clickBrick prompt engineering: optimizing large language model performance in clinical psychiatry
Falk Gerrik Verhees, Fabian Huth, Vincent Meyer, Fabian Wolf, Michael Bauer, Andrea Pfennig, Philipp Ritter, Jakob Nikolas Kather, Isabella Catharina Wiest, Pavol Mikolas · 2026
Read clickBrick prompt engineering: optimizing large language model performance in clinical psychiatry by Falk Gerrik Verhees, Fabian Huth, Vincent Meyer, Fabian Wolf, Michael Bauer, Andrea Pfennig, Philipp Ritter, Jakob Nikolas Kather, Isabella Catharina Wiest, Pavol Mikolas (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/PMC13304116.txt.
Description
Falk Gerrik Verhees and colleagues present clickBrick, a 2026 prompt-building method that helps language models read psychiatric records. Well-built prompts can raise model accuracy on clinical tasks, but gains in mental health records were untested.
Charts for 100 patients were scored by experts on 12 cross-diagnosis mental state features, then two local models tried the same task with prompts of growing structure. Balanced accuracy ranged widely across features, and structured prompts lifted the most responsive model a great deal. Features pulled with the best prompts also predicted final diagnoses in a past group of 1,692 patients.
Catalog record
| CREATOR | Falk Gerrik Verhees, Fabian Huth, Vincent Meyer, Fabian Wolf, Michael Bauer, Andrea Pfennig, Philipp Ritter, Jakob Nikolas Kather, Isabella Catharina Wiest, Pavol Mikolas |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33047 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |