On Nov. 9, 2008, University of Texas Southwestern University Hospitals (Dallas) implemented several EMR components: pharmacy information system, electronic medication administration record (eMAR), clerk order entry, and emergency department record (with ED physician order entry, patient tracking, and nursing and physician documentation). This marks the completion of the first phase of EMR implementation; the second phase, which includes computerized physician order entry (CPOE) and nursing documentation, will be completed in 2009.

The implementation of the new integrated EMR system from Epic Systems (Verona, Wis.) was overwhelmingly successful, with positive user response from across the organization, including pharmacy, nursing, and ED. The implementation required a 24-hour cutover process that called for the back-loading of active patient medication and ancillary orders into the Epic system from legacy systems. Additional staffing in pharmacy and nursing helped facilitate the data entry and reconciliation.

On the morning of the implementation, users were managing medication orders and administrations within the system. Unit clerks were able to enter their orders for laboratory, radiology and other ancillary services. The ED was fully automated from triage to the rooms, with all providers having visibility to the care of patients on electronic trackboards.

The implementation, however, wasn't without its major issues. First, there were significant problems in the first 48 hours with the interfaces between the pharmacy information system and the Pyxis dispensing cabinets. Moreover, there were also interface issues between the Siemens ADT (Admission, Discharge, and Transfer) system and our Epic EMR system. Both of these interfaces were thoroughly tested before our implementation, but new problems were uncovered after go-live. With the Pyxis interface, errors were made internally in moving code from our test environment to our production environment. With the ADT interface, test scripts were not written with enough nuances to replicate some of the more complex patient flows in the hospital. These issues have subsequently been resolved, but they served as lessons for the future.

Another major challenge after go-live was workflow on our units around the collection of lab specimens by nurses. In the past, lab orders were placed in the computer system, but the specimen collection time and collecting person were written on paper and transported to the lab with the specimen. With the EMR, the specimen collection time and collecting person are entered into the system and the lab order is not released until the nurse documents this information. After the go-live, the specimen was often transported before the nurse entered the information, which resulted in the lab having specimens before it had an order. There has been significant retraining of the users on this workflow, but it provides a great example of the need for process reengineering within an implementation.

It is entirely too early to determine the full impact of the first phase of the EMR implementation on the organization. Clearly, the system now includes a critical mass of information that makes it an invaluable tool to our clinicians: it contains all clinic notes, all inpatient and outpatient results, all operative notes, all emergency department notes, all physician dictations and all active orders on inpatients. However, it remains to be seen how well the organization, floor by floor, nurses by nurse, and physician by physician takes advantage of the functionality within the system to reconcile medications appropriately, follow-up on ancillary orders, and document all care in the ED. The single greatest barrier to these goals will be ease of use, and there is a strong commitment to exploring ways within the system to make it easier, as well as train users on tips that will allow them to master the system.

There were several critical success factors in the implementation of the electronic record at UT Southwestern University Hospitals:

Superusers. A superuser program was implemented in which staff was recruited from every unit to undergo additional training on the system and new workflows. These superusers were available on every floor to support their peers.

Additional Staffing. The patient care areas increased their staffing for one whole week after go-live to ensure that clinicians did not feel overwhelmed by patient care responsibilities with the additional burden of learning a new information system. This additional staffing was separate from the IT and vendor floor support that was available on each unit. The new EMR system tremendously slowed down our ED; there really is no substitute for repetition and practice before users feel comfortable with the system, and we see improvements every week.

Importance of “by-the-elbow” support. It has been established at UT Southwestern that on-site EMR support is superior to phone support. Clinicians are far more likely to feel positive about the system if they are able to have a support person “show them” how it is done. Even after the go-live period, there are plans to have permanent on-site support for our EMR application.

Importance of nurse informaticists. To support the EMR efforts in the organization, University Hospitals have created a department of nurse informaticists (recruited from our own nursing ranks, and educated on information systems) that lead the EMR implementation from a system design and nursing education perspective. Combined with the physician champions, these nurse informaticists provide a practical perspective on how to best utilize the EMR to drive improvements to patient safety and clinician efficiency.

Collaboration with ancillary departments. The EMR makes visible many of the processes that are involved with hand-offs between patient care units and ancillary departments. It is crucial that there are regular checkpoints for weeks after go-live to ensure that proper handoffs are happening as expected with ancillary departments such as laboratory, radiology, respiratory therapy and nutrition services.

UT Southwestern University Hospitals have made tremendous strides over the past year to improve the clinical information systems capabilities of the entire health system. A great deal has been learned that can be applied to the upcoming implementation of CPOE and nursing documentation systems.