Data Management: A Key Part of EHR Implementation
Oct 30, 2013 09:03AM ● Published by MED Editor
Have you deployed an EHR (Electronic Health Record) system? Are you planning on deploying one? In either case, how are you planning to manage the data that it contains or is generating to ensure performance and availability? Was that part of your process and design requirements?
Getting the new system online and in production is one element; but ensuring EHR application performance and is another. There are many facets of data management that needs to be addressed to ensure that the system delivery of patient services and practice management is timely and highly available.
First, policies need to be developed, documented, and communicated to the physicians and staff so that efficient and required processes are followed in the handling of various data types. For example, a complete data handling process definition should identify each type of data as a part of a specific category of data. Each of those categories should have very clear requirements for accessibility, reproduction, storage, and disposal.
Practice marketing data will have very different handling and performance requirements than patient records, just as insurance claim filing data will vary greatly from office inventory data. Compliance to those policies should be a part of the reporting feature-set of the EHR system chosen or as an additional purchased or custom extension. Applicable federal, state, and association compliance definitions should be referenced to develop the appropriate requirements of an EHR deployment.
Secondarily, data read and write performance management for the EHR system also has to be addressed. The primary copy of data needs to be managed for optimal access during day-to-day operations. Traditionally, data storage systems used only spinning magnetic disk drives for storing information. All data was treated relatively the same and, thus, created dramatic inefficiencies due to the great differences in access and write requirements of the information being processed.
Data storage technologies such as flash, solid-state, and various spinning disk storage can now be enabled in a wide array of vendor solutions, price points, as well as performance and capacity levels to meet the needs of the practice or provider facility. Dynamic data tiering, included in some vendor solutions, manages each of those storage technologies in one system to optimize data access and write performance. This is accomplished by placing the specific data on the storage technology best matched to that information’s performance needs. Frequently accessed or updated data is placed in memory, on solid-state drives or flash technology while less frequently accessed data is placed on slower, cheaper, and higher-capacity spinning disk drives.
This optimization of the technology thereby optimizes the various practice functions’ access to needed information or data entry performance. Additionally, an effective tiering capability lowers the overall cost of data storage through the efficient placement of the various practice data based on that data’s frequency of access, age, as well as type of data. The storage technology used to store the EHR data should support these features just as the EHR should have the capability to take advantage of these optimized data access approaches.
Finally, data availability of the EHR data should be planned, utilizing methods for secondary, or backup, copies as well as a recovery strategy in the event the primary copy is lost or corrupted. There are many approaches to ensure that the data is available for operations and patient services.
In the past, the general approach was to back up the primary data to a tape backup. The problem with that approach is that tape recovery is typically slow and is susceptible to potential errors. Tapes can be defective, are timely to access, and are difficult to manage.
Newer methods allow for backups to be written to disk systems that are much faster and require less oversight and management. Like tape backup systems, the secondary disk copy can also be located at another remote site, on a different power system and different network to provide higher availability. Those can be mirrored disk systems for near-instant recovery of operations or can be recovery-based once the primary system is available. Mirrored systems can be synchronous, which means that the write is confirmed as consistent on both systems before committing and allowing for completion of that task.
Asynchronous solutions allow batched, or multiple write transactions, to be sent to the secondary system after the primary copy is updated and committed. The benefit of mirrored data is the potential for near-immediate transition to the secondary copy to minimize any disruption to patient activities. In many cases, the mirrored or primary copy is also backed up off-site to another copy for recovery in a more significant disaster event. Those backups can be based on a complete, or full, copy.
Typically, however, full backup is performed less frequently followed by a series of ‘incremental’ backups that contain only the changes to the full copy. Those incremental backups provide a method of change control without replicating the same unchanged data over and over. While it requires far less storage space, it is also much faster to apply in a recovery scenario. This method also provides numerous options in terms of the number of versions of the data.
Recovery point objective (RPO) determines the amount of data loss a practice is willing or able to tolerate. Recovery time objective (RTO) defines the amount of time a practice is willing or able to wait before data is available for continued operations. Both mirroring and backup methods should be chosen based on the practice requirements for recovery point and recovery time requirements and objectives.
While an EHR system will obviously provide great efficiencies and benefits to both the healthcare provider and patients, the EHR application is not the end of the story. The information that is at the core of this transition to electronic record and practice management requires significant care and attention. Using the methods and technologies mentioned above will ensure that those benefits can be realized in compliancy and with limited disruption. This puts the focus right back where it started: on the patient care.
Wade Hoffman, MBA, is an owner at EarthBend and a founding member of Tonka-Tek.