4 ms·
Ask HN: Is there a Comma.ai but for health?
Allow me to expand on the weird question: comma.ai has innovated in a very regulated space by crowdsourcing data through sharing apps with enthusiastic users and using the driving data to build useful self driving models and technology.
What if someone did this, but for your own crowdsourced biometric data? I've long wanted a preventative health system, and an open source implementation like comma.ai would be ideal i think?
- grizzles 8y agoI made something in this vein last year. My idea was to create a simple open src db that natively exported time series vectors, one hot encodings and so on. Built on postgres it did pchr and had strong anonymity/deniability for non pii. Because I can't think of a single EHR app that does this, the big expensive systems such as Cerner included. They mostly care about other stupid things like billing. I lost the code due to a somewhat cavalier personal backup policy. Fortunately I don't sweat that kind of thing anymore because I found out that zfs is awesome. It was a fairly minor setback but I ended up deciding to do something else. If you take it up I don't think it would be hard to find users, before I abandoned it I found it easy to recruit a few random quants and a handful of (non US) doctors for the trial. You could probably get US doctors too. My thing converted data into a format that was consistent with some major open health datasets that have been released by eg. the states of Florida & Texas. If those datasets are HIPAA compliant then so would have been my thing. I thought that would have been a very persuasive argument for why doctors should use it. => "This thing can be used because if what it does is illegal then these widely published datasets by state government must also be illegal." Assuming everyone's consented all round of course.