Skill based assessmentBMJ 2003; 326 doi: https://doi.org/10.1136/bmj.326.7391.703 (Published 29 March 2003) Cite this as: BMJ 2003;326:703
- Sydney Smee
Skill based assessments are designed to measure the knowledge, skills, and judgment required for competency in a given domain. Assessment of clinical skills has formed a key part of medical education for hundreds of years. However, the basic requirements for reliability and validity have not always been achieved in traditional “long case” and “short case” assessments. Skill based assessments have to contend with case specificity, which is the variance in performance that occurs over different cases or problems. In other words, case specificity means that performance with one patient related problem does not reliably predict performance with subsequent problems.
For a reliable measure of clinical skills, performance has to be sampled across a range of patient problems. This is the basic principle underlying the development of objective structured clinical examinations (OSCEs). Several other structured clinical examinations have been developed in recent years, including modified OSCEs—such as the Royal College of Physicians' Practical Assessment of Clinical Examination Skills (PACES) and the objective structured long case (OSLER). This article focuses mainly on OSCEs to illustrate the principles of skill based assessment.
The objective structured clinical examination (OSCE) was introduced over 30 years ago as a reliable approach to assessing basic clinical skills. It is a flexible test format based on a circuit of patient based “stations.”
At each station, trainees interact with a patient or a “standardised patient” to demonstrate specified skills. Standardised patients are lay people trained to present patient problems realistically. The validity of interactions with real patients, however, may be higher than that with standardised patients, but standardised patients are particularly valuable when communication skills are being tested.
OSCE stations may be short (for eample, five minutes) …