These are not station scripts. They are the methods the stations keep asking for: how to answer when you do not know, how to take consent for any procedure, what every station costs the patient, and how to tell whether you have finished or merely stopped.
A candidate is shown a nerve in the axilla. She does not recognise it. She has four seconds of silence in her head, decides that silence is worse than a guess, and says "the ulnar nerve." It is the thoracodorsal nerve. The examiner moves on.
She has just lost more than one identification. She has lost the knowledge mark for that structure. She has lost professionalism marks, because a confident wrong statement about a patient's anatomy is a safety behaviour, not a trivia failure. And she has lost the thing that would have saved her: the examiner can lead a candidate who is reasoning towards an answer, and cannot lead one who has already committed to the wrong one.
If you have already sat this exam and failed a station you thought you knew the content of, this is probably how. If you are sitting it for the first time, the useful thing to notice is that nothing in that account was a knowledge problem. She was not asked anything unreasonable. She lost the station on the four seconds, and the four seconds are trainable.
In every station type in this exam, a bounded honest answer outscores a confident wrong one.
Not because examiners reward humility, and not as consolation for candidates who did not revise. It is a scoring mechanism, and it follows from the fact that this exam marks several domains at once rather than a single answer. A wrong answer given confidently scores zero in the domain it was aimed at and negative elsewhere. The same gap, handled as a bounded answer, scores in communication and professionalism, keeps the knowledge domain live because you are still reasoning inside it, and leaves the examiner able to prompt you.
Three moves, in order, every time.
State the limit, briefly. One clause, not a paragraph. "I can't identify that with certainty." Not an apology, not an explanation of what you did and didn't revise, and never a performance of anxiety.
Give what you do have. This is the part candidates omit and it is where the marks are. You almost always have something: a location, a description, a relation, a range, a principle, a route to the answer. Say it in the vocabulary of the station.
Offer a reasoned differential or a route. Name the two or three possibilities and say which you favour and why, or say how the answer would be established and by whom. Then stop, and let the examiner take it.
Never silence. Never a guess presented as fact. Never a number you invented.
The worked examples in this chapter look like one skill in several costumes, and mostly they are. But the limits behind them are not all the same kind of thing, and treating them as though they were will eventually produce a wrong answer in the case built specifically to punish it.
A limit you feel. You do not know, and you know that you do not know. This is the ordinary case — the unidentified nerve, the unfamiliar specimen — and it is where most candidates' instincts are already roughly correct, if untrained.
A limit you recognise. You may know perfectly well what you think is true, and saying it plainly would be wrong — not because you are unsure, but because it is not yours to say. This is a boundary rather than a gap, and it requires the opposite instinct from the first kind.
A limit that belongs to your instrument, not to you. Sometimes the honest answer is not "I don't know" but "the thing I am using to find out cannot tell me, regardless of how carefully I use it." You may be entirely correct in your technique and still be unable to get an answer, because the method itself has a ceiling.
A limit on your certainty that you will not experience as a limit at all. This is the least intuitive of the four — the case where a candidate who is not looking for it will simply never notice it is there, because from the inside, everything feels finished and correct.
Sometimes the answer is bounded because you do not know it. That is the anatomy and pathology case, and the recovery is reasoning out loud.
Sometimes it is bounded because it is not yours to say. That is the duty of candour case, and occasionally the consent case. Here you may know perfectly well what you think happened, and saying it would be wrong. The move looks identical from outside — a stated limit, then what you can give, then a route — but the limit is a boundary rather than a gap, and the tone that goes with it is different. Reasoning out loud is right for a gap and wrong for a boundary. If you speculate helpfully about who left the swab in, you have demonstrated exactly the failure the station was built to detect.
Candidates who learn only the first kind apply it to the second and talk themselves into trouble while sounding thoughtful.
This is genuinely different from not knowing something, and the difference matters because the wrong corrective here is actively dangerous rather than merely weak.
A common instinct is to name what would settle the question — to reach for a further test or investigation as the answer to your own uncertainty. This is unsafe as a general instruction. There are situations in which naming a further test is precisely the wrong answer, because the delay involved in obtaining it is itself the harm — a clinical scenario in which the correct response to diagnostic uncertainty is to act immediately, not to request more information, is the clearest possible case, and a candidate who reaches reflexively for "I'd get a scan" in that situation has just described a course of action that would cost the patient the outcome the station exists to test for.
The safer instruction is broader: name the limit, name what would follow from it, and name what that following costs. Sometimes what follows is indeed a better test — and you say so, along with what it would show and how long it takes. Sometimes what follows is a decision that cannot wait for a better test at all, and you say that instead, together with what you are choosing to do in the meantime and why. Occasionally what follows is not available to you right now under any circumstances, because the object you would need to examine further has already been removed from consideration, and the honest answer is that the question will have to be settled by re-examination later, on different terms.
The unifying instruction is not "ask for more information." It is: say precisely what your instrument cannot tell you, and say what happens next as a result — including, honestly, what that next step costs, in time, in risk, or in the loss of an option.
This is the case this method is least naturally built to describe, because everything else assumes you know you have hit a limit. This one is different: you are correct, you are certain, and the discipline is still required — not because you might be wrong, but because describing your reasoning before committing to a conclusion is itself part of what is being marked, independently of whether your conclusion turns out to be right.
Put plainly: describe, even when you are sure — particularly when you are sure. A candidate who is confident and skips straight to the answer has not made an error of fact. They have skipped a step that was never contingent on their being uncertain in the first place. The examiner cannot tell, from a correct answer delivered instantly, whether you reasoned your way there or recognised a pattern from memory — and in several of this book's station types, the reasoning is what is actually being scored, with the correct answer treated as a necessary but insufficient condition for full marks.
Prosection, an unidentified structure. (First kind.)
"I can't identify that with certainty. It's a cord-sized nerve lying posterior to the axillary artery and running towards the posterior wall of the axilla, so the possibilities are the thoracodorsal nerve or the lower subscapular nerve. I'd favour the thoracodorsal, because it's running with the subscapular vessels towards latissimus dorsi."
A limit, a location and calibre, a two-item differential, a reason. The examiner now has three places to prompt from.
Pathology, an unrecognised specimen. (First kind.)
"I'll describe it before I commit. This is a segment of colon with a raised, ulcerated lesion about four centimetres across, with rolled everted edges, extending through the wall. That appearance and that depth would make an adenocarcinoma most likely, with a chronic inflammatory ulcer the alternative I'd want to exclude."
Description first, diagnosis second. A candidate who leads with the diagnosis and is wrong has nothing left to recover with. A candidate who describes accurately can be led to the right answer, and has already scored for the description.
Consent, a figure you don't have. (Third kind, borderline first.)
"I don't want to give you a number I'm not sure of. It's uncommon rather than rare, and I'll find the exact figure from our unit's results and confirm it with you before the operation."
In some settings this reads as weakness. Here it is a professionalism mark. Inventing a percentage is one of the fastest ways to fail a consent station, and candidates do it because they believe a specific number sounds authoritative. It sounds authoritative to the patient. It does not to the examiner, who knows the figure.
Duty of candour, a question you must not answer. (Second kind.)
"I don't want to guess at how it happened, because that wouldn't be fair to you or to anyone else involved. What I can tell you is what I know: a cut was made on the wrong side before the team realised. It's being formally investigated, and I'll make sure you're told what that investigation finds."
Same three moves. But notice that no reasoning is offered, and that this is deliberate. The limit here is a boundary. Filling it with plausible explanation would score worse than saying nothing.
History, a case where the usual instrument-limit answer would be actively dangerous. (Third kind, the unsafe-instrument case.)
"I'm not going to wait on a scan for this. Torsion is a clinical diagnosis, imaging isn't reliable enough to rule it out, and waiting for one costs the testis. Based on the time of onset he's given me, this needs a surgical opinion now."
This is deliberately the exception the other examples might tempt you to generalise from. Everywhere else, naming the further test is the marked response to uncertainty. Here it is not: the correct instrument-limit answer is to name the limit of the examination and explicitly decline the investigation, because the investigation's delay is the harm. A candidate who reaches for "I'd get an ultrasound to confirm" in this specific scenario has just described the answer that would cost the outcome the station exists to test for. (See the Testicular Pain history station for the full worked case this example is drawn from.)
Examination, a lump you recognise at a glance. (Fourth kind.)
"This is a two-centimetre, round, smooth lump on the scalp, with a central punctum visible and no fluctuance change on transillumination. It's mobile over deep structures but the skin cannot be moved independently over it — it's tethered to the overlying skin, which is the finding that separates it from a lipoma. Consistent with a sebaceous cyst."
This is the case candidates most often skip straight past, because skipping feels efficient rather than reckless. A candidate who says only "sebaceous cyst" has given a correct answer and lost marks anyway, because the examiner cannot tell whether that word came from the findings or from a glance and a guess. The instruction is unconditional on how confident you are: describe before you name, particularly when you are sure, because the certainty is exactly the condition under which the discipline is easiest to skip and hardest to notice yourself skipping. (See the Superficial Lesion or Ulcer station for the governing instruction this example draws from: "describe it even when you are sure, particularly when you are sure.")
The examples above all show a candidate correctly recognising a limit and handling it well. There is a second, equally important failure direction.
The failure in the other direction is deploying this method where no real limit exists — using the language and cadence of a bounded answer to disguise a gap that revision, not honesty, should have closed. A candidate who says "I can't be certain, but I'd favour X" about a structure they simply never learned is not demonstrating professional humility; they are using this chapter's own vocabulary as camouflage, and an examiner who has seen this pattern before is entitled to notice the difference. The tell is usually in the differential offered: a genuine bounded answer produces two or three specific, locally reasoned possibilities; a disguised gap produces either no real differential at all, or one so generic it could apply to half the syllabus.
This method rescues the edges of your knowledge. It does nothing for the middle, and pretending otherwise, fluently, is worse than an honest "I don't know this," because it consumes the examiner's patience and the professionalism domain's goodwill on a performance rather than a genuine limit.
Everything above assumes a gap that is visible — you know you have hit a limit, and the only question is how you handle it. There is a related failure worth naming separately, because it is not a limit at all.
Burying a gap in volume — talking at length, correctly, about everything you do know, so that the one thing you have not established passes unnoticed inside the recitation. In a telephone referral, this looks like giving a full, accurate, unhurried history that never quite states the one observation the listener actually needed, and relying on the sheer completeness of everything else to prevent them from noticing.
This looks similar to a candidate reasoning aloud around a genuine limit, but it is a different mechanism entirely. Nothing is hidden from you by uncertainty here; completeness is being mistaken for, and used as a substitute for, disclosure. The candidate is not bounding an honest gap. They are concealing one, fluently, by volume. The corrective is not any of the three moves given above — it is simply: say the missing thing plainly, near the start, rather than trusting that thoroughness elsewhere will excuse its absence.
It is not a substitute for knowing things. The method rescues the edges of your knowledge; it does nothing for the middle. A candidate who deploys it on the contents of the femoral triangle has not demonstrated a professional limit, they have demonstrated that they did not revise, and the examiner can tell the difference instantly.
It is not hedging. Bounding one answer scores. Bounding every answer reads as a candidate with no confident knowledge at all, and the communication domain will register it.
It is not slow. The whole thing takes about eight seconds. Practise it aloud until it is one movement, because the four seconds it replaces are not a thinking pause — they are a panic, and you cannot compose during a panic. The wording has to be there already.
The only useful drill is adversarial. Have someone point at structures you have not revised, ask for figures you do not have, and ask you questions you should refuse. On your own you will unconsciously choose gaps you can nearly fill, which trains the wrong thing.
The drill should specifically include at least one instrument-limit case and one confident-and-correct case, not only a gap and a boundary case — otherwise you can complete the whole exercise having only ever practised half of what this chapter teaches.
If you take nothing else from this chapter into the exam, take the order: limit, what you have, route.
Applies to any station where a procedure is being consented for, regardless of which procedure — the four-item core (procedure, alternatives including doing nothing, material risks, honest benefit) is deliberately built to be filled for any operation rather than any specific one.
A candidate is asked to consent a patient for a laparoscopic cholecystectomy. He knows the operation. He lists fourteen risks, in order, accurately, including two the examiner had not expected. He finishes with ninety seconds to spare and asks whether the patient has any questions.
He fails.
He never established what she already understood, so he pitched the whole conversation at a level he guessed. He never asked what mattered to her, so he gave equal weight to a risk she would have dismissed and a risk that would have changed her decision. He used "conversion to open" twice without translating it. He offered no alternative, including doing nothing. And the invitation to ask questions arrived after the decision had effectively been made, which is not an invitation.
This is the commonest way to fail a consent station and it is available only to candidates who know the operation well. The ones who know it less well are forced to have a conversation instead, and often score better.
Consent stations mark a process. The procedure is the variable.
The risk list changes between cholecystectomy, hernia repair, thyroidectomy and mastectomy. Nothing else does. The sequence, the domain weighting, the opening, the checking, the documentation and every behaviour that earns marks are constant across all of them.
Which means this is learnable once and carried into any station, including one for a procedure this book never covered — and the exam can hand you exactly that.
Consent is the most evenly balanced station type in the exam. Clinical knowledge carries the procedure description and the risks. Communication carries whether the patient could repeat back what you said. Professionalism carries whether you checked understanding, worked inside your competence, and documented. Technical skill is minimal.
Candidates pour everything into the risk list because it is the part that feels like knowledge and the part they revised. That single section is at most a quarter of the marks, and spending seven of eight minutes on it forfeits the rest. You cannot pass this station by being encyclopaedic, and the better you know the operation the more tempting it is to try.
Ten moves. The same ten, every procedure.
1. Introduce and confirm. Name, grade, patient identity, and that you are here to talk about the operation.
2. Establish what they already understand. "Before I go through the details, can you tell me what you've been told so far?" This earns communication marks and tells you where to pitch everything after it. Never skip it to save time; it saves time.
3. Establish what matters to them. A risk is material if a reasonable person in this patient's position would attach significance to it, or if you are or should be aware that this particular patient would (per Montgomery v Lanarkshire Health Board). That is the legal standard and it has a practical consequence: you cannot know which risks are material until you know something about their life. A self-employed builder and a retired teacher need the same list weighted differently.
4. Describe the procedure in plain language. No jargon, or jargon immediately translated. Where you will operate, what you will do, how long, what they will have afterwards — drains, dressings, scars, a catheter.
5. Benefits, stated honestly. Including the honest limit: what the operation is likely to fix, and what it will not. This is where the operation's own failure lives, and it is not a risk. A hernia can recur, a repair can fail to relieve the pain it was meant for, a cosmetic result can disappoint — with every step done correctly and none of the risks below having occurred. That is a limit on the benefit, stated here. Bleeding, infection and nerve injury are hazards the operation can cause; a benefit not landing is not a hazard, it is the absence of a benefit, and folding it into the risk list at step 6 blurs a distinction patients are entitled to have kept clear.
6. Risks in three tiers. Common and expected. Less common. Rare but serious. Then the procedure-specific ones weighted by what you learned at step 3. Use frequency words with a consistent meaning and do not invent percentages — see below.
7. Alternatives, including doing nothing. What happens if they decline, or wait. This is the step most often omitted entirely and it is a marked one.
8. Anaesthetic and recovery. Who will see them about the anaesthetic, how long in hospital, when they can drive, lift, return to work. Recovery is often what the patient actually wanted to know.
9. Check understanding, properly. Not "does that all make sense?" — everyone says yes. Ask them to tell you in their own words what they will have done and what the main risks are. If they cannot, you have not consented them, you have talked at them.
10. Questions, documentation and the form. Invite questions with time left to answer them. Say what goes in the notes, that they get a copy, and that consent is confirmed again on the day.
This is the transferable core. To consent for an operation you have never consented for, you need four things and no more:
If you can fill those four, you can run the sequence. If you cannot fill number three, you are not the right person to be consenting this patient, which is itself a scoring answer.
Every operation carries bleeding, infection, pain, scarring, anaesthetic risk and venous thromboembolism. Reciting them demonstrates nothing, because they are the same six every time.
What distinguishes a candidate is the short list that belongs to this operation and no other:
Procedure
The specific ones
Laparoscopic cholecystectomy
Bile duct injury, conversion to open, retained stones, bile leak
Open inguinal hernia repair
Chronic groin pain, recurrence, ilioinguinal nerve injury, testicular ischaemia
Thyroidectomy
Recurrent laryngeal nerve injury and voice change, hypocalcaemia, need for lifelong replacement
Axillary node clearance
Lymphoedema, numbness of the inner upper arm, shoulder stiffness
Learn four operations to that depth and the pattern generalises: ask yourself what this operation can damage that nothing else can, and what it can fail to achieve.
Do not invent a percentage. Inventing one is among the fastest ways to fail this station, and candidates do it because a specific figure sounds authoritative. It sounds authoritative to the patient and not to the examiner, who knows the real figure.
Use frequency language consistently — common, uncommon, rare — and give real numbers only where you actually know them, ideally your own unit's. When you do not know:
"I don't want to give you a number I'm not sure of. It's uncommon rather than rare, and I'll find our unit's figure and confirm it with you before the operation."
That is the consent application of the method in The Bounded Answer, and it scores.
Consent should be taken by the person performing the procedure, or by someone suitably trained with sufficient knowledge of it and its risks, per current General Medical Council guidance on decision-making and consent.
If the station hands you a procedure outside your competence, the marked answer is to say so and say who should do it. It is a professionalism probe and candidates who bluff it lose the domain outright. Saying "this should be consented by someone who performs this operation, and I'd ask my registrar to see her today" is a pass, not a failure.
Two related things the sequence assumes and does not test: that the patient has capacity, and that they are deciding freely. If either is in doubt the station has become a different one — see the Capacity and Best Interests station.
It is not a script to recite. A candidate reciting the ten moves in order without listening scores badly in exactly the domain the sequence exists to serve.
It is not a substitute for knowing the operations you assist at. The method tells you what four things to find out. Finding them out is still your job, and a candidate who fills them from nowhere on the day will be caught at the first follow-up.
It is not a risk-disclosure exercise. The point of the conversation is that the patient makes a decision. If they could not have decided differently after speaking to you, you did not consent them.
Take an operation you have never consented for, fill the four things from the operating list you are on this week, and run the ten moves aloud in eight minutes. Then have someone interrupt at step 6 to ask what happens if she does nothing, which is where most candidates lose the thread.
The point of learning this sequence here is that it works in a station this book does not contain.
Every station in this book asks the patient in front of you to give something up so that you can find something out. Most of the time nobody says so, including you. The examination that finds the diagnosis is also, at the same moment, costing the person on the couch something — comfort, dignity, time, patience — and the station is not testing whether you noticed the diagnosis. It is testing whether you noticed the cost.
This chapter is about that cost: what it is, why you will usually not be told when you have taken too much of it, and what to do about that fact before you start rather than after.
Four currencies recur across the stations in this book. They are different enough that no single instruction covers all of them, which is why this chapter is organised as a table rather than a rule.
Currency
The signal that you have asked too much
The corrective when there is no signal
Pain
A wince. Self-regulating — the patient tells you.
Warn, act, stop when told, and go back where it matters.
Dignity
None. The patient will not object.
Decide in advance what the examination requires, and explain the reason before you ask.
Time and awkwardness
A clock, which points the wrong way.
Accept the pace as correct rather than as failure.
A different kind of exposure — being asked open questions that invite more than a fact
Variable, and easy to mistake for willingness.
Supply the frame before the question, so the patient knows how much is being asked of them.
Read the middle column before the corrective. It is the argument of this chapter, compressed into four cells.
Three of the four currencies fail invisibly. A tentative test on a painful joint records an untested endpoint and looks, on paper, exactly like a properly performed one. An examination that took dignity nobody was prepared to give records an unremarkable finding and nothing else. An open question asked without warning that the patient did not want to answer records whatever they said, with no note that it cost them something to say it. Only pain produces a clean, visible failure signal — the patient winces, or says stop, and you know immediately. For the other three, the station will look identical whether you spent the currency correctly or overspent it. You will not be told in the room.
This is the chapter's whole argument: because you will not be told, the discipline has to be exercised before you ask, not adjusted in response to what happens after you do. Decide the pace, the framing, the amount of exposure required, in advance — and treat the absence of an objection as confirmation of your plan, not as permission to have skipped the planning.
The table's middle column names the signal. It does not yet say what happens when the signal is absent and you overspend anyway. There are three distinct costs, not one, and a candidate needs to know which one is in play because the corrective is different for each.
A false finding. A tentative examination — pressing gently because you sense reluctance, or rushing because the patient seems tired — does not simply fail to find something. It records a negative, and a negative that was never properly tested is worse than no finding at all, because it will be trusted. This is the commonest cost and the one most stations are built to expose.
A gap. An unearned "she understands" — moving on from an explanation because the patient nodded, without checking — records comprehension that was never established. An unasked red-flag question in a history records "no red flags" when the true state is "not asked." The station shows a completed field with nothing behind it.
The loss of the option itself. This is the least intuitive of the three and the one worth stating carefully, because it does not look like the other two. Some overspends do not merely produce a bad record — they remove your ability to correct course at all. Committing to a step you cannot undo before you have confirmed it was safe to take does not produce a false finding or a gap; it removes the option to have checked, permanently, at the moment you acted.
These three sort into two forms, and they run in opposite directions. In the first two — the false finding and the gap — the window closes by acting carelessly: you did the wrong thing, or failed to do the right one, and a corrective exists because you could, in principle, have noticed and stopped. In the third, the window closes by waiting: the cost of inaction accumulates silently until the option to act is simply gone. The corrective in both cases has the same shape — know where the edge is before you reach it — but it resolves to opposite instructions. Where the window closes by acting: check before you commit. Where it closes by waiting: commit before you cannot.
A candidate who has only been taught the first form will, correctly, become cautious — and in the situations governed by the third form, caution is exactly the wrong instinct. Knowing which of the two you are in is not a refinement of this chapter's argument. It is required to apply it at all.
Everything said about pain so far treats it as a single line — a station either does or does not cause it, and the patient either does or does not signal. Examination stations that test movement split this into two separate questions, and a candidate who has only thought about pain as one thing will not notice the second.
Who supplies the force, and who reads the result. In some manoeuvres the patient can do the work themselves: asked to bend a joint or lift a limb, they will stop at their own pain threshold, and the currency is spent entirely on their own terms. In others, this cannot be arranged. A ligament test that relies on feeling an endpoint, or a rotation performed against resistance to find where movement stops, only works if the examiner supplies the force — which means the patient's stopping point is no longer theirs to set. The exchange cannot be handed over, and pretending otherwise is the failure, not a considerate accommodation of it: a manoeuvre performed too gently to avoid causing discomfort produces a false negative rather than a kind examination, because the endpoint that was never reached was the finding.
The corrective is not to avoid these manoeuvres, and not to perform them at full force out of a fear of under-testing. It is to warn specifically, perform the manoeuvre once, to the point that actually produces the finding, and stop — the same three-part shape (warn, act, stop) as the pain row generally, but applied to a case where the patient's own signal cannot end the manoeuvre early without also destroying the reason it was being done.
And where a comparison between an active and a passive version of the same movement is itself the diagnosis — whether a restriction is muscular or structural, for instance, as in the knee and shoulder examination stations — both halves of the exchange must be completed for the station to have happened at all. Skipping the examiner-supplied half to spare the patient discomfort does not reduce what was spent; it produces a station with no finding in it, which is a worse outcome for the patient than the discomfort would have been, since it is the reason they were referred in the first place.
Currencies are not always spent one at a time. When two are spent together, only one of the two signals tends to be visible, and a candidate who is watching correctly for the visible one will still get this wrong.
Consider an examination that requires both physical exposure and sustained standing. Fatigue is visible — the patient sways, or asks to sit. Embarrassment at the exposure is not; a patient who is uncomfortable being looked at rarely says so. A candidate registers, correctly, that something is costing too much, and responds to the signal they can see: they shorten the physically demanding part of the examination. But the shortening does not relieve the cost that was actually driving the discomfort — the discomfort was the exposure, not the standing — so the wrong currency gets protected and the wrong one gets cut.
The corrective is not to watch harder for the invisible signal, which by definition you cannot do. It is to offer the relief that addresses the visible currency without assuming it addresses the invisible one — offer a seat, but do not treat offering a seat as having solved the problem. Offer the rest, not the shortcut.
This framework describes a candidate initiating a cost that is borne by the patient, inside an encounter. Some stations sit outside it, and it is worth being able to say why rather than simply listing exceptions, because the two ways a station can sit outside this chapter are structurally different from each other.
Displacement. There is a cost, and it is real, but it does not land on a patient — it lands somewhere this framework does not measure. A station built around managing your own composure under provocation (the angry relative station, for instance) is spending something, but what it is spending is your composure, not a patient's currency. A station built around a colleague's standing (escalation and raising a concern) displaces the cost onto a third party and a professional relationship, not onto a person in the room with you. These stations are not exceptions to the argument; they are instances of a cost being paid by someone other than the person you are questioning, and this chapter's table does not extend to them, because its corrective — decide in advance, explain before asking — assumes the person paying is also the person you are speaking to.
Absence. Some stations have no encounter at all. Where the candidate is answering to an examiner about a scenario or a specimen, with nobody present who is spending anything, there is no cost to plan for and no signal, present or absent, to fail to notice. This applies to the A–E assessment scenario (answered to an examiner, not performed on a present patient in the sense this chapter describes) and to the anatomy prosection and radiological anatomy stations — the specimen was donated or imaged, the person is not present, and there is no encounter in which a cost is incurred.
And the anatomy area itself splits on this, which is worth stating precisely rather than leaving as a blanket exclusion. The absence applies to the prosection and radiological stations, where the specimen or image is not a person in the room. It does not apply to the surface anatomy station, where a living volunteer is examined directly and spends the same currency — dignity, with the same absent signal — that this chapter describes everywhere else. The line is between a donated or imaged specimen and a person present in the room, not between anatomy and everything else, and a reader who assumed the whole area was exempt would be wrong about a quarter of it. Indeed, the surface anatomy station is the natural comparison point for anyone wanting to see this chapter's logic applied to anatomy directly: the same currency, the same absent signal, reappears the moment a living person replaces the donated specimen.
This chapter is not a script for consent, and not a general theory of communication. It does not replace the four-item structure used when consenting a patient for a procedure, which has its own logic built for a different purpose — securing a decision, rather than managing an ongoing cost during an encounter. Nor is it a rule about kindness. A candidate who is gentle without having decided in advance what the examination actually requires is not applying this chapter; they are being nice, which is a different and less reliable thing.
Equally, this chapter is not a licence to slow every station down uniformly. The corrective for time and awkwardness is specifically to accept the pace as correct — not to manufacture delay as a performance of care. Treating every station as though it required maximal caution is itself a failure to apply the table, since the table's whole purpose is to tell you which currency is actually at stake in front of you, not to install a single tempo for every encounter regardless of what it costs.
Some of the worst answers in this exam are given by candidates who did everything they had planned to do, competently, in the right order, and then stopped — because the plan was finished, not because the problem was. This chapter is about the gap between those two things, and about the specific, recognisable sources that produce a false sense of having finished.
The failure this chapter describes does not look like a failure while it is happening. It looks like competence. That is what makes it worth a chapter of its own rather than a line in a checklist: a checklist tells you what to do, and this failure occurs precisely in candidates who have done everything the checklist told them to.
Rather than a list of situations to memorise, this chapter gives you one question to ask yourself whenever a station starts to feel finished before you have actually reached an answer:
Where is my sense of completion coming from, and what does that source require?
The question works because there is a small, recurring set of answers to it, each with its own corrective. Knowing the set means you can diagnose a new station you have never seen, rather than only recognising the specific examples given here.
Source of the false boundary
Corrective
A checklist you brought
Reorder. Ask what discriminates first, commit to a hypothesis out loud, use the rest to test it.
A scope the question set
Widen past what you were asked. The referral names the leg; the disease is everywhere else.
A framework that is correct and mandatory
Act at each step rather than after it, and reassess. The structure is right; finishing it is not the task.
A structure you announced aloud
The structure was a promise about how you would answer, not about what the answer is. Before you stop, check it against the question rather than against the plan.
Each row is worth taking in turn.
You arrive at some stations already carrying a mental checklist — a structured approach you were taught, or drilled, for exactly this kind of presentation. The checklist is not wrong. The failure is running it in the order you learned it rather than the order the situation demands, and treating its completion as the end of your job.
The corrective is to reorder rather than abandon: identify, as early as possible, whichever item would actually discriminate between your live possibilities, commit to a working hypothesis out loud, and then use the remaining items on your checklist to test that hypothesis specifically — rather than working through them in sequence regardless of whether they bear on the question in front of you. The abdominal pain history station is the clearest instance of this failure: a candidate who runs the full pain-characterisation sequence in fixed order, regardless of what the patient has just told them, has completed the structure without doing the diagnostic work it exists to enable.
Sometimes the boundary is not one you brought yourself; it is one the question appeared to set for you, and you accepted it without noticing you had a choice. A referral about a leg describes a leg. The disease that produced the referral is not confined to it — the claudication history station makes this point directly, since a candidate who characterises leg pain thoroughly but never reaches the smoking history, the chest, or the wider vascular risk picture has answered the question as set rather than the question as it actually is.
The corrective here is to widen deliberately past the literal scope of what you were asked, once you have reason to think the true problem sits outside it — not by ignoring the question, but by treating it as the entry point to a larger one rather than as the entire brief.
This is the least intuitive source, because the framework in question is not wrong, not something you invented, and not something you are free to abandon. It is the correct, taught, required structure — and following it correctly can still produce the false sense of having finished, because completing the structure and solving the problem are not the same act, and a candidate under pressure will reliably substitute the first for the second.
The clearest instance of this source is the A–E assessment of a deteriorating patient, run letter by letter through a fixed, mandatory sequence. A candidate who works through it accurately and in order has done something right, and feels finished — five letters, five answered. What they have not done is treat anything, reassess anything, or commit to anything, and no amount of care taken over the survey supplies those. This is the hardest of the four sources to work against, because nobody can reasonably be told to abandon the framework itself; the correction has to happen inside it.
The corrective is not to depart from the framework. It has three concrete parts:
Act at each step before moving on, rather than recording a finding on the way past it. A low reading is not data to carry forward; it is an action, taken immediately, after which the step is finished.
After every intervention, go back and look again. The value of an action is the change it produces, and a candidate who intervenes without saying what they would recheck has performed a gesture rather than completed a step.
State a working conclusion out loud, early, and say what would change it — not an exhaustive list of possibilities, but one or two live ones, with the finding that would separate them, offered while still acting rather than saved for the end.
This is the fourth source, and it behaves differently enough from the other three that it needs to be understood on its own terms rather than treated as a variation on them.
Two things distinguish this source from the previous three. First, most of the time the structure being announced was not actually composed by the candidate — it is very often a brought checklist, of exactly the kind described in the first row above, simply delivered as if it were bespoke to this question. Naming this source "self-imposed" would therefore be misleading; the content usually was not self-imposed at all.
What is genuinely distinct is the act of saying it out loud. Having announced — even to yourself — that you will address a question in a certain number of parts, delivering all of those parts produces a strong feeling of having kept your word, which is not the same feeling as having answered the question, but is very easily mistaken for it. Because the structure came from your own mouth moments earlier, it cannot be recognised and questioned in the way an externally imposed checklist can be; it does not feel like a scaffold you adopted, it feels like your own reasoning.
The corrective for a brought checklist — reordering — does nothing here. The problem is not sequencing. The problem is that discharging a self-announced structure has been mistaken for answering the question the structure was meant to serve.
The neoplasia and oncology station is the clearest instance: a candidate who opens by announcing a shape — the mechanism, the staging, and the treatment implications — and then delivers all three has kept a promise, and mistaken keeping it for having answered the actual question, which was about mechanism rather than about completing a structure.
The structure was a promise about how you would answer, not a promise about what the answer is. Before you stop, check it against the question rather than against the plan.
This row does not sort the same way as the other three. The first three rows above are provenances — they describe where the false boundary came from. This fourth row is not a provenance in the same sense; it is a mode — it describes how the structure comes to bind you, regardless of where the structure itself originated. The generative question at the top of this chapter still returns a usable answer for all four rows, but the four answers are not all the same kind of answer, and forcing this row into the shape of the other three — asking "where did this structure come from?" as though that were the operative question — misses the actual mechanism, which is about commitment rather than origin.
This chapter's discipline — knowing the source of your sense of completion before it misleads you — has a close counterpart in the currency chapter elsewhere in this book, which describes a candidate overspending a patient's trust without any signal that they have done so. Both chapters describe a failure that is invisible from the inside and only correctable by deciding your approach in advance, before you can be misled by how finished something feels.
Whether that resemblance reflects a genuine, general property of what this exam rewards — that premature closure and invisible overspending are really the same underlying failure wearing different clothes — or is simply two useful chapters that happen to share a family resemblance, is left as an open question for you to test against your own experience of the stations in this book, rather than settled here. Either way, the practical instruction is the same: decide your approach before you start, because neither failure announces itself while it is happening.
This chapter is not a licence for indecision. The corrective in every row above is to act — reorder and commit to a hypothesis, widen the scope and keep working, act inside the framework rather than after it, check your announced structure against the question and finish addressing it. None of the four correctives is "take longer" or "be less certain." The failure this chapter describes is premature closure, and the corrective to premature closure is not permanent openness; it is closure at the right point, driven by the actual question, rather than closure at whatever point a borrowed or self-announced structure happens to run out.
Nor is this chapter about a checklist being a bad thing. Checklists, frameworks, and even self-announced structures for organising an answer are, in general, good practice and are taught for good reason. This chapter exists because good practice, executed faithfully, is a plausible-looking substitute for actually answering the question, and the substitution is worth being able to recognise precisely because the thing substituted for it looks so much like competence rather than error.
Part I — Approaching Any Station
Every station in this book states a task in one or two sentences. Read it as a specific instruction, not a general topic. "Assess this patient's fluid status" is not an invitation to discuss fluid physiology; it asks for an assessment, of this patient, now. Confusing the topic for the task is a common source of answers that are accurate but don't actually respond to what was asked.
Several of this book's stations depend on stating your reasoning aloud rather than working silently and announcing a conclusion — see The Bounded Answer for the full method, and every knowledge-exchange station in this book (the anatomy prosection stations, the pathology stations) for worked examples of it in practice.
This is the subject of an entire chapter — The Bounded Answer — and is not repeated here. Read it before any station, since its method (state the limit, give what you have, offer a reasoned route) recurs across more stations in this book than any other single skill.
Consent has its own chapter (Consenting for any procedure) and its own dedicated stations (Part VI). It is mentioned here only to flag that its four-item structure — procedure, alternatives, risks, honest benefit — is referenced by name in stations outside Part VI as well, wherever a station touches on informing a patient about an intervention.
The following patterns were found by analysing the actual “Common Failures” section of the stations we have drafted. Each appears independently in five or more stations, across at least five different parts of the book — evidence that these are genuine recurring failure modes rather than restatements of one station's specific content.
Found in: abdominal examination, A–E assessment, arterial examination, burns assessment, knee examination, transfusion decision — six stations across three different parts.
The specific step skipped differs by station (inspection, disability, Buerger's test, safety checks), but the mechanism is the same: a systematic sequence has a step that feels optional or slow, and under time pressure it is the first thing dropped. The corrective is the same regardless of station type: treat every step in a stated sequence as mandatory, not optional, until you have positively decided — and can state why — that it doesn't apply to this specific case.
Found in: the angry relative station, breaking bad news, duty of candour, escalation, fluids and electrolytes — five stations, entirely within Communication and Applied Science.
A conversation or a management plan that is otherwise well-executed can still end without a stated action — no follow-up appointment, no monitoring parameter, no named next step. The corrective: before ending any station, ask yourself whether you have said what happens next, specifically, not just described the current situation well.
Found in: three anatomy prosection stations, the gross specimen station, and the transfusion decision station — five stations, spanning Anatomy, Pathology, and Applied Science.
This is the single most consistent pattern in the book, appearing wherever a candidate is asked to identify or state something specific. It is also the subject of this book's first chapter, The Bounded Answer, and its recurrence across five independently-drafted stations is direct evidence that the chapter's central claim — that this failure is general, not specific to any one station type — holds up against real content.
Found in: duty of candour, escalation, gross specimen, the abdominal pain history, mediastinum prosection, neoplasia principles — six stations across five parts.
Distinct from confident misidentification, this is the failure of giving an answer that is not wrong, but is too general to demonstrate the knowledge or judgement actually being tested — "there was a complication" instead of naming it; "a nerve near the artery" instead of stating the precise relation. The corrective: before answering, ask whether a more specific version of the same answer is available to you, and give that version.
Found explicitly in the abdominal pain history station, and present in weaker form in the A–E assessment and knee examination stations wherever a structure exists but isn't allowed to change the order of what follows.
This is What is making you feel finished's "checklist you brought" pattern, found directly in station content rather than only in the chapter that names it. The corrective: a structure tells you what to cover, not always in what order — let what you find redirect what you ask or do next.
These five patterns, not the specific content of any one station, are probably the most efficient thing to revise in the final days before the exam. Each one recurs across station types you might not expect to share a failure mode — a prosection station and a transfusion decision share pattern 3; a communication station and a fluid-balance station share pattern 2. Revising the pattern, rather than only the individual station content, transfers further than either alone.
Part II — Anatomy Stations