Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept study

Abstract In this proof of concept study, we demonstrated how Large Language Models (LLMs) can automate the conversion of unstructured case reports into clinical ratings. By leveraging instructions from a standardized clinical rating scale and evaluating the LLM’s confidence in its outputs, we aimed...

Full description

Saved in:
Bibliographic Details
Main Authors: Sebastian Volkmer, Alina Glück, Andreas Meyer-Lindenberg, Emanuel Schwarz, Dusan Hirjak
Format: Article
Language:English
Published: Nature Portfolio 2025-03-01
Series:Schizophrenia
Online Access:https://doi.org/10.1038/s41537-025-00601-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850095007597330432
author Sebastian Volkmer
Alina Glück
Andreas Meyer-Lindenberg
Emanuel Schwarz
Dusan Hirjak
author_facet Sebastian Volkmer
Alina Glück
Andreas Meyer-Lindenberg
Emanuel Schwarz
Dusan Hirjak
author_sort Sebastian Volkmer
collection DOAJ
description Abstract In this proof of concept study, we demonstrated how Large Language Models (LLMs) can automate the conversion of unstructured case reports into clinical ratings. By leveraging instructions from a standardized clinical rating scale and evaluating the LLM’s confidence in its outputs, we aimed to refine prompting strategies and enhance reproducibility. Using this strategy and case reports of drug-induced Parkinsonism, we showed that LLM-extracted data closely align with clinical rater manual extraction, achieving an accuracy of 90%.
format Article
id doaj-art-81de6d847b9f40c1a5b0ad8f09d8d005
institution DOAJ
issn 2754-6993
language English
publishDate 2025-03-01
publisher Nature Portfolio
record_format Article
series Schizophrenia
spelling doaj-art-81de6d847b9f40c1a5b0ad8f09d8d0052025-08-20T02:41:32ZengNature PortfolioSchizophrenia2754-69932025-03-011111410.1038/s41537-025-00601-5Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept studySebastian Volkmer0Alina Glück1Andreas Meyer-Lindenberg2Emanuel Schwarz3Dusan Hirjak4Department of Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg UniversityDepartment of Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg UniversityDepartment of Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg UniversityDepartment of Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg UniversityDepartment of Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg UniversityAbstract In this proof of concept study, we demonstrated how Large Language Models (LLMs) can automate the conversion of unstructured case reports into clinical ratings. By leveraging instructions from a standardized clinical rating scale and evaluating the LLM’s confidence in its outputs, we aimed to refine prompting strategies and enhance reproducibility. Using this strategy and case reports of drug-induced Parkinsonism, we showed that LLM-extracted data closely align with clinical rater manual extraction, achieving an accuracy of 90%.https://doi.org/10.1038/s41537-025-00601-5
spellingShingle Sebastian Volkmer
Alina Glück
Andreas Meyer-Lindenberg
Emanuel Schwarz
Dusan Hirjak
Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept study
Schizophrenia
title Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept study
title_full Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept study
title_fullStr Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept study
title_full_unstemmed Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept study
title_short Validating large language models against manual information extraction from case reports of drug-induced parkinsonism in patients with schizophrenia spectrum and mood disorders: a proof of concept study
title_sort validating large language models against manual information extraction from case reports of drug induced parkinsonism in patients with schizophrenia spectrum and mood disorders a proof of concept study
url https://doi.org/10.1038/s41537-025-00601-5
work_keys_str_mv AT sebastianvolkmer validatinglargelanguagemodelsagainstmanualinformationextractionfromcasereportsofdruginducedparkinsonisminpatientswithschizophreniaspectrumandmooddisordersaproofofconceptstudy
AT alinagluck validatinglargelanguagemodelsagainstmanualinformationextractionfromcasereportsofdruginducedparkinsonisminpatientswithschizophreniaspectrumandmooddisordersaproofofconceptstudy
AT andreasmeyerlindenberg validatinglargelanguagemodelsagainstmanualinformationextractionfromcasereportsofdruginducedparkinsonisminpatientswithschizophreniaspectrumandmooddisordersaproofofconceptstudy
AT emanuelschwarz validatinglargelanguagemodelsagainstmanualinformationextractionfromcasereportsofdruginducedparkinsonisminpatientswithschizophreniaspectrumandmooddisordersaproofofconceptstudy
AT dusanhirjak validatinglargelanguagemodelsagainstmanualinformationextractionfromcasereportsofdruginducedparkinsonisminpatientswithschizophreniaspectrumandmooddisordersaproofofconceptstudy