I think the "burden of proof" part is exactly what's holding him up. As a skeptic, and an analytically minded person, and perhaps even as a blogger - he's uncom

I think the "burden of proof" part is exactly what's holding him up. As a skeptic, and an analytically minded person, and perhaps even as a blogger - he's uncomfortable writing "Mr Smith has chronic back pain" without a little more concrete evidence. I'm arguing that as a doctor he shouldn't require more evidence in this case. It's fine to write "Mr Smith has chronic back pain and I recommend he can use his own chair" and be right only 80% of the time, and introduce a process whereby sneaky chair-seeking miscreants can reliably get a doctor's note even if they don't have back pain. It's fine because the consequences of being wrong are extremely low. In practice doctors should be, and often are, OK with making diagnoses of much more serious conditions that are only 80% or 90% or 99% likely to be correct, even when some other test could reduce the uncertainty further. If he's waiting to be 100% certain a patient has bipolar disorder or anorexia before prescribing any treatment, he's doing them a disservice. The level of certainty required should be in line with the severity of the condition being diagnosed, and of the recommended treatment. And when writing a note to an employer, it's correct to say "Mr Smith has cancer" or just "Mr Smith has a medical condition", not "I'm 94% sure Mr Smith has cancer, perhaps the next scan will be more conclusive".