Randomised clinical trials with clinician-preferred treatment.

The standard design for randomised clinical trials may be inappropriate when the clinician believes that one of the treatments being tested is superior for the patient, or when the clinician has a preference for one of the treatments. For such instances the suggestion is that the patient is randomly allocated to treatment only when there is clinical disagreement about treatment of choice for that patient, and then the patient is assigned to a clinician who had thought that the regimen allocated is the one most appropriate for that patient.