Found in 1 comment on Hacker News
entee · 2020-10-19 · Original thread
Any technological or other intervention in the healthcare system must answer the following questions:

1.) Who pays for it? Doctors? Insurance? Patient?

2.) Who benefits from it? How much? Are they the same as the people who pay for it?

3.) How do you prove it has this benefit? Does the FDA need to approve it? (spoiler: almost always yes if a patient is involved directly and you claim a strong medical benefit)

4.) Whose pocket are you taking money from? Efficiency is just another word for less money sent to someone who profits from the inefficiency.

5.) Who is liable if your technology makes a mistake or leads to someone making a mistake?

6.) How do you integrate with current systems? Are you an app that runs on EMR or other patient data systems? If not, how do you get the data and do you have a HIPAA problem? If so, how do you get those people to go through pain it takes to integrate your system into their 20 year old, barely been updated ever system?

None of these are strictly technical problems, though 6 has a technical component. Having worked at a company that pivoted through a couple of the ideas you list, our failures were often heavily influenced by some combination of the above. Maybe a technology that provides a radical boost to some outcome, or an extremely large cost reduction could overcome these hurdles by virtue of their excellence. That's very rare, and any technology company in healthcare MUST have a strategy around these questions.

Taking your first example:

Check symptoms before visit. How do you ensure measurements are reliable, does every patient need a blood pressure cuff? Does this need to be an FDA approved device? How do we know they used it correctly? Even if the measurements are fine and we need no special equipment, maybe doctors get a payment for doing a vitals check (I don't know for sure), and if they do, why should they promote something that takes, say, $10 out of their pocket? Maybe the doctor can't prescribe a given pill without having run a specific test, so they need to check the symptom directly in the office anyway. Maybe the insurers would like some savings, but how do they force the adoption of something that's not all that widespread in the doctor's patient population? Will they pay for the devices the patient needs at home? How will that money get to the patient?

None of these things are insurmountable, but they are tricky to navigate. Off the top of my head I can find similar concerns to all of your options. Many of your ideas have had technological proofs of concept developed in an academic context stretching back years. They're not widely deployed because the healthcare system is incredibly fragmented with many players who have competing and often misaligned incentives. And there's a ton of regulation layered on top of that: #7, all your disease specific options, and perhaps some versions of #2 and #3 would need to go to the FDA before you could market your technology as solving the problem.

That's not to say your ideas are not good ideas, they absolutely are! But the reason they haven't been deployed is not because we just lack the technology. Those who have successfully deployed technologies in healthcare (for example https://www.virtahealth.com/) have also spent an enormous effort answering the above questions.

For a great book that breaks down a number of issues with tech in healthcare see:

https://www.amazon.com/dp/1449305024/ref=cm_sw_em_r_mt_dp_yu...