Statistics from Altmetric.com
I read with interest the article by Lewis et al about the effects of sleep deprivation on on-call medical staff.1 Besides the admitted weakness of difficulty in coming up with a suitable design comparing the performance between normal and sleep deprived states, the study is contentious in at least two other aspects. Firstly, the usage of factual medical knowledge is inappropriate as a tool for the assessment of clinical performance because the delivery of clinical care involves not only correlation of complex clinical, laboratory, and radiological clues but also crucial decision making as well as interaction with patients and colleagues. Secondly, even after accepting that the factual knowledge is the most important element in clinical performance, testing one’s knowledge in the multiple choice format for this purpose is questionable because at the bedside we do not have choices to choose from.
Thank you for asking us to comment on our study. We have chosen what is a simple model of assessment of medical knowledge, which is used both in medical student finals and the membership examination and hence has a long history of acceptability as an assessment tool. We fully accept that it is in no way an assessment of widespread clinical performance and indeed clinical care does involve summation of a number of issues before a clinical diagnosis can be reached. However, it is not possible to test such a complex model.
Again I would agree that the knowledge base could be considered to have a questionable purpose at the bedside but unless one has a firm knowledge base it is not possible to make any decisions.
I think the criticisms which have been raised are not specific of our study but are general criticisms of the validity of assessment of doctors.