Finding the Right Data Before Imaging Starts
Clinicians are drowning in a sea of data from electronic health records; the more sophisticated systems have a zillion structured fields and a growing mound of unstructured data, says Michael Ethan Zalis, M.D., an associate radiologist at Massachusetts General Hospital.
Clinicians are drowning in a sea of data from electronic health records; the more sophisticated systems have a zillion structured fields and a growing mound of unstructured data, says Michael Ethan Zalis, M.D., an associate radiologist at Massachusetts General Hospital.
At RSNA 2014, Nov. 30-Dec. 5 in Chicago, Zalis will speak about use of natural language processing technology to help radiologists more easily find the information they need when reading diagnostic imaging exams. Zalis also serves as a co-founder and chief medical officer at QPID Health, a spin-off of NLP technology developed at Massachusetts General.
When reading images, radiologists cant sift through all the information in an EHR to get the essential pieces of the patients narrative, he explains. They have to hunt for a problem list and last note, and they may be outdated.
Natural language processing enables physicians to express concepts when conducting an exam. A radiologist may be reading a CT scan and think about referring the patient for an MRI. But before doing that, he or she needs to know if the patient has body incompatibilities such as a pacemaker, nerve stimulator or prosthesis that would make the procedure unsafe. But emerging NLP products are beginning to enable clinicians to ask the computer: Does the patient have this, this or this?
That capability will enable a radiologist to conduct automated searches for specific information in an EHR before the start of an exam, Zalis says. Programming the searches would enable a radiologist to identify before imaging starts if there could be airway or heart issue, or if a specific drug could harm the kidneys. The NLP technology will trove through records and identify the words and phrases that suggest the presence of concepts being searched. So, you could program ahead of time the same questions you want to know, categorized by the types of systems you have. Massachusetts General, using the technology developed in-house and now commercially available, has been doing this for several years, he adds.
Radiologists need this technology in the emerging fee for value environment, Zalis contends. It will turn a radiologist from a commodity image reader into someone who really knows whats going on with the patients. There are emerging tools that can transform a radiologists ability to become an informed and efficient consultant.
The session, Natural Language Processing to Solve Problems in Clinical Practice, is the first of a group of NLP presentations starting at 2:30 on Dec. 4.
At RSNA 2014, Nov. 30-Dec. 5 in Chicago, Zalis will speak about use of natural language processing technology to help radiologists more easily find the information they need when reading diagnostic imaging exams. Zalis also serves as a co-founder and chief medical officer at QPID Health, a spin-off of NLP technology developed at Massachusetts General.
When reading images, radiologists cant sift through all the information in an EHR to get the essential pieces of the patients narrative, he explains. They have to hunt for a problem list and last note, and they may be outdated.
Natural language processing enables physicians to express concepts when conducting an exam. A radiologist may be reading a CT scan and think about referring the patient for an MRI. But before doing that, he or she needs to know if the patient has body incompatibilities such as a pacemaker, nerve stimulator or prosthesis that would make the procedure unsafe. But emerging NLP products are beginning to enable clinicians to ask the computer: Does the patient have this, this or this?
That capability will enable a radiologist to conduct automated searches for specific information in an EHR before the start of an exam, Zalis says. Programming the searches would enable a radiologist to identify before imaging starts if there could be airway or heart issue, or if a specific drug could harm the kidneys. The NLP technology will trove through records and identify the words and phrases that suggest the presence of concepts being searched. So, you could program ahead of time the same questions you want to know, categorized by the types of systems you have. Massachusetts General, using the technology developed in-house and now commercially available, has been doing this for several years, he adds.
Radiologists need this technology in the emerging fee for value environment, Zalis contends. It will turn a radiologist from a commodity image reader into someone who really knows whats going on with the patients. There are emerging tools that can transform a radiologists ability to become an informed and efficient consultant.
The session, Natural Language Processing to Solve Problems in Clinical Practice, is the first of a group of NLP presentations starting at 2:30 on Dec. 4.
More for you
Loading data for hdm_tax_topic #better-outcomes...