subject 1 • • .4. All of the above studies confirmed one thing: measuring pressure between the body and an actual support is difficult. Difficulties in measuring interface pressure include its variability over a small area, particularly over bony prominences with high pressure gradients, and the difficulty of avoiding artefacts due to the presence of the sensor itself 5 . The movement of soft tissues (to which the sensors are attached) in relation to bony prominences is also difficult to control. Krouskop and Garber 6 showed that interface pressure measurements are not absolute or comparable between different instruments or investigators. They are often not sufficiently repeatable to indicate a statistically significant difference between surfaces clinically understood to be very different. Bader and Hawken 7 showed large variations in pressure measurements when the subject was reseated between pressure readings. Finally, since the soft tissues of the body and the materials of the mattress are both to some extent viscoelastic, it is difficult to isolate the time­ dependencies in such a test, and apportion them correctly to tissues and mattress. As yet there are only a small number of randomised controlled trials (RCT's) for the evaluation of support systems, and the results of these have been inconclusive 8 . The reason for this is largely one of population numbers. RCT's have only provided statistically reliable information in a few instances, perhaps where very big actual differences existed between support systems in the trial. One example was conducted by Allman et az9, who used an entire hospital population to show a difference in outcomes between conventional mattresses and air-fluidised therapy. Another was by Ferrel 10 who conducted a large study in a nursing home to demonstrate the superiority of the low-air-loss bed system over a standard mattress. A further problem with clinical trials is that of blinding: whatever care is taken in the random allocation of patients to different surfaces, it is obvious to staff, patients, and researchers which type of surface the patient is on. There is no guarantee that this knowledge will not affect the patient's care, and so the outcome.