Hacker News Reader

Comment by hwc | Hacker News Reader

I had the exact same experience.

The company in question is Epic Systems, maker of very popular medical records software.

I’m not sure what it is about medical systems, but the next biggest player in this space in the US I believe is Cerner, and it’s largely written in a different oddball language. I used to work at a hospital system as sort of a general purpose developer. Had things gone a bit differently, I very well could still be there as a CCL (Cerner Command Language) guy. It’s a strange hybrid of declarative SQL-like syntax and half-baked procedural syntax.

Cut my teeth writing CCL for several years at Cerner. It has its rough points, but I learned a lot of valuable SQL optimization experience.

Looking back I can see why post-processing SQL query results using a procedural section at least gives you an escape hatch to do business reports or other iterative tasks.

I don't miss it, but I can see why CCL was designed and written at the time.

Most of why I disliked working with CCL as a client developer was because it was a pretty poor developer experience. The tooling was cumbersome, the feedback loop was long, no easy way to do source control (I think they were working on improving that aspect), and of course the limited user base comes with its difficulties. Not that any of this is particularly surprising for a homegrown programming environment.

Hey, what if they wanted to use another database than Oracle and not have to rewrite all the SQL.

The tell was "The lucky one got to program Visual Basic."

You can tell the author wasn't one of the lucky ones because MUMPS is a joy to work with compared to VB6.

I was going to comment “He must have worked for Epic”

My country's medical system is made by Epic, should I be scared?

My interactions with it as a patient have been quite good. Both my previous provider and current provider use it. All the doctors are trained on it, as well as all the staff. It makes all sorts of things quite simple, for example tranferring all my records from my previous provider to my current one was one click by my doctor. later, we could review my medical imagery in the office as needed.

From what I understand, getting to that point required a huge amount of effort on the provider's part (many thousands of hours of contractors configuring systems deep in the bowels of hospitals or data centers). But that's not surprising: anything regarding health and IT in the US eventually grows to consume our entire economy.

Same here. The hospital that my current doctors work for is on an Epic system. From the patient POV, it's quite good. Very easy to figure out where to find stuff. I've seen far, far worse UIs.

Nah, you'll be fine. Most hate for Epic stems from 3 places: users who don't want to learn new systems, the growing pains of adapting Epic to new regulatory systems, and the general fact that an EMR let's regulators and administrators demand too much from their clinicians that couldn't be done with paper charts.

This is Epic's common strategy of blaming all criticism on "resistance to change". The thing is, their software is simply bad, and its implementation in the Nordics has been a disaster.

At the end of the day, it's complex software for doing complex tasks. Even with 0 bugs, and a "perfect UI", people would by annoyed to use it simply because the tasks that need to be done are not what they signed up for when entering healthcare. If your job made you use a eye poking machine, you'd be upset even if it was the least painful eye poking machine on the market.

The only answer is automation, and in a safety critical role like healthcare, automation has to be done with great care.

That's of course true in the abstract and general, but I'm not talking about that. A lot of the criticism in Denmark has been quite quantifiable: things like increased number of clicks to perform similar tasks, and instances of losing work and being forced to start over if you, say, picked the wrong option early in a process, increased number of steps for no benefit. And also the presence of many steps and decisions that relate to billing, which of course should not need to be made here, since no bills are ever written. A bunch of things that add up to more time spent on things that the old systems made simple.

But from Epic and the higher-ups (who have to defend their expensive acquisition), it's always the same reply even to the most concrete of criticism: you just don't like change! At some point it really moved beyond just a bad excuse and became victim blaming.

(Please know that the context here - and I think in our neighbouring countries as well - is not that there was no automation, rather there were a set of mostly functional solutions in each region that were replaced by a single top-down mandated country-wide system. There was a book published recently about the process that led to the implementation of Epic's system from which I did not get the impression that this was based on merit as much politics.)

Horsefeathers. It is an objectively bad interface.

The newly released, expensive, hated and bug-ridden system here in Norway ("helseplattformen") is by Epic, written in MUMPS/caché.

So yeah, be worried.

Yes.

Medical software dev in the US is a dumping ground for the under-qualified or people looking to coast the balance of their career.

Please tell us more!

Reply on news.ycombinator.com