A useful OSCE partner is someone with whom you can agree a specific task, practise consistently and examine mistakes without turning feedback into a contest. Matching examination dates is helpful, but it does not resolve differences in case preparation, role-play or expectations. Agree those details before the first timed encounter.
The agreement below is an original study aid, not an official examination document. It can be adapted for encounter-based preparation, but SCA, PLAB 2, CPSA, CASC, SOO, NAC, MRCOG Part 3, AMC Clinical and other assessments should not be treated as identical formats.
Start with the task you both intend to practise
Write the examination, component and type of encounter. Then agree what this particular session is for. It might be opening a consultation, exploring a concern, explaining a plan or recovering after a misunderstanding. "Do some OSCEs" is too vague to evaluate afterwards.
Use the current examining body's material to set any formal timing or task conditions. For example, the GMC's PLAB 2 guide, checked on 19 September 2026, provides its own assessment context. Do not transfer that context automatically to a different examination because both involve simulated encounters.
Decide whether this is a learning attempt, which permits pauses, or an uninterrupted rehearsal. Neither is inherently superior. They answer different questions: what needs clarification during learning, and what can you currently perform without help?
Agree roles before agreeing a score
The candidate needs a task sheet. The patient or colleague role needs information about what they know, what they are concerned about and what they should reveal only when appropriately asked. An observer needs a way to record what happened.
With only two people, observation is limited because the person acting also has to attend to the interaction. A recording may help when both participants explicitly agree and storage and deletion arrangements are clear. Recording should never be assumed or used to share someone else's performance without permission.
Do not invent an official pass mark for a home-written station. A useful early session may have no numerical score at all. It can identify one behaviour that helped and one omission worth addressing.
A reusable practice-partner agreement
| Item to agree | Write your decision before starting |
|---|---|
| Assessment and task | Examination, component and purpose of this encounter |
| Timing | Learning mode or uninterrupted mode, with the source of any exam timing |
| Roles | Candidate, patient or colleague, observer where available |
| Case provenance | Original material or resource used with appropriate permission |
| Preparation | Who supplies each role sheet and when |
| Feedback | Specific observations before interpretations or scores |
| Disagreement | Check the task and source, then mark unresolved points for review |
| Recording | Consent, access, storage and deletion, or no recording |
| Scheduling | Availability, cancellation notice and rescheduling approach |
| Follow-up | One agreed learning task and the purpose of the next session |
This agreement should be short enough to reuse. It is not a contract, a professional assessment or a way to authorise practice beyond the participants' competence.
Leave room to change the arrangement after the first session. A partner may be reliable and thoughtful but available at the wrong times. Another may be a useful source of feedback while preparing for a different examination. Name those limits rather than assuming the partnership must cover every learning need.
A complete original starter exercise
For the candidate: "You are speaking to a fictional patient after a previous explanation left them uncertain. Establish what they understood, clarify their main concern and summarise the next question that needs answering. Do not invent clinical findings or prescribe a treatment."
For the patient role: you heard that an investigation was being considered and assumed this meant a serious diagnosis had been confirmed. You are embarrassed to ask because the earlier clinician seemed busy. If the candidate asks an open question about your understanding, explain the assumption. If they immediately repeat a long explanation, say you are still unsure rather than inventing new symptoms.
For the observer: record whether the candidate asked what the patient understood, discovered the assumption, distinguished a proposed investigation from an established diagnosis and checked the revised understanding. Do not mark down a candidate merely for using a different reasonable phrase.
The exercise is deliberately about clarification. It is not a substitute for an examination-specific case or a clinical management protocol. Its advantage is that both participants know what successful clarification would look like.
Give feedback in the order the evidence permits
Start with the candidate's account of the difficulty. Then report a specific observation. For example: "You explained the investigation before asking what the patient thought it meant." Only then discuss the interpretation and a possible alternative approach.
Avoid unsupported conclusions such as "You have poor empathy". A short fictional encounter may reveal a missed opportunity; it does not establish a fixed personal trait. Likewise, a confident actor should not use their own preferred consultation style as an undisclosed marking standard.
When the patient role improvised a fact that materially changed the case, acknowledge it. The candidate should not be judged against a hidden answer based on a scenario they were no longer given.
Resolve disagreement without rewriting the encounter
Return to the actual instruction and the recorded behaviour. Was the disputed point part of the task? Was the relevant information available? Is the criticism about a clinical omission, a communication preference or a factual misunderstanding?
Check the appropriate current source for a clinical dispute. When neither partner can resolve it confidently, retain the uncertainty and take it to a suitable supervisor or teacher. Agreement between two learners does not independently validate the answer.
For the next session, change the task in a way that tests the proposed improvement. In the starter exercise, give the patient a different misunderstanding. A candidate who simply memorised yesterday's clarifying sentence should have to listen again.
Use solo practice between shared sessions
This article is published by iatroX and includes its simulations as one option alongside human practice. Under the September 2026 product description, relevant tracks offer voice and text encounters, coached and uninterrupted modes, transcript-linked feedback and Tutor-led follow-up. They do not make peer observation or specialist teaching unnecessary.
Use solo rehearsal for repetition when schedules do not align. Bring a specific uncertainty to the next partner session rather than asking your partner to endorse an AI score. No automatic sharing, partner-matching service or multi-person simulation integration is claimed here.
The published UK subscription in September 2026 is £99 paid upfront for a year, equivalent to £8.25 a month billed annually, or £29 monthly, with simulations included alongside question banks, Tutor, planning and CPD tools. The value is repeated relevant practice, not access to unrelated examinations.
For useful human feedback, keep the partner. For technical physical skills, arrange supervised practice. For additional supported solo encounters, inspect an available simulation. A well-organised combination is more useful than expecting one partner or product to do everything.
Frequently asked questions
Must my OSCE partner be preparing for exactly the same examination?
Not necessarily, but shared examination context reduces the amount you must explain and check. Agree the actual task and use the correct official requirements rather than mixing formats silently.
Should we assign a pass or fail to every practice encounter?
Only use a formal judgement when you have an appropriate basis for it. Specific observations and a next practice goal are often more defensible than a made-up pass threshold.
Can iatroX replace a practice partner completely?
Its September 2026 simulation description supports useful solo rehearsal, but that does not reproduce every benefit of human observation or supervised practice. Use each format for the task it can actually support.