Using large language models to enhance clinically-driven missing data recovery algorithms in electronic health records
Sarah C Lotspeich, Abbey N Collins, Brian J Wells, Ashish K Khanna, Joseph Rigdon, Lucy D’Agostino McGowan · 2026
Read Using large language models to enhance clinically-driven missing data recovery algorithms in electronic health records by Sarah C Lotspeich, Abbey N Collins, Brian J Wells, Ashish K Khanna, Joseph Rigdon, Lucy D’Agostino McGowan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13228140.txt.
Description
Sarah C Lotspeich and colleagues test a 2026 way to heal gaps in electronic health records without paying clinicians to re-read every chart. Their earlier method used a clinician-built roadmap of related diagnoses to recover missing values, but manual review scales badly. Large language models were asked to expand the roadmap, with clinicians approving each addition, and the resulting algorithm was checked against one hundred hand-reviewed patients.
Those charts had 413 missing values. Expert reviewers recovered about twelve percent of them, while language-model-enhanced algorithms proposed roughly twenty percent, at similar accuracy. The clinician-approved final version landed in between. Applied to a thousand patients, the method raised the median count of available values per patient, showing a feasible route to cleaner research data.
Catalog record
| CREATOR | Sarah C Lotspeich, Abbey N Collins, Brian J Wells, Ashish K Khanna, Joseph Rigdon, Lucy D’Agostino McGowan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |