• Free for qualified executives and consultants to industry

  • Receive quarterly issues of Area Development Magazine and special market report and directory issues


First Person: R. Stephen Spinazzola, PE; Vice President; RTKL Associates Inc.

  • R. Stephen Spinazzola, P.E., Vice President, RTKL Associates Inc.
Aug/Sep 09
Electronic medical records (EMRs) play a significant role in both economic stimulus plans and healthcare reform concepts being discussed in Washington. As with other applications of technology, EMRs require a lot of computing and data-handling power. Do American healthcare providers currently have the infrastructure they need to implement these technologies on the kind of scale that proponents envision?
Spinazzola: What's deployed right now in healthcare data centers is a decimal-point fraction of what we're going to need to build out in the next five years. I would say we're less than 5 percent deployed across the country as a whole.

If they need that much more infrastructure, it sounds like providers are going to be doing a lot more than simply replacing their paper patient charts with computerized medical records. What is the broader picture?
Spinazzola: At the end of the day, an EMR means not only what was in paper form; if I have an MRI done, that's also a part of my EMR now. Every time I talk to clients there's a new application being developed. An example is ICU monitoring. That's done at a patient's bedside, but the data is sent to a monitoring station and to a data center, and the data is trended so that the doctor can see what has happened over time. It's endless.

Describe what a typical healthcare data center will be like. What are the typical requirements and attributes of the facility itself?
Spinazzola: Most hospitals are building Tier III data centers, with redundant components that are concurrently maintainable. Generally you have one utility coming in and parallel to that you have an onsite generator. You must have redundant cooling, and that cooling is on redundant power. The building is going to have to have the raised flooring. All of a sudden what seems like it would be a 10,000-square-foot building becomes a 30,000-square-foot building. A Tier III facility has to have available more than twice the power than you actually need. It gets very expensive to deploy these facilities because of the power.

How much money are we talking about? Federal funds have been made available to help some healthcare providers implement EMR systems, but will hospitals be getting any economic stimulus money to pay for these data centers?
Spinazzola: Depending on the size of the institution and where they are on the continuum of deploying applications, the typical data center is going to be in the $20 million to $60 million range. The stimulus is funding hardware and software and brainpower, but there's no money for the data center. Data centers don't have a cash flow analysis - there's no money coming out of them. It's about risk management and fulfilling an IT need.

Many providers these days are suffering from low margins and reduced capital budgets. If the stimulus packages have left out money for data centers, how will providers be able to afford this kind of development?
Spinazzola: I think it's going to drive them into alternative funding schemes. Ever since I've been in the business, hospitals have used developers to build medical office buildings. They'll provide the land, and the developer will build an office building and lease it back to doctors. I see that as a potential opportunity for data centers. You would have third parties do design-build-leaseback deals. However, there's going to be a smorgasbord of situations - it won't be one size fits all. Some hospitals have the money and some don't.

Does every hospital need its own data center? In some industries, dedicated data centers are mostly for the bigger players, while smaller companies may opt to collocate their data operations with others, or make use of third-party hosting services. Will that model hold true in the healthcare industry as well?
Spinazzola: I haven't seen any hospital that's willing to take the leap and say, "I'm going to collocate." Bigger hospitals are going to want their own data centers, and if you have a hospital with all of these applications, you also need a help desk. However, I see third parties, maybe the companies that make imaging equipment, for example, offering it as a service. Smaller hospitals may want to seek a third party to help them with it.

When you've seen major hospitals planning data centers lately, do the new centers tend to replace existing data operations, or do they supplement and expand what's already in place? And are they going up within the walls of the hospital, on the hospital campus, or further away?
Spinazzola: The ones we're been working with right now tend to do two data centers, because they need a disaster recovery plan. They'll do a primary data center and a backup data center. They usually want to put the new data center off-campus. It's easier to make the data center already on-campus the redundant site. They need the offsite center to be close enough to have real-time reporting. They tend to want it to be less than 20 miles away.

Are there good reasons for pulling the plug entirely on the on-campus data operation that already is in existence and just moving data operations totally offsite?
Spinazzola: Some are using this as an opportunity to free up space in the main hospital campus. Then they can put in another revenue-generating function.

How long will it take for healthcare providers to catch up with their data center infrastructure needs? Will the work ever be truly done?
Spinazzola: I think it's going to evolve over the next five years, maybe even longer, maybe seven years. I think it's going to take that long for the marketplace to absorb the technology. Then the building of the facilities will slow down. You hope that you've done your planning well and have a 15-year facility. But by then we'll already be in the second generation of equipment refresh.

Follow Area Development