Here's a new term for many of you: "cognitive informatics." It's an approach to information processing that builds technology around how users--clinicians, in the case of EMRs--think and make decisions.
"We will devise guidelines and software products for vendors to meet minimum standards in decision support to ensure that the data is usable, to ensure the best possible representation of patient history and a quick and easy way to visualize this history so they don't have to do a lot of research," Vimla Patel, co-director of the Center for Cognitive Informatics and Decision Making at the University of Texas Health Science Center's School of Biomedical Informatics in Houston, tells Federal Times. "What we need is an automated system that can summarize all of a patient's history and his current state,"
Patel is leading one of four research teams nationally that are part of the Office of the National Coordinator for Health Information Technology's stimulus-funded Strategic Health IT Advanced Research Projects (SHARP) program. The Center for Cognitive Informatics and Decison Making is working with EMR vendor Cerner to devise ways of mining the vast storehouses of patient data typical of EMRs to deliver only the most pertinent elements to physicians and other busy healthcare professionals.
Cerner has helped customers create about 20 disease-specific summary pages that let users pull up data relevant to individual patients and easily enter new information as necessary. The Kansas City, Mo.-based vendor also is working on a "semantic search" function that VP of Medical Informatics Dr. David McCallie calls "Google for the health chart." Explains McCallie, "You can enter a key word or phrase to search a patient's entire history record and get just that information you are looking for. If, for example, you search 'rash,' it understands 50 or 60 different conditions that qualify as rash and reports all of them."
Much of this work, McCallie says, can apply to what SHARP is trying to accomplish.
Another SHARP team, at Mayo Clinic, is looking to add "metadata" to EMRs, applying natural language processing to map physician notes to a machine-readable, coded language to assist in data mining.
"We will be able to learn how the population was identified and when. This is additional information about the data. With a list of patients and this information about them, it may be possible to find anomalies and to study them without having had to do a manual chart review," Mayo's SHARP program director, Lacey Hart, is quoted as saying.
"This would be quick and efficient and, from a patient-care perspective, it means we would have information in near-real time and can make accurate decisions regarding such things as treatment or research," Hart explains.
To delve deeper into this fascinating subject:
- read this Federal Times story
- check out this story about UT's Center for Cognitive Informatics and Decision Making
HHS awards $144M for advanced HIT research, workforce training
ONC to outsource HIT program assessment and management
Privacy, security 'tiger team' told granular consent technology is flawed