This is the requirements list, not the buying advice. Copy the four blocks below into a tender document or a scoring sheet and ask each vendor to answer line by line. The reasoning behind the choice and how to run the process are in "How to Choose OSCE Software." This page is the artifact that the guide sends you to get.
The list is grouped by stage of the exam cycle, because that is how the gaps appear. Most products handle one or two stages well. The stage a school discovers late is the last one, after the marks are in.
Blueprint and station bank
Scheduling and exam day
Marking and standard setting
Analysis, results and feedback
The requirement schools forget
Line two of the marking block appears to be a detail. It is the strongest finding in this area. In the Swiss Federal Licensing Examination, 377 examiners across five faculties marked on paper and on screen. On paper, 14.2 percent of 1,428 checklists returned at least one missing rating. On screen, where completion was required before a sheet could be submitted, there were none. Examiner preference for the electronic form rose from 54.5 percent in the first year to 89.8 percent in the second.
A missed rating is not a small problem. It removes a candidate from that item's denominator, which moves the station statistics and can move a borderline decision. The mechanism that fixes it is a screen that refuses an incomplete sheet, which is why the requirement is worded as a refusal rather than a reminder.
A second finding is worth putting in the tender. In a comparison of tablet and paper marking, the share of encounters carrying an examiner comment rose from 41 percent to 51 percent, and the increase was largest for borderline candidates, who are exactly the ones whose feedback matters most.
Questions to ask the vendor
Ask this | Why it matters | What a good answer sounds like |
|---|---|---|
Show me the item analysis from a real sitting. | Separates a scheduling tool from an assessment platform. | A live report with facility, discrimination, and reliability, not a slide. |
Which standard-setting methods run inside the product? | If none do, you will export every exam to a spreadsheet. | Named methods, run on the exam data, with the output shown. |
What happens when the network drops mid-station? | Exam day is the only day this question gets answered. | Marking continues on the device and syncs when the link returns. |
Can we get every score out, per item, per candidate? | You need the raw data for your own analysis and your archive. | A per-item export in a spreadsheet format, available the same day. |
Who can change a score after marking closes, and is it logged? | Boards are asked this during accreditation reviews. | A named role, a reason field, and a record that cannot be edited. |
How are results released to candidates? | Release is a policy decision that the product must adhere to. | A visibility control the exam office sets, not an automatic email. |
What the checklist deliberately leaves out
It says nothing about hosting, price or contract length. Those belong in the commercial section of your tender, and they differ too much between schools to template. It also says nothing about what the software should look like. Examiners judge that in ten minutes at a pilot, and their verdict is worth more than a row in a scoring sheet.
Frequently asked questions
Can we run an OSCE without dedicated software?
Yes, and many schools do. Paper marking works, and the free calculators cover the statistics. What you take on is transcription, the missed-rating rate described above, and a results timeline measured in weeks. Software buys back time and removes one error class. It does not make the exam valid.
How many of these 32 requirements are essential?
Score them rather than count them. The marking block and the analysis block are where exams fail, so weight those. A product that covers the bank and the timetable but exports nothing useful after the exam leaves you where you started.
What is the difference between a scheduler and an assessment platform?
A scheduler builds the circuit and prints the timetable. An assessment platform also collects the marks, applies the standard and computes the statistics. Both are useful. Only one produces the evidence your quality report needs.
Should we ask for a pilot before signing?
Yes, with your own station and your own examiners. A demonstration shows the product working on data designed for it. A pilot shows it working on yours, which is the only question that matters.
Who should own this checklist internally?
The exam office, with the clinical lead signing off the marking and standard-setting blocks and the data protection officer reading the audit and export lines. A tender answered by one person alone tends to score highly on the stage that person knows.