With this Oracle roll-out in Sweden, and the Epic roll-outs in Denmark, Finland, and Norway, I wonder if anyone has…
Morning Headlines 3/23/23
Veterans Affairs shares statement on electronic health records rollout
Mann-Grandstaff VA Medical Center (WA) shares improvements made to the VA’s Oracle Cerner system, and stresses that the VA remains committed to fully implementing the same system as that used by the DoD and Coast Guard.
SonderMind, a mental healthcare company that helps match patients with therapists, acquires the technology of virtual mental healthcare provider Mindstrong, which is in the midst of winding down operations.
There in the sky! Not a bird, not a plane, it’s VA’s old VistA system!
The VA will shift its legacy VistA system to the cloud to ensure its functionality and accessibility while the new Oracle Cerner system is optimized and implemented across the department’s remaining facilities.
Re: VIsta ‘lift and shift’ (spell check let another spelling through that is probably a better description) — so WHICH cloud platform are they moving to. Hopefully not the Oracle cloud — that would put a perverse incentive to not finish the migration.
But on the technical end, lifting and shifting (remembering the “f”) does not enable interoperability or cross VISN communication. The individual VISNs have diverged too much for a user to be able to use their solution and bring a patient record over from another VISN — the variance is difficult enough using the clinical data for research, where they take the time to map all the relevant variances — but real time migration of a patient chart?
If I were the VA — and I am not — I would ask the vendor to prove that 100 records, of at least 10 years of treatment, can be migrated from a single VISN to every other VISN. Rinse and repeat with each VISN. Validate their receipt against the original record for: accuracy, completeness, and functional usability upon receipt. Take ten records from each VISN, and transport them serially to every other VISN in the network. Repeat the validation specified above.
It will fail, and the number of defects generated will shut off the “interoperability” component of the proposal. If they did run this testing battery they may learn more about why the Oracle/Cerner implementation is so troubled.
But, it will cost more as the VistA architecture was never designed for a cloud solution.
I believe that AWS is the VA’s cloud provider for VistA. It will be interesting to see how they implement the VISN-customized versions of VistA in what is ordinarily a single-instance environment. I doubt they will attempt to standardize all over again with VistA the way they are attempting to do with Oracle Cerner since that would not only be redeveloping VistA, but putting hospitals through a painful process conversion even before they get to Oracle Cerner.
From the article:
“Reginald Cummings, the deputy chief information officer for infrastructure operations explained on the panel, he’s moving VistA to the cloud. That’s significant because the VistA system has traditionally operated as multiple instances, each running separately at each VA facility. More than simply “lifting and shifting” VistA to a cloud, to use the slightly disparaging cloud-talk term, Cummings said, they’re also “containerizing” it. Containerizing refers to a way of packaging an application together with the resources it needs to operate, such as the operating system itself, the storage and interfaces such that it can run in a more secure and portable manner.
In the cloud, any VA facility could, in theory, access any VistA instance more easily than it can now.”
That last statement seems to me to imply that they would at a minimum provide access. But, I think you are right, they aren’t calling out “one patient, one record” in any of the articles I read.
It would still be an interesting exercise
I read the FNN article by Tom Temin. That’s not what I got out of it. They aren’t attempting to improve interoperability, not at all.
“… the VistA cloud move meant VA would not be updating or modernizing VistA.”
It’s very possible to move multiple VistA instances from their legacy hardware hosting to the cloud. Likewise, containerizing them should be relatively easy. The key concept is this: You don’t mess with the system’s internals! Likewise, you don’t attempt to merge the multiple VistA instances together. You don’t reconcile them, and you don’t change the way they interface with each other.
The project Best Practice is this: Save your energy needed to reorganize the insides of a system, for the new system (Cerner). The legacy system should have the absolute minimum possible done to it.
I’ve done work like this and am familiar with the issues involved. You wouldn’t believe how much effort was required to confirm, with 100% assurance, that a logical data element was stored in a given field, in a multi-instance system. In IT we often do things to a 95% confidence level and then respond to any issues that arise. For a system like this? You don’t dare do that.
Leave the legacy system internals alone. Messing around with those just stirs up legacy issues and does not position you for the future.