It’s really hard to persuade people that devices need evaluation like drugs do. This might be to do with the physical nature of a device: after all, if you can see the new cannula attachment, or special breathing mask, you know what it’s doing and what it must be making. Or it might be in the immediate satisfaction of a result, seeing the saturations rise or the respiratory rate decrease. Or it might be in the overwhelming common sense of the physiology/biomechanics/clinicopathological reasoning. But whatever the cause, it’s really tricky to have such things evaluated.

There’s a device which takes special weight-bearing X-rays of the spine for scoliosis assessment, and provides fewer x-rays than a CT scanner. It’s terribly expensive. And the evaluations which have been undertaken of it are …troublesome. They establish neatly its technical success, but don’t go on to assess if it actually makes a difference to patients [1] Of course, the failure to use patient-important outcomes and an obsession with measuring the measurable is not exclusive to x-ray machines or devices [2]  but we really should remember to be increasingly cautious the more we move away from a real-person outcome and the closer we get to a machine-read outcome.

– Archi