Performance of Zero-Shot Classifiers for Categorizing RCT Abstracts by Intervention Type: Validation Study
JMIR Medical Informatics · 18.06.2026 · 14 · e77943-e77943
DOI: 10.2196/77943 ↗Kurz gefasst
Validierungsstudie in der Zeitschrift JMIR Medical Informatics, online erschienen am 18.06.2026. Verfasst von Diana Buitrago-Garcia und 4 weiteren Autor:innen. Beteiligte Einrichtung: Division of Rheumatology, Geneva University Hospitals and University of Geneva, Rue Alcide-Jentzer 22, Genève, 1205, Switzerland, 41 022 372 36 78. Hauptthemen laut MEDLINE: Abstracting and Indexing, Randomized Controlled Trials as Topic, Artificial Intelligence. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: Swiss National Science Foundation.
Abstract (Originalsprache)
Background
Artificial intelligence has gained relevance due to its potential to reduce the workload in evidence synthesis or bibliometric projects. While the main focus has been lately on the use of instruction-tuned large language models, zero-shot classification models have not been tested for such task. These models are large language models trained on large datasets of labeled data able to categorize text among proposed labels, irrespective of the text domain or the topic. They are relatively small, able to run on consumer-grade computers, and almost hyperparameter-free.
Objective
In our study, we use abstracts of randomized clinical trials in rheumatology as a case example to evaluate the performance of openly available, generalist, zero-shot classification models in classifying types of interventions against a human gold standard.
Methods
We classified all rheumatology RCT abstracts published between 2009 and 2022 (n=1,054) as “drug” or “non-drug” using two zero-shot text classification models (DeBERTa and BART) and few-shot prompting using Llama3 8B. Different labeling of categories provided to the zero-shot classification models and different prompts provided to Llama3 8B were tested. Performance was evaluated using accuracy and predictive value of both categories against a human gold standard.
Results
Most randomized controlled trials, RCTs (452/1054, 42.9%) assessed drug interventions. The DeBERTa model achieved the highest accuracy (929/1054, 88.1%; 95% CI 86%‐90%) when using the “drug” and “non-drug” labels. Llama3 8B and few-shot prompting had slightly higher accuracy and predictive values. Both zero-shot and Llama3 8B models had performance on par with a human without experience in evidence synthesis (905/1054, 85.9%; 95% CI 83.6%‐87.8% accuracy). Misclassifications occurred for trials where the intervention was harder to classify, such as procedures (eg, intra-articular injections), food compounds, vitamins, supplements, or biological treatments.
Conclusions
This study shows the potential of zero-shot classification models for simple classification tasks, demonstrating accuracy comparable to that of an untrained human. These models are potential tools to streamline systematic review tasks for bibliometric studies in classifying abstracts by supplementing one reviewer.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Division of Rheumatology, Geneva University Hospitals and University of Geneva, Rue Alcide-Jentzer 22, Genève, 1205, Switzerland, 41 022 372 36 78.
- Surgery Department, Geneva University Hospitals and University of Geneva, Geneva, Geneva, Switzerland.
ORCID-Profile: Diana Buitrago-Garcia, Delphine S Courvoisier, Michele Iudici, Denis Mongin
Förderung
- Swiss National Science Foundation
Zitieren
APA
Buitrago-Garcia, D., Courvoisier, D. S., Capderou, S., Iudici, M., & Mongin, D. (2026). Performance of Zero-Shot Classifiers for Categorizing RCT Abstracts by Intervention Type: Validation Study. JMIR Medical Informatics, 14, e77943-e77943. https://doi.org/10.2196/77943
Vancouver
Buitrago-Garcia D, Courvoisier DS, Capderou S, Iudici M, Mongin D. Performance of Zero-Shot Classifiers for Categorizing RCT Abstracts by Intervention Type: Validation Study. JMIR Medical Informatics. 2026;14:e77943-e77943. doi:10.2196/77943
BibTeX
@article{BuitragoGarcia2026,
author = {Buitrago-Garcia, Diana and Courvoisier, Delphine S and Capderou, Sami and Iudici, Michele and Mongin, Denis},
title = {Performance of Zero-Shot Classifiers for Categorizing RCT Abstracts by Intervention Type: Validation Study},
journal = {JMIR Medical Informatics},
year = {2026},
volume = {14},
pages = {e77943--e77943},
doi = {10.2196/77943}
}
EUVIMED