OSCE preparation
AI OSCE prep for medical schools
Students rehearse OSCE communication stations out loud with an AI patient, as often as they need, between scheduled SP encounters. Each score quotes the student's own words or notes what was missing, so faculty can coach to what was said.
Last updated: November 2026
Why students reach the OSCE underpracticed
A standardized patient session needs a trained actor, a room, and often a faculty observer, so a student gets only a few tries at each skill before the OSCE. That is enough to assess a communication skill, but it rarely gives a student the repetition needed to build one.
In the US, Step 2 CS ended and no national exam replaced it, so each school now decides on its own how to assess clinical communication. Schools in the UK and elsewhere that run OSCE stations in English can use the same practice.
Some students practice with a general AI assistant instead. It can improvise replies, but it isn't written as a patient case with a history to uncover, it often runs as typed text, and it doesn't score the conversation against your checklist. Students here practice with AI patients built for spoken, scored conversations.
How OSCE practice works in ClinicalSim
- Faculty choose the station. Faculty pick a case from the catalog, or approve one built around the school's own checklist, before any student sees it.
- The student opens it in a browser. The student starts the station on a phone, tablet, or computer and talks with the AI patient out loud.
- The AI patient answers in character. The patient responds as the case is written and shares only what a real patient would share when asked, so the student has to draw out the history.
- The student reads the report. The report scores each checklist item or framework step and quotes the transcript lines behind each score, so the student sees what went well and what to practice next.
- The student repeats the station. The student runs the station again, and faculty see completion and repeated attempts across the cohort under the access rules the school sets.
OSCE stations students can rehearse
Each row pairs a common OSCE communication station with published frameworks a school can score it against, and cases already in the catalog. Each case names the framework it uses, and a school can swap in its own checklist.
| Station | Frameworks | Cases in the catalog |
|---|---|---|
| History taking | Calgary-Cambridge, SEGUE, or Kalamazoo (KEECC-A) | History-Taking Practice: Acute Abdominal Pain |
| Adolescent psychosocial history | HEEADSSS | Confidential Adolescent Psychosocial Interview; Confidential HEEADSSS Assessment With a Risky Teen |
| Explaining a diagnosis or plan | Ask, tell, ask and teach back | Limited English Proficiency Discharge Counseling After PICU Asthma Admission |
| Delivering serious news | SPIKES, with NURSE responses reported | New Diagnosis: Delivering Serious News Adult EM |
| Informed consent | AMA consent elements | Informed Consent for Laparoscopic Cholecystectomy; Why Does My Baby Need a Spinal Tap? |
| Counseling and behavior change | SEGUE, or motivational interviewing (MITI) | Pediatric Vaccine Hesitancy Counseling; But Antibiotics Always Help Him |
| Error disclosure | AHRQ CANDOR | Medication Error Disclosure Without Harm: Duplicate Ceftriaxone Dose |
| Your school's own checklist | Your rubric | Built with your faculty |
Scoring students and faculty can check
- Every score traces to a word-for-word line from the transcript, or to a note that the behavior did not occur.
- Each communication framework and the school's own checklist are scored separately.
- Each UME competency is recorded as demonstrated, partially demonstrated, or not demonstrated. Entrustment stays a program decision that this evidence informs.
- Competencies map to the AAMC Foundational Competencies, as on the medical school page. A school outside the US can score against its own checklist instead.
One scored line from an example report
In this vaccine hesitancy conversation, the learner, a pediatrics resident, scored 5 of 5 on the SEGUE task set "Understand the patient's perspective." This is the first line the report quotes as evidence.
"I hear the worry, um, in your voice, and I also hear how important it is to protect Ava"
The same report scores the close of the visit 2 of 5 and names what to add next time: a summary of the plan, return precautions, and a check for remaining questions.
Extend your SP program, don't replace it
ClinicalSim gives students repetitions between SP sessions. Your faculty and standardized patients keep the graded OSCE.
| What students get | SP session | General AI assistant | ClinicalSim |
|---|---|---|---|
| Spoken conversation | Yes, in person | Often typed | Yes |
| Available any time | No, scheduled | Yes | Yes |
| Scored against a named framework | By a faculty observer | No | Yes, with quoted evidence |
| Cohort view for faculty | Manual | No | Yes |
| Runs the graded OSCE | Yes | No | No, that stays with faculty |
What your program sees
- Participation and repeated practice by cohort, with each piece of feedback linked to the transcript line behind it.
- Access roles your institution sets before launch, so students know who can open a report.
- A cohort-level comparison of first and later attempts, without ranking individual students.
- Practice scores are formative. They do not establish OSCE readiness or predict a passing result.
Getting started
- Start with one cohort and a few stations, and have faculty approve each case before students open it.
- Students need a browser and a microphone on a phone, tablet, or computer. Nothing needs installing.
- Practice is audio only and in English for now, so physical examination and nonverbal skills stay with live encounters.
- Institutional IT may need to allow the voice service before the first session.
Frequently asked questions
Can AI standardized patients help students prepare for OSCEs?
Yes, for the communication stations. Students rehearse taking a history, explaining a plan, delivering serious news, or getting consent out loud with an AI patient, then review feedback that quotes their own words under each score. They can repeat a station between scheduled SP sessions as often as they need. Practice scores are formative and do not establish OSCE readiness or predict a pass.
What AI virtual patient tools do medical schools use for OSCE prep?
A medical school choosing an AI virtual patient tool for OSCE prep should check four things: students speak rather than type, physicians wrote the cases, scoring follows the school's checklist or a named framework such as Calgary-Cambridge or SEGUE, and faculty can see the evidence behind each score. In ClinicalSim, students talk with an AI patient and every score quotes their own words. Feedback is formative, and the graded OSCE stays with faculty.
How is this different from practicing an OSCE station with ChatGPT or Claude?
A general AI assistant can improvise a conversation, but it isn't built as a patient case. Here the AI patient follows a physician-authored case and shares history only when the student asks for it. The encounter is spoken, and the report scores it against your checklist or a named framework, with transcript evidence behind each score that faculty can review.
Does ClinicalSim replace standardized patients in OSCE prep?
No. It extends your SP program. Students use AI patients for the repetitions between scheduled SP sessions, while standardized patients and faculty keep live coaching, the graded OSCE, and summative assessment. Physical examination, eye contact, and body language still need an in-person encounter.
Can we upload our own OSCE checklist?
Yes. Your school supplies its checklist with its own element names and definitions. Faculty and our team then agree which items a spoken case lets the student show, and how each is scored, before practice starts. Physical exam and nonverbal items need another method. Faculty-rater validation on your own rubric hasn't been done, so scores stay formative.
Can ClinicalSim run the graded OSCE?
No. Its feedback is formative practice between SP sessions, and the graded OSCE stays with your faculty and standardized patients. A practice score doesn't establish OSCE readiness, predict a passing result, or settle a progression decision. Programs can use the reports to choose a coaching focus before the real station, and students use them to plan their next attempt.
Does AI OSCE practice work for medical schools outside the US?
Yes, wherever stations run in English. A school in the UK or elsewhere can score practice against its own OSCE checklist, or against a published framework such as Calgary-Cambridge or SEGUE. Practice runs in English only for now, and the school's IT team may need to allow the voice service before students start.
Which communication frameworks can OSCE practice be scored against?
Calgary-Cambridge, SEGUE, and the Kalamazoo checklist (KEECC-A) for history taking, HEEADSSS for adolescent interviews, SPIKES for serious news (with NURSE responses reported), ask, tell, ask and teach back for explanations, motivational interviewing (MITI) for counseling, the AMA consent elements, and AHRQ CANDOR for disclosure. A school's own checklist can sit alongside any of them and is scored separately.
Does OSCE practice score tone of voice or the physical exam?
No. Students speak out loud, and scoring reads the transcript of what they said rather than the sound of their voice. The platform doesn't cover physical examination and can't assess eye contact or body language, so those parts of an OSCE station stay with standardized patients and faculty observers.
Who sees a student's OSCE practice results?
The people in the roles your institution sets before launch. Students can repeat a station in private first, before a report is shared with a faculty member, coach, or OSCE lead. Faculty review cohort patterns without ranking individual students, and because practice results are formative, they shouldn't feed a grading or progression decision.
References
- Harden RM, Stevenson M, Downie WW, Wilson GM. Assessment of clinical competence using objective structured examination. BMJ. 1975. doi:10.1136/bmj.1.5955.447
- Makoul G. The SEGUE Framework for teaching and assessing communication skills. Patient Education and Counseling. 2001. doi:10.1016/S0738-3991(01)00136-7
- Kogan JR, Hauer KE, Holmboe ES. The Dissolution of the Step 2 Clinical Skills Examination and the Duty of Medical Educators to Step Up the Effectiveness of Clinical Skills Assessment. Academic Medicine. 2021. doi:10.1097/ACM.0000000000004216
Related insights
How to design an OSCE case that shows what a learner can do
OSCE case design starts with the decision the station should support, then observable behaviors, SP training, and a scoring pilot before the station counts.
What an OSCE score predicts, and what it doesn't
Among 3,424 physicians followed for up to 12 years, lower national clinical skills exam scores went with more retained complaints. Not a read on one resident.
What 12 medical students want from AI patient simulation
Twelve clinical-year students in interviews and three codesign workshops put feedback, case quality, and faculty involvement ahead of novelty.
Why standardized patient programs run out of capacity
A standardized patient program's capacity is set by how many trained people it can put in a room, not by how many rooms it has.
Related use cases
- Undergraduate medical educationFrom history-taking to delivering a diagnosis, across four years
- Breaking bad newsSerious news practiced out loud before the first real family
- Informed consentScoring the conversation behind the signature
- Residency and fellowshipA longitudinal communication curriculum across PGY years
What this page does not claim
ClinicalSim is not endorsed by, affiliated with, or approved by the AAMC, the AMA, or AHRQ. Scores map to published frameworks; they do not establish compliance with any standard, which a surveyor determines against the institution's own policy.
ClinicalSim output is formative. It is evidence that informs program judgment, and it is not intended for employment, credentialing, privileging, licensure, or other high-stakes decisions.
Consistency between model runs is not the same thing as agreement with expert human raters. Faculty-rater validation on a customer's own rubric is the gate before any high-stakes use, and that work has not been done.
Practice scores do not establish OSCE readiness or predict a passing OSCE result.
A question for leadersHow many times does a student rehearse a station before the real OSCE?
See a scored OSCE practice conversation
We'll walk through a station, the checklist it is scored against, and the feedback a student sees.