EUVIMEDEUVIMED entdecken
VergleichsstudieOpen Access

Constructing synthetic datasets with generative artificial intelligence to train large language models to classify acute renal failure from clinical notes

Onkar Litake, Brian H Park, Jeffrey L Tully iD, Rodney A Gabriel

Journal of the American Medical Informatics Association · 2024

DOI: 10.1093/jamia/ocae081 ↗

Kurz gefasst

Vergleichsstudie in der Zeitschrift Journal of the American Medical Informatics Association, 2024. Verfasst von Onkar Litake und 3 weiteren Autor:innen. Beteiligte Einrichtung: Division of Perioperative Informatics, Department of Anesthesiology, University of California, San Diego, La Jolla, CA 92037, United States. Hauptthemen laut MEDLINE: Acute Kidney Injury, Artificial Intelligence, Electronic Health Records. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 14 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NIGMS NIH HHS.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objectives

To compare performances of a classifier that leverages language models when trained on synthetic versus authentic clinical notes. Materials and Methods A classifier using language models was developed to identify acute renal failure. Four types of training data were compared: (1) notes from MIMIC-III; and (2, 3, and 4) synthetic notes generated by ChatGPT of varied text lengths of 15 (GPT-15 sentences), 30 (GPT-30 sentences), and 45 (GPT-45 sentences) sentences, respectively. The area under the receiver operating characteristics curve (AUC) was calculated from a test set from MIMIC-III.

Results

With RoBERTa, the AUCs were 0.84, 0.80, 0.84, and 0.76 for the MIMIC-III, GPT-15, GPT-30- and GPT-45 sentences training sets, respectively.

Discussion

Training language models to detect acute renal failure from clinical notes resulted in similar performances when using synthetic versus authentic training data.

Conclusion

The use of training data derived from protected health information may not be needed.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Acute Kidney InjuryArtificial IntelligenceElectronic Health RecordsHumansROC CurveNatural Language ProcessingArea Under CurveDatasets as Topic

Stichwörter der Autor:innen

ChatGPTartificial intelligencegenerative AIlarge language models

Autor:innen und Institutionen

  • Division of Perioperative Informatics, Department of Anesthesiology, University of California, San Diego, La Jolla, CA 92037, United States.
  • Department of Biomedical Informatics, University of California, San Diego Health, La Jolla, CA 92037, United States.

ORCID-Profile: Jeffrey L Tully

Förderung

  • NIGMS NIH HHS

Zitieren

APA

Litake, O., Park, B. H., Tully, J. L., & Gabriel, R. A. (2024). Constructing synthetic datasets with generative artificial intelligence to train large language models to classify acute renal failure from clinical notes. Journal of the American Medical Informatics Association. https://doi.org/10.1093/jamia/ocae081

Vancouver

Litake O, Park BH, Tully JL, Gabriel RA. Constructing synthetic datasets with generative artificial intelligence to train large language models to classify acute renal failure from clinical notes. Journal of the American Medical Informatics Association. 2024. doi:10.1093/jamia/ocae081

BibTeX
@article{Litake2024,
  author = {Litake, Onkar and Park, Brian H and Tully, Jeffrey L and Gabriel, Rodney A},
  title = {Constructing synthetic datasets with generative artificial intelligence to train large language models to classify acute renal failure from clinical notes},
  journal = {Journal of the American Medical Informatics Association},
  year = {2024},
  doi = {10.1093/jamia/ocae081}
}
Kostenlos starten →