---
title: "Training a simulated patient for one case: script, standardization, the feedback role and SP ratings"
description: "How to train a simulated patient for one OSCE case: read-through, clinician rehearsal, the unscripted-question rule, video check, and the rating role."
canonical: https://studydrome.com/docs/osce/people/training-an-sp-for-a-case/
updated: 2026-10-03
---

# Training a simulated patient for one case: script, standardization, the feedback role and SP ratings

How to train a simulated patient for one OSCE case: read-through, clinician rehearsal, the unscripted-question rule, video check, and the rating role.

Training a simulated patient for a single case consists of five steps. First, a read-through of the script. Then a rehearsal with a clinician playing the candidate. Then a rule for every question the script does not cover. Then a recorded run checked against the script. Last, a drift check on the day. If the simulated patient also gives feedback or completes a rating, that is trained separately. The [program](/docs/osce/people/simulated-patient-program/) sets the standard; this page trains one person for one case.

## Why one case at a time

A simulated patient is not typically trained. They are trained to play this case, for this station, for this cohort, the same way every time. The script is the [simulated patient part of the station template](/docs/osce/design/writing-a-station/). It holds identity, background, presenting complaint, history, and concerns. It holds what to say and not say, the questions to ask, and the standardized answers. Training turns that document into a performance that survives forty repetitions.

The AMEE guide on OSCE organization cites an estimate of up to 15 hours to adequately train a simulated patient. Not every case needs that. A short history with no signs takes less. A role with distress, physical findings, or a critical item that depends on an answer takes more. The trainer plans the hours from the script.

## The five steps

| Step | What happens | Output |
| --- | --- | --- |
| 1. Read-through | The simulated patient reads the script with the trainer; questions about the person, the illness, and the words are answered, and the answers written into the script | An amended script; a list of gaps |
| 2. Clinician rehearsal | A clinician plays the candidate: a good one, a poor one, and a borderline one; the trainer watches for answers that vary | Timing checked; hard questions found |
| 3. The unscripted-question rule | The script's "if asked anything else" section is agreed line by line: what the patient does not know, what they deflect, what they invent never | A closed script |
| 4. Recorded run and check | A full run is recorded and a trainer or second clinician scores the portrayal against the script, item by item, including manner and expression | A signed-off portrayal, or a second rehearsal |
| 5. Drift check on the day | The trainer watches part of a station early and again late in the session, against the same items | A note in the incident log if the role has moved |

Step 4 is the one the evidence insists on. A study recorded four stations of a high-stakes OSCE, with four simulated patients per case. The words were consistent across most items. The facial expressions were significantly different. An emergency station that depended on physical presentation differed across all four. Words are trained by reading. Expression and physical presentation are trained by watching a recording of yourself. Then someone tells you what to change.

## What does the rehearsal test?

Three things the read-through cannot.

- **Timing.** Does the history the script gives fit the station's minutes when a competent candidate asks for it in order? If the script is too rich, the candidate never reaches the part the marking scheme rewards.
- **The critical line.** Some critical items depend on an answer, such as allergies or anticoagulants. The rehearsal checks that the answer comes only when asked, and comes the same way every time. The [critical-error policy](/docs/osce/design/critical-errors-and-must-pass/) depends on that line.
- **The poor candidate.** A simulated patient trained only on good candidates does not know what to do when no one asks the right question. The rehearsal with a poor candidate teaches them to wait and not to help.

## What are the limits of the role?

They are written into the script and repeated in training. Which examinations are permitted and how far they go. What the simulated patient will not say or do. How to stop a station. The [program's](/docs/osce/people/simulated-patient-program/) safe work environment domain sets the rule; the case training applies it to this role. A simulated patient who plays a distressing role is trained in how to exit it. A debrief is scheduled, not offered.

## When simulated patients give feedback

In formative OSCEs, the simulated patient may provide the candidate with feedback from the patient's perspective. It is a separate skill, trained separately. The training covers what to comment on: how the candidate made the patient feel, what was clear and what was not. It covers what not to comment on: clinical correctness and the marks. It covers how long, and in what words. The ASPE standards of best practice list training for feedback as its own element for that reason. A simulated patient who has not been trained for feedback does not give it.

## When simulated patients complete a rating

Some exams ask the simulated patient to rate the candidate's communication on a short scale. The rating is a separate instrument with its own training, and the evidence says it measures something different from the examiner's judgment. A 2023 study found that simulated patients were rated higher than trained observers on 20 of the 21 items in which the two differed. The gap was widest on empathy and nonverbal items. A 2006 study found that simulated patient examiners scored students higher than physicians did, with a weak correlation between the two groups. The AMEE guide reports that simulated patient scores agree with physicians' checklists but not with global scoring.

So the training for the rating covers the anchors, the same way examiner training does. The exam reports the rating on its own. It is never added into the examiner's global rating, and the [quality section](/docs/osce/quality/) reads it as its own metric.

## Limits

The portrayal-accuracy evidence comes from a single high-stakes program involving experienced simulated patients. The rating studies are single-site. The training steps are the practice the AMEE guide and the ASPE standards describe. They are not a tested protocol. The recorded run is the check that tells a school whether its own training worked.

## Frequently asked questions

### How do you train a simulated patient for an OSCE station?

In five steps: a read-through of the script with the trainer, a rehearsal with a clinician playing good, poor, and borderline candidates, an agreed rule for every unscripted question, a recorded run scored against the script, and a drift check during the exam. Feedback and rating roles are trained separately.

### How long does it take to train a simulated patient for one case?

The AMEE organization guide cites an estimate of up to 15 hours for adequate training. A short history with no physical signs takes less; a role with distress, findings, or a line that decides a critical item takes more. The trainer plans the hours from the script, and the training time is paid.

### What should a simulated patient do when asked something not in the script?

Follow the agreed rule, usually to say they do not know or to deflect, and never to invent. The rule is settled line by line in training so that every candidate meets the same case. An unscripted answer that helps one candidate and not the next has changed the station.

### Should simulated patients give feedback to candidates?

In formative OSCEs, yes, if trained for it. The feedback covers how the candidate made the patient feel and what was clear, not clinical correctness or the marks. The ASPE standards of best practice treat feedback training as its own element. An untrained simulated patient does not give feedback.

### Are simulated patient ratings of candidates reliable?

They measure something different from the examiner's judgment. Studies have found that simulated patients rate higher than trained observers and physicians, with the largest differences on empathy and nonverbal items, and agreement with physicians on checklists but not on global scoring. Train the rating separately and report it separately.

## Sources

**Sources**

1. [Khan KZ, Gaunt K, Ramachandran S, Pushkar P. The Objective Structured Clinical Examination (OSCE): AMEE Guide No. 81. Part II: organisation and administration](https://doi.org/10.3109/0142159X.2013.818635) (Medical Teacher, 2013)
2. [Lewis KL, Bohnert CA, Gammon WL, et al. The Association of Standardized Patient Educators (ASPE) Standards of Best Practice (SOBP)](https://doi.org/10.1186/s41077-017-0043-4) (Advances in Simulation, 2017)
3. [Baig LA, Beran TN, Vallevand A, Baig ZA, Monroy-Cuadros M. Accuracy of portrayal by standardized patients: results from four OSCE stations conducted for high stakes examinations](https://doi.org/10.1186/1472-6920-14-97) (BMC Medical Education, 2014)
4. [Wollney EN, Vasquez TS, Stalvey C, et al. Are evaluations in simulated medical encounters reliable among rater types? A comparison between standardized patient and outside observer ratings of OSCEs](https://doi.org/10.1016/j.pecinn.2023.100125) (PEC Innovation, 2023)
5. [McLaughlin K, Gregor L, Jones A, Coderre S. Can standardized patients replace physicians as OSCE examiners?](https://doi.org/10.1186/1472-6920-6-12) (BMC Medical Education, 2006)

> [!TIP]
> **In practice**
> A simulated patient who completes a rating needs a role of their own on the scoring system, separate from the examiner's, so the two instruments never merge. To see how StudyDrome's examiner app gives simulated patients and observers their own role, read /docs/osce/people/in-studydrome-examiner-app/.