In this post, I am not going to distinguish between a doctor's "helping" a patient kill herself and the doctor's killing the patient herself, since in both cases the doctor kills, in the former case in concert with the patient and the latter alone. To help an assassin pull the trigger or aim the gun is to be a co-assassin. There are two different kinds of reasons given for allowing doctors to kill suffering terminally ill patients: care and autonomy. Those who focus on care base their argument on the patient's suffering and the physician's task in relieving that suffering. One difficulty with justing the euthanasia on such grounds is that once one sees euthanasia as a part of the physician's task of relieving the suffering, then it would become the physician's job to euthanize an incompetent patient (an eight-year-old?) who is suffering, terminally ill and yet begs to live, but where either no proxy is available or the proxy consents, just as it would be the physician's job to do life-saving surgery on an incompetent patient who resists the surgery. But even a lot of supporters of physician-assisted suicide will say that this is going too far.
Suppose instead that we base the killing of the patient on autonomy considerations: the patient chooses to be killed. Here, it is not clear what role in the justification is played by the fact that the patient is terminally ill and suffering, except maybe an epistemic role in providing evidence that the patient is not insane to request killing. After all, if the point is that people have the right to make deep decisions about life and death matters, then it seems that it would be equally the right of a patient who is quite well physically and mentally but who wishes to avoid creditors to request being killed by a physician, whereas surely this is not a request a doctor should accede to.
Now, a defender of euthanasia might object that a limiting condition on a physician's following of patient instructions is the good of the patient, and the person seeking to escape creditors would do better to declare bankrupcy than to be killed. Thus, just as on the first view, euthanasia was justified by care with consent being a limiting condition, on this view euthanasia is justified by consent with care being a limiting condition. However, this is problematic in a different way: It misunderstands the doctor's role. While those who took the first view were wrong in thinking that care calls for killing, they were right that care is the doctor's task.
Consider a surgeon who removes a malignant tumor. It would surely be very strange to say: "In order to honor the patient's choice to determine what is and what is not a part of his body, Dr. Magrodska removed Mr. Jones' tumor." Surely the right thing to say is that Dr. Magrodska removed the tumor because it was malignant, with Mr. Jones' consent being a mere necessary condition (and one that could be satisfied in other ways were Mr. Jones to be incompetent). Professionals are not servants of their clients' wishes, and physicians are professionals par excellance.
Now, there may be things that only a physician is qualified to do and which are done primarily as a response to the patient's wishes. Certain kinds of elective cosmetic surgery are such. But I think it is correct to say that in such a case the physician is not acting as a physician. Rather, she is acting as a medically-trained beautician. And even there she risks losing her status as a professional altogether if she does not act from a belief that the surgery makes the patient more beautiful. That a task requires medical training does not make the task a medical one (not that killing people painlessly requires medical training—see my previous post). Someone with medical training may be needed as consultant for a film set in a hospital—but such consulting is not a medical task.
There is thus a tension between the aspects of care and autonomy in the justification of killing terminal patients. If one focuses on care, then consent becomes a mere limiting condition and one arrives at abhorrent conclusions about killing vociferously protesting children. But if one focuses on autonomy, then physician-assisted suicide ceases to be a medical task.