In an earlier article, we examined the reasons why an early EMR adopter might need to migrate to an entirely new clinical information system and some of the important issues that need to be addressed in managing that migration. In this article, we show how one practice overcame these challenges.
In the first article in this series, we examined the reasons why an early EMR adopter might need to migrate to an entirely new clinical information system and some of the important issues that need to be addressed in managing that migration ("Is it time to make the switch to a new EMR?" Urology Times, July 2009). In this installment, we will show how one practice overcame these challenges.
Once the new product and vendor were selected, UANT faced the daunting task of migrating to the new system without losing access to and functionality of the older information. Several decisions were made on the basis of the principles that no information would be left behind and all information would be functionally accessible from the new system.
These principles drove the migration strategy-discrete information in the older system would need to be identified, classified, and assigned a new destination in the new system. For example, problems that had been "assessed" during patient encounters existed as discrete data elements (checkboxes and "yes or no" questions on EMR screens) in the older system that could be identified using either a description, ICD9 code, synonym, or some combination of these. Each problem so identified in the old system was mapped to its equivalent discrete description in the new system to create a list of active problems available to the user to assess. The same logic and methods were used to map current medications, past medications, and lab results, since they existed as discrete origins and destinations in the systems.
Some information existed as text data in the older system, but was used to populate discrete data elements in the new system. For example, the data extraction engine was able to identify use of a text checkmark in the old system that was labeled "appendectomy" and map it to a Past Surgical History Problem with the same label. In this way, non-discrete data elements were used to create discrete data. This mapping involved hundreds of hours and thousands of spreadsheet rows, but was necessary to make available clinical information exist as discrete data and allow meaningful use of this information in the future.
The strategy to migrate drug allergy information included a decision to map discrete allergy information in the old system to its equivalent in the new system, as well as map non-discrete allergy information (that data entered as text in certain places in the old record) to discrete allergies. It also involved not mapping "no known allergies" and not assuming that the absence of allergy data in the old system meant the patient had no allergies. When the first contact with the patient occurred after the conversion, the system prompted the clinical staff to update allergy information, including a prompt if the patient had no allergy information. This strategy combined the preservation of information that had been gathered before with a safety check in the new system.
Timing the transition
As with migration from paper to EMR, moving from one EMR to another may result in current clinical information on a patient being located in more than one system at the same time. UANT accomplished its migration in waves over several months, so that a physician was signed in to one system only at any given time. Many patients in the organization are referred within the group to specialists in centers of excellence, such as urologic cancer, erectile dysfunction, and pelvic floor dysfunction. Consequently, a patient could have records in both EMRs at the same time. This potential safety issue was managed by:
In addition, information was copied from the old system to the new one immediately prior to each subsequent go-live wave based on date of the last "gap conversion" minus any dual resulted information. A line was drawn on a date the last data was transferred. Changes in data, including additions, deletions, and revisions that occur between the date of the last conversion and the current date are migrated as "gap" information.
After the final gap conversion, it was important to ensure that no additional new information was entered inadvertently into the legacy system. This was accomplished by turning off demographic and lab interfaces and changing permissions for users to read-only privileges after ensuring by careful audit that all documents, scan documents, and lab results had been properly reviewed.
Bottom line: With a sound, careful approach, an organization can manage the transition from one electronic medical record system to another without losing important clinical information and/or functionality of older data. It is a daunting task, but following the principles we have described here can make it a manageable project.
Neil H. Baum, MD Dr. Baum is a urologist in private practice in New Orleans. He is the author of Marketing Your Clinical Practice-Ethically, Effectively, and Economically.
Robert A. Dowling, MD Dr. Dowling is medical director of Urology Associates of North Texas, a 49-physician, community- based, single-specialty group in the Dallas-Fort Worth metroplex.