An interesting article on medical technology:
A DIY ‘bionic pancreas’ is changing diabetes care — what’s next?
It describes two opposing views on how medical technology can be designed and deployed:
- the dominant top-down approach, with experts designing and other experts approving technology that’s supposed to be good and safe for a large number of users
- the hacker/open/free/malleable bottom-up approach, in which become sufficiently expert to fine-tune technology to their needs
My mother has type-1 diabetes (T1D), so I know a bit about this universe from the “bottom” side. There are enormous variations in lots of parameters between T1D patients, which are not very well understood. Even the best experts cannot design a treatment protocol that patients could blindly follow and live well. There are patients who have to do that, in particular older patients or patients with cognitive problems. They survive, but they don’t live well. Good treatment must let the patients monitor and adjust their own treatment.
And then, automation happens. The artificial pancreas, a cyber-physical system, containing a processor and lots of software. Again, no expert can design an algorithm that will be good for everyone. We simply don’t understand T1D well enough for that. My prediction is that black-box systems of this kind will do more harm than good. Which is why we need malleable systems.