Friday, November 30, 2007
MedQuest operates 92 imaging centers around the South and West, and they went looking for a PACS to improve productivity for their rads and clinicians. MedQuest is owned, by the way, by Novant which owns a significant chunk of healthcare facilities.
MedQuest operates totally in the outpatient space. I'm sure they were hurt by DRA-2005, but obviously, news of the demise of such operations is a bit premature. As was predicted by many folks, DRA simply forced such companies to operate more efficiently, and they chose Amicas to help with that task.
I wonder if MedQuest is considering running all 92 sites from one server? It can be done, you know....
Agfa got a big one too, but it is a little more in the realm of a potential big deal. This article from HealthTech Wire notes the "three year, multisource" contract to provide Enterprise Image Management Solutions (EIMS) to the "alliance's more than 1,500 member hospitals and 49,000 other healthcare sites." The relationship goes back to 1995, the Dark Ages of PACS.
If I understand correctly how Premier works, this is not a guaranteed sale of 50,500 installs (I don't think all the vendors put together could handle that), but rather simply a pricing deal, offering significant discounts to the member hospitals. Feel free to comment if I'm wrong about that. Still, it's nice to be on someone's good list for a change, right?
Tuesday, November 27, 2007
Software-wise, Symbia uses the Leonardo, I mean eSoft, no, Syngo, well, today it's the Molecular Imaging Workstation, MIW, for processing. I gave the Siemens folks grief about the lack of mouse wheel support, and they say it is coming. Eventually. I also groused about some problems I've been having with the PET/CT software, and they suggested I use the more generic 3D Fusion program instead of the older nuclear med version, and I will give that a try. All of this applies to SPECT/CT as well. I inquired about a thin-client, so I could properly look at PET/CT or SPECT/CT when I get a panic call from one of my colleagues. (It's good to be loved, and to have job security to boot!) So far, there really aren't any good (or cheap) options. One involves creating a virtual machine-type access to the Leonardo (I'm sticking to that term, sorry) and another involves loading the viewing software thick-client style onto the remote computer. A thin-client is in the works, but not available for the foreseeable future. So, I guess I'll have to rely on Agfa to do this, or perhaps we'll get TeraRecon, or Voxar thin-clients which will accomplish the same thing, as well as giving us a cardiac CT solution.
Coming to the Amicas booth is a little like going home; most of the faces are familiar, although the products continue to evolve and grow into great things. Version 5.0 has been released, and I've played with it on my test-server, although we can't put it on the production server until our Amicas hospital does the same. The 5.0 upgrade involved changes in server software, moving up to Microsoft Windows Server 2003, but this paved the way for the move to 5.5, and ultimately 6.0.
The big news in 5.5, due out early 2008, is a huge improvement in worklist functionality, and in particular, the way status is handled. Right now, an exam can have "status" in one dimension. It is unread, STAT, reported, whatever, but it can have only one such designation. The Amicas folks tell me that some discussion in the last Advisory Committee meeting led them to the fact that status doesn't have to be uni-dimensional, but it would make more sense to mold it to our workflow. Something can be STAT, and need extra images, and be from the ER. Thus, we can assign as many dimensions as we might need to cover the various intertwining workflows we follow daily. This is fully configurable by the site, and thus can be tailored to one's particular needs.
There are several new worklist-related tools for working with the ED, including an ED button on the worklist to go directly to the ER's preliminary report (if you can get your ER docs to cooperate, which we can't) and then there is an included mechanism to flag a disagreement. This then goes into its own worklist block, so the ER doc can be alerted, and maybe embarrassed too.
But the new worklist approach goes way beyond taking care of the ER docs. It works exquisitely well with RadStream, which I mentioned in an earlier post. Let me elaborate, based on a discussion with Dr. Mark Halsted, the radiologist from Cincinnati Children's Hospital who developed the program. (For more details, have a look at this article from Imaging Economics.) When the tech verifies a study, he or she has to fill out a brief checklist that helps assign an "acuity score" which is calculated by the program. The worklist then automatically sorts the studies based on this acuity score. The best part is that the more urgent studies percolate to the top of the list, and if I'm reading from a long worklist, the next most urgent study will come up after I close the last. And, if I'm using the Worklist Accelerator, which caches stuff in the background, I am able to read seamlessly, not missing a beat while progressing to the next-most-critical exam.
Now, it's after I'm done reading that the real impressive stuff occurs. If I need to convey a critical result, I simply click the "go to next study" button, instead of the "mark dictated and go to next study" button. A window pops up allowing me to declare this a study in need of rapid reporting. The study then drops into a worklist display block on MY computer showing that it is in progress of being called to the ordering doc (or however one wants it to read), and once taken care of, it will appear in a block noting that the critical results have been communicated. Of course, there is the complete audit trail for JCAHO's perusal, or (Heaven Forbid) that of lawyers. This is not an idle point: I have heard of cases of rads being sued for supposedly not telling the referring docs that there was finding on a patient's study. A preliminary entered into PACS might be helpful here, and no doubt that was done in these cases. But keep in mind that even calling the doc, telling him the findings, and documenting this in his report might not be enough; the doc could simply call him a liar. You see, even with a phone-call, supposedly the epitome of service, there would be no independant audit trail.
Some people prefer automation, some prefer the human touch. While Vocada Veriphy uses the former, RadStream uses the latter. The Vocada rep was almost snide in his support of automation over humans, but RadStream's use of a human in the communication equation makes rather good sense. Yes, one can automate the rules to contact Dr. Big Referrer, and Vocada will follow them. But, there is a limit to what a computer can do, and how much your referrers will appreciate the computer texting them or emailing them, or whatever. Frankly, this could contribute to the impression that radiologists are lazy, and would rather push a button and have the computer do the dirty work, instead of picking up the phone and calling them. With a human Communication Specialist (RadStream's fancy word for operator), there is the possibility of human interaction, and some on-the-fly decision making such as the referrer himself or herself needs to hear this report, not just his nurse, and the like. Yes, Vocada's automation might have some fallback rules for what to do if no one responds to the message, but it might not be enough.
While there is no set rule at this point, I believe that JCAHO prefers critical communications to be from person to person, not machine to machine. It certainly tells the clinicians that we are making every effort to get them the report as fast as possible, and as intelligently as possible. And it keeps everyone honest.
Amicas has a complementary product called VisionReach that takes care of the less critical results. Basically, the clinician can be emailed or faxed or whatever he/she chooses with the results, and the contact method can be chosen by the clinicians themselves. The emailed report can include JPEG's of key images, and there is also a link to the viewer itself. There is even a tab to allow the clinician to order a followup exam. Nothing like making it easy for them to send stuff to your site!
Carried over from Version 5.0 are a couple of additional features. The embedded Voxar 3D program has more functionality, with color presets and the ability to capture images back to the PACS system. There is also a way to easily cycle through prior studies by hitting the "-" key. Amicas only allows one prior to be visible at a time, and currently the only way to cycle is to bring up the prior list and select the one you want. This new little function will speed things along significantly.
Reading digital mammography is now FDA-approved with 5.0, but Amicas presently does not have dedicated mammo tools. Rather, one can use the regular viewer (and 5MP monitors, of course), or purchase the iRead program from Cedara. With Merge about to go under, I'm not sure how good an idea that really is at the moment.
I did get an extensive tour of Version 6.0 (I almost slipped and called it Impax 6, which would have been totally unforgivable.) I won't reveal the specifics as yet; you'll just have to wait and see. But do trust me on this: Version 6 will be one of the finest PACS GUI's ever offered, combining the easy usability of its ancestors with one heck of a lot of power. Believe it.
My advice to Amicas is quite simple...don't release this until it is ready for prime-time (unlike another Version 6 I know of), but get it ready soon. And don't sell it to GE.
The true test of a product, I think, is whether you would buy it again, having lived with it for a while. I've lived with Amicas since 2004, and I would buy it again in a heartbeat. It works, and it keeps getting better. No, I am most emphatically not on the Amicas payroll, but I am very enamoured with their products. Because they let me work the way I want to work, not the way they want me to work. It's that simple.
Monday, November 26, 2007
My friend Rick from Intelerad was gracious enough to show me some of the improvements to be found in the latest and greatest version of Intelepacs. There has been no huge change, but a nice system is even nicer today.
Presently, my group uses the InteleViewer stand-alone program at our site that has no PACS. My boys are very pleased with it. It helps that one of them trained at ProScan, and learned speed-reading from Dr. Pomeranz himself. Dr. P. is an Intelerad user, and even was featured doing live readings from the Intelerad booth. I wonder if any fellows back in Cinci overread his live readings...
The Intelepacs worklist is a little spartan, being mainly a table of values, but there is color coding for status, and the colors can be changed at the user level. Great way to confuse your partners if they are foolish enough not to sign out of their account! Each column can be moved about, and one can sort by the column value or search for something specific within. To actually read a block of studies, simply click and drag over the ones you want, and they are sequestered to you. One can apply various worklist filters, although there is no natural division that I could see, such as a divider between one day and the next as Amicas does it. Still, this is a very usable setup.
The Intelerad folks have done something impossible; they have topped the Amicas spine labelling program. Well, it isn't a huge triumph, but they have developed a way for the computer to automatically realize if you are advancing cranially or caudally (up or down for those in Rio Linda) after starting at a particular level (which you still have to specify.) It works well.
Tab thumbnails have been added for displaying prior studies, a nice touch. There is again Voxar or TeraRecon integration. Presentation states can be saved, although one customer voiced concern about referring docs seeing how he left the window/level settings. Well, just because you're paranoid doesn't mean they aren't out to get you, I suppose, but I'm really not too worried about that.
At first, I thought the hanging protocol setup was as user unfriendly as Sectra's, but then Rick realized that I wasn't familiar with the "Capture" function. This simply captures (duh) the on-screen layout, and then you can tweak it with a complex system that is pretty much like Sectra's. But it works, and in the end, it's pretty much like Amicas' which I'm very fond of.
There is a method for ER reports to go back and forth, and an audit trail thereof. Voice clips are now included in the interface.
Bottom line, this good program has become even better, which would help to explain it's #1 KLAS rating. Not that I believe KLAS ratings, but.....
We had the chance to look over some of the new stuff going into Impax 6.4, which we should receive at the end of January, 2008. There are some nice interval tweaks and fixes, which will improve the Impax experience. First and foremost is a marked improvement in the CT/MR multiplanar navigation. This will include auto-linking, and active targeting. This eliminates a significant complaint we had with 6.2, where triangulation was trying at best. Window/level and zoom parameters can also be linked. Version 6.4 will support Vista, but only if you don't have any Barco monitors. Likely a driver issue at this point. Comments, Barco? The simple search function is tweaked, and is now a bit more customizable, and usable overall. A glitch in deployment of Voxar 3D has been fixed, and I will no longer accidentally activate Voxar 39 times during the course of my reading. There is also a nice new toy, Smart seek web access. This appears as a tab or tabs in the study information are of the information screen. Basically, it is a smart browser, connecting one to various search engines such as Google or MyPACS. It is set to search on the body part or history, or what-have-you of the case up on the viewer. Could be a time saver in practice.
Agfa is working on an EMR display, called the Impax Clinical Module (we were assured that the button labeled ICM could be renamed), which requires a server box called the Enterprise Clinical Dashboard. No, I don't know how much that might cost.
A rather important new option is the incorporation of Vocada's Veriphy critical result reporting program. What Vocada (now owned by Nuance/Dictaphone) lacks in spelling ability, they make up in usefulness of their product. Veriphy is an automated system that will follow pre-set rules to notify our illustrious referring clinicians of a critical result. One can set the system to send a text with a phone number to call to get a voice clip, for example, or even send an email. This step is documented, as is the clinician's response, so the JCAHO requirement for an audit trail is fulfilled. Since this is about the only way to graft such a system (including voice clips) onto Impax, I was impressed. There are similarities to Amicas' RadStream, which the Vocada salesman poo-pooed, but in some ways RadStream works better, as it reprioritizes cases based on pre-test criteria. But you can't get RadStream for any other PACS at the moment, so it is somewhat of a moot point. Anyway, while Veriphy is a separate program, Impax will automatically log you into it upon startup.
We were also shown Works In Progress, many of which looked quite promising. Dr. K, one of our finest MSK rads (all of our rads are fine, by the way) was quite intrigued with ORTHOGON, a wizard based orthopedic measurement program. Basically, one selects the measurement to be taken, say a Cobb angle, and the program shows via a big red dot where to put the cursor next. It's easy enough for a caveman Nuclear Radiologist to use. "Follow the bouncing red dot" was the watchword.
Agfa is developing its own PET/CT and Virtual Colonography programs, and both look very promising. The PET/CT will perform autoalignment, and should work with any two datasets, no matter their modality. There is a very nice method of growing a volume of interest about a lesion, which seemed to work quite well for SUV's, as well as volume measurements. There is a report generator module as well. I have two complaints about their implementation, though. First, the blending control involves clicking over the word "Blending" at the bottom of the viewport, then dragging to control. Give us a button or a slider, please. Secondly, they are using hot-spots, areas on the viewport that will control various functions. Again, I prefer buttons. I need to see where I'm going.
The Virtual Colon piece is nicely done too, incorporating features that I have seen in some high-end products like Viatronix. In particular, there is an indication of what parts of the colon you have seen, and what you haven't, which could be critical.
AND......coming in 6.5......spine labelling! We saw a brief video of how it will work, and it will be OK, rather similar to Amicas' version, but not quite there. I asked why it has taken several years to reach this point, and was told that it took a great deal of programming to detect interspaces and such. Which is funny, because most other spine labelling programs just interpolate. Don't overthink things, guys.
So, there are many improvements coming for us Impax users. We're still toggling, though....
Sunday, November 25, 2007
Anyway, IDS7DX, the radiologist version of the new software, looks uncannily like IDS5, which is the version I saw at SCAR (back then it was SCAR) in 2003. Charlotte, the apps person conducting the demo, told us that this was deliberate. Sectra wanted to preserve the look and feel of the product that was familiar to its users, but wished to update the code to .NET, and freshen and "Windowize" the GUI. This they have done, and apparently they have won an award for excellent .NET implementation. I have said disparaging things about Agfa doing similar things, carrying look-and-feel forward in version after version. I'm complementing Sectra on their efforts because they haven't recapitulated esoteric functionality, as Agfa did. Some things work, some don't, and you have to keep very close tabs on your users to find out which are which.
Anyway, as a .NET program, the new Sectra client may be downloaded via a single click, from anywhere. The same client works for everyone. No separate web-appliances need be used. Preserved is the open API construction that allows Sectra to use a large number of third-party modules, such as Voxar 3D and the like. JPEG 2000 compression is utilized as well, and for the Sectra authored volume rendering program, still in beta, Open GL rendering is performed. Nothing like using a $300 commercial card instead of a $2,000 proprietary version!
Worklists are definable at the system, role, and user level (which is what Impax 6.x was supposed to do, but doesn't.) Very complex worklists can be built through drop-downs. The worklist pane of the extensive (and to tell the truth somewhat busy) information screen does not seem to have state-indication (ala Amicas) but you have the entire list of worklists in the pane to the left of the active worklist, and a study can be dragged-and-dropped from one worklist to another. All panes on the information screen can be resized. There is a study-dependent document or comment field. There is also a blank window that can be filled with the html document of your choice, such as a calendar, stock ticker, or Dalai's PACS Blog.
The viewer component is quite similar to the 2003 version, although it seems able to handle large studies with ease. There is the ubiquitious (these days) extensive right-click menu which can be configured to your heart's content, and one can even change the alt-keystrokes for shortcuts. I must admit I still have some problems with this concept, as I did in my old review of IntegradWeb from the late, great DI. Perhaps it benefits those who have a dozen PACS systems and would like to force the keyboard shortcuts and whatnot to match each other, but maybe the industry should find a standard approach? Oh well, I digress.
The viewer is as usable as any, and more so than many. Hanging protocols, however, need work. There is a way to set up custom protocols, but it is rather cumbersome, requiring the use of a series of drop-down menus that contain every possible series description, and then dragging the various series representations to a mock display. Trust me, it is not user-friendly. Sectra apparently assumes that radiologists won't want to do this themselves, and they are correct on that.
Bottom line is that for Sectra, the port to .NET was successful, given that they did not otherwise massively alter a GUI that wasn't bad to begin with. The worklist display needs some updating, and the hanging-protocol setup needs to be revamped, but otherwise, I could like this. More than some other systems I know altogether too well....
I have yet to attend an educational session, but I'm headed straight to the Cardiac MR talk as soon as I finish this little entry.
One major change I've seen at RSNA is literally around my neck; the badges now have RFID tags, which lets the powers that be do some sort of monitoring of the attendees. No doubt this information will be used for proper ends, such as making sure that we actually do attend the educational stuff for which we are requesting credit. Personally, I'm worried that GE is using the tags to locate me when their snipers are in position. Nah, that would be too easy.
So far, I've had a quick look at the Philips Precedence SPECT/CT scanner. The darn thing is huge, and I'm not sure it will fit the room where it would need to go. I'll see the Siemens Symbia in detail on Wednesday, but I already know it's considerably smaller.
I stopped in at Amicas for a brief chat, and got to see the latest implementation of RadStream, the critical results software. It is very well done, and solves a lot of reporting problems we have at most hospitals. I've gone from a skeptic to a true believer on this one. My real appointment with Amicas is on Tuesday, so I'll have more to report then.
I wandered by the Dynamic Imaging booth, noting with sadness that this is the last time it will be a separate installation. I had a chance to talk with several of my friends, who now sport GE badges. I didn't see any weapons in view, so I accepted the invitation to see some of their latest and greatest stuff. Now, I'm not going to do anyone's product justice in my RSNA reports, since I'm going to try to write them on the fly as it were. But I can tell you that their version 3.7 has some great new stuff. There is a PET/CT reading module that is really well done, basically one of the best implementations of such software I have seen, and I can't wait to get my hands on it at our GE/Centricity site. There has also been improvements in the hanging protocol setup, and there are new overlay options that allow, for example, more optimal display of MR Spectroscopy. I even saw a peek of I5 in alpha form, which uses the same viewer, but tries to do what Agfa did on the information screen, i.e., provide all important info at a glance. They even have a "Coverflow" like display of image series. Frankly, this prototype was a little "busy" for me, but they are on the right track.
As for the integration of DI products with GE, this is still in flux, but DI promised me that they will give me a letter to post that will outline the approach they will take. I know what I'd like them to say, but..... Actually, the fellow in charge of integration of the two product lines was sitting at the table, listening to me wax poetic (that means babble) about the joys of Centricity, and especially Centricity Web. I get the feeling the GE brass hasn't a clue about some of the problems I deal with daily with respect (not) to their products. Now they do.
I've been here at McCormack for about four hours now, and I've met about two dozen folks that read my blog. Many of them had never met me before, and I hope they weren't too disappointed. Frankly, it always surprises me that anyone reads this thing. I'm just posting what I see, filtered through my biased keyboarding fingers. Really, if I've made someone think, or at least laugh, then I've done my job. If the PACS industry is hanging on my every word, well....no one would be more surprised than me.
More to come from cold, cold Chicago!
Friday, November 23, 2007
I have loved Swiss Army Knives from the first time I saw one as a child. They are fascinating devices, and of course, the more blades and tools, the better.
You may know that there are two competing companies making these little jewels, Wenger and Victorinox. Click HERE for a full history of the Wenger operation, and HERE for the Victorinox story.
Wenger periodically offers an Elite version for collectors and so forth, and this year, their masterpiece is like nothing you've ever seen, at least available for sale. In fact, it is noted in the Guiness Book of World Records:
Here are the specs:
- 85 Implements110 Functions
2.5" 60% Serrated locking blade
Nail file, nail cleaner
Adjustable pliers with wire crimper and cutter
Removable screwdriver bit adapter
2.5" Blade for Official World Scout Knife
Spring-loaded, locking needle-nose pliers with wire cutter
Removable screwdriver bit holder
Phillips head screwdriver bit 0
Phillips head screwdriver bit 1
Phillips head screwdriver bit 2
Flat head screwdriver bit 0.5mm x 3.5mm
Flat head screwdriver bit 0.6mm x 4.0mm
Flat head screwdriver bit 1.0mm x 6.5mm
Magnetized recessed bit holder
Double-cut wood saw with ruler (inch & cm)
Bike chain rivet setter, removable 5mm allen wrench, screwdriver for slotted and philips head screws
Removable tool for adjusting bike spokes, 10mm hexagonal key for nuts
Removable 4mm curved allen wrench with philips head screwdriver
Removable 10mm hexagonal key
Patented locking philips head screwdriver
2.4" Springless scissors with serrated, self-sharpening design
1.65" Clip point utility blade
Philips head screwdriver
2.5" Clip point blade
Golf club face cleaner
2.4" Round tip blade
Patented locking screwdriver, cap lifter, can opener
Golf shoe spike wrench
Golf divot repair tool
4mm allen wrench
Fine metal file with precision screwdriver
Double-cut wood saw
Cupped cigar cutter with double-honed edges
12/20-Guage choke tube tool
Watch caseback opening tool
Mineral crystal magnifier with precision screwdriver
Compass, straight edge, ruler (in./cm)
Fish scaler, hook disgorger, line guide
Shortix laboratory key
Micro tool holder
Micro tool adapter
Micro scraper - straight
Micro scraper - curved
Laser pointer with 300 ft. range
Metal saw, metal file
Micro tool holder
Philips head screwdriver 1.5mm
Fine fork for watch spring bars
Pin punch 1.2mm
Pin punch .8mm
Round needle file
Removable tool holder with expandable receptacle
Removable tool holder
Special self-centering screwdriver for gunsights
Flat philips head screwdriver
Mineral crystal magnifier, fork for watch spring bars, small ruler
Spring-loaded, locking flat nose-nose pliers with wire cutter
Removable screwdriver bit holder
Phillips head screwdriver bit 0
Phillips head screwdriver bit 1
Phillips head screwdriver bit 2
Flat head screwdriver bit 0.5mm x 3.5mm
Flat head screwdriver bit 0.6mm x 4.0mm
Flat head screwdriver bit 1.0mm x 6.5mm
Magnetized recessed bit holder
Tire tread gauge
Fiber optic tool holder
Patented locking screwdriver, cap lifter, wire stripper
Actual Size: 8.75. W x 3.25. L
Weight: 2lbs 11oz
Limited lifetime warranty
Made in Switzerland
So, GE and Agfa...if you're in the process of selecting my Christmas gift, this is what I really, really, REALLY want. Remember, it's the Wenger Giant Swiss Army Knife™ V1.0, Model #16999. Oh, the price? Only $1,200.00. But I'm worth it, right?
Sunday, November 18, 2007
I had known this patient was to be scanned, and she had been extensively counselled. I didn't realize, however, that she was on schedule for this particular day. Imagine my loss of intestinal control when I started scrolling through this study.
While I don't know for certain, this has to be one of the only, if not THE only, PET/CT scans depicting a twin gestation in the literature.
Saturday, November 10, 2007
Above, you will find the uncorrected and the corrected coronal images from a recent PET/CT scan performed on a GE Discovery ST. The patient is a young man in his 20's with newly-diagnosed lymphoma. There is some distortion in height due to the way I saved it, so ignore that. Notice, however, the intense, diffuse, homogeneous skin activity. This is not unusual for a non-attenuation corrected image, but we should not see this degree of activity in the corrected version. The study was repeated the next day with the same result. The patient had not eaten, wasn't cold, and had even been given Ativan. GE says nothing is technically wrong, and the patient must have something wrong with his skin. I asked the Siemens apps people, and they said there was obviously something wrong with the processing. Do I detect a little bias?
I'm at a loss here, folks. Anyone have any ideas???
Wednesday, November 07, 2007
This is a wonderfully unfortunate demonstration of Dalai's Laws of PACS, in particular, Laws I and IV. Law I, "PACS is the radiology department" is quite obvious when PACS is dead. All we can do is look at the CT console for those direst of emergency cases that can't be shipped out. I suppose we could film the hundreds of CT slices....nah. Because Dalai's Fourth Law says, "Once PACS, never back." We are not going back to film. Ain't gonna happen.
We can do little but wait for AT&T to fix the communication problem. We were told that could take from 2 to 8 hours. Right now, we're at 9 hours and counting. What should we have done here? This is a situation that is foreseeable, but not easily solvable. We have backups upon backups, but we don't have a backup for the most critical component of all in a web-based system: the Network! While it is the Internet that failed here, we would be just as dead if our LAN failed, so I'll group them together. In today's case, however, it was obviously the WAN that croaked. At this point in time, there aren't too many alternatives for broadband access. In my town, we have DSL, run mainly by AT&T, and cable, run mainly by Time Warner. There are various permutations with MetroE's and Frame Relays and such, but ultimately, all of these require physical lines, wires, fiber optic cables, or whatever, and those can be cut. The only other possibilities would be satellite internet, which is relatively slow and expensive, or a direct microwave connection to an ISP, which would be fast but still expensive. (No, I don't have time to pull the numbers, being on call and all, but trust me on this.)
I'm thinking WIMAX, the promised nationwide WiFi system, might solve this problem. Put an antenna on the roof, and you're connected. It would certainly have avoided an entire hospital being shut down by a ditch digger. Although I suppose a lightning strike or other such event could knock out WIMAX as well. Maybe we need to go back to dial-up?
The only good to come of today's experience is that we are able to tell the ER docs to think long and hard before they order useless studies. OK, we didn't use the word "useless" but the message got across. Maybe the mentality will sink in. Probably not.