Data center upgrade essential to enhanced patient care: as digital applications expand to physicians, nurses and patients, healthcare providers require greater data availability and security to reduce error rates, risks and costs

Health Management Technology, Oct, 2004 by Johnathan Donovan

Healthcare IT has been evolving rapidly in recent years from its use only in administrative applications to playing a central role in re-engineering the way patient care is delivered. This sea change has created a strategic challenge for CIOs as they prepare business continuity and disaster recovery plans.

Now, as IT infrastructure costs have climbed to more than a third of the average hospital's capital budget, there is even more pressure on CIOs to ensure the reliability, availability and integrity of these proliferating applications. It also has become increasingly urgent for them to rethink their approach to such "critical-path" elements as network infrastructure and data centers. In doing so, they will need to consider the following:

* Recognize how traditional data center strategies come up short in terms of uptime, power backup, cooling and space planning, and how those limitations can be better solved through a more innovative approach.

* Deploy strategies for "right-sizing" vs. oversizing cooling and power backup.

* Reduce capital and operating costs by eliminating expensive raised floors in data center design through use of a self-contained data center.

* Design a better way to cool and power today's highly dense blade servers.

To their credit, healthcare IT leaders are working diligently to ensure that application, network operating system and hardware infrastructure layers of the OSI model that support patient care are designed with adequate fault tolerance, redundancy, failover and hot-swappable components. IT also is committed to enforcing the standardization of components and systems in order to reduce human error, create consistency across the network, quickly detect root causes of problems and simplify the repair/recovery when faults occur.

The problem, however, is that the hardware layer and everything above it are dependent upon the physical layer--i.e., the power, cooling, cabling, distribution, fire and security systems--and without adequate attention to this critical area, the likelihood that patient care delivery applications will be highly available is extremely low.

Data Center Challenge

As today's cutting-edge IT applications such as PACS (picture archiving communications systems), VoIP, wireless badges, telemedicine and CPOE are rolled out to physicians and nurses, there is a growing expectation that this data should be accessible on demand.

This proliferation of IT infrastructure often begins to place a burden upon the physical plant and facilities. Outdated buildings and infrastructure, plain old bad wiring, faulty cooling systems and "cable chaos" permeate today's healthcare data centers. Historically, many viewed medical data centers as high-cost, low-return investments. More often than not, their development was overly time-consuming and complex due to lack of standardization in the process.

Although most IT processes today are designed in a systematic and standardized way, unfortunately, most data centers are not. Rather, they are built-out with a series of distinct components cobbled together at the customer site. A data center designed in Boston, for example may be completely different from a data center designed in New York depending on which outsourcing firm built it. As one observer put it, "More pre-engineering goes into a $25,000 Toyota than a $1 million data center."

As they contemplate solutions, providers will need to focus on three principal objectives: 1) ensure data availability by designing data centers and IT spaces upon standardized designs and pre-engineered components that are designed to function as a self-contained, modular "system"; 2) create flexibility by right-sizing system capacity, optimizing floor space and reducing turnaround times; and 3) reduce capital and operating costs by upgrading pre-engineering and design requirements, improving capacity utilization and simplifying self-service.

Network-Critical Physical Infrastructure

In a world that expects zero downtime, the best way to ensure optimal availability is to have a strong network-critical physical infrastructure (NCPI) as the foundation for a highly reliable network. Solutions must be open, adaptable and allow deployment of NCPI quickly and easily by using a standard building block design for power, cooling, racks and cabling (Figure 1).

[FIGURE 1 OMITTED]

In contrast to the costly and inefficient legacy approach, today's architectural standards also should include pretested, standardized and hot-swappable components that offer built-in redundancy to eliminate traditional points of failure and make possible lower costs, higher quality and easier serviceability. These new architectural designs should result in substantially lower human error rates, which normally account for more than 60 percent of system downtime.

The scalable nature of NCPI-based data centers also enables IT managers to build only what they need for the near term. Adopting a right-sized modular approach allows them to easily add to initial deployments to gain greater capacity as their needs increase (Figure 2). In addition, the approach allows providers to convert existing space to a data center in less than a few weeks, compared with a three-to-six-month time frame to deploy most legacy systems, plus avoid the unnecessary expense of maintaining unused equipment and facility space.