There are two reasons for this: one, medical terminals are super, hilariously locked down. The medical systems I've worked on are ultra-hostile instances that are often centered around detecting and eradicating non-explicitly-authorized applications, even to the point of kneecapping actually-needed software.
Second and more practically, malpractice suits are vicious wars of discovery and bitter, knock-down-drag-out compliance fights. As with so many things, the legal fiction of accessibility software sounds nice but would not stand up to the scouring scrutiny of legal counsel and defendants with critical reasoning skills and powers of discussion (and deduction) above that of policy and compliance controls. I don't mean to belittle your idea at all, but much like the doctrine of Sovereign Citizens, all is well and good until you actually encounter a judge or human with reasoning skills and the ability to trivially perceive what's /actually/ going on, and then it all falls apart.
Not disagreeing at all! My suggestion was predicated on the idea that this particular medical regulation might be a "Type II" law. To explain — I feel like there are two types of law:
• Type I: law that has a logical "spirit" behind it (usually because it was drafted all-at-once by a small number of authors) where judges will interpret the law and violations against it as reasonable human beings do;
• Type II: law (usually more "regulation", but let's call it all "law" here) that is effectively the output of a continuously-iterating bureaucratic process; or law with a "spirit" that is obviously contravened by newer laws, but which still stands because no case has yet come to knock it down; or where there's no "spirit" to the law at all, just an infinite stream of under-the-table negotiated compromises, encoded as a huge book of requirements and exceptions.
Type II law includes, for examples: tariffs; statutory crimes; and created-by-fiat licensing schemes. Judges can't think like humans when deciding whether someone has violated a Type II law, because there's no human-legible rationale behind why the law exists. They have to just plug in the formula specified in the legal code, and see what happens.
I suspect that much of medical regulation is Type II law. Insofar as that's true, you could get away with a lot as long as you still adhere to the precise wording of the relevant requirements.
This particular case (EMR software) may indeed fall more under the provision of a Type I law/regulation, though.
Would it be possible for an EMR software vendor to ship both a "canonical" client interface that meets all the specs/requirements; and then a separate, secondary client interface, for the same backend, which isn't included in the "package" of EMR software sold to the clinic/hospital/etc., but rather is instead a "bottom-up adoption" play, with download links to it being passed around under-the-table between doctors?
If there's any potential for doctors being liable for such-and-such if they were found to not be "using" the primary client, then said secondary interface could be implemented as macro+screen-scraping software, that operates the primary client-interface underneath—much like Mint used to do before banks supported OFX export. Then just get it classified as accessibility software, like a screen reader, and let those doctors claim a need for it :)