I wasn't sure how to reply to this because what you're describing absolutely doesn't sound like Epic. And then...
And now it makes sense.
The previous generation of EMRs (ECW, Athena, Centricity, Meditech, etc.) were woefully inadequate with data sharing. What you're describing, Les, sounds like a bunch of places using ECW. Every practice had its own instance, and ne'er the twain shall meet.
Obamacare, for all its warts, helped by taking a big hammer to this problem and requiring connectivity. Epic came along. Athena got better. Cerner came along. And suddenly the HL7 standard was in real use and systems could talk to each other. For example - if you go to almost any hospital in eastern mass, they're on Epic and can very easily query patient data from the other Epic shops. Patient's doctor's office on Athena? Well, it's a bit tougher but my hospital is about to go live with a data archive project designed to pull legacy data from each practice into the Epic record (a monumental undertaking, I can't wait to not have to work on it anymore).
Most of the complaints with Epic are that it's too workflow-based, and providers and other staff that came from paper or other EMRs suddenly had to rethink their basic daily tasks. An order triggers a scheduling request. A scheduling request triggers an appointment, an appointment triggers pre-visit charting, etc. If the doctor/NP/whomever didn't put in that referral order, nothing else can happen. A lot of our providers were ridiculously obstinate and refused to do this, and blamed the system.
Is this the best way to do it? I have no idea. But from what I've seen from the previous generations of EMRs, it's a monumental step in the right direction. I'm hopeful that the next generation brings another substantial leap forward and we can kind of go back to that patient experience where the provider is actually paying attention to them and not clacking away at the computer.