With a federal mandate calling for the nation's hospitals and physician practices to move towards Electronic Medical Records (EMRs) by the year 2014, administrators and practitioners are left wondering how to implement the complex system and, more importantly, how to keep it running once it's put into place. And that poses some difficult challenges for designers.

What is EMR?

EMR is a software and database application for the collection of a patient's medical information in an electronic form that can be viewed on a computer and easily shared by the patient's caregivers. It acts as the single and central data collection of all information related to the patient, including patient orders, drug records, diagnostic records, and associated images. EMR accomplishes this by pulling information from other systems, and interfaces and assembles it into a single entry and viewport portal.

The benefits of EMR are immense in terms of raising the quality of care by creating a single record where data entry errors are minimized or eliminated. Additionally, the patient's information is available across the health network, allowing seamless portability. Within a few more years, advancing technology will likely allow for a total and ubiquitously available health record, a far-reaching but achievable end-goal of the industry.

However, fewer than 25% of the nation's hospitals have implemented EMR to date. The reasons? It is a daunting task. The lack of interoperability standards between systems, the high cost of infrastructure, and the high cost in general of the application's development are all barriers to entry. But with some serious planning-especially in relation to infrastructure development-healthcare administrators can begin to prepare their facilities for EMR's eventual arrival.

Getting started

An “EMR charter” is the first step in planning for an electronic medical records system. Such a charter spells out specific needs and goals of the particular health system or hospital and outlines how the effectiveness of EMR will be measured. The EMR charter is really the first step in understanding how EMR will affect the hospital in terms of operational flow and infrastructure/architecture planning.

Once the EMR charter is accomplished, the focus turns to infrastructure planning and the development of an implementation plan. One of the key words that must integrate itself into EMR planning conversations is redundancy. EMR is essentially a critical infrastructure in and of itself that supports patient care. If it goes down, patient care is severely impacted. EMR depends on the availability of three critical resources: power, HVAC, and an operating data network.

EMR is dependent on the 24/7 operation of servers and both wired and wireless network infrastructure, even when utility power fails. Power to all EMR components, infrastructure, and any related systems should be on the emergency power source. The level of emergency power on which it should be located-such as the NEC-defined “critical branch” or the “equipment branch”-should be discussed with the owner. In the case of a multiple generator site with paralleling switchgear and load shedding, the automatic transfer switch serving the EMR loads should have a higher, rather than lower, load priority.

An uninterruptible power supply (UPS) should also be considered for server and network equipment, including wireless network access points that are used for wireless access to EMR. Desktop related stations tied to EMR should also be on emergency power.

Mechanical equipment that cools the IT rooms serving EMR servers and network infrastructure must work in order for EMR to be available. It must also offer 24/7 redundancy in terms of cooling equipment and power supply. Loads can be minimized by installing a small process cooling loop with redundant equipment. This way, the load can be confined to cooling critical equipment. Since the load is smaller, it also makes it more feasible to put this process cooling loop and its equipment on emergency power. The cooling equipment should have a high load shedding priority on emergency power, as it supports healthcare delivery and EMR.

The data infrastructure supporting EMR must be developed from outside the campus and directed into the hospital. EMR data is typically connected to the outside world for sharing within the hospital healthcare network. Some hospitals can even keep their EMR database in an off-site data center, or use off-site back-up.

Connecting to the rest of the world

So the question becomes: What is your connection to the outside world? Typically, hospital campuses have fiber infrastructure coming into the campus from a service provider. The path of this infrastructure, once it leaves the hospital property, is seldom under the hospital's influence and therefore the risk of service disruption is not within the hospital's control. One very large hospital in the Pacific Northwest has their main data center off-site about five miles away from the hospital. The fiber link travels on above-grade poles so any car that hits a pole in an accident can take down their link to their data.

The way to mitigate these kinds of risk is to consider multiple service providers for your campus so that fully redundant fiber pathways in and out of the campus are ensured. This is easier to do in developed metropolitan areas, as there are usually multiple vendor choices for site connectivity. In less developed areas, alternatives such as redundant lines from a single provider coming from different routes to the hospital are an option.

Data distribution within the hospital is another area of focus. Most hospitals are dealing with an existing copper infrastructure and have internal fiber distribution within IT closets. Vertical distribution of data occurs between floors through the IT closets. Failure or disruption of the vertical data path is possible due to a number of causes. It is prudent to design redundant vertical data pathways that are physically separated from each other so that if one loop fails, there is no disruption in service for EMR or other data applications. This can be accomplished by running the data loops in the same closet but with some separation, or by running them in different areas of the hospital.

One must decide what bandwidth requirements are needed for both wired and wireless capabilities, and what standards need to be applied to the facility so that EMR will work effectively. Additionally, it is wise to examine other technologies such as PACS, Voice over IP telephony, and other applications that use the common infrastructure, and decide how these technologies and EMR will impact the facility's data distribution and bandwidth. It really gets down to how much bandwidth is needed at the desktop for the wired infrastructure. For most applications today, CAT 5e wiring works well; however CAT 6 wiring should be considered for future applications.

Most hospitals have an 802.x-based wireless LAN that serves as wireless infrastructure for various applications. The addition of EMR for mobile-based access will not cause an excessive use of bandwidth on the WLAN. What does cause bandwidth issues is pushing graphical images across the WLAN. Obtaining a level coverage that does not provide gaps in service to mobile EMR users is a real design concern that should be addressed. In general, putting Wireless Access Points (WAPs) on 40-foot center coverage grid gives good WLAN coverage. It is also important to provide a “service loop” in the cabling that serves a WAP. That way, if the WAP needs to move to fill coverage, there is extra cable length to accommodate the move.

Systems integration

The power of EMR extends beyond the realm of a patient's medical records. An EMR network is also able to integrate other information and communications systems in a healthcare facility into one portal-these systems including:

  • “smart” medical equipment

  • voice over IP (VoIP) telephony

  • personal communicating/texting devices

  • nurse call systems for admissions, discharge, and transfer data

  • radiology PACS

  • location systems for staff, patients, and assets

  • biomedical systems equipment

  • patient bed interface

By integrating with other systems and sharing a common portal, EMR is able to pull all patient-care-related data and supporting systems together into one record. For example, integrating with medical equipment allows real-time patient monitoring data to be imported into the patient's record for real-time viewing. Integration with communication systems and texting devices can allow for sending messages and communications directly from EMR to the physician and caregivers from the bedside. Integration with systems such as nurse call allow the nurse call system to mine the patient's admissions data, discharge, and transfers so that the nurse call information system knows the exact status of the patient. This same integration can also let housekeeping know when the room has to be cleaned.

There is a potentially endless web of information that can be shared between the EMR and other systems, and vice versa. But the key word here is potential. Some of the interface applications are well developed, such as the admissions; discharge and transfer; and PACS, texting, and VoIP. It depends on the EMR vendor and the middleware or interface engines that have been developed. Many of the interfaces are customized software that bridge EMR and the systems.

From a planning standpoint, it is imperative that systems within the hospital share a common database platform-such as SQL Server or Oracle, among others-so that data transfer can take place as seamlessly as possible. From an infrastructure planning point of view, each system and, depending on its manufacturer, its interface to EMR, might consist of dedicated servers. For example, some nurse call systems will require separate interface servers for ADT, another for communications interface with VoIP, texting and personal communicators. This means that the number of servers planned for the IT rooms or main server room can increase quite quickly, so one must look at space for servers and increased HVAC requirements (all on emergency power) to support the eventual number of servers.

Part of the up-front planning procedure is to discuss what interfaces to EMR are beneficial and then creating a mapping diagram of information flow. From there, it is necessary to start looking at the individual interfaces, identifying additional servers and server rack space, discussing database sharing, and figuring out how they will connect to each other.

Creating and implementing a plan for an EMR rollout takes time, considerable infrastructure considerations, and-above all-a hefty budget. Rarely can a hospital implement EMR in one fell swoop; it is too costly. However, a well thought-out plan that considers infrastructure requirements and meaningful integration with existing systems could result in the beginning of an affordable and highly functional EMR system. HD

Tom Leonidas, Jr., PE, is a healthcare practice leader and vice-president at Sparling, a technology and electrical consulting firm that specializes in healthcare planning and design. He can be reached at

tleonidas@sparling.com. Healthcare Design 2009 September;9(9):52-56