The problem is that risk assessment becomes paramount, and in stretched services it may be the main purpose of contact with patients. No health professional wants to bear the heavy burden of death, possibly compounded by an inquiry and fear of loss of livelihood. This isn't just wanting to keep out of trouble, though. We don't forget the people who die; we can't suffer the pain of their families, but there will always be regret and sorrow.
When I learned about depressive disorder, I heard about melancholia. I read that people slow down, look sad or flat, their movements ponderous. They talk quietly, often with gaps. Sometimes they are agitated. Their sleep is disturbed, they may barely eat, and their thoughts are black. We have all seen these people, and they are often very unwell. But what of the smiling depressives?
It can be more complicated to assess their risk: a rigid approach may lead the psychiatrist to believe that it is negligible, and it will look negligible when recorded. But a broader, less focused conversation might reveal more. It's just difficult for the psychiatrist, with limited time, to achieve that.
My own experience is certainly that what you see is not always what you get. I feel depressed before my appearance follows suit. And at that time, if I say I feel depressed, the response will be: Well, you look fine, you seem really well. I know that people want me to be well, and that it's often no more than that, but it's quite discombobulating. Later, I may feel no worse, but haven't bothered to put makeup on, and am told: You look dreadful. That's not cheering either.