Even if it's more like Thalmic Labs or 23andMe, I'm worried about whether fundamental infrastructure (or a leap-frogging equivalent, like cell phones) will be widespread enough for this sort of research/work to be useful to the people who need it most.

Like, there are feel-good stories like Dean Kamen's Slingshot (IIRC, he partnered with Coca-Cola to distribute the things) but I have absolutely no idea how well such things work in the field and the papers written about things like the Roundabout PlayPump are pretty damning.