How to Structure Clinical Reasoning in a Viva

To structure clinical reasoning in a viva, use the same spoken sequence every time: frame the problem, state immediate risk, give a weighted differential, name the discriminators, then commit to a management plan. That clinical reasoning structure makes you sound safe, organised, and decisive even when the stem is unfamiliar.

A simple structure also stops two common failures. First, you drift into a fact dump. Second, you keep gathering information but never show the examiner what you think is going on.

Why structure matters to examiners

In official marking language, current spoken assessments do not reward diagnosis alone. MRCP(UK) Part 2 Clinical Examination (PACES23) consultation encounters assess clinical communication, differential diagnosis, clinical judgement, managing patients concerns, and maintaining patient welfare; the MRCGP Simulated Consultation Assessment separates data gathering and diagnosis from clinical management and medical complexity and relating to others; MRCOG Part 3 combines patient safety, information gathering, communication, and applied clinical knowledge. Examiners are listening for how you think, not just whether you can name a condition.

A scattered answer can hide good knowledge. A structured answer lets the examiner follow your prioritisation, see that you recognise risk, and hear how you would move from uncertainty to action. That is what sounds senior.

A clinical reasoning structure you can use in any viva

Use this six-part spine in almost any oral station. If you keep these headings in the same order, your viva answer structure stays stable under pressure.

• Problem representation: Give a one-line summary of the case in clinical language. Age, tempo, key features, and context.
• Immediate risk: State what is dangerous right now. Show the examiner what you would not miss.
• Weighted differential: Give your leading diagnosis first, then two or three realistic alternatives in order of probability or consequence.
• Discriminators: Name the history points, examination findings, bedside tests, or investigations that would separate those options.
• Initial plan: Say what you would do now, not after a perfect work-up. Include treatment, escalation, and monitoring where relevant.
• Review and revision: Show that you will reassess when new information arrives and change course if the picture shifts.

In practice, it can sound like this:

• This is a 67-year-old with sudden pleuritic chest pain and hypoxia, so my main concern is an acute cardiorespiratory cause.
• My immediate priority is to identify instability and exclude life-threatening diagnoses.
• My leading diagnosis is pulmonary embolism, but I would also consider pneumonia and acute coronary syndrome.
• The key discriminators are haemodynamic status, ECG, chest radiograph, blood gas, troponin, and imaging for venous thromboembolism if clinically appropriate.
• While clarifying the diagnosis, I would give oxygen if needed, monitor closely, and escalate senior help early if unstable.
• I would then review the results and re-rank the differential.

If you freeze, go to the next heading. Do not restart from the beginning.

A short rescue script

For a short case or when the examiner interrupts early, one line per heading is enough:

• Most likely diagnosis.
• Immediate risk not to miss.
• Two alternatives.
• Key discriminator.
• Initial management.
• Review point.

That is often enough to recover control of the station.

What good candidates sound like

Strong candidates do three things aloud. They rank. They justify. They commit.

They do not give a shopping list of ten causes with no order. They say which diagnosis is most likely, why it fits, what is dangerous, and what evidence would change their mind. That is what turns knowledge into reasoning.

Worked example: acute medicine

A 67-year-old with central chest pain, diaphoresis, and nausea arrives in the emergency department.

A structured answer might sound like this:

This is most likely an acute coronary syndrome until proven otherwise. My immediate concern is haemodynamic instability or an arrhythmia, and I also need to exclude other life-threatening causes of chest pain such as aortic dissection and pulmonary embolism. The features supporting acute coronary syndrome are the central heavy pain, autonomic symptoms, and age. The next key discriminators are vital signs, ECG, serial troponins, and examination for heart failure or alternative pathology. While investigating, I would start time-critical acute coronary syndrome management according to the clinical picture, obtain senior support early if unstable, and reassess after the first ECG and blood results. If the history or examination suggested tearing pain, pulse deficit, or focal neurology, I would immediately re-rank dissection higher.

Notice what that answer does. It gives a lead diagnosis, shows danger awareness, explains why, links investigations to decisions, and leaves room to change course.

Adapt the same structure to different station types

In a consultation station, the skeleton stays the same but the language changes. You still move from working diagnosis to management, but you translate it into plain English, share risk, and negotiate a plan. The RCGP SCA framework and consultation overview make that progression explicit, linking data gathering and diagnosis to safe patient-centred management, follow-up, and safety-netting.

In a structured discussion with an examiner, be explicit about time frames: what I would do now, in the next hour, and after initial results. In exams such as MRCOG Part 3, applied clinical knowledge is assessed alongside patient safety, information gathering, and communication, so even a technically correct answer must still sound safe and usable.

In an ethics or communication stem, use the same order. Define the problem, identify the immediate risk, state the relevant principles, outline options, make a recommendation, and explain when you would escalate.

How to handle uncertainty without sounding vague

Examiners do not expect certainty from limited information. They do expect you to manage uncertainty safely.

Useful phrases include:

• At this stage, my leading diagnosis is...
• The immediate alternative I must not miss is...
• What would most change my management is...
• Pending that information, I would...
• If the patient deteriorates or this result is abnormal, I would re-rank the differential and escalate...

That language sounds mature because it separates probability from urgency.

Take a surgical example: a patient becomes tachycardic a few hours after an operation. A weak answer lists pain, sepsis, bleeding, pulmonary embolism, dehydration, and anxiety. A better answer says haemorrhage is the immediate priority because it is time-critical and dangerous, but I would also consider sepsis, pulmonary embolism, uncontrolled pain, and hypovolaemia from other causes. My next discriminators are observations, wound and drain review, blood loss history, bedside examination, ECG, and urgent bloods, and I would start resuscitation while clarifying the cause.

That is the difference between knowing possibilities and showing reasoning.

Common mistakes

Official domains across current spoken assessments repeatedly separate data gathering, clinical judgement or management, patient safety, and the handling of patient concerns. Most weak answers fail because one of those elements disappears when pressure rises.

• Giving an unranked list of differentials instead of a weighted differential.
• Naming investigations without saying what question each test is answering.
• Jumping straight to definitive treatment without first stating immediate risk or instability.
• Sounding certain when the evidence is thin, then collapsing when challenged.
• Hiding behind vague phrases such as I would do some bloods and monitor.
• Forgetting the patient perspective, including concerns, explanation, consent, and safety-netting.
• Describing an ideal textbook pathway rather than what you would do first in the real station.
• Speaking for too long before making a diagnostic commitment.

A good self-check is simple: have I said what I think, why I think it, what else matters, and what I would do now?

Practice workflow for mock viva sessions

Use a tight rehearsal loop rather than long passive revision blocks. You are training a speaking skill.

• Pick one short stem.
• Take 30 to 60 seconds to jot the six headings.
• Answer aloud for 60 to 90 seconds.
• Let your partner interrupt with probing questions.
• Debrief immediately.
• Repeat with a new case.

Three cases done properly are worth more than twenty cases skimmed in your head.

Debrief questions that actually help

• Did I state a leading diagnosis early?
• Did I identify the immediate danger?
• Did I link investigations to decisions?
• Did I give an initial management plan with review or escalation?

If you are practising alone, record yourself. If you are using AI rehearsal, ask it to interrupt, challenge your prioritisation, and force you to justify why one diagnosis is above another. The goal is not longer answers. It is cleaner thinking.

Summary

• Use the same six-step clinical reasoning structure every time.
• Rank your differential instead of listing everything you know.
• Link each investigation to the uncertainty it resolves.
• Commit to an initial plan early, including escalation and review.
• Practise aloud until the structure survives stress and interruption.

References

• https://www.mrcpuk.org/sites/default/files/documents/PACES23%20Consultation%20scenario%20writing%20guidance.pdf
• https://www.rcgp.org.uk/mrcgp-exams/simulated-consultation-assessment/marking-and-results
• https://www-preview.rcog.org.uk/careers-and-training/exams/mrcog-our-specialty-training-exam/mrcog-part-3/mrcog-part-3-format/