I do have to agree with you insofar as the studies' particular results go: in both scenarios, the patient suffered: in the first scenario because of a lifetime of depression, in the second because of the same + the trauma of the early life experiences. So, if we follow your dictum:

> We feel empathy for what they went through.

(which I mostly agree with) —then yes, the relatively higher portion of empathy should go for the character in the latter scenario (which matches the results of the study).

That said, if we generalize the study and apply it to the setting it's concerned with (medical practitioners empathizing with patients)—it's not a matter of relative measures between multiple patients; it's a question of whether the empathy a patient may (or may not) inspire is sufficient to cross some threshold for a medical practitioner.

If the fact that the cause is of biological origin is sufficient to drop the empathy response beneath that threshold —then it's problematic (the paper points out supporting research for just how problematic it is). After all, the person who has suffered a lifetime of depression does deserve empathy from a doctor/therapist, even if not as much (I guess) for somebody who suffers additional afflictions on top.

Actually that makes me wonder how much dealing with many patients who are very badly off (and deserving much empathy) moves the threshold for the doctor, so that maybe they can't empathize so much for patients with situations which are legitimately bad—but not as bad as what they've seen a lot of...

OK, I think I see your point. Maybe we can train the professionals to act out of reasons other than empathy? Or be more attuned to the current state and not the history. (Don't know if that is possible, or what specifically needs to be done).

The real issues at hand with this research are 1) prescribing medicine too often and 2) interacting with patients too little.

If you take a look at the original research (linked to from the posted article), then you can skip to the discussion section to see the authors wording of what you are taking about. They are saying the same thing as you, essentially, in a more specific and sophisticated way (page 4 of 5)

> Or be more attuned to the current state

That's a good suggestion I hadn't thought of. Really not sure aside from that—I've got some vague ideas about certain 'empathy trainings' that could be done, but I'd be concerned about throwing a random factor into how the rest of their work is done...

This topic encompasses one of the proposed advantages of using machine learning in medicine - data driven and unbiased diagnostics.

The authors of the original research discuss the evidence in the context of medical practice generally relying upon biological evidence and that such a reliance is linked to a lack of empathy and a corresponding lack of belief that psychotherapy would be a helpful treatment. That dissuasion from psychotherapy, they argue, could lead to worse outcomes for patients that would have benefited from psychotherapy.