Annotated OET role-play: Band A vs Band C

If you have read our post on how OET Speaking is actually scored, you have the rubric. The next step that most candidates struggle with is applying it to their own speaking. "Empathy" and "patient perspective" sound clear in the abstract; in a 5-minute role-play with a clock running, the line between Band A and Band C is much harder to see.

This post is the rubric applied. Two full role-plays — one Nursing, one Medicine — at the realistic exam length. For each, the same candidate card is answered two ways: the way a Band A candidate would handle it, and the way a Band C candidate would. After every candidate turn, we add an examiner-style note explaining which criterion that turn is scoring on, and what the difference is.

The point of reading these is not to memorise lines. The role-play patient in your exam will be different. The point is to internalise the difference so when you are in front of the examiner, the right instinct fires automatically.

The role-plays below are realistic but synthetic. Real OET role-cards are confidential. The format, length, task structure, and criteria are accurate to the live exam — only the specific clinical content has been adapted.

Role-play 1: Nursing — acute lower back pain

The candidate card

You are a registered nurse in a primary-care clinic. The patient, Ms Hannah Lee, age 29, is an administrator who comes in with new-onset lower back pain after lifting a heavy box at work yesterday afternoon. She is anxious about losing income if she can't work.

Your tasks:

1. Take a focused history of the pain (onset, location, severity, red flags).
2. Reassure her that acute non-specific lower back pain is common and typically self-limiting.
3. Explain why imaging (X-ray, MRI) is not indicated at this stage.
4. Give clear self-management advice (movement, pain relief, posture).
5. Address her concern about returning to work.

You have 5 minutes for the role-play. The roleplayer will respond as the patient would.

Band C version

What this version sounds like: clinically competent but emotionally flat, with weak organisation and patient-perspective issues.

Candidate: "Good morning, Ms Lee. I'm the nurse. So, you have back pain? When did it start?"

Examiner note: opening is functional but cold. Drops patient perspective and appropriateness marks at the first opportunity — there is no acknowledgement that this is a person walking into a clinic anxious, just an immediate jump to information gathering.

Patient: "Yesterday afternoon. I was lifting a box at work and felt a sharp pain in my lower back. It's been bad ever since."

Candidate: "Okay. Is the pain on one side or both? Does it go down your leg?"

Examiner note: efficient clinical history-taking — this turn does fine on clinical accuracy of information. But the second sentence is two questions stacked together, which makes it hard for the patient to answer fully. Organisation of talk takes a small hit.

Patient: "It's mostly in the middle. Sometimes it goes down to my right thigh but not below the knee."

Candidate: "Any numbness, tingling, weakness? Any trouble with bladder or bowels?"

Examiner note: covers the red-flag checks, scores on clinical accuracy. But the candidate has now asked four screening questions in a row with zero conversational warmth. The patient has no idea why these questions matter, or what the candidate is concluding from the answers.

Patient: "No, none of those. Just the pain. It's a 7 out of 10 right now."

Candidate: "Right. So based on what you've said, this is acute non-specific lower back pain. It's very common and it usually resolves on its own within two to four weeks. You don't need an X-ray or an MRI."

Examiner note: this is where the Band C/D split gets sharp. The candidate has gone from history-taking straight to diagnosis-and-plan with no signposting and no check-in. The patient has not been told why an MRI isn't indicated — they have been told it isn't. Information giving is technically present but uninterpreted. Patient perspective scores low because the candidate has not paused to ask what she is worried about.

Patient: "Oh, okay. But shouldn't we check, just to be safe?"

Candidate: "No, imaging in the first episode of back pain like this doesn't change what we do. The scans are normal in most cases. You'd just be exposed to radiation for no benefit."

Examiner note: clinically correct, but the wording — "doesn't change what we do" — sounds dismissive when paired with the earlier flat tone. Appropriateness of language drops. The same content delivered with even one acknowledgement phrase ("I understand wanting to be sure — let me explain why we don't usually scan at this stage") would have scored a full band higher.

Patient: "Okay. So what should I do?"

Candidate: "Take paracetamol or ibuprofen as needed. Keep moving — bed rest actually slows recovery. Use a heat pack if it helps. Try to return to work in a couple of days, no heavy lifting. Come back if you get any numbness, tingling, or bladder problems."

Examiner note: covers tasks 4 and 5 in one breath, which is organisation failure number two — multiple distinct pieces of advice are flattened into a single list with no structure. Vocabulary is plain, which is good. But there is no check for understanding, and the candidate has still not addressed the patient's stated concern about losing income.

Patient: "But I was worried about losing income. I can't afford to be off work."

Candidate: "You should be able to go back in a couple of days. Just no heavy lifting."

Examiner note: the patient explicitly raised her main concern, and the candidate gave a one-sentence answer with no empathy. This is the biggest single mark loss in the role-play. Empathy and patient perspective both drop a band.

Result: a Band C candidate has covered the clinical content but has missed the conversational structure that the OET is actually testing.

Band A version

Same candidate card. Same patient. Different criteria scores.

Candidate: "Good morning, Ms Lee, I'm so sorry to hear you're going through this. Before we look at the back pain itself, can I just check — what's worrying you most about this right now?"

Examiner note: 28 seconds in, and the candidate has already scored marks on patient perspective (asked open-endedly what worries the patient), empathy (acknowledged the situation), and appropriateness of language (warm but professional opener). This is the opening that separates A from C.

Patient: "Honestly, it's the work side. I'm an administrator — I'm on my feet a lot, and if I'm off work I lose income. Also, I'm scared this might be something serious."

Candidate: "Right — so two real worries: the impact on your work, and whether something serious is going on. Both of those make sense. Let me work through this with you so by the end of today you have a clear sense of both. First, can you walk me through what happened — when the pain started, and what you've noticed since?"

Examiner note: patient perspective scoring on naming both concerns back to the patient (a textbook reflective-listening move) and organisation of talk scoring on signposting the structure ("by the end of today you have a clear sense of both"). Note that the actual clinical work — taking the history — is now framed as for her, not as a checklist the candidate has to complete.

Patient: "I was lifting a box at work yesterday afternoon and felt a sharp pain in my lower back. It's been bad ever since. About a 7 out of 10."

Candidate: "Sharp pain, lifting a box yesterday, 7 out of 10. Can you tell me where exactly it is? Does it stay in the lower back, or does it travel anywhere?"

Examiner note: the brief repeat-back ("Sharp pain, lifting a box yesterday, 7 out of 10") is doing important work — it scores on patient perspective (showing she's been heard), and it gives the candidate a moment to consolidate the history. Then a single, well-formed question — not three stacked into one.

Patient: "Mostly the middle of the lower back. Sometimes goes down into my right thigh, but stops above the knee."

Candidate: "Okay — that's helpful. I'm going to ask a few more questions, mostly to rule out anything that would change what we do next. Have you had any numbness or tingling in your legs or feet, any weakness, or any change in your bladder or bowels?"

Examiner note: this is the same red-flag screen as the Band C version, but two things are different. First, the candidate signposts the why ("to rule out anything that would change what we do next") — that's a big clinical accuracy of information marker because it teaches the patient how clinical reasoning works. Second, the questions are clustered cleanly with conjunctions, not stacked.

Patient: "No to all of those. Just the pain."

Candidate: "Good — that's reassuring. So from what you've described, this looks like what we call acute non-specific lower back pain. That's the most common kind of back pain we see — it usually comes from a muscle or soft-tissue strain, like lifting a box that was heavier than it should have been. The good news is the body heals this kind of injury well on its own. Most people are significantly better within two to four weeks."

Examiner note: this turn covers task 2 (reassurance) cleanly. Notice how the explanation is layered — first the medical name, then plain-English translation, then the prognosis. Vocabulary scores on the controlled use of medical terminology with plain-English follow-up ("acute non-specific lower back pain… most common kind of back pain we see"). Information giving scores on framing the prognosis with a specific timeline.

Patient: "Oh — okay. But shouldn't we get a scan or X-ray, just to be safe?"

Candidate: "That's a fair question and a lot of people ask it. I'll explain. For a first episode of back pain like this, without any of those red-flag signs we just checked for, scans actually don't help us — they don't change what we'd do. In fact, they often pick up small changes that are unrelated to the pain and that worry people unnecessarily. If something does change — if you develop numbness, weakness, or bladder issues — we'd absolutely scan then. But not at this stage."

Examiner note: this is the turn that decides task 3 (explaining why imaging isn't indicated). Compare to the Band C version: "scans are normal in most cases" vs. the Band A "scans actually don't help us — they don't change what we'd do… they often pick up small changes that are unrelated and worry people unnecessarily". Same fact, but the Band A version explains the clinical reasoning. Information giving scores higher because the candidate is teaching the patient, not just informing them. The conditional at the end — "if something does change, we'd absolutely scan then" — also scores on empathy because it shows the candidate respects her instinct to want to be sure.

Patient: "That makes sense. So what should I actually do?"

Candidate: "Let me put this in a clear plan. Three parts: pain management, movement, and work. For pain, simple painkillers — paracetamol around the clock for a few days, ibuprofen if your stomach is okay with it. A hot pack on the lower back can help too. For movement, this surprises a lot of people: bed rest actually slows recovery. Keep moving as much as the pain allows. Gentle walking is good. For work — and I want to come back to this because I know it's your main worry — you can return to your administrator role within a couple of days, as long as you avoid heavy lifting. Ask a colleague to take the boxes for the next two weeks, and set up your workstation so you're not seated for long stretches."

Examiner note: a Band A signal. The candidate explicitly structures the answer ("three parts: pain management, movement, and work"), covers tasks 4 and 5 in clearly delineated sections, and — critically — calls back to the patient's stated concern about work ("I want to come back to this because I know it's your main worry"). That single sentence is doing the work of two criteria at once: organisation of talk (signposting the return to a previous topic) and patient perspective (showing the concern was heard and is being addressed deliberately).

Patient: "Okay. So back to work, no heavy lifting. That feels doable."

Candidate: "Good. Before we finish, let me just check — does that all make sense, and is there anything you'd like me to go over again?"

Examiner note: classic organisation closing move (check for understanding) and patient perspective marker (inviting questions). Even if she says no, the candidate has scored those marks.

Patient: "It all makes sense, thank you."

Candidate: "One last thing — if any of these happen, I want you to come back straight away rather than waiting: new numbness or weakness in the legs, any problem with bladder or bowel control, or fever. Otherwise I'll see you in two weeks for a follow-up. Does that sound okay?"

Examiner note: safety-netting at the close. Clinical accuracy of information scores on naming the specific red flags. Empathy scores on the calm tone of "I want you to come back straight away" — not alarmist, but clearly serious.

Result: same case, same clinical facts, ~50 extra seconds of speech total. The candidate has covered all 5 tasks with structure, addressed the patient's stated concern explicitly, and shown the examiner clear instances of each scoring criterion.

The five patterns that separated A from C above

If you read both transcripts back-to-back and ask "what specifically did the Band A do that the Band C didn't?", five concrete moves account for almost the entire difference:

1. Opened with an empathy-and-perspective question, not a clinical question. The Band A version's first move was "what's worrying you most?" The Band C version's first move was "when did it start?"
2. Signposted structure. The Band A version named the structure of the conversation ("two real worries: work and whether something serious is going on") and named the structure of the answer ("three parts: pain management, movement, and work"). The Band C version did neither.
3. Explained reasoning, not just facts. When asked about imaging, the Band A version explained why scans don't help here. The Band C version stated they don't help.
4. Called back to stated concerns. The Band A version explicitly returned to the work concern at the end ("I want to come back to this because I know it's your main worry"). The Band C version answered the work concern with a single dismissive sentence.
5. Closed with a check. The Band A version invited questions before ending and added safety-netting. The Band C version stopped speaking when it ran out of things to say.

Notice that none of these five moves require deeper clinical knowledge. They are all conversational structure. That is the OET Speaking exam in a sentence: it is a test of structured clinical communication, not of clinical knowledge.

Role-play 2: Medicine — newly diagnosed hypertension

The candidate card

You are a doctor in a primary-care clinic. The patient, Mr Daniel Okafor, age 52, is a project manager. His average blood pressure across three recent readings was 158/96. He has no symptoms, no other medical conditions, and a moderate family history of cardiovascular disease. He is sceptical of taking long-term medication.

Your tasks:

1. Explain the diagnosis of hypertension in plain English.
2. Explain why it matters even though he has no symptoms.
3. Discuss lifestyle modifications (diet, exercise, salt, alcohol).
4. Discuss when medication is recommended.
5. Address his concern about long-term medication.

Band C version (compressed for length — same pattern as above)

Candidate: "Mr Okafor, your blood pressure readings have been high. Your average is 158/96, which is stage 2 hypertension. We need to address this."

Examiner note: opens with a label, no context, no acknowledgement. Patient perspective misses immediately because the candidate hasn't checked what Mr Okafor already knows or thinks.

Patient: "But I feel completely fine. Are you sure?"

Candidate: "Yes. Hypertension often has no symptoms — that's why it's called the 'silent killer'. If left untreated, it increases your risk of heart attack, stroke, kidney disease, and heart failure."

Examiner note: clinical accuracy is fine but the phrasing "silent killer" is jarring and the list of consequences is too dense to land. Appropriateness of language drops, and the patient has now heard four scary words in two sentences without any reassurance.

Patient: "Okay. So what now?"

Candidate: "Lifestyle changes first. Reduce salt, no more than 5 grams per day. Exercise 30 minutes most days. Reduce alcohol to no more than two units per day. If your blood pressure is still elevated after three months, we'll start medication. The first-line agent for someone your age is usually a calcium channel blocker like amlodipine."

Examiner note: covers tasks 3 and 4 in one breath. Organisation of talk failure: the lifestyle advice should have its own section, with the patient's input invited. The casual mention of medication doesn't address his stated reluctance to take long-term drugs.

Patient: "I really don't want to be on medication for the rest of my life."

Candidate: "I understand, but if your lifestyle changes don't bring it down, medication is necessary. Most people tolerate amlodipine well."

Examiner note: claims to understand without showing understanding. Empathy drops because there's no exploration of why he doesn't want medication. He might be worried about side effects, dependence, cost, or his sense of identity as a healthy person — the candidate has assumed.

Band A version (compressed — same pattern as Nursing example)

Candidate: "Mr Okafor, thanks for coming in. Before I share what I'm seeing in your numbers, can I ask — what's your sense of what we're looking at today? And is there anything in particular you're worried about?"

Examiner note: textbook open. Patient perspective scoring on inviting his understanding first. Empathy scoring on the warmth of "before I share what I'm seeing".

Patient: "I've been told my blood pressure is high. I feel fine though. I'm a bit worried I'll be put on medication for the rest of my life, honestly — my father was on five different pills by the time he was 60 and I really don't want that."

Candidate: "That's really helpful context — thank you. Two things I'm hearing: you feel fine, which makes the diagnosis feel almost theoretical; and you've watched a family member end up on a long list of medications and you don't want to go down that path. Both of those are completely understandable. Let me explain what we're actually looking at, and then we'll talk about what your options are — including the lifestyle-first approach, which I'd like to try first if you're willing."

Examiner note: this single turn does the work of three criteria. Patient perspective: named both concerns back. Empathy: validated both as reasonable. Organisation of talk: announced the structure of what's coming and previewed the candidate's recommendation (lifestyle first), which directly addresses his medication concern before he has to ask again.

Patient: "Okay. Tell me what's going on."

Candidate: "Your blood pressure average across three readings is 158 over 96. To translate: the top number is the pressure when the heart is pumping, the bottom is when it's resting between beats. Normal is below 130 over 80. Yours is above 140 over 90 consistently, which we call hypertension. It's not an emergency, but it's a real diagnosis. Now, you said you feel fine — and that's the tricky thing about high blood pressure. It very rarely causes symptoms. The reason we treat it isn't to make you feel better today; it's to protect your heart, brain, and kidneys over the next 10, 20, 30 years. With a number like yours, untreated, the average person's risk of a heart attack or stroke in the next decade is meaningfully higher than someone with normal blood pressure — and the risk goes back down to baseline if we control it."

Examiner note: high-quality clinical translation. Information giving scores heavily because the candidate explains both the numbers and why they matter. Note the phrase "untreated, the average person's risk… is meaningfully higher" — careful not to give a specific number that might be wrong, but quantitative enough to land. Vocabulary scores on the controlled use of "hypertension" with plain-English follow-up.

Patient: "So I have to take medication."

Candidate: "Not necessarily, and certainly not today. Here's where I'd like to spend most of our time: lifestyle changes. For someone like you — middle-aged, no other conditions, blood pressure in the stage-1-to-2 range — lifestyle can bring the numbers down significantly. If we get a real change there, we may not need medication at all, or we may need a much lower dose if we eventually do. Specifically: reducing salt to under about a teaspoon a day across all your food — and most of that hides in bread, sauces, and restaurant meals, not the salt shaker — is probably the single highest-leverage change. After that, 30 minutes of moderate exercise — walking briskly counts — five days a week. And bringing alcohol down to no more than two units most days, with two alcohol-free days a week. Those three together, in someone your age, can drop systolic pressure by 10 to 20 points."

Examiner note: structured into three named interventions, each with a specific number and a practical note. The "hidden salt" detail isn't strictly necessary clinically — but it scores on appropriateness because it shows the candidate understands the patient is going to leave the room and try to actually do this, not just hear the advice. Clinical accuracy of information is high because the candidate gives a realistic effect size ("10 to 20 points").

Patient: "Okay. And if that doesn't work?"

Candidate: "Fair question. The deal I'd propose is this: three months of focused lifestyle work, with a check-in halfway through. If at three months your average pressure is still above 140 over 90 — even though you've made real changes — then we'd talk about adding a single medication. The newer first-line agents are well-tolerated and they're typically one pill, not five. They're not chemotherapy — they're more like a tax on your morning routine. But I want to be honest with you: if we end up there, it's because lifestyle alone wasn't enough, not because anyone's giving up. About a third of people in your situation can manage with lifestyle alone. A third end up on one pill. A third end up on two. Knowing which category you're in takes a few months to figure out."

Examiner note: the candidate explicitly addresses his stated concern about ending up like his father ("one pill, not five… they're not chemotherapy — they're more like a tax on your morning routine"). Empathy and patient perspective are both scoring heavily here. The probability distribution at the end ("about a third…") is a sophisticated move — it scores on clinical accuracy of information because it gives the patient a realistic sense of his odds without false certainty.

Patient: "That feels much more manageable. Let me try the lifestyle stuff."

Candidate: "Good — I think that's the right call for you. Before you go, let me put down the three things in writing for you, and let's book a check-in for six weeks from now to see how it's going. Does anything I've said feel unclear or worth going over again?"

Examiner note: closes the loop. Check for understanding, concrete next step, follow-up date.

The same five patterns, again

Read the medicine role-play back-to-back. The five moves that separated A from C in the nursing case are exactly the same here:

1. Open with an empathy-and-perspective question. ("Before I share my numbers, what's your sense of what we're looking at?")
2. Signpost structure. ("Two things I'm hearing… let me explain what we're actually looking at, and then we'll talk about your options.")
3. Explain reasoning, not just facts. ("The reason we treat it isn't to make you feel better today; it's to protect your heart over 20 years.")
4. Call back to stated concerns. ("They're typically one pill, not five — they're not chemotherapy.")
5. Close with a check. ("Does anything I've said feel unclear or worth going over again?")

If you do nothing else in your preparation, drill these five moves into automaticity. They are the highest-leverage practice you can do.

How to use these transcripts in your own practice

Two specific drills:

Drill 1 — Shadow the Band A. Read the Band A transcript out loud, slowly, with intonation. Then re-read it without looking, paraphrasing if you need to. You're not memorising; you're letting the rhythm of the structured turns seep in. Do this twice a day for a week.

Drill 2 — Diagnose your own gap. Record yourself doing the same role-play (use the candidate cards above). Play it back. For each of the five A-vs-C patterns, check: did you do it, partially do it, or skip it? Whichever pattern you skip most often is your single highest-leverage area to drill.

If you want a feedback loop tighter than self-scoring, run the recording through any rubric-aligned scoring tool — we built one at OET Live, but the practice is what matters, not the tool. The recordings only become useful if you compare them against the Band A pattern explicitly. Vague self-assessment ("that was alright I think") is the most common reason candidates plateau.

---

These two transcripts are not magic templates. The next exam will give you a different candidate card, a different patient, possibly a profession outside nursing and medicine. What's transferable across all of those is the structure: the five moves that separated A from C above will separate A from C in any clinical role-play you sit. Practise the structure, not the words.

For more on what each criterion actually rewards, see how OET Speaking is actually scored. For where these patterns most commonly break down under exam stress, see top OET Speaking mistakes and how to fix them. For a structured 8-week plan that ends in being able to produce the Band A version reliably, see the realistic 8-week OET Speaking roadmap.