Sunday, August 18, 2013

Demos Among Friends

We got word, rather late in the game, that one of our referring clinician groups was thinking of changing their PACS. With a little nosing around, I was able to get myself invited to their meeting, where they would hear from the two vendors selected by their head of imaging. The group was only dimly aware that we have our own Merge PACS to sell, which happens to be the same system that is used at our common hospital.

The night of the demo came. I won't name the two vendors because one has a reputation, deserved or not, for taking legal action against those who speak ill of it, and I might as well not name the second. I sat beside the clinicians, their administrator and their IT guy, and pretended to be an interested bystander. In fact, even the clinicians didn't know I was there with a PACS in my pocket.

I'm going to skip over one presentation, as it was kind of lackluster, and didn't strike any chords with the clinicians. The presenters were knowledgeable, but there were certain gaps in the information, such as how the software actually runs on the client machine. Thin client? Thick client? Java? dotNet? Ummm, we'll have to check on that.

The other vendor did a much better job, a fantastic job, really, presenting the product from head to toe, and enticing the clinicians in the process. More about that in a moment.

After the salesfolk left the building, I had the chance to somewhat sneakily and sheepishly present the possibility of selling space on our Merge PACS. By this point, the air conditioning had been shut off, and the computers were about to automatically shut down. I had a brief, rather amateurish powerpoint, that I went through in 30 seconds. I then asked if the docs had any questions about Merge. They were familiar with it after all, as it is also used in our common hospital.

"No. We don't need to see a demo. We know that system. And we hate it."

Aghast, I asked why. Well, it doesn't do this. And it's hard to do that. And this looks funny.

And none of these were at all accurate. The complaints with the hospital's PACS all stemmed from the fact that these guys had NEVER been trained. Ever. Not for three minutes. Before the systems went into hibernation that night, I was able to show them levels of functionality on the Merge PACS that they had never seen before. And they were impressed.  I've got my foot in the door, at least, for their PACS services.

But let's get back to the vendor with the impressive presentation. Behind the slick, well-rehearsed demo is a product that tries very hard to be all things to all people. I frankly cannot think of a function it doesn't have.

I've used the Swiss-Army Knife analogy before, but...

Below is a photo of the Wenger Giant (which I managed to find on eBay for a fraction of its original price):

Swiss Army Knife

Here is the Victorinox SwissChamp XAVT, which I also found at deep discount on eBay:


Both of these are huge, have just about every tool available from their respective manufacturers, and are as solid as a battleship. You wouldn't want to actually carry them in your pocket (not even a possibility for the Wenger) but if for some reason you had to press them into service, you would have no doubt of their functioning perfectly.

Contrast the Swiss versions with this one from China. This multitool can be had in bulk for $1. Yes, $1.00.

15 IN 1 multi knife,multi function knife,pocket knife,survival knife tools,promotional gifts,multi knives,knife w/ torch,camping

Does it work? Well, probably. Most of the time. I'm sure the blades and the tools and such do their job, until they don't. What did you want for $1? (The Swiss museum pieces go for $2,000 and $500 retail, respectively, although smaller and more functional versions are more on the order of $50-$100.)

The properly-demoed PACS cited above is impressive for two reasons. First, it is cheaper than many, and second, it has every function, and every element of customizability known to man. The customization pages contained literally hundreds, if not thousands, of possible settings and changes.

But, remember Feature Fatigue. We end up using only a small subgroup of that vast feature set about 90++% of the time. The rest has the potential to get in our way.

And, in a mission-critical piece of machinery and software, we expect 100% uptime. While that might be unreasonable, PACS is used for emergent patient care, and it darn well better work at least 99.99999% of the time.

Sadly, the particular vendor with the flashy, inexpensive product does not have the best track record, nor should we really expect this with their price-point, although I will be so bold as to expect it anyway.

Some reviews of this particular product on AuntMinnie.com, Reddit, and personal anecdotes have not been particularly positive:
PLEASE have the admin or the clinical lead on the PACS purchase contact me directly. I'm sure the facade looks good...in fact it was this and the cheap price that sold us...

However, when you clicked a button on the RIS or PACS, many of them didn't work. Their main tech/engineer left them over a year ago and he was the only one who knew what was going on with the product as he wrote most of their code.

... I could tell you about an event caused entirely by this vendor that would make you shudder, but (they have) shown me they are much more concerned about litigation than service, so I'll save myself a frivolous lawsuit and leave it at that....
I have used no less than 6 PACS and (this vendor) is by far the worst. It's a shame too, because I prefer the customizable PACS... The idea is great but the execution is just awful. I keep waiting for a PACS that behaves like Windows---with a solid default structure with a good amount of features BUT the ability for tons of personalization, but in an intuitive manner. (This vendor) fails on intuitive and fails on solid default structure.
And there is a rather long message on Reddit which I've truncated:
Please forgive the ranting. I'm STILL livid about this and it's been roughly a year now, since the incident.

Fine when it works. Worst. Disaster Recovery. Ever. By any vendor, in my entire 20+ year career in general IT...

Overall, they have been fine. Not the highest end system, but fine for basic PACS/RIS operations. The admin interfaces are fairly simple, the user interface is also fairly simple and easy-to-use. We've not had any interoperability issues with it and it seems to comply with all PACS/RIS standards...

My experience with them train wrecked, when they outright lost patient data, then stonewalled my entire organization RE: the status of recovering data that they lost and answering any questions even vaguely related to what happened, why and how. As far as I know, it was an anecdotal experience, but they handled the entire thing very poorly...

The backup mechanism did work partially and we spent 2 weeks reloading 2 TB of data onto the server, through a USB 2.0 connection. While I'm not sure of the exact count, we did lose thousands of studies, because not only did the server-to-server sync fail, but the offsite sync failed, "at some point, prior to that" and wasn't noticed by anyone at (the company). AWESOME!

...They actually argued, "Yes, we dropped the ball entirely and lost data and kept you down for weeks due to the magnificence of our negligence, but we did support the site during the past 3 months, so we feel only a partial refund of the monthly fees are appropriate". There were other costs they did agree to reimburse the hospital for, but I have no idea of the specifics, I just know they argued the nickels, while holding their hats in their hands.

I personally and professionally fault them for the piss-poor way they handled the situation, more than the situation itself. Imo, it was grossly negligent, but ultimately, things do happen. How we handle these things is much more telling of an entity's true nature and whomever made the decisions in my situation showed me that, as a company, they were much, much more concerned about legal liabilities and how much it was going to cost them, than recovering the site/data, as fast as possible.
Note to legal department of this PACS company:  I am only copying and anonymizing information about your product posted by others. I don't have personal experience with your product, nor do I want to have any.

The clinicians are still deciding. But here again, a decision is being made about PACS systems by those who either don't use them (IT, administration), or use them in a secondary capacity (clinicians). We, the radiologists who will be reading studies off of the contraption selected need to have a say. Now, granted, I'm biased in this particular case in that I want to sell a system. But I've chosen for myself something that works well for me and the members of my group, some of whom are not quite as computer-savvy as I am, and I think I chose well.

The slicker vendor took great pains to emphasize how his company "listens to radiologists" and remakes the product to fit their needs. Which leads to 3,000 buttons and 5,000 options. I've seen this before. And I've also seen situations wherein PACS companies listen to their loudest and largest customers, and alter the product to fit that particular customer, screwing the rest of us. None of this needs to happen.

In my next article, I'm going to build on the work published by Brad Levin from Visage in on the SiiM website, and I think you'll find it is quite germane to this discussion.

2 comments :

Dr. Sardonicus said...

This is why purchasing expensive software makes me very nervous. The sales guys can hide anything and you didn't know it was there (like bad service).

You say system #2 was overly customizable. Recently on AM, a poster put a list of features he would want in his software. To my mind these were all quite desireable and beyond the capabilities of any PACS I am aware of (my sample size is quite small). Does system #2 address any of these?

http://www.auntminnie.com/forum/tm.aspx?m=394532

Anonymous said...

Brad Levin said.....Great post Dalai, and I selfishly look forward to seeing the followup! Software will never be perfect, so it's healthy to have skepticism after you've narrowed your selections. The kicker should be, make the vendors prove that they can do what they do with a demo on YOUR network, with YOUR data. We recommend customers forget about doing an arduous RFP. Bring in vendors to present, narrow the list of viable vendors to 2, and make them perform a live pilot to make your choice. It'll save you months of back and forth, and you'll have a better, more confident outcome in the end. Also, the link to the SIIM website in your post goes to my original commentary earlier this year. The most recent commentary, SIIM survey results and infographic may be found via this link: http://siimcenter.org/commentary/2013/08/survey-results-paint-mosaic-problems-and-opportunities-radiology. Thanks!