A steelmanned accurate take is that they are not biologically objectively confirmable. Which is more of an actual problem in psychiatry than other areas of medicine. Looking at say past x-rays and noting a tumor that went uncommented upon at the time on malpractice lawsuits for instance.

But notably also not done the majority of time in practice. It is seldom done to the level of say verifying your flu strain, let alone other more expensive and/or invasive tests which do not serve the patient's medical needs.