Minnesota-based Mayo Clinic is developing medical informatics tools that could give healthcare providers new methods of collecting data from electronic health records (EHRs), reports Information Week.
The nonprofit healthcare provider is using a Strategic Health IT Advanced Research Projects (SHARP) grant to develop its health information exchange, which is based on the government’s open-source CONNECT software platform.
“We’ll use [the new system] to do population health management, aggregate outcomes analysis and comparative effectiveness research pooled across healthcare encounters by different providers with the same patients,” said Christopher Chute, MD, the Mayo Clinic’s primary SHARP grant investigator, quoted by Information Week.
A primary goal of the project is to enable medical professionals to access patient data from multiple electronic sources. One challenge the researchers face is the lack of a standardized format for medical records. In an effort to address this problem, the clinic is developing an open-source natural language processing software program. The software can interpret data from a variety of sources and decipher the semantics of the language to categorize the patient data needed by physicians more effectively.
Lacey Hart, a SHARP administrator for the Mayo Clinic, said “this gets to the heart of meaningful use. It’s one thing to meet the government requirement that you should have an electronic record but it’s another thing, once you have that record, to make meaning out of it.”
At present, data from the medical records of 30 patients who have been diagnosed with diabetes has been used to test the language processing tools. Using systems that were developed in partnership with IBM’s Watson Research Center, those patient records were processed and subdivided into more than 134 billion individual pieces of data that can be processed, stored and organized for easy retrieval.
Information Week reports that the process has been tested successfully, and that research into how the data can be normalized across several EHR formats is ongoing.