The recent actions in Pakistan brought back memories of the 9/11 events when I was literally a mile away from the Pentagon. I was on the 11th floor at a DICOM working group meeting with a clear view of most of the city and could clearly see what was going on. I saw first hand the confusion, smoke, and semi-panic in Washington DC on that early morning. The first reaction of course was to call home, as I knew they would be trying to find out if I was safe. Of course, cell phones did not work as the circuits were over-loaded. It took me about an hour to realize that the landline might work so I went back to my hotel and made my calls to my family from there.
The second challenge was to get back to Dallas from DC before the weekend. I thought, no problem, we also have trains don't we? I could not get through to Amtrak and decided to walk to the main station. Well, I found out that there were trains, if I was willing to be on the stand-by list for up to three weeks from that date. That left me with no other option than to rent a car. There were plenty available, although I had to negotiate a one-way rental penalty of $1,000 or so when I turned in the car in Dallas. Lessons learned: have a back-up for your back-up, in my case, renting a car for the train back-up. That is why I have a back-up of my laptop computer and also keep a copy off-site somewhere in a cloud just in case.
Many institutions keep multiple copies of their images available for the first days to weeks at a modality. Some PACS architectures have gateways that store images for a certain period of time. Many also archive a copy off-site at a Vendor Neutral Archive, and then also have a tape-back-up that they put into a vault off-site. Determining what to duplicate where, when, and for how long should be based on a risk analysis. This analysis should include the unimaginable.
This is especially important for those components that directly impact critical care such as the ER. Most ER’s by now have multiple CR systems, and if the volume does not justify large units, a single plate reader should be sufficient as a back-up. You might even keep a laser printer, and make sure there is sufficient film to go with it. If the printer is connected to a network, keep a direct patch cable around that allows the CR to connect directly to the printer in case there is a major network failure. Another solution is to have a CD burner in the ER that allows images to be stored on exchange media that can be transmitted over the "sneakernet" to the radiologist who can review them temporarily on his or her workstation.
One should use common sense, however, and do not go overboard, which is where the risk analysis comes in. I have seen institutions where the images are still archived on removable disks on the CT or MRI, just because it makes the technologists “feel safe” or just because it “has always been done that way.” If over a period of, let’s say one year, no one ever asked for these to be retrieved, one might rethink this and possibly eliminate it from the workflow. I also have seen an institution where there were five copies available: one copy at a gateway, one at a local server, one at the archive, one at the web server and one at an electronic medical record server, which was functioning as a Vendor Neutral Archive provider. This also could use some analysis.
In conclusion, make sure you have a back-up and redundancy so you are never “stuck,” and also make sure that there is a back-up for your back-up in case the first back-up fails as well, however, don’t go over board. This is what I learned when traveling, and this is what you should consider when doing a risk analysis so you won’t be caught scrambling for a solution.