The problem with broad definitions is that it causes false negatives in potential treatments. If you have a treatment that's effective on 10% of your group and useless on the rest you need a sample ten times as large to find an effect.
Or, consider TPA. It is an extremely dangerous drug (used correctly it still has greater than a 1% chance of killing the patient), if you administer it to "stroke" patients you almost certainly do more harm than good. The reality is that if you use it on a bleed type stroke it can (and likely will) only make things worse. Use it on a clot type stroke and you might save them. We of course know this and only use it when a CAT has confirmed it's a clot--but what would happen if we didn't distinguish the two cases?
> The vast amount of treatment for autism is therapies.
> It really doesn't matter if the underlying cause is very different in terms of treatment because a speech therapist works the same with a kid with autism as they do with a kid with down syndrome.
> If there were more pharmaceutical interventions then I might care a bit more. But there's just not.
There are a limited set of drugs involved in autism therapies, but they are general drugs that would be applied regardless the diagnosis (for example, antidepressants.)
In fact, it's much like depression in that sense in that it's a very broad definition with a number of causes. Some antidepressants even work better or worse for people.
Or, consider TPA. It is an extremely dangerous drug (used correctly it still has greater than a 1% chance of killing the patient), if you administer it to "stroke" patients you almost certainly do more harm than good. The reality is that if you use it on a bleed type stroke it can (and likely will) only make things worse. Use it on a clot type stroke and you might save them. We of course know this and only use it when a CAT has confirmed it's a clot--but what would happen if we didn't distinguish the two cases?