PACS investment gains you a foothold on competition.

Find out who, what, when, where, why and how before you purchase a PAC system or you may end up with an expensive system better suited for playing games rather than diagnosing patients and transmitting files.

There several things to keep in mind when deciding the who, what, when, where, why and how of buying picture archiving communications systems (PACS) and teleradiology. The adage: "He who hesitates is lost," more than applies to the purchase of PACS and teleradiology today.

Why is PACS practical today?--Managed care and capitation are playing an incredibly strong role in healthcare today. Providers are demanding more bang for their buck. Gone are the days of 24- to 48-hour report turnarounds. Two hours is the new norm. Couriers can't provide that response time and double shooting film is expensive in an era of cost containment. It makes little sense to buy a spiral computed tomography scanner (CT) that can do 50 patients a day if the referring clinician can't get a timely interpretation (or even have access to the images).

PACS and more directly, teleradiology, solve that problem. Why the change? Money, pure and simple. The development of moderate cost (less than $60,000) systems that can transmit diagnostic quality CT, ultrasound and general radiographs allow clinicians to net an average of $300 to $400 per day per site in diagnostic fees for a typical rural (or smaller urban) hospital.

How can I afford PACS?--A small hospital generates on average 10 to 12 radiographic series (typically two-view films) per day, a couple of CTs and a couple of ultrasounds--nothing earth shattering and definitely not enough volume to staff a full-time or even a half-time radiologist. If each plain-film study generates $20 per case, each CT $100 and each ultrasound about $65, then each small hospital site would generate around $530 per day. Amortizing the cost of the $60,000 system over three years (786 days, based on a five-day work week), and you get a cost of just under $80 day including communications costs using an ISDN (integrated services digital network) line that will transmit a nominally compressed 2,000-line resolution chest film in under two minutes. Net profit per hospital: $118,000.

Cut the number of studies in half (six plain films, one CT and one ultrasound per day) and you still make a tidy profit. Multiply that revenue generation buy the five, six or even seven hospitals in the network and you can see where teleradiology has the potential to make big bucks.

It is important to note we haven't factored in the cost of the dedicated receive station, which we're assuming in this model already exists. However, even if we factor in $150,000 for the cost of two 2,000 x 2,500 line resolution display workstations--with all the bells and whistles--and divide the cost by five transmit sites, it declines the revenue stream by a mere $30 per site per day. Adjust these numbers accordingly for your situation and it's a win-win situation.

When should I invest in PACS?--The "When" is now (if not yesterday) for the obvious reasons of revenue generation. There is another equally important reason for buying now--your competition. By buying now you gain a foothold on your competition. In three years when managed care and capitation really start to take hold (if they haven't already in your region) the cost of a $60,000 transmit site (or even a fully optimized $450,000 system) will be completely amortized. Others bidding for the same capitated heads or managed-care contracts using teleradiology will have to factor in their equipment costs that you have already amortized. At $550 day (five sites at $80/site plus $150/day for the receive site), that translates to nearly $150,000 per year-- more than enough to make or break a large-scale contract (or show the difference between profitability and loss). Even if the contract comes in the next fiscal year when your equipment is only one-year old, having the equipment in place and operational provides a distinct advantage.

Many people have held out hoping equipment costs will decrease even more. Equipment costs will probably not go down appreciably even though hardware performance, as in previous years, will rise.

This brings us to another "how," as in how much is too much. I own three computers--a 100 MHz multimedia Pentium with more bells and whistles than I know what to do with, a 486 DX4 laptop that can give the Pentium a run for its money and my glorified typewriter, a 386/25 running DOS 3.3. My 386 can run circles around my other machines for the application it serves: typing. While my Pentium is searching for DLL's, INI's and similar files that are necessary to run Windows, my 386 has been waiting patiently for my fingers to strike the keyboard.

Likewise, "more is better" does not necessarily hold true for medical imaging, although most medical imaging workstations do require a tad more horsepower than your standard multimedia PC (and considerably more than my 386). The line between Pentium-based PCs and their UNIX-running Sun, HP and other counterparts is becoming thinner and thinner, especially with the development of 32 and 64 bit operating systems that run on PC platforms. Still, in high-volume applications, the edge is given to the Sun SPARC-based workstations because of their ability to perform true multitasking operations. In moderate-to low-volume applications, however, PCs will work just fine and result in a considerable cost savings.

Where should I use DICOM?--A more difficult question deals with the use of industry standards--specifically the DICOM (Digital Imaging and Communications in Medicine) standard. When do you need DICOM and when don't you? The rule of thumb has always been you can't go wrong with a standards-based solution. For the most part, that holds true. However, with DICOM there are exceptions. The cost of implementing 12-bit DICOM is about twice as high as implementing a non-DICOM (proprietary) network. The reason for this, or so we're told, is to recoup the cost to develop the interfaces to the hundreds of imaging modalities, each of which requires a specific DICOM interface. In a market where standards implementation has taken over a decade to formulate, one would think that vendors would bend over backward to promote the standards-based solution.

The reality of DICOM vs. non-DICOM comes into play when a hospital is looking to import or export images (creating a teaching file, for example), archive images or establish a baseline PACS that can be grown outward or upward (i.e., start with ICU/CCU viewing, then bring in modality archiving, computed or digital radiography). DICOM images include information in an area called the header, which eliminates redundant (and oftentimes inaccurate) data entry on the part of the technologist. In addition to patient attributes (name, ID, birth date, sex, etc.), there are also both study and modality specific attributes included in this header information. This information is critical when establishing an archive, as data will often be prestaged (i.e., transferred from the archive to the requesting workstation) based on a series of criteria found in the header. This can be done as a simple single field request or a more complex Boolean search. If a hospital is going to archive all images and not make hard copies, than a DICOM-based image archive is mandated.

The use of DICOM is less defined in other applications, including ICU/CCU viewing and teleradiology. The established guidelines for PACS and teleradiology, including the ACR (American College of Radiology) Standard for Teleradiology, do not dictate file formats that need to be subscribed to, but only resolution requirements for particular images. Most importantly, within the DICOM standard there is a provision for image transmission and archiving in an eight-bit image file format. Although the image file format is the same as what is used in both ultrasound and nuclear medicine applications, a true DICOM interface to ultrasound and nuclear medicine will also include the patient, general, and modality specific attributes as well.

What is a JPEG?--In a teleradiology scenario there is little, if any, diagnostic significance for interpreting an eight-bit image frame grabbed from the imaging modality and the same 12-bit DICOM data set. This statement holds true if the image is formatted in a ".acr" file format (eight- or 12-bit) or other file formats, including desktop publishing standard-based formats such as JPEG (Joint Photographic Experts Group), TIFF (Tagged Image File Format), and others. Most clinical radiologists (with teaching institutions being an exception) have expressed preferences to have the window and level settings established for them by the technologists and presented so they can read online, just like the film.

So when do you buy DICOM and when do you not? In an intranet type environment when you are connecting multiple hospitals together to do primary or preliminary interpretation and no archiving is planned, I recommend most facilities consider the most cost-effective solution, DICOM or non-DICOM, including using PCs and frame grabbers. I typically request, however, that images be transmitted and saved in a .acr file extension, using the eight-bit DICOM file format output. Many companies have begun to offer this feature recently. If DICOM is desired near-term and budget constraints are an issue, hybrid solutions are also available, provided the facility is willing to "write off" the cost of software and the frame grabber boards. Because most DICOM solutions are software-based, a hospital could start with a frame grabber solution then grow the system into a DICOM-based solution, migrating PC-based capture terminals to workstations with appropriate hardware and software upgrades. A given in this scenario is that a software escrow account be established as well as a clause in the vendor contract that will allow other providers access to the selected vendor's DICOM conformance statement.

Other areas to consider are "who" and once again "how." How is easy-- just go slowly and carefully. A formal needs assessment process is a must and will serve as the template for your RFP (request for proposal) for both vendor quotes and the final contract.

"Who" is the $64,000 question-- There are a plethora of vendors in this marketplace--some quite good while others make Alicia Silverstone (star of the movie "Clueless") look absolutely brilliant. Finding out which vendor is the right vendor for your application is not easy, especially since no one vendor offers the right solution for all applications (or even most applications, although frankly there is no reason why one vendor couldn't be the end all/be all of the universe. No one in this market is even close.) I've worked with clients whose decisions were to buy systems from Agfa, E-Med, Cemax/Icon, Olicon and several other major players. I've also had clients make decisions to install systems from smaller regional players. Most of the systems went in without a hitch, although a few redefined the "Maalox moment." Choosing a vendor often boils down to a gut feeling-- nothing more, nothing less. It's not an exact science. And when the numbers are in and the cards are on the table, one player wins and several others lose. If you've done your homework, the loser won't be you.

Michael J. Cannavo is the president of Image Management Consultants, a division of Healthcare Imaging Specialists, Inc.