Comparative evaluation of handwriting recognition by large language models (LLMs) in interpreting handwritten medication lists
Tom G Jacobs, Laize Sílvia dos Anjos Botas Beca, Paul D van der Linden, Merel van Nuland · 2026
Read Comparative evaluation of handwriting recognition by large language models (LLMs) in interpreting handwritten medication lists by Tom G Jacobs, Laize Sílvia dos Anjos Botas Beca, Paul D van der Linden, Merel van Nuland (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-13241658.txt.
Description
Tom G Jacobs and co-authors report a 2026 comparison of large language models reading handwritten medication lists. Handwritten bedside lists remain common, especially where resources are limited. Reading them wrongly can cause medication errors.
Volunteers handwrote a standard list of ten medicines with dosing instructions, in Dutch. Three GPT-based models transcribed the lists, and scoring covered drug name, dose, frequency, and route. The largest model scored highest, and accuracy rose with human-rated legibility. Printed writing and blue ink were easier for the models than cursive writing and black ink. Complex dosing instructions caused the most errors. The authors see promise for digitization, provided a human still checks the output.
Catalog record
| CREATOR | Tom G Jacobs, Laize Sílvia dos Anjos Botas Beca, Paul D van der Linden, Merel van Nuland |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 15 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |