Cosign. I worked in healthcare for the better part of a decade and spent just over five years of that doing data interchange (mostly HL7).
It's always somewhat amusing when threads about it pop up on HN because it almost invariably winds up with smart programmers insisting that these have to be simple problems that would be easily solved if only we were using the right API, while having zero understanding of the business and regulatory requirements that, in reality, make many of the things attached to it very difficult to get right in a field where you literally have the potential of killing somebody with a severe enough mistake.
Totally. You have to work in Health IT to know just how fucked it is. Before being in the industry I was convinced it was low quality, shitty tech people.
In reality it’s massive monolithic slow moving orgs that data block, create black boxes, have poor documentation, and just generally don’t want to move forward.
Specifically I’m referring to EHR vendors and other legacy systems. Not necessarily providers or even payers.
"it almost invariably winds up with smart programmers insisting that these have to be simple problems that would be easily solved if only we were using the right API"
Because they are. Make the common case fast. There is no reason to expose all possible complexity to every transaction.
>Make the common case fast. There is no reason to expose all possible complexity to every transaction.
Hint: addressing that is exactly what HL7 has been doing for over 30 years.
The easy parts are easily solved because they're easy, and the easy parts are 90%+ of any given integration.
Those remaining hard parts are where the time and effort goes, and a lot of them are very fundamental things that seem easy until you try to actually do them.
"Simple" things like ensuring that you properly match a patient to their preexisting medical record are incredibly fraught with unexpected complexity that you can't handwave away with an interchange format.
Agreed, 100%! It always reminds me of the episode of "The Office" where the warehouse wins the lottery, everyone quits, and the office staff are conscripted to load the truck with paper. Searching for a way to load the truck more efficiently, one of the main characters is astonished that the warehouse workers never tried backing the truck into the warehouse to be closer to the paper. Believing them to be morons because they hadn't thought of this, he implements his solution and causes a fair amount of damage to the truck and the building because he didn't stop to think that maybe, just maybe, the people who had been doing this work for years (work he had no real experience with or insight into) might have had a good reason or two why they hadn't already implemented this solution.
It's a silly example, sure- all I'm trying to say is that there are usually a few pretty good reasons why "the solution is so simple, just do x" comments haven't already been implemented and the commenters don't usually have the context to know that. It's been my experience that a lot of problems with seemingly simple or obvious solutions haven't been solved that way for good reason.
Cosign. I worked in healthcare for the better part of a decade and spent just over five years of that doing data interchange (mostly HL7).
It's always somewhat amusing when threads about it pop up on HN because it almost invariably winds up with smart programmers insisting that these have to be simple problems that would be easily solved if only we were using the right API, while having zero understanding of the business and regulatory requirements that, in reality, make many of the things attached to it very difficult to get right in a field where you literally have the potential of killing somebody with a severe enough mistake.