EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Knowledge augmented causal discovery through large language models and Knowledge Graphs: Application in chronic low back pain

Damon Lin, Marzieh Mussavi Rizi, Conor O'Neill, Jeffrey C Lotz, Paul Anderson, Abel Torres-Espin

Journal of Biomedical Informatics · 2026

DOI: 10.1016/j.jbi.2026.105102 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Biomedical Informatics, 2026. Verfasst von Damon Lin und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science and Software Engineering, California Polytechnic State University, San Luis Obispo, CA, USA. Hauptthemen laut MEDLINE: Large Language Models, Low Back Pain, Chronic Pain, Knowledge Discovery. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Institute of Arthritis and Musculoskeletal and Skin Diseases.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

OBJECTIVE: Causal discovery from observational data is limited by structural constraints of available datasets, the absence of causal logic, and the lack of external domain knowledge. We propose a Knowledge-Augmented Causal Discovery (KACD) framework that combines data-driven structural learning with LLM-based knowledge systems to overcome these limitations, with application to chronic low back pain (cLBP). METHODS: We evaluated KACD in three configurations of increasing complexity: a vanilla LLM, an LLM augmented with retrieval-augmented generation (RAG), and an LLM augmented with a Knowledge Graph-based RAG system (KG-RAG). Each configuration was benchmarked against a clinically derived, expert-validated causal graph for cLBP. Causal querying employed a decomposed two-phase prompting strategy - separately querying plausibility, statistical association, and temporal precedence - inspired by the methodology used by domain experts during ground truth construction. Performance was evaluated using F1-score, true positive rate, false discovery rate, and structural Hamming distance. RESULTS: KG-RAG with plausibility-based prompting achieved the highest F1 score (0.697), outperforming standard RAG (F1 0.689), vanilla LLM (F1 0.649), and data-driven structural learning alone (F1 0.396), however this was backbone-dependent with RAG outperforming KG-RAG depending on the LLM used. Decomposed causal prompting outperformed direct causal queries across all knowledge system configurations. Combining the data-driven proto-model with any knowledge system consistently raised the performance floor of either method individually across all prompting strategies and system configurations. CONCLUSION: LLM-driven knowledge systems can effectively augment data-driven causal discovery for biomedical causal modeling, with KG-RAG providing superior performance through cross-document synthesis and entity-centered retrieval. These findings support KACD's potential as a decision-support tool for researchers constructing causal models in complex, multi-disciplinary health domains, pending further validation of its use in practice.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Large Language ModelsLow Back PainChronic PainKnowledge DiscoveryMedical InformaticsHumansData MiningAlgorithms

Stichwörter der Autor:innen

Causal discoveryChronic lower back painKnowledge graphLarge language model

Autor:innen und Institutionen

  • Department of Computer Science and Software Engineering, California Polytechnic State University, San Luis Obispo, CA, USA.
  • School of Public Health Sciences, University of Waterloo, Waterloo, Ontario, Canada.
  • Department of Orthopaedic Surgery, University of California San Francisco, San Francisco, CA, USA.
  • Department of Computer Science and Software Engineering, California Polytechnic State University, San Luis Obispo, CA, USA. Electronic address: pander14@calpoly.edu.
  • School of Public Health Sciences, University of Waterloo, Waterloo, Ontario, Canada; Department of Neurological Surgery, University of California San Francisco, San Francisco, CA, USA. Electronic address: abel.torresesp…

Förderung

  • National Institute of Arthritis and Musculoskeletal and Skin Diseases

Zitieren

APA

Lin, D., Rizi, M. M., O'Neill, C., Lotz, J. C., Anderson, P., & Torres-Espin, A. (2026). Knowledge augmented causal discovery through large language models and Knowledge Graphs: Application in chronic low back pain. Journal of Biomedical Informatics. https://doi.org/10.1016/j.jbi.2026.105102

Vancouver

Lin D, Rizi MM, O'Neill C, Lotz JC, Anderson P, Torres-Espin A. Knowledge augmented causal discovery through large language models and Knowledge Graphs: Application in chronic low back pain. Journal of Biomedical Informatics. 2026. doi:10.1016/j.jbi.2026.105102

BibTeX
@article{Lin2026,
  author = {Lin, Damon and Rizi, Marzieh Mussavi and O'Neill, Conor and Lotz, Jeffrey C and Anderson, Paul and Torres-Espin, Abel},
  title = {Knowledge augmented causal discovery through large language models and Knowledge Graphs: Application in chronic low back pain},
  journal = {Journal of Biomedical Informatics},
  year = {2026},
  doi = {10.1016/j.jbi.2026.105102}
}
Kostenlos starten →