CTSA News

The Electronic Medical Record Search Engine (EMERSE) empowers researchers to extract essential information from the free-text of electronic medical records with previously unimaginable speed and accuracy.

EMERSE is housed at the Michigan Institute for Clinical and Health Research (MICHR), which is part of the National Institutes of Health (NIH) National Center for Advancing Translational Sciences (NCATS) Clinical and Translational Science Award (CTSA) Program. Translational science efforts at MICHR focus on streamlining the research process, helping studies run more efficiently, deliver better results, and create greater impact. Since its development in 2005, EMERSE has been a prime example of translational science in action – evolving to open research doors previously locked by the complexity of free-text clinical notes.

Cracking the Code of Clinical Notes

Researchers often rely on standardized fields, such as diagnosis codes or medication lists, to guide their studies. But a goldmine of information hides in clinician notes: the unstructured, free-text documentation of patient encounters. These notes are where physicians may narrate symptoms and describe rare conditions, for example.

Searching this free-text used to be, in the words of Dr. David Hanauer, Informatics Faculty Lead at MICHR and the original developer of EMERSE, a Sisyphean task: “There were just tons of people here at the university whose job it was to go through charts and find information, and they would often say, ‘I’m looking for this one bit, I don’t know if they even have it, and I can’t find it anywhere,’ and they would open up note after note after note.”

Researchers could spend hours, days, or even months searching free-text notes and still potentially miss important data. This inefficiency was a major barrier, especially for research on rare diseases lacking standardized billing codes, where vital details can only be found in healthcare providers’ notes.

Intuitive Searching with No Tech Expertise Needed

With a user-friendly interface, authorized researchers can search clinical notes simply by entering a word or phrase. EMERSE doesn’t just match terms; it includes thousands of synonyms, covering alternatives for clinical concepts and both generic and brand names of medications, ensuring researchers don’t miss critical data due to wording differences.

With the proper oversight and approvals, researchers can securely search across lists of patient medical record numbers, or hunt for patients who mention specific terms—whether or not they know who those patients are. EMERSE turns what once consumed weeks or even months into a matter of minutes.

What makes EMERSE truly unique is its ongoing development. Twice a year, the project team holds community meetings to gather user feedback. The real-world insights from these meetings help EMERSE continually expand and improve, adding features as needs arise.

From Painstaking to Painless and Powerful Searches

Before EMERSE, simple word searches might fail simply due to phrasing. For example, a search for “headache” would return both patients with headaches and those with notes like “patient reports no headache.” EMERSE tackled this challenge with natural language processing, which recognizes context and negations. Future upgrades will soon include optical character recognition, meaning EMERSE can now sift through scanned, handwritten, or uploaded documents.

Today, EMERSE isn’t just changing research at Michigan. It’s open-source and free to use, and currently supports research at almost 20 sites nationwide, including nearly 700 studies here at Michigan, turbocharging research that had previously been impossible or painfully slow.

By turning user feedback into practical features, continuous improvement, and new avenues of discovery, EMERSE embodies translational science by enabling researchers to ask and answer questions that once seemed out of reach, speeding research across every disease and condition.