So long as people can freely choose whether they will do it or not I don't see a moral problem. There would be a very big problem if healthcare providers were mandated to provide such a service. And note that while the evaluation certainly needs to be by a doctor a nurse is quite capable of doing it. Look at the Canadian method--for the most part it's something that's actually done quite routinely in emergency rooms across the world. Sedation followed by a paralytic. Usually that's a prelude to intubation but if you walk away in the middle it kills. Canada then pushes potassium chloride just in case as the paralytics wear off pretty fast.
And we are better off as individuals if we have the option of having external providers do it as that removes any dependency on actually being able to do things. There also is the benefit that it brings an external evaluation into the system that can recognize that maybe the evaluation was wrong. (I'm thinking of a case I heard about--woman thought she had lung cancer, chose to not treat it, simply work until she dropped. Autopsy said TB, not cancer.)
You completely missed the point. This discussion is about dementia. The assisted suicide laws in Canada and other countries generally require the patient to be of sound mind, as evaluated by a qualified clinician. The laws don't apply to patients with severe dementia.
In the comment above, @thinkingtoilet apparently wants someone to kill them if they ever have severe dementia. Presumably that desire would be expressed in some sort of "living will" type document. If the patient meets the criteria then should a healthcare provider strap them down and kill them, even if in the moment the patient says then don't want to die? That seems ethically dubious. It essentially puts providers in the position of being serial killers.
Canada has also had some serious abuses and ethically questionable situations. They are not necessarily a model to emulate.
Of course there are edge cases. Reality is continuous, not disjoint, and thus any attempt to impose a line will inherently create edge cases. What you are missing is that the case of inaction (not permitting it) also creates bad things. You can't make a situation without bad, all you can do is attempt to minimize the bad. Note that the "ideal" (as in maximum social benefit) amount of bad things happening is not zero. Preventing bad things always comes with a cost, there will always come a point where additional preventing of bad things causes net harm.
Consider, for example, nuclear power. It has basically been regulated out of existence in the US because of the standard that radiation exposure must be as low as reasonably achievable. The problem with this is that it doesn't result in safer nuclear plants, it results in plants that run on different power sources. Natural gas? Approximately 10x the risk (and that's not counting climate effects.) Oil? Approximately 10x the risk of gas, thus 100x the risk of nuclear. Coal? Approximately 10x the risk of oil, thus 1000x the risk of nuclear. The expected (and observed) safety benefit of the regulations is negative.
And to preempt the inevitable "Fukushima!", that was political. The expected death toll of staying put was approximately zero. The city was evacuated, killing hundreds, for no good reason.
I'm not missing anything. Medical ethics as commonly understood in modern western civilization imposes a clear line between withholding care versus actively killing someone who is unable to give informed consent to the procedure. The topic under discussion isn't even close to being an edge case. Minimizing the bad is not the goal. Nuclear power has zero relevance here and bringing it up is just an attempt to confuse the issue.
It used to be that medical "ethics" precluded actively killing someone. That's a religious legacy that more and more countries are coming to recognize is wrong. Fundamentally, this reduces to whether quality of life can be negative.
If quality of life can be negative then there will be cases where the humane act is to provide someone with a comfortable death.
And my point about nuclear power is that excessive regulation actually is counterproductive at maximizing human benefit.
And we are better off as individuals if we have the option of having external providers do it as that removes any dependency on actually being able to do things. There also is the benefit that it brings an external evaluation into the system that can recognize that maybe the evaluation was wrong. (I'm thinking of a case I heard about--woman thought she had lung cancer, chose to not treat it, simply work until she dropped. Autopsy said TB, not cancer.)