10 Tips to Successful Data Migration During an EHR Change
Guest post by Calvin Chock, vice president, product management and engineering, McKesson Specialty Health.
Shifting from one electronic health record (EHR) system to another can be a highly disruptive and anxiety-filled process for a health system. Often, among the largest obstacles encountered is the need to migrate legacy EHR data between the old and new EHR systems. But a good understanding of this data migration process — and a strong technology vendor relationship — can help overcome this challenge and lead to a successful EHR transition.
There can be many reasons for a health system to transition between EHR systems. The original EHR system could be missing key features, or it might have reached its end-of-life, or perhaps it is not certified to meet evolving meaningful use requirements. The old EHR may not have kept up with new population management requirements on health systems, such as the need for more support for value-based care models. Whatever the reason, all health systems will want to find a solution that not only meets their projected operational and patient care needs, but which also minimizes disruption to the health system during the transition.
Conceptually, most EHRs capture the same types of information. However, when early (and still market-dominant) EHRs were first introduced, there were very few mature medical content standards, even for important categories of medical content, like diagnoses, lab results, medications and allergies. As a result, many of these early systems created their own proprietary terminology. When current standards (as incorporated into Unified Medical Language System – UMLS) began to coalesce, these early EHR systems typically struggled to migrate to the new terminology standards. Often this resulted in vast corpuses of legacy/non-standard historical patient chart data remaining in the EHR. Understanding and mapping this blend of standard and proprietary legacy data, for the purpose of coalescing each patient’s history into a new EHR, can be a tremendous challenge. Trying to fully automate this mapping process with a high level of accuracy and with larger patient volumes is a still greater challenge.
To give a specific example of this mapping problem, an older EHR may have a patient listed as having a “seafood allergy.” In meaningful use certified EHRs today, that “seafood allergy” description as such might not exist; that patient’s allergy entry would need to be accurately translated & codified into a new standardized term, perhaps referencing a either a specific shellfish allergy, or perhaps for any number of non-shellfish seafood allergies. Trying to faithfully automate terminology mapping decisions like this when there isn’t enough information to make an accurate determination can be nearly impossible. Yet not properly translating and mapping this allergy means the new EHR cannot properly use this information to trigger important patient safety system alerts (e.g., a drug-allergy interaction alert). Further, without accurate translation & coding, the new EHR will not be able to properly transfer this important part of a patient’s record to another healthcare system, like a patient portal, a clinical decision support service, an HIE, or another EHR.
In my experience, it is possible to identify three types of data migrations.
The first type of migration involves shifting to a newer version of the same product, perhaps running on a new technology platform. This type of data migration will typically allow the most comprehensive transfer of data with the least amount of disruption.
Another type of migration involves switching to a new EHR from another vendor, but with the aid of a cooperating & generally supportive EHR vendor. Although migration from one vendor’s EHR system to another’s is more complicated, if the original EHR’s vendor is willing to share information about their data structure and ontology with the new vendor, it will typically ensure access to a larger subset of key data elements are and more reliable data mapping. Look for established migration pathways, since EHR-to-EHR migration processes often improve each time they occur. This class of migration is often suggested by an older EHR vendor when they decide to sunset or end-of-life an EHR product.
Finally, the most difficult type of migration, involves a move to an EHR with very limited inter-vendor cooperation. This also happens to be the most common type of migration, especially when a health system chooses to migrate to an EHR vendor’s competitor’s EHR. The vendor of the current EMH system is often not willing to share more than the legally-required level of information, so the new vendor must rely on either a proprietary data extract or a batch of patient CCDA files (CCDA is an export format for patient summary data, which all EHR vendors must supply.). In either case, these patient data extracts will then need to be manually loaded, reviewed, and electronically (or manually!) reconciled for every patient chart.
Health systems planning this third type of migration should allot many additional months to work with the new vendor’s IT staff as they determine how to transform and load the old EHR’s patient data. Further, the health system must be aware that the transition will require a much higher investment of staff time for the data to be cleaned up and reconciled. This type of transition can take months or even years to complete.
If upgrading to a newer version of a current is not a viable option, a health system should follow these general guidelines for a successful transition to a new system:
- Find an experienced vendor partner.
In addition to finding an EHR system that meets your health system’s operational/functional needs, try to find a vendor partner with experience migrating data from your current EHR vendor.
- Focus on your staff workflows.
The new EHR will not do everything the same way as the old EHR. Your goal is to ensure new EHR vendor understands and supports the various workflows in your health system: billing, scheduling, pharmacist, physicians, etc. Ensure the new system can load enough data from your old system that the impact to your staff’s workflows will be tolerable.
- Create an internal implementation team with a blended skill set.
The best EHR transitions occur in health systems that assemble an internal team with members representing all disciplines & workflows within the organization. This transition team will typically: