hh.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Leveraging LLMs for ICD Coding and Uncertainty Estimation: Can the model's awareness of the hierarchical structureof ICD-10 codes impact its prediction performance?
Halmstad University, School of Information Technology.
2025 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Manual ICD coding of clinical notes with International Classificationof Diseases (ICD) codes is a time-consuming and error-prone task,yet critical for effective healthcare management, billing, and research.This thesis investigates the use of Large Language Models (LLMs) toautomate the prediction and assignment of ICD codes from clinicalnotes, aiming to enhance both the efficiency and accuracy of the pro-cess while reducing the reliance on human coders.We developed a novel approach to leverage the hierarchical struc-ture of ICD codes, which reflects the nested and interrelated natureof diagnoses, to improve the predictive capabilities of LLMs. By incor-porating this hierarchical awareness, the model is better equipped tocapture the dependencies among medical codes, thereby minimizingerrors associated with irrelevant or less likely disease classifications.Additionally, we implemented uncertainty estimation techniquesduring the inference phase to assess the confidence of the model’spredictions. By quantifying prediction uncertainty, we can identifycases where the model may be less certain, providing valuable in-sights into areas where human oversight might still be necessary.Our results show that integrating hierarchical structures reducesthe likelihood of predicting irrelevant diseases that are distant fromthe actual disease in the hierarchical tree, highlighting the potentialof this approach to improve ICD code assignment

Place, publisher, year, edition, pages
2025.
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:hh:diva-55372OAI: oai:DiVA.org:hh-55372DiVA, id: diva2:1933618
Educational program
Master's Programme in Information Technology, 120 credits
Examiners
Available from: 2025-01-31 Created: 2025-01-31 Last updated: 2025-10-01Bibliographically approved

Open Access in DiVA

fulltext(1342 kB)470 downloads
File information
File name FULLTEXT02.pdfFile size 1342 kBChecksum SHA-512
fdc061d5175cd7b8c58d37388e3881a6a573d9b04777c14e2a143e66ecea26c4f2fd5459caee41271ac751dd68907e538151dfde705e9c7f82d7d84643e6af30
Type fulltextMimetype application/pdf

By organisation
School of Information Technology
Computer and Information Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 471 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 970 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf