Business
Researchers to study data from VA EMR system
■ The initiative will allow VA-affiliated physicians to discover better ways to handle various diseases, including cancer and congestive heart failure.
By Pamela Lewis Dolan — Posted June 8, 2009
- WITH THIS STORY:
- » Related content
The U.S. Dept. Of Veterans Affairs for the first time is opening up its electronic medical records to allow researchers from across the system to look at the data.
The de-identified, aggregated data of veterans will allow researchers to pinpoint the most effective treatments for specific conditions, including posttraumatic stress disorder and antibiotic-resistant staph infection.
The VA says the result will be broader clinical studies that will provide physicians, both inside and out of the system, with better data on the best treatment methods for various conditions. The project will also show how the same data-mining methods could be used in other health information exchanges, including the national health information network, once it's fully up and running.
Matthew Samore, MD, an infectious disease and epidemiology physician from the VA Salt Lake City Health Care System, who is heading Utah's part in the project, said researchers with the VA, like those within other health care networks, have traditionally only had access to the data from their local facilities.
In addition to posttraumatic stress disorder and methicillin-resistant Staphylococcus aureus, researchers with the project, called the Consortium for Healthcare Informatics Research, also will be collecting chest x-ray reports, and data on patients with cancer or congestive heart failure.
With the VA being one of the largest health care systems in the nation, research on the clinical processes used on its millions of patients could prove very valuable, experts say. While data will only be accessible by those affiliated with the VA, the research will be made publicly available for use by physicians everywhere through published research and white papers.
Dr. Samore said the VA, like many other health care systems, spends millions per year on the manual retrieval of clinical process data from its EMR system. With an open system that uses natural language processing (a way for computers to understand and process actual language), researchers will be able to take a large set of de-identified data and analyze it a more timely and cost-effective manner.
Dr. Samore said many will be looking at the VA as a testing ground for how these systems of data retrieval can be used elsewhere, once more hospitals become connected to health information exchanges.
As the rest of the health care industry is in the early stages of EMR adoption, it is a long way off from using these systems and HIEs in this way, said Dr. Samore. And systems that are technically able to collect data this way have a limited data pool to collect from. The VA is considered among the most advanced when it comes to electronic medical record implementation, making it a logical site, and testing ground, for a program such as this.
But implementing lessons learned from the VA at other health information exchanges may prove challenging, experts say.
"The VA is a closed system. They know that [EMR] system inside and out. When you take that and apply it to the commercially available products, yes they are EMRs but their data model, their software model may be different," said Pam Matthews, senior director of health care information systems for the Healthcare Information and Management Systems Society.
Matthews said while research may help the VA redesign some of the clinical decision support tools within its EMR system, it won't be that easy to implement changes across an HIE made up of multiple types of systems and various methods of implementation.
But, she said, "There are definitely things we can learn from their experiences to take outside the VA system. There's always something to learn."
Dr. Samore said the biggest issue for the VA has been making sure patient information is kept secure. Because health information is protected, it's been difficult for researchers and physicians to use the data in a useful way.
The VA has also fallen under scrutiny for its ability to protect patient privacy after a series of events, including the theft of a laptop computer containing the data of 26 million veterans. Therefore, it's being very cautious about how the consortium is implemented. Every person who will be granted access to the data not only has to be affiliated with the VA system but has to go through training on how to handle and safely access the data. Additionally, data will not be available for download on local hard drives, but will be stored in one central repository that users can access electronically.
Dr. Samore said the impact of this project might not be considered revolutionary by many people's standards. For many physicians, he said, the project is too "behind the scenes" to mean anything. But he expects the impact to be realized over the next four to five years.
"It will improve the scope, the scale, the quality and I think ultimately that it will not only inform new guidelines but help resolve some conflicts in current guidelines," he said.