There seems to be no accepted way of surgically assessing patients with sleep disordered breathing (SDB). Because of this, clinicians fall roughly into three camps: those who just use one operation for all patients, those who have given up surgery for SDB, and those who decide to use one of these controversial forms of assessment. Drug Induced Sedation Endoscopy (DISE) is probably the most widely used form of assessment. Aside from the criticism that this isn’t ‘natural sleep’, the experience of many surgeons who start surgery for SDB is that the information gained from DISE seems not to translate to improved outcomes for their patients. There seems to be no consensus on what and much of anaesthetic agents to use, or how to interpret the findings of DISE. The VOTE system of interpretation of DISE is the one with the most studies behind it, and is used mainly in Europe. This article describes the difficulty of interpreting DISE even with the VOTE system to aid it. Thirty-one videos of DISE were compared by a senior surgeon, and again separately by a junior doctor. The anaesthetic method for DISE was relatively standard, with a target controlled infusion of propofol, and only the video used for the between the two observers. This was to avoid the problem of interpreting DISE at the wrong depth of sedation. The junior observer seemed to consistently over-emphasise the importance of all levels of obstruction except the tongue base, which was consistently under-emphasised. The learning curve for DISE seems to suggest that clinicians require specific training for this technique. I would also suggest that perhaps the inter-observer reliability of the VOTE system should be called into question. Another problem is that we are basing this study on the experience of the senior observer, but there is no way really of showing that what he decided was the correct interpretation of the actual problem. Would patients have been better treated with the junior doctor’s interpretation or a different junior doctor’s interpretation? Are we testing the experience of the observer or the VOTE system; is DISE a help at all? Surgical trials are difficult to arrange, but without good quality research in this field we are going to keep coming up with the same questions rather than answers.