Mike Gray continues to evangelize DICOM Archives (his latest here and here-registration required) - which I think has interesting parallels with medical devices at the point of care.

Imaging modalities are obviously the medical devices, similar to patient monitors, point of care diagnostics, pumps, vents, etc. at the point of care. The Picture Archiving and Communications System is equivalent to the servers that many of these medical devices communicate, storing data for surveillance, alarm notification and retrospective analysis. The medical device data stored on some device servers are of little value beyond the immediate future (like full disclosure waveform data), while other data has greater long term value (like the CQI data collected by "smart" pumps).

As the servers connected to medical devices proliferate, hospitals will increasingly face the prospect of migrating data from one vendor's server to another. I would imagine some vendors will refuse to offer any migration assistance - what's in it for them? In the radiology world, many migrations are handled by third parties. Here Mike describes the latest evolution in the PACS world:

There are a few "Enhanced" DICOM Archives entering the market. Their
enhancement is the ability to modify the DICOM header created by PACS A
to make it compatible with PACS B, and vice versa. The library of
translations used to accomplish this multi-vendor exchange is
ironically the result of real-world data migration experience. (Isn't
it nice to see that there is another return on the painful investment
in data migration?)

Mike uses the term "tag morphing" to describe this translation between proprietary data formats contained in otherwise DICOM compliant records. Imagine tag morphing between smart pump servers for migration. Even more radical, imagine a tag morphing server that simply talks to multiple vendor's medical devices just like a PACS.

It gets interesting when a different PACS queries and retrieves that
same study data. The Enhanced DICOM Archive is capable of translating
that study data to make it compatible with the requesting PACS. Most of
the Enhanced DICOM Archives that I have looked at are also capable of
using the new study Order to trigger a relevant prior pre-fetch and
auto-route the priors to a PACS that is incapable of automatic Q/R from
a foreign archive. The Enhanced DICOM Archive is also capable of
setting a study status tag, so the target PACS will understand that
this is a prior and not a new study. You wouldn't want that mistake.

So the working combination of a PACS from one vendor and an Archive
from another is real. But why would anyone want to do this? Well that
translation capability would come in handy if the Health System found
itself owning several PACS from different vendors. It would also come
in handy when it came time to replace a PACS, because data migration
would not be necessary. And that makes the enhancement worth the
investment and then some.