Comments on: How can EHR be made to work?

By: Why I differ with Mr. Halamka’s EHR strategy « EHR: How difficult is it without the correct strategy?

Thu, 01 Oct 2009 17:20:04 +0000

[...] comments HealthBlawg on How can EHR be made to wo…Paul Roemer on What should be the role of the…Paul Roemer on When you’re in a hole, [...]

By: HealthBlawg

HealthBlawg — Wed, 30 Sep 2009 15:01:56 +0000

Whither meaningful use of the certified, interoperable EHR?…

Paul Roemer asks: Have you ever been a part of a successful launch of a national IT system that: required a hundred thousand or so implementations of a parochial system? has been designed by 400 vendors? had 400 applications based……

By: Scott Nicol

Scott Nicol — Sun, 27 Sep 2009 22:53:52 +0000

Have you ever been a part of a successful launch of a national IT system that:
* required a hundred thousand or so implementations of a parochial system
* has been designed by 400 vendors
* has 400 applications based on their own standards
* has to transport different versions of health records in and out of hundreds of different regional health information networks
* has to be interoperable
* may result in someone’s death if it fails

Yikes, I’m almost tempted to inquire about that ad in the back of Popular Science on Locksmithing, but I do like a challenge.

Kinda brings it all home. I’m wondering what the folks at NextGen and Greenway are telling their clients when it comes to the upload part of all this. Am I buying the capability as they envision it today or as it *will be* at some future set of temporal coordinates. Perhaps that’s part of the wonderful 20% service contract.

Probably the best example of what works today in healthcare is the various registries. These proprietary systems all have one thing in common; they accept a single data set in a very rigid format. If you want to play, it has to be by their rules. If your data is hosed up, it gets bounced. There are tools that help you scrub it before the upload, so rejection is rare. Of course this is all batch-based, and an EMR to EHR transfer should be more of a real-time single record arrangement – but other than that it’s quite similar.

As far as examples of a good transport system, I think you must turn to a Google or some other entity that can handle mass transactions with low risk of failure. The Automated Clearing House (ACH) is an electronic network for financial transactions in the United States. This system moves more than 10 billion transactions per year – albeit in batch. But if you look at the proprietary sub systems like the ATM networks, thy work very well and appears to be real-time, otherwise you could take a trip around town to 10 different ATM’s and clean out your account a few times over. I’ve never had a bogus transaction via ATM in more than 25 years. Credit cards are another story of course.

Yes it can be done well and at a reasonable cost, but rigid standards and simplicity in the design of the transport and storage architecture must be at the top of the priority list.

By: squarie

squarie — Sun, 27 Sep 2009 20:49:42 +0000

my 2 cents
Phased approach
Start with the data points which are available in existing data streams (ie HL7)** yes here someone will have to decide on the format. we know each facility is a “snowflake” but the data will have to be standardized in/out of the PTP nw
Feed a peer-to-peer network which is searchable based on several layers of matching criteria (exact and fuzzy)
Authentication and logging is handled within each facility with proper audit measures in place.
What’s next? Ideally something like medication or radiology reports, not sure which would be easier!
labs data
cardiology waveforms
other stuff i know nothing about!
free-text narratives.

I’m not sure a nationwide repository is necessary,it would be a duplication of data. Perhaps it would open a business model for EHR backup services.
yup, still lots of unanswered questions.