It's in every hospital. In each operating room. In each patient's room. It's even in the hallways. Find a point of care, and it's there. It's the data and communications network, and it's transforming healthcare facilities into massive data centers dependent on continuous availability of technology systems. As technological processes such as electronic medical records, digital imaging, and wireless communications continue to streamline the healthcare industry and improve overall outcomes for patients, the criticality of the network and its support systems cannot be overlooked. This requires an increased emphasis on infrastructural factors such as backup power, precision cooling, and system monitoring and management both in the data center and in the hospital. Collectively, these support systems keep the network running, improve the efficiency of facility operations, assist in maintaining medical staff satisfaction levels, and support overall facility credentials for national rankings and evaluations.

Power protection


Healthcare facility designers are fully aware that healthcare environments are not required by the NFPA or the Joint Commission to use uninterruptible power supply (UPS) systems. Yet at the most basic level, the impact of not using a UPS is felt during a utility power failure or when generators are tested bimonthly. In many instances, this is a brief interruption to the power stream that often lasts approximately 10 seconds, or the time it takes generators to start. The true impact of the outage is felt during the several minutes it takes for servers to reboot, establish connectivity and reacquire data. This downtime results in lost revenue from equipment unavailability, dissatisfied doctors or patients, and increased risk exposure for the facility.

Amid efforts to prevent downtime altogether, many healthcare IT, facility, and data center managers are installing individual UPS systems to protect specific workstations or devices, such as MRI machines or nurses' stations. This is similar to 10 to 15 years ago when servers were first being deployed in data centers and it was common for each server or server rack to be configured with its own UPS. As servers continued to proliferate, this approach became less practical because of the growing cost of adding multiple UPS systems over time. In addition to the cost factor, this device-level approach also creates a maintenance nightmare for IT managers who are tasked with testing and monitoring potentially hundreds of individual UPS systems. This distributed UPS strategy also fails to properly address the problem of capacity, as more and more loads in the hospital are IT-enabled and require uninterruptible power.

Instead of providing protection at the device level, modern healthcare facilities are achieving much higher power availability and simpler management by moving power protection upstream and implementing a centralized UPS system. For smaller facilities this may mean providing a single, large UPS directly downstream of the automatic transfer switch of the facility's critical branch; larger facilities are being designed with multiple critical branches, each having a single, large UPS. This centralized approach supports the ability to add loads for future IT innovations found in operating rooms and other critical spaces.

One option both the hospital facility manager and healthcare data center managers should consider for lowering cost and meeting environmental goals is using a flywheel system to complement or replace the battery system that fuels UPS backup power. The flywheel system provides efficient, short-term backup power and is ideal for use in applications with fast-start generators. Flywheel units can provide instant, ride-through power of 190 kilowatts for 13 seconds-more than enough time to switch to generator-and are parallelable for additional capacity. Flywheel energy storage systems have proven to be approximately 10 times more reliable than VRLA batteries in similar applications. They have a low installation cost and take up minimal floor space; roughly the size of an IT rack. Additionally, flywheel units do not need special ventilation, and have no special disposal requirements because they do not contain hazardous materials.

Precision Cooling

Deployment of new technologies such as servers, switches, and routers brings new capabilities to healthcare facilities. It also brings more heat to network closets and data centers. Without appropriate cooling considerations, even the most robust backup power system can fail. Imagine dozens of network closets in a healthcare facility, each housing servers and wiring that give medical personnel access to patient data at the touch of the button through wireless technology. This technological convenience requires lots of equipment running in small spaces, resulting in higher temperatures. When sensitive electronics operate at higher than normal temperatures, their short-term reliability is compromised and long-term viability is significantly decreased. Proper environmental control-temperature, humidity, and air quality-plays a key role in overall continuity.

In network closets, remote facilities, or standalone doctors' offices, it's important for healthcare facility designers to determine feasibility for ventilation ductwork and whether or not the heat can be exhausted outside of the facility. If the closet is cooled by the building's air-conditioning system, cooling is probably insufficient because building air-conditioning systems cycle on and off and are not designed to regulate the precise temperature and humidity requirements needed by computer equipment.

For equipment loads of 100 to 1,000 watts-typically one or two servers-ventilating the closet space with intake and exhaust grills should be adequate. This must be accomplished without compromising the physical security of the space; leaving a door open for ventilation purposes is never an appropriate solution. Rooms producing more than 1,000 watts of heat-typically a few tower servers and a UPS system-will need some form of dedicated precision cooling. Self-contained cabinets or rack enclosures are available that include support systems such as precision air conditioning, monitoring capabilities, security features, and may be a simple installation depending on the room.

In larger network closets (several full racks) and small data center settings, installing specialized computer room air conditioning (CRAC) units and using the hot-aisle/cold-aisle configuration is the best solution. In high heat-density environments, which involve racks of blade servers, cooling system scalability can be accomplished via supplemental refrigerant or chilled water cooling systems, which bring cooling closer to the source of heat and can be mounted overhead or in the aisle. This allows cooling to be focused where it is needed most and can be added to, rather than displace, existing cooling systems. The latest innovations in precision cooling technology include digital scroll compressors, plug fans, and integrated monitoring and controls, which all improve the efficiency and performance of the data center.

Monitoring

The third line of defense in ensuring continuity in a healthcare facility is effective monitoring and systems management via data analysis. A critical systems management approach to monitoring makes sense, especially in healthcare facilities, as it involves establishing a management strategy and technology platform for critical support systems. Basic capabilities of a monitoring system should include real-time monitoring, locally or remotely; alarm notifications; event escalation and tracking toward root cause; and historical data collection and archiving, including performance tracking. IT and availability performance tracking is particularly beneficial because, through both short- and long-term performance measurements, it allows preventive measures to be taken before a unit actually fails.

Similar to implementing a centralized UPS system, using a centralized approach to monitoring can increase visibility into system performance, enable more effective preventive maintenance, and free up people resources to more productive endeavors. This lowers overall cost to the facility, as it's less expensive to monitor and maintain existing equipment than to continually repair or replace malfunctioning items.

Best Practices

One example of a larger, existing healthcare data center that has undergone a recent power and cooling renovation is Norton Healthcare in Louisville, Kentucky. The facility's 35-year-old data center was supporting four hospitals, 10 urgent care centers and more than 60 physicians' offices, and running more than 400 servers, including 160 virtual servers and 140 blade servers. As the IT demands of the hospital increased, the data center could no longer keep pace with supporting the hospital's more than 400 business-critical applications, such as financial management, scheduling, electronic health records, and digital medical imaging. The decision was made to renovate the data center to meet the growth demands and criticality levels needed in today's top healthcare facilities.

Norton understood the criticality of its operations and, when given the opportunity to upgrade the facility's power system design, two 400 kVA UPS units in a dual bus configuration were implemented to provide the data center with completely redundant power. A static switch was also installed to provide a simple, reliable response to brief overload conditions and to manage the two UPS units. Also, during the course of the renovation, new electrical, data distribution, fire suppression, cabinets, cable management, security, and monitoring capabilities were installed.

A reliable precision cooling infrastructure was created with 11 CRAC units arranged in a hot-aisle/cold-aisle configuration of three rows that were spaced six feet apart on a new raised floor. This approach created cooling efficiencies that were not previously available in the existing data center.

The data center's Network Operations Center (NOC) was also completely remodeled, and its operations were enhanced through the implementation of monitoring software. This technology provided the NOC with centralized monitoring and control of SNMP devices through existing network management systems. It also sends e-mail alerts and local notifications when environmental conditions or power capacities change.

As a result of the renovation, Norton Healthcare transformed its data center into an efficient, high-availability operation, which supports the hospital's more critical applications. The scalable and redundant data center infrastructure has already allowed for the addition of several hundred new servers running the latest medical applications.

Conclusions

As hospitals and other healthcare facilities become more reliant on technology-based, point-of-care offerings, the IT infrastructure to support those technologies is just as critical as the care itself. Simply put, hospitals and data centers must be viewed holistically, and both must implement the best practices of proven power protection, precision cooling, and monitoring strategies to help improve patient care, increase efficiency, and ensure the continuous availability of life-critical information. HD

Dan Draper is in Liebert products healthcare marketing for Emerson Network Power.

For more information, visit http://www.emersonnetworkpower.com.

Healthcare Design 2009 August;9(8):77-86