Health and Wellness Informatics News
Within a few minutes, the software has identified 54000 patients with the condition. It is a process that would have taken years for physicians to perform manually.
Healthcare is quite squarely in the age of big data and data analytics. It is still quite difficult in clinical research to identify patients accurately with various complex conditions like valvular heart disease with medical records. However, the implementation of AI and NLP can further help to organize the system.
In case the researchers cannot identify these patients, they cannot study them or track the practice patterns or conduct any population management.
The part of this problem is that the current methods were necessary to identify highly specific conditions like valvular heart disease diagnosis and procedure codes. These were in use for billing purposes and often not much useful for clinical care. These can be quite non-specific and may not include detailed data regarding the condition.
A patient with moderate or severe aortic stenosis is entirely different from a patient with mild valve disease. Yet, some use the codes for aortic valve disease that can apply to an entirely different clinical problem. Without any accurate or systematic case identification and population management and research for valvular heart, conditions are not possible.
The data needs to identify patients with valvular heart disease from a report easily.
NLP is a branch from AI where a complex set of rules can read free-text reports. It then created a structured and systematic database. Upon the accomplishment, there is potential for both the studying of population and to perform high-quality population management.
The hospital used software applications as the form of an architecture to build and validate NLP tools. Then they used the algorithms to apply to their entire dataset with EHR. It also involves data organization from the backend EHR systems. Then it runs formatted data through the software to create an organized and structured dataset.
They gained first success with the development and validation of the technology. They took the next step to apply the technology to their echocardiography database within Kaiser Permanente Northern California. It includes one million echocardiography reports from the past decade.
It helped them to identify the patient’s hand to extract all the key elements from each of the echocardiography reports. They are now using it to examine the practice patterns and the outcome for their patients.