I'm going to call the R&D costs a wash because that cost should also be factored into the price of the diabetic injector that forms the basis of the $30 homebrew design.
From your list, that leaves IP licensing and regulatory compliance which are both imposed by the state.
I believe we would still see plenty of (and actually probably more) innovation with a much more relaxed IP policy that favored small, incremental improvements and the diversity of product variations that competition would bring. In short, copyright terms should be 5-10 years max and should only be granted in cases where there's proof of significant research. Simply, if the ability to monopolize your 'idea' for the better part of a decade doesn't incentivize you sufficiently--it's likewise not worth depriving the public over it.
With regards to regulatory compliance, I believe we're seeing a case of the perfect becoming the enemy of the good. Imagine if the state extended the good-Samaritan policy to all medical providers--legally protecting anyone who exercises good-faith and due-diligence towards improving the health of their fellow man. Also, replace the old-west era FDA with information-age search engines, customer reviews and a plethora of private Consumer Reports -type organizations. In this scenario, there are several auto-injectors available on the market ranging from a $30 model which has failure rate of 50% to a $600 model that's been tested out the wazoo and only fails 1% of the time. Even under this very pessimistic scenario (50% seems pretty conservative to me), many lives are saved because there's a long tail of people who can't afford the $600 model but more than half of them end up surviving an episode of anaphylaxis because they can access the low and mid -range models.