New text medical notes search tool for researchers

The University at Buffalo Institute for Healthcare Informatics (IHI), a UB Clinical and Translational Science Institute (CTSI) core facility, has developed a new tool that allows researchers to search text medical notes in the IHI’s electronic health records (EHRs). 

The IHI is dedicated to improving the health of communities locally, nationally, and globally through the secure storage, aggregation, provision, and innovative analysis of healthcare data. The institute collaborates with Buffalo Translational Consortium investigators to provide them with the necessary tools and support for their research. 

“This new search tool serves as an entry point for exploring more intricate text indexing methodologies, including Natural Language Processing and the integration of large language models,” says UB Chief Research Information Officer Jonathan Blaisure. “As the University at Buffalo enhances its artificial intelligence capabilities, we intend to implement these techniques in the future.”

“Researchers can ask us to do specific keyword searches in text medical notes,” says IHI Data Steward Viraj Kothari. “For example, if a researcher is interested to see how many notes or patients have specific keywords, like a medical notation, some simple keyword examples would be ‘diabetes’ or ‘blood pressure.’”

Kothari says the tool could be beneficial for researchers seeking to find patients that are not properly coded through medical codes. “Or, when there are some keywords a researcher believes would be more likely to be included in medical notes, the tool could help them to get much narrower results to their inquiry.”

To access the new search tool, investigators must put in a ticket through the IHI request form found here. 

“We will then collaborate with the researcher to carry out the search,” says Kothari. “Once completed, we will provide them with a breakdown overview of the result. If they have proper documentation, we can also work with them to set up a virtual machine for them to access those notes for review.”

The IHI’s Clinically Integrated Healthcare Data Repository (CIDR) holds electronic health record and claims data for most of the Western New York population. The CIDR currently holds fully identified EHR data from Kaleida Health and UB|MD for more than 1 million individuals. This data is readily available through TriNetX, a state-of-the-art end-user query tool that allows users to search de-identified clinical patient data for feasibility, recruitment, and project enrollment.

“By leveraging the CIDR, researchers can gain access to comprehensive and diverse datasets that enable a wide range of research endeavors,” says IHI Executive Director Peter Winkelstein, MD, MS, MBA, Chief Medical Informatics Officer, UBMD; Chief Medical Informatics Officer, Kaleida Health; and Professor, Department of Pediatrics, Jacobs School.

Visit the IHI core pages on the CTSI website to learn more about the institute and how it can support your research, and watch the CTSI’s Translational Spotlight newsletter for updates on new data additions to the IHI repository. See a July 2024 CTSI feature for additional information on EHR datasets available through the IHI.