Machine learning to extract communication and history‐taking skills in OSCE transcripts

We report the first application of ML in the context of student-SP OSCEs. We demonstrated that several MLMs automatically labelled OSCE transcripts for a range of interview content and some clinical communications skills. Some MLMs achieved greater performance and transferability. Optimised MLMs could provide automated and accurate assessment of OSCEs with potential to track student progress and identify areas for further practice.