?
• What should I try next time?

Those questions can turn a vague judgement into useful feedback.

---

Do not overload the learner

Doctors can be extraordinarily efficient at identifying faults. Once invited to give feedback, we may suddenly notice everything. The introduction was too long, the differential diagnosis was incomplete, the examination sequence was slightly odd, the explanation contained jargon, the learner forgot to summarise, their eye contact could have been better or their management plan missed two possibilities. All of those observations might be true, but giving them simultaneously does not necessarily help.

Learners have limited capacity to work on multiple things at once, particularly when they are still developing a complex skill. Someone learning a procedure may already be concentrating on equipment, sequence, anatomy, communication and their own movements. Adding eight corrections can leave them with no idea where to begin.

Hence you need to prioritise. Ask yourself which change would make the greatest difference to performance. Sometimes there is one underlying issue responsible for several smaller problems, and improving that may resolve several things at once. Feedback is not an audit report. Its purpose is to help somebody take the next useful step.

---

Positive feedback should be more than praise

We tend to think of feedback as something that happens when a learner needs to improve. When somebody performs well, it is tempting simply to say “Excellent”, “Well done” or “You handled that really well” and move on.

Praise has value. It can encourage learners and reassure them that they are progressing. But praise and feedback are not quite the same thing. “You were brilliant” communicates approval; it does not necessarily teach the learner anything about why their performance was successful.

This matters because getting something right does not necessarily mean that the learner understands why they got it right.

A learner may reach the correct diagnosis by chance. They may choose the right management option while using flawed reasoning. They may handle a difficult conversation successfully because they happened to copy something they once saw another clinician do. Even when their underlying judgement was sound, they may not yet be consciously aware of what they did that made the interaction work.

If we simply say “Excellent” and move on, we may miss an important learning opportunity. Instead, explore successful performance in much the same way that you would explore something that went wrong. If a learner reaches the correct clinical decision, ask what led them there. If they handled an angry patient particularly well, identify what they did. Perhaps they allowed the patient to explain the problem without interruption, acknowledged their frustration before trying to solve it and then summarised the concern accurately. If they gave an unusually clear presentation, identify the decisions that made it clear: perhaps they established the main argument early, selected only the information that supported it and signposted the audience through the discussion. This allows you to establish whether the successful outcome resulted from understanding, deliberate skill or simply good fortune. More importantly, it helps the learner understand their own successful performance.

There is an important difference between getting something right and knowing why it was right. The second is much more likely to be reproduced when the learner encounters a different patient, problem or situation.

Positive feedback can therefore do more than make somebody feel good. It can help turn an accidental or intuitive success into something deliberate and transferable. Rather than simply reinforcing the outcome, you are helping the learner identify the reasoning, decisions and behaviours that produced it.

This is particularly valuable for experienced learners. Senior trainees and colleagues may receive remarkably little meaningful positive feedback because competent performance has become expected. People notice when something goes wrong but say nothing when a difficult situation is handled exceptionally well.

Yet experienced clinicians may also have developed effective behaviours so gradually that they are barely conscious of them. Specific positive feedback can make those strengths visible. It tells someone not merely that they performed well, but what they should continue doing and why.

Good positive feedback therefore answers much the same questions as corrective feedback. It tells the learner what they did, why it worked and what they should keep on doing.

Feedback can range from a fifteen-second comment after a clinical encounter to a much more substantial discussion about performance. The approach should be proportionate to what you are trying to achieve.

---

Understand whether you are giving feedback, coaching or assessing

Feedback, coaching and assessment often overlap in medical education, but they are not identical.

• Assessment makes a judgement about performance against some form of standard.
• Feedback provides information that can help somebody understand and improve their performance.
• Coaching goes further by helping the learner think through how they might develop.

The distinction matters because learners behave differently depending on what they believe is happening. Imagine that a trainee has just completed a consultation. If they believe you are making a formal judgement about whether they are competent, they have a strong incentive to demonstrate what they know and conceal what they do not. If they believe the purpose is genuinely developmental, they may be much more willing to say, “I wasn't sure what to do at that point.”

Both situations are legitimate. Medicine needs assessment because we have to establish whether people can practise safely. But learning also requires opportunities where people can expose uncertainty without feeling that every mistake is immediately becoming evidence against them.

As a teacher or supervisor, make the purpose of the conversation reasonably clear. Are you assessing performance? Are you helping somebody improve? Are you doing both? You cannot always remove the tension, particularly when you are simultaneously a supervisor and assessor, but acknowledging it is better than pretending it does not exist.

---

Feedback does not always need to be formal

Medical education has accumulated a considerable number of feedback frameworks. These can be useful, particularly when you are learning to give feedback and want some structure for the conversation.

The danger is turning every interaction into a ceremony. If a trainee has just performed a procedure and one small adjustment would improve their technique, you may only need to say what you noticed, explain why it matters and let them try again. A lengthy structured feedback conversation would add little.

The complexity of the feedback should be proportionate to the situation. A brief teaching encounter may need thirty seconds. A workplace-based assessment may justify a more structured conversation. Repeated concerns about a trainee's performance may require protected time, careful documentation and an agreed plan for what happens next.

Good feedback is not defined by whether you followed every stage of a model. It is defined by whether the conversation helped the learner understand and improve their performance.

---

Learn to give useful feedback in thirty seconds

Clinical medicine contains hundreds of small opportunities for feedback: after presenting a patient, taking a history, explaining something to a relative, performing part of an examination or contributing to a handover.

A simple way to think about these brief interactions is:

• What did I notice
• Why did it matter
• What should the learner try next time?

For example: “You gave me all the relevant information, but I couldn't tell which problem you thought was most important. Next time, start by telling me your overall impression and then give me the evidence that supports it.”

That may take fifteen seconds, but it contains an observation, explains its significance and gives the learner something specific to try.

These small conversations matter because feedback becomes part of ordinary clinical work rather than something saved for formal assessments. A learner who receives frequent, specific, low-stakes feedback has repeated opportunities to adjust their performance.

It also makes difficult feedback less unusual. If the only time a supervisor ever gives meaningful feedback is when something has gone seriously wrong, the word itself starts to acquire rather ominous associations.

Feedback models can be useful, particularly when you are new to teaching. They provide a structure for conversations that might otherwise feel awkward and can remind you to involve the learner rather than simply delivering a judgement.

The problem comes when the model becomes more important than the conversation. Learners do not all need the same type of feedback, and not every situation fits neatly into a predetermined sequence. A framework should help you think; it should not stop you thinking.

---

Pendleton’s rules

Pendleton’s rules are widely taught in medical education. In their familiar form, the learner is first invited to identify what went well, the teacher adds their observations, the learner then identifies what could be improved and the teacher adds further suggestions.

There are obvious strengths to this approach. It involves the learner, encourages reflection and prevents the conversation from becoming a list of criticisms delivered by the teacher. For somebody learning to give feedback, the structure can also provide reassurance about how to begin and move through the discussion.

The difficulty is that it can become predictable and artificial. Learners who know the model may realise that the positive discussion is simply the stage they have to get through before discovering what the teacher actually thinks went wrong. More importantly, the sequence may not fit the educational problem in front of you.

If a learner already knows exactly what went wrong, there may be little value in making them work through what went well before discussing it. If they have completely misunderstood their performance, you may need to challenge that misunderstanding rather than wait for them to discover it themselves. And if there is a serious concern about patient safety or professional behaviour, clarity is more important than faithfully following a feedback model.

Pendleton’s rules are therefore useful as a structure, particularly for less experienced teachers, but they should not become a script.

---

The feedback sandwich

The feedback sandwich is probably the best-known feedback technique: say something positive, give the criticism and then finish with something positive. It feels kinder than delivering criticism directly and can remind teachers that feedback should not focus exclusively on faults. But it has significant limitations.

However, if used routinely, learners quickly recognise the pattern. The first positive comment can become something they simply wait through because they know the criticism is coming next. The final positive comment may then feel like an attempt to soften what has just been said rather than meaningful feedback in its own right.

It can also dilute important messages. If somebody needs to understand that a particular behaviour was unsafe or unacceptable, surrounding that message with reassuring comments may create ambiguity about how seriously they should take it.

None of this means that positive feedback should disappear. As we have already discussed, recognising what somebody did well is an important part of learning. The problem is using praise primarily as packaging for criticism.

If something was done well, explain what was good about it because that information is educationally useful. If something needs to improve, explain clearly what you observed, why it matters and what the learner should do differently. There is no need to hide one inside the other.

---

Ask-tell-ask

Ask-tell-ask offers a more flexible structure. You begin by asking the learner for their perspective. You then tell them what you observed and add information, correction or guidance where necessary. Finally, you ask again to check their understanding, explore what they will do next or encourage them to identify how they might apply the feedback.

For example, after observing a consultation you might ask: “How did you think that went?” The learner may identify that they struggled to explain the management plan. You can then add your observation: “I agree. I noticed that you explained each option accurately, but you gave the patient a lot of information before checking what mattered most to them.” You might then ask, “How could you approach that differently next time?”

The strength of this approach is that it creates a conversation rather than a sequence of teacher judgements. It also gives you information about the learner's understanding before you decide what they need from you.

Even ask-tell-ask should not become compulsory. Sometimes the learner needs a direct correction. Sometimes they need a question. Sometimes they need reassurance, demonstration, practice or a much longer conversation.

The most useful approach is therefore not to search for the perfect feedback model. Learn a few structures, understand what they are trying to achieve and then use them intelligently. The model should serve the conversation, not the conversation the model.

Feedback is not particularly useful if it remains a discussion about something that has already happened. Its educational value comes from what the learner does with it afterwards.

A good feedback conversation should therefore end with some sense of direction. The learner should leave knowing what they need to work on, what they might do differently and, ideally, when they will have an opportunity to try it again.

---

Think about feedforward, not just feedback

The word feedback naturally directs our attention backwards. We discuss something that has already happened: the consultation, presentation, procedure or decision the learner has just made.

But the educational purpose is almost entirely about the future. If a trainee's patient presentation was poorly structured, explaining why it was difficult to follow is useful. The more important question is what they should do differently when they present the next patient.

This is sometimes described as feedforward: using what we have learned from past performance to shape future performance.

The distinction matters because it is possible to give very accurate feedback that does not actually help somebody improve. “Your presentation was too long and contained a lot of irrelevant detail” may be entirely correct. But unless the learner understands how to decide what is relevant and what to leave out next time, they may simply try to speak faster or remove information at random.

A more useful conversation might explore how they could identify the main clinical problem first, decide which information supports their assessment and structure the presentation around that. The learner now has something to try.

Where possible, make the next step practical and observable. “Be more confident” is difficult to practise. “Give me your clinical recommendation before explaining your reasoning” is something the learner can deliberately try in their next case.

The best feedback therefore does not simply ascertain what went wrong but also what should we try next time?

---

Help learners become better at seeking feedback

Learners also have a role in making feedback useful. A question such as “Any feedback?” places almost all of the responsibility on the supervisor. It may produce something useful, but it also makes it very easy to receive a generic response such as “That was fine” or “Just keep practising”.

More focused questions tend to produce more useful information. A learner might ask, “Was my explanation of the treatment options clear?” or “I am trying to make my patient presentations more concise. Was there anything there that you thought I could have left out?” Someone practising leadership might ask, “When I led that discussion, did I make the decision clear enough at the end?”

This does not mean that learners should only request feedback on things they already know they need to improve. Supervisors may have noticed something entirely different. But giving the observer some indication of what you are working on can make feedback more focused and increase the chance that somebody actually observes the relevant behaviour.

It also helps learners become less passive recipients of feedback. Rather than waiting for somebody else to identify their weaknesses, they begin to think deliberately about their own performance, identify areas they want to develop and seek information that helps them judge their progress.

This is an important part of becoming an independent professional. Consultants do not usually have somebody standing beside them after every difficult consultation offering structured feedback. Over time, doctors need to become increasingly capable of monitoring their own performance, recognising uncertainty and actively seeking another perspective when they need it.

---

Judge feedback by what happens afterwards

A feedback conversation can feel excellent and still achieve very little. The teacher may have been supportive, the learner may have agreed with everything that was said and both may leave feeling that the discussion went well. The more important question is what happens when the learner encounters a similar situation again. Did their performance change?

This is one reason repeated observation matters. If you have advised a trainee to make their clinical recommendation earlier in a patient presentation, try to hear another presentation. If somebody is working on explaining risk more clearly, observe them doing it again. If a learner has changed their approach to a procedure, give them an opportunity to practise the change.

Without this, we can easily mistake agreement for learning. A learner saying “Yes, that makes sense” tells you that they have understood the conversation at some level. It does not tell you that they can translate that understanding into better performance.

If the same problem persists despite repeated feedback, that is also useful information. Perhaps the advice was too vague. Perhaps the learner understood what to do but lacks the skill to do it. Perhaps they need demonstration or supervised practice rather than further explanation. Or perhaps the original diagnosis of the problem was wrong.

Feedback therefore works best as a cycle: observe, understand, discuss, agree what to try and then observe again.

Ultimately, the quality of feedback should not be judged by how sophisticated the conversation sounded. The real test is whether it helped the learner understand their performance and do something better next time.

Most feedback conversations are relatively straightforward and developmental. Others are more difficult because the learner disagrees, the relationship is complicated, the feedback is uncomfortable to hear or the concern has implications for professional standards or patient safety. These situations require the same principles as ordinary feedback, but greater care about clarity, context and responsibility.

---

Difficult feedback requires clarity as well as kindness

Feedback becomes harder when performance is poor, particularly when the learner may not recognise the problem. There can be a temptation to soften the message so much that it becomes difficult to understand. We say things such as “There are perhaps one or two areas you might want to think about” when what we actually mean is that there is a significant concern about performance.

Kindness matters, but ambiguity is not necessarily kind. If a learner needs to improve something important, they need to understand that clearly. Describe what you observed, explain why it matters and distinguish between a developmental suggestion and something that genuinely needs to change. This does not require aggression. In fact, difficult feedback is often more effective when it is calm, factual and specific.

Compare “Your attitude with that patient was unacceptable” with: “When the patient challenged the plan, you interrupted them several times and then said, ‘I've already explained this.’ They became visibly more upset and stopped engaging with the discussion. We need to look at how you handle disagreement in consultations”. The second message is still uncomfortable, but it gives the learner something concrete to understand and discuss.

---

What if the learner disagrees with your feedback?

Disagreement does not automatically mean that the learner is defensive or lacks insight. They may have information you do not. They may have interpreted the situation differently. They may have been trying to apply previous feedback from another supervisor. Or they may simply disagree with your judgement.

Start by exploring their perspective. What did they think was happening at that point? As them to talk through why they approached it that way. It may reveal something important. Sometimes you will change your view. Sometimes you will not.

If the learner continues to disagree, you do not necessarily need to persuade them that your interpretation is the only possible one. But you should be clear about the standard and the impact of the behaviour you observed. There is a difference between saying, “You must agree with my opinion” and saying, “You may see this differently, but this is the concern I observed, this is why it matters and this is the standard expected”. That distinction becomes particularly important when the issue concerns patient safety, professional behaviour or repeated poor performance.

---

Psychological safety does not mean lowering standards

Learners are more likely to ask questions, admit uncertainty and discuss mistakes when they feel psychologically safe. That matters enormously in clinical education. But psychological safety is sometimes misunderstood as meaning that learners should never feel uncomfortable.

Learning can be uncomfortable. Discovering that your clinical reasoning was weak, that your communication upset a patient or that your performance was below the expected standard is unlikely to feel pleasant.

The aim is not to remove that discomfort. It is to create an environment in which the learner can engage with it without feeling humiliated, dismissed or personally attacked. You can therefore be both supportive and demanding. A psychologically safe learning environment allows somebody to say, “I didn't know what to do,” while still expecting them to develop the knowledge or skill they need. It allows mistakes to be discussed openly while still taking patient safety seriously. Safety should make difficult conversations possible, not prevent them from happening.

---

Feedback is influenced by the relationship and the environment

It is tempting to think that good feedback is simply a matter of saying the right thing in the right way. The relationship between the people involved matters just as much.

The same comment can be interpreted very differently depending on whether it comes from a supervisor the learner trusts, somebody they barely know or somebody they believe has already formed a negative opinion of them.

Credibility matters too. Learners are more likely to engage with feedback when they believe the person giving it understands their work, has observed enough to make a reasonable judgement and genuinely wants them to improve.

The wider learning environment also influences what happens. In a department where mistakes are routinely discussed constructively, feedback may feel normal. In an environment where criticism is associated with blame or humiliation, even carefully worded developmental feedback can feel threatening.

This is one reason effective feedback cannot be reduced entirely to a communication technique. The culture around the conversation helps determine whether the learner is able to use what they hear.

---

Adapt your feedback to the person receiving it

People respond differently to feedback. One learner may appreciate you getting straight to the point, while another finds the same approach abrupt or discouraging. Some people want examples and evidence before they accept a criticism. Others need to talk through what happened. Some are comfortable being challenged, while others may become defensive if they feel that the conversation has become confrontational.

This does not mean that we should change the standard depending on the learner. It means recognising that the same message can sometimes be communicated in different ways.

Think about the people you work with. You probably know colleagues who would rather hear, “The main problem was X, and this is what you need to do differently next time,” than spend ten minutes gradually working towards the point. Other people want to understand exactly how you reached your judgement and will ask for examples. Some naturally want to discuss the situation and explore possibilities, while others need a little reassurance that criticism of one aspect of their performance is not a judgement on their overall ability.

Various models have been developed to describe these differences. One commonly used example is Social Styles, which broadly describes people as tending towards Driver, Analytical, Amiable or Expressive ways of communicating. You do not need to know the model to apply the underlying principle, and people rarely fit neatly into one category. The value is simply in recognising that different people may find different approaches to feedback easier to hear and use.

Someone who likes directness and results (the Driver or Red in the Social Styles model) may prefer you to identify the main issue quickly and concentrate on what needs to change. Too much preliminary discussion may frustrate them.

Someone who values detail and evidence (The Analytical or Blue in the social Styles model) may want you to explain exactly what you observed, why it mattered and what standard you were using. A vague comment such as “You need to be more confident” may be particularly unhelpful to them.

Someone who places considerable importance on relationships and harmony (The Amiable or Green under the Social Styles model) may find very blunt feedback harder to process. You do not need to dilute the message, but making your supportive intent clear and allowing some space for discussion may help them engage with it.

Someone who naturally likes to talk through ideas and possibilities (The Expressive or Yellow under the Social Styles model) may respond well to a more conversational approach. The challenge may then be ensuring that the discussion ends with one or two concrete actions rather than several interesting ideas about what they might do differently.

Your own preferences matter too. A naturally direct person may think they are simply being clear when somebody else experiences them as abrupt. Someone who dislikes conflict may soften a difficult message so much that the learner does not realise there is a significant problem. Someone who values detail may give an extremely thorough analysis when the learner really needs to know the two things that matter most.

This is why becoming good at feedback involves developing some flexibility in how you communicate. You do not need to analyse somebody's personality or put them into a category. Pay attention to how they respond. Ask questions. Notice whether they need greater clarity, more evidence, more opportunity to talk or simply a more direct conversation.

Most importantly, adapt the delivery, not the standard. If somebody's clinical practice is unsafe, their communication is ineffective or their professional behaviour needs to change, that message should remain clear. Adapting your feedback means finding the best way of helping that particular person understand and act on it, not making the feedback easier simply because they find it uncomfortable.

---

Be particularly careful when there is a power imbalance

Feedback between colleagues of similar seniority is different from feedback given by somebody who assesses, supervises or influences the career progression of the learner. A trainee may appear to agree with you because disagreeing feels unsafe. They may say the feedback was helpful because you will shortly be completing their assessment. Silence does not necessarily mean agreement. That makes it even more important to invite their perspective genuinely rather than rhetorically.

Your language matters too. Feedback should focus on observed behaviour and professional expectations rather than becoming personal. Where there are serious concerns, learners should understand what has been documented, what standards are expected, what support is available and what will happen next. The greater your authority over somebody, the greater your responsibility to use it carefully.

---

Remember that people may interpret feedback differently

Even carefully delivered feedback can be interpreted differently from the way you intended it.

Language, culture, previous educational experiences and expectations of hierarchy can all influence how people respond to criticism, praise and directness. A style that one learner experiences as refreshingly straightforward may feel unusually confrontational to another. Equally, very indirect feedback may be interpreted as suggesting that an issue is optional when you intended it to be taken seriously.

This is another reason to check understanding rather than assuming that because you said something clearly, it was necessarily heard in the way you intended. You might ask, “What are you taking away from this conversation?” or “What do you think the main thing to work on is?” Their answer may surprise you.

Communication is not complete simply because the teacher has spoken. The important question is what the learner understood.

---

Consider whether the problem is really the learner

When somebody struggles, our attention naturally goes to the individual. Sometimes that is where the problem lies. Sometimes it is not.

A trainee who repeatedly fails to perform a procedure independently may simply not have had enough opportunities to practise. Someone who seems disengaged may be working in an environment where teaching is repeatedly cancelled because of clinical pressure. A learner who performs poorly with one supervisor but well elsewhere may be responding to something within that particular relationship or setting.

There may also be personal circumstances affecting performance that you know nothing about.

This does not mean explaining away poor performance. It means remaining curious about its cause.

Educational supervision involves asking not only “What does this learner need to change?” but occasionally “What needs to change around this learner?”

---

Give feedback close enough to the event to be meaningful

Feedback is generally easier to understand when the event is still fresh. If you wait three months to tell somebody that their communication in a particular consultation was problematic, neither of you may remember enough detail to have a useful discussion.

That does not mean feedback must always be immediate. Sometimes emotions are high, the clinical environment is inappropriate or you need time to gather information. Difficult feedback may be better given privately once both people can concentrate properly.

There is a difference between timely and instantaneous. Choose a point when the learner can remember what happened but is also able to engage properly with the conversation.

---

Do not save important feedback for the end of a placement

One of the least useful times for a learner to discover an important problem is at their final assessment. If you have been concerned for several weeks that a trainee's clinical reasoning, communication or professional behaviour is below the expected standard, telling them at the end of the placement gives them little opportunity to respond.

That does not mean making premature judgements after one difficult day. You may need further observation or another opinion. But once a meaningful pattern becomes apparent, delaying the conversation because it feels uncomfortable is rarely helpful to the learner.

Important concerns should be raised early enough for the learner to understand them and, where possible, do something about them. There should therefore be as few surprises as possible in a formal end-of-placement assessment. If a significant concern appears there, the learner should ordinarily already know that it exists and what they have been expected to do about it.

---

When poor performance needs more than feedback

Most feedback is developmental. Sometimes, however, you are dealing with performance that raises concerns about competence, professionalism or patient safety.

At that point, feedback is only one part of your responsibility. If the concern is significant, persistent or potentially harmful, you may need to document what happened, agree clear expectations, arrange further supervision or assessment and involve the appropriate educational or clinical lead. The exact process will depend on the learner's role and the organisation in which you work.

Do not promise confidentiality if the issue is something you may need to escalate. Equally, avoid turning an isolated mistake into a judgement about somebody's overall competence. Context matters. People have bad days, unfamiliar situations expose weaknesses and even experienced clinicians make mistakes.

The task is to distinguish a developmental problem that can be addressed through ordinary teaching from a concern that requires a more formal response.

Giving good feedback is itself a skill. Like any other teaching skill, it improves when you pay attention to what happens during the conversation, notice how people respond and reflect on whether your approach actually helped.

Part of that development involves becoming more comfortable receiving feedback yourself.

---

Learn to receive feedback as well as give it

Doctors who teach should also become comfortable receiving feedback themselves. Ask learners what helped, what remained unclear and what they would change. If you teach regularly, look for patterns rather than reacting dramatically to individual comments. Peer observation can also be valuable because another teacher may notice habits that learners do not mention or that you cannot see yourself.

Feedback about teaching needs interpretation just like feedback about clinical performance. A numerical rating tells you relatively little on its own. A comment from one learner may reflect a genuine problem, an individual preference or a mismatch between their experience and the intended level of the session.

That does not mean dismissing uncomfortable feedback. It means examining it intelligently. The most useful teachers are not those who receive universally glowing evaluations. They are those who remain interested enough in their own performance to notice what could be better.

---

Becoming better at feedback takes practice

Giving useful feedback is harder than it first appears because it requires several skills simultaneously. You need to observe accurately, distinguish important issues from minor ones, understand why the learner may be struggling, communicate your judgement clearly and adapt the conversation to the person in front of you.

You will not get every conversation right. Sometimes you will realise afterwards that you were too vague. Sometimes you will say too much. You may avoid an issue because it felt uncomfortable or discover that something you intended as constructive was heard very differently by the learner.

Reflect on those conversations in the same way you would reflect on any other part of your teaching. What did the learner understand? What did they do afterwards? Was your feedback specific enough? Did you address the real problem? Would you approach the conversation differently next time?

Becoming good at feedback is itself an iterative process.

Giving effective feedback is only one part of becoming a good clinical teacher. The same principles connect with many other aspects of medical education: understanding how people learn, choosing appropriate teaching methods, supervising learners, assessing performance and developing the confidence to teach in different clinical settings.

For a broader introduction, read How to Develop Your Teaching Skills as a Doctor: From Imposter Syndrome to Confident Medical Educator. It looks at how doctors can develop as teachers throughout their careers and brings together many of the practical skills involved in clinical education.

If you want to explore the educational theory behind different approaches to teaching, Teaching Methods for Doctors: A Practical Guide to Clinical Teaching Techniques looks in more depth at areas such as constructive alignment, Bloom's taxonomy, Miller's pyramid, experiential learning, cognitive load, scaffolding and simulation.

Doctors who want structured practical training can also consider our Teach the Teacher Course for Doctors and Healthcare Professionals. The course explores how to plan and deliver effective teaching, engage different learners, give useful feedback and develop your own approach as a clinical educator.

If presenting is a particular challenge, our Public Speaking & Advanced Presentation Skills Course for Doctors and Healthcare Professionals focuses specifically on developing confidence and communicating effectively when presenting to groups.

For doctors considering a more formal qualification in medical education, you may also find Is a PGCert in Medical Education Worth It? My Experience as a Doctor and Clinical Teacher useful. It explores the experience of completing a postgraduate certificate alongside clinical work, including the workload, educational benefits and practical considerations involved.

Good feedback ultimately comes back to a relatively simple idea. The aim is not to demonstrate how accurately you can identify somebody else's weaknesses. It is to help another person understand their performance well enough to improve it.

That requires more than a feedback model. You need to know what good looks like, observe what actually happened, understand why it happened and communicate what matters in a way that the learner can use. And then you need to see what happens next.