This is part of our data migration blog series – a range of topics intended to help organizations who are migrating from one EHR to another.
As a data migration project nears its final phases, the most important tasks still remain: migrating the data into the production environment in preparation for go-live. Hopefully, after months of testing and validation, and multiple dry-runs in test environments, the go-live process will be simple and uneventful. To ensure the process executes smoothly and successfully here are a few–of the many–technical considerations to keep in mind.
One of the first tasks with any data migration project is establishing appropriate access to the systems and environments; the project can’t start until access is granted to all systems involved in the migration. The same is true for the go-live (or production) portion of the project – it can’t start until appropriate access is provisioned. The types of systems involved in the go-live process will be the same types of systems encountered during the rounds of testing, but the production systems are installed in different environments with different servers and configurations. Making sure the migration analysts have the appropriate credentials to perform each of the data migration tasks in the production environment and systems will be critical to the success of the go-live.
There are many technical components that will need to be accessed in a migration – servers, databases, file shares, interfaces, applications, job services, etc. Overlooking access to one of these is easy to do and will likely delay the project. Confirm appropriate access to each of the technical components well in advance of the go-live. Keep track of each access point required during migration testing and make sure each analyst is provisioned appropriate access in the production environment.
Depending on the size and scope of the project, there may be a need to start the production data loads weeks in advance of the data migration go-live date, rather than performing the entire data migration over a single weekend right before go-live. Timing each of the steps in the ETL (Extract, Transform, Load) process and understanding the expected throughput for each step should be one of the goals of testing, specifically the full-scale round of testing. This will help the technical analysts make recommendations on the go-live schedule.
Take into consideration that certain data elements of the migration process will require more time to complete than others, so the timing of each ETL step should be broken down for each data element. For example, we have found that the overall ETL process for CCD messages in a data migration can execute in a much shorter time than the process for scanned images. This often results in scanned images loading weeks in advance with gap loads leading up to go-live, while CCD message can sometimes be migrated right before go-live without the need for gap loads. Determining the expected throughput for these items, and any of the other data elements within scope, will allow the project team to efficiently schedule the migration of each data element, minimizing the number of gap loads required before go-live, and allowing adequate time for each item to migrate.
Preparing the production environment for the data migration is vitally important to the success of the migration. In some cases, the test environments and production environments may have different versions of the systems installed, or different hotfixes or updates applied. These differences can affect the behavior of the migrated data, and, in some cases, may prevent data from loading successfully. Before starting the go-live phase of the migration, confirm the specific versions, hotfixes, and updates that have been applied to the production environment, and make sure they match the test environment.
Additionally, when preparing the production environment, confirm that all the technical components that are used in the migration of data (such as interfaces, APIs, services) are built and configured the same way as the testing environments. Also, confirm that all their dependencies (such as dictionary builds, tables, mappings) have been copied or rebuilt in production. Any differences in these components between the test and production environments can result in unexpected behavior when migrating the data. Establish a process to keep these environments in sync and validate the accuracy of the synchronization before migrating any data.
The data migration process can be long and arduous; however, the most critical tasks of the project are in the final phases, so careful preparation is even more prudent and will be important to the successful completion of the project. Following the technical considerations discussed above will increase the chances of a successful migration, which will ultimately satisfy everyone affected by the migration, including stakeholders, the project team, and patients.
To learn more about the technical considerations for data migration go-lives or the data migration process and how it can positively affect you, please contact Galen Healthcare or Tyler Suacci below: