Table 1. Uptime Institute. Cost Model: dollars per kW plus dollars per square foot of computer floor.


IT is becoming an increasingly important part of health-care delivery, and the need for reliable IT infrastructure in support of patient care has become critical. With moves toward digital imaging and digital pharmacy, and ever-growing patient record databases, the need for real-time communication throughout the medical professions is reaching a threshold. Hospitals continue to move away from paper forms in favor of PC tablets and away from X-rays in favor of digital imaging, increasing the need for a reliable electronic repository (i.e., data center).

Where hospitals once stored all patient information as paper records in thousands of folders, patient information can be digitized and stored electronically at a consolidated facility that can be, but does not necessarily need to be, located on site. According to a study published in the New England Journal of Medicine in March 2009, only 1.5 percent of nearly 3,000 non-federal hospitals surveyed have a comprehensive electronic health record system. Estimates range widely as to the expected annual savings when patient records move completely online, but it could be as high as $77 billion.

As storage media such as SANs and tape storage become more affordable, the business case for removing hard copies improves. In addition, as the use of these storage technologies increases, multiple hospitals can access, retrieve, and cross reference patient data using network software.

Patient records can be retrieved and documented by a physician on a tablet at a moment’s notice. Although some departments have kept electronic patient records for decades, the cross pollination of information drives its further use. Overlay software that can gather information from radiology and pathology departments on a common system allows for accurate and fast communication between departments. No longer does a physician have to wait for the file to be transferred from Radiology to Oncology, because information retrieval can be transparent and almost immediate.

Figure 1. Typical data hot-hot data center configuration supports 2N to the mission.


IT use in health-care departments such Radiology, Pathology, Cardiology, Oncology, and Pharmacy, to name just a few, is also expected to increase. The IT facility needs to be capable of storing these digital images but also make them accessible to physicians at a moment’s notice. The PC tablet is an example of relatively new approach to point-of use-patient care that enables physicians to track and document notes and recommendations.

Managing the ever-growing volume of health-care financials and reimbursement claims is an equally urgent task for health-care providers. Storing more patient records in electronic databases tends to reduce billing errors. The reduced error rate increases efficiency and revenue. Electronic doctors’ notes, digital pharmacy costs, radiology cost, etc., can all be easily linked and accurately summarized with each statement.

Johns Hopkins Hospital has pioneered the use of electronic patient records (EPR) and deployment of high-reliability mission critical space to house the EPR. Johns Hopkins has used EPR since 1998 for basic information and lab results. The next generation of EPR allows fast and easy on-line access to patient records in all areas across the Johns Hopkins continuum of care, thus enhancing patient care. Through EPR, the user can view patient data integrated from multiple sources, including lab results, radiology reports, ECG reports, sonogram reports, discharge summaries, and operation notes.

The Johns Hopkins Hospital has developed an EPR interface engine, an application providing patient medical records across its network of hospitals and medical facilities. Eventually clinical context object workgroup (CCOW) technology will enable EPR to share patient and user context with other CCOW-enabled applications. Some of the functionality of EPRs can be web enabled with read/write or both access depending on user privileges. Write-enabled users can add all patient problems, allergies, or medication lists.

Figure 2. Traditional power distribution to critical hospital loads.


The Johns Hopkins interface engine allowed the hospital to develop a single application where all data can be accessed on a common platform. Before that, multiple applications were required to document patient status and notes about the patient. This development, along with growth of EPR, has spurred growth in IT. Patient data retention in hospitals generates a tremendous volume of data that needs to be stored and accessible. According to Edward Fields of The Johns Hopkins Hospital, data centers are an integral part of operations of the hospital.

Historically, hospitals met their IT needs with small server closets or, in the case of older facilities, housed servers within mechanical/electrical rooms that are often located in the basement of the building where piping or high voltage power lines originate. The location of the IT room was dictated by point of presence on a hospital campus (often on the lower levels spaces of the hospitals). The technology available at the time also often dictated location. The change from coaxial cables in the 1980s to fiber and wireless in modern facilities allowed more flexibility in the location of the data center.

Since IT was considered non-critical overhead to hospital administrators, infrastructure support and funding was limited. Traditional hospital IT rooms had some of the features of a data center and many undesirable features of modern data center facilities. These traditional data centers were space limited, with limited infrastructure capacity and redundancy (typically Tier I or less). Backup generation capacity was only available after remaining needs of the critical and revenue generation spaces. Electrical systems often did not include UPS system.

In addition, IT equipment was haphazardly installed. Inefficient space utilization and lack of cooling and power created havoc for the IT manager. Outages were all too frequent and easily accepted. Hospitals could continue to operate because they did not rely too much on IT and the loss of IT infrastructure did not impact the mission of the hospital as compared to losing the ICUs that generate revenue.

Until recently, maintaining the integrity of the IT network and infrastructure was considered a luxury, and not a necessity, as this had little or no impact on day-to-day operation of the hospital.

In many cases, the increased demand for IT is creating a burden on the existing computing capabilities for hospitals. Today interrupting communication will hamper hospital operation and reduce efficiencies gained by the digital communication. In order to maintain a smooth, efficient, and well-run facility, reliable IT infrastructure (i.e., data center) will be necessary.

As the need for a dedicated data center pushes its way to the forefront at more hospitals, it is critical that planning for these facilities considers both current demand as well as projected growth. According to IBM’s Vision For The New Enterprise Data Center – IBM Point of View (October 2008), “Medical images that used to be two-dimensional and 1 MB in size a few years ago, are now four-dimensional and 1TB in size! By 2010, it is estimated that 30 percent of the world’s digital storage will consist of these medical images.” In the case of hospitals that are affiliated with universities, medical research (human genome) and academic research (physics, engineering, weather, and climate change) extends the demands on the data center even further.

IBM recently introduced its “Blue Gene” supercomputer, capable of providing up to 20 peta flops. Supercomputers such as these require significant power (up to 35 kilowatts [kW] per cabinet) and cooling, in an extremely compact footprint (less than 25 square feet each).

Consolidation of systems also drives data center growth in health-care facilities. The consolidation helps increase reliability and standardization and takes advantage of economies of scale.

The growth in health-care IT is generating a need for larger and more reliable infrastructure to support its data centers. This strains the whole facility as an indirect revenue generation component of the hospital that can consume significant financial resources.

The loads to power and cool a modern data center can be as much as twice that of the critical (IT) load. Amplifying the cost of the infrastructure requirements is the need for reliability, which often results in redundant equipment and multiple paths of distribution, and so the cost of supporting mission-critical space can be extremely high compared to traditional commercial construction. The usual metrics for construction such as dollar per square foot are typically not relevant for data centers. Instead, the construction cost is a function of level of reliability and the proposed critical load density within a data center. The growth factors that influence the facility are critical load, HVAC systems to support the critical load, and system inefficiencies. As such, the preferred metric for data center construction costs should be dollars per kilowatt (critical power).

The table illustrates the sliding scale for construction costs for a typical data center with 200 kW of critical power located on 2,000 square feet of raised floor (100 W/sq ft).

Data Center Mission

A typical Tier II data center can cost approximately $12,500/kW to build, and a Tier IV facility can be twice as costly. Even with the high construction cost, Tier IV sites still fail. Some facilities utilize a remote disaster recovery site (leased or owned facility), which as the name implies, is a smaller facility located offsite that stores only the most critical data in case of a prolonged outage at the main facility.

Figure 3. Hospitals find that their modern IT operations require more redundancy and complexity.


Due to the high cost of Tier IV facilities, the “hot-hot” solution is becoming more common. Hot-hot solutions include two Tier II sites or a Tier III and a Tier I site located in two distinct facilities, having redundant communication between the sites and redundant communication to the hospital. The hot-hot solution provides backup for critical systems while still maintaining a fully redundant solution to the hospital. The solution helps to reduce the initial capital cost of providing higher reliability facilities, while proving OPex savings (which can be as high as $100,000/month per site).



The need for more IT in health care is becoming greater and affects not only large institutions but smaller medical providers as well. Media storage and high-density computing is becoming integral to day-to-day operation of many hospitals. This increased growth in high-density computing, and medical imaging is creating a greater demand on hospital capital and operating expenditures, where traditionally the data center has been considered an overhead component of hospital operation with little operational impact. IT growth also affects capital and operational budgets. Both facilities and IT should evaluate existing data center operations to determine the impact of the data center on the hospital operation. The teams should address challenges and problems in a collaborative effort as IT is a relatively new concern for hospitals and health-care providers. A systematic approach in evaluating existing facility, projected growth, capacity, and fiscal planning is key in order to stay ahead of the health-care IT growth curve.