Ore. hospital loses 5,000 X-rays in hard-drive crash. Apparently 4 out of 5 hard drives failed, and the back-up wasn’t backed up. Still that’s nothing—I lost a whole year of downloaded soccer videos when I knocked over my external hard drive. And I still was too cheap to buy a back-up to the back-up.
Of course some of you might think that the radiology images are slightly more important than say Chelsea’s game with West Brom….
Categories: Uncategorized
I’m not so sure they’re distrustful of the capabilities — they just don’t want to pay for redundancy, and when some clueless MCSE will tell them that with RAID5 they already have redundancy, they conclude they don’t need any more.
See, with “transparent pricing” in PACS systems, the cost of everything is apparent, but of course the benefits are not. So it is true they can have a system that “works” for one price, and they can have one that will work longer (alothough we can’t say exactly how much longer) for a bigger price. So lots of `em will buy the one that “works”.
In this particular case it looks like the hospital had bought a redundancy solution, but hadn’t done a proper job of acceptance testing.
t
I design PACS implementations for my company’s customers and this is one of the things I always try to drive home. It frequently amazes me how much trust people give hard drives and how distrustful they are of data redundancy/back-up systems.
I have taken to putting disclaimers about data redundancy and back-up systems in large type, just below where the final price is and above where the customer signs, on projects where someone at the customer’s site has forced the removal of those aspects of the project. (Unless the customer has a data redundancy/back-up schema or system and it isn’t purchased from me, then I reference it so everyone knows what is to do what.)
I found a cached version of the story that says
He goes on to say “It wasn’t our fault”. But of course it was — his CIO never tested the system to be sure it was operating correctly. We are reassured now that “The hospital now receives daily reports on the status of its archiving system”. Hmmmm. A little horse and barn-door.
I used to work on systems like this, and exactly this possibility kept me awake nights. There is a great deal that can go wrong with data migration, but it is pretty easy to tell when it has. There is always a window during which the system is vulnerable to data loss; when there truly is only one copy in one physical location. This can’t be avoided. But the window is measured in minutes or hours, not weeks or months. It looks to me like this train wreck was caused by two groups asleep at the switch: GE and the hospital.
t