Feature: The quandary of data management post-integration
From November 2007 to July 2008, Piedmont Heart Institute (PHI) in Atlanta merged with three private cardiovascular groups to form a 62-physician practice. During that process, one priority was figuring out how to make the disparate practice and hospital systems "talk" to each other. This process became a nightmare, Mark L. Cohen, chief of quality, informatics-IT and research from PHI, said during an interview.
Because each independent group had different IT infrastructures, computer networks, hardware and software components and EMRs, staff had to figure out the glue that would keep these various systems together, and for that consolidation was key, Cohen offered.
Practices moving to the hospital side have two options in terms of data management—make disparate systems talk between the hospital and the practice or consolidate patient data into one system. While systems can be run in parallel, it is often difficult and time-consuming to do so, Cohen said.
Staff must find a way to make the systems talk on the back-end while still making the process seem seamless for the front-end user. Some have turned to private health information exchanges (HIE) to connect PC infrastructures, EMRs and network data while others have decided to transfer all patient-based charts from the practice to the hospital system. But this is no easy feat.
Running two disparate systems simultaneously can be pricey. In fact, some have estimated interfaces to meld data to cost an upward of $30,000, not including vendor patches, software upgrades and fixes to implement meaningful use requirements.
“Don’t stay isolated,” Cohen suggested. “Leaving our systems as separate entities would really impede our quality and efficiency down the road.” Therefore PHI turned to consolidating practice data with the hospital data.
First things first
Cohen and staff first performed an evaluation of EMR systems and what information was necessary to make the system run flawlessly with all the proper and necessary data. Once staff chose a system, hammering out inconsistencies was key. For example, the first problem was how to handle paper-based records to ensure the information was moved over to the electronic system.
“What we almost instantly realized was that it is nearly impossible to transfer data electronically from one EMR to another,” Cohen said.
Maintaining the practice database for 10 years (the number of years hospitals/practices generally are required to maintain patient documents under federal and state laws) would have been very expensive. Instead, Cohen and staff chose to undertake the tedious process to hand move nearly one million patient documents in one fell swoop.
However, Cohen said the hospital did not realize the number of hours and staff it would take. When all was said and done, it was estimated that it took 30 minutes per patient to transfer patient data into the new electronic system.
“Even though there were documents that transferred and you could access the documents, you still must have these data in the right places,” Cohen said.
While some suggest that running both the practice system and the hospital EMR system may be more efficient and cost-effective, Cohen disagreed. “If you are running these two systems you will still need to perform maintenance and make sure the system is up to date, which will be costly,” Cohen said. That may include upgrading the system to meet clinical trending and meaningful use requirements, he said.
Cohen said that making the old practice systems virtual is important. “With virtualization, the computer system is moved onto a different server structure and the old system is frozen,” he said. “These data never change and give you the ability to keep the system and documents for a long period of time without having to constantly run the old system and maintain the upkeep.”
The process is like attempting to transfer old phone contacts from a Blackberry to an iPhone, Cohen said. “It just doesn't work. So you have to start over from scratch.”
He suggested that practices look at this project and say, “I can either do the work now and get paid, or I can do the work 10 years down the road.” Cohen summed that it would be much more expensive and painful to wait, so “just get it over with upfront.”
For more tips on how to successfully maintain patient data post-integration, check out the feature article in the upcoming February issue of Cardiovascular Business.