If you look at the Supreme Court of Canada cases, the two leading ones in which the original doctrine was established—Reibl v. Hughes and Hopp v. Lepp—what they require is that all reasonably indicated treatments for the patient for the condition they have must be disclosed to the patient, and the benefits, risks, and harms of each of them disclosed, including the option of having no treatment at all. If the patient chooses them, then they'll have to be made available; otherwise, the patient's consent to the treatment they're given—and let's assume we're talking here about what I would call the “non-treatment of euthanasia”—would not be validly consented to. There wouldn't be an informed consent to that if the patient hadn't been offered all reasonable alternatives.
You can't impose those other alternatives, of course, but your legal obligation and your ethical obligation is to offer them to the patient. Look at the work of Dr. Harvey Max Chochinov. There's some very good work on this. Even people who have requested euthanasia change their minds very often when offered and given good palliative care and pain management.
I've done a lot of work in this area. I've actually researched it for about 35 years. In 1993 I was asked to give the opening keynote address at the international pain conference in Paris. I proposed that access to proper, fully adequate pain management was actually a fundamental human right. That has now been encapsulated in what is called the “Declaration of Montreal”, which was passed in 2010. It says that for a health care professional not to respond with reasonable alacrity to people in serious pain is a breach of human rights.