JAN: Joint Attention Networks for Automatic ICD Coding

Wu, Yuzhou; Chen, Zhigang; Yao, Xin; Chen, Xuechen; Zhou, Zeren; Xue, Jian

doi:10.1109/jbhi.2022.3189404

Cited by 13 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, other multilabel classifications were not affected by the faux label-wise attention. In [2], a brand-new end to end Joint Attention Network was created to integrate label correlation with long-tailed label dissemination. JAN collected the semantic data from the label description and the text of the clinical document in accordance.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Nevertheless, the pseudo label wise attention could not be applied to other multi-label categories. In order to include longtailed label, a new end to end Joint Attention Network (JAN) was created in [2]. Crossregional biological contacts are promoted, medical payment methods are supported, and medical disorders are diagnosed using International Classification of Diseases (ICD) codes.…”

Section: International Classification Of Disease Codingmentioning

confidence: 99%

See 1 more Smart Citation

Analysis on Automatic International Classification of Disease Coding with Medical Records

Joseph,

P Vijayan

2024

E3S Web of Conf.

View full text Add to dashboard Cite

The clinical concepts in the information gathered from the healthcare services are categorized and standardized using medical coding. The International Classification of Diseases (ICD) includes codes for various diseases that have an impact on financing, reporting, and research. In order to provide patient care and billing, medical coding allocates a subset of ICD codes to each patient visit. Medical personnel must spend a lot of time and effort on manual medical coding, which can lead to missed revenue and claim denials. Different studies on machine learning achieved promising performance for automated medical coding. Many researchers carried out their research on ICD. But, heterogeneous mode of operations by doctors and diagnosis methods makes the medical coding as more complex one. Furthermore, the current ICD approaches did not reduce computational complexity or increase accuracy. To address these problems, a range of deep learning and machine learning approaches are tested for ICD.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Section: International Classification Of Disease Codingmentioning

confidence: 99%

Analysis on Automatic International Classification of Disease Coding with Medical Records

Joseph,

P Vijayan

2024

E3S Web of Conf.

View full text Add to dashboard Cite

show abstract

“…Nevertheless, these methods still rely on dictionary mapping for coding. Additionally, deep learning models [18,19] have been employed to address the aforementioned issues by modeling large amounts of historical coding data. However, the complexity of application coding rules poses challenges for such methods, and the requirement for a substantial volume of historical coding data makes it less feasible for early adoption of new disease classification standards.…”

Section: Introductionmentioning

confidence: 99%

Automated Generation of ICD-11 Cluster Codes for Precision Medical Record Classification

Feng,

Zhang,

Chen

et al. 2024

INT J COMPUT COMMUN, Int. J. Comput. Commun. Control

View full text Add to dashboard Cite

Accurate clinical coding using the International Classification of Diseases (ICD) standard is essential for healthcare analytics. ICD-11 introduces new coding guidelines and cluster structures, posing challenges for existing coding tools. This research presents an automated approach to generate valid ICD-11 cluster codes from medical text. Natural language records are represented as vectors and compared to an ICD-11 corpus using cosine similarity. A bidirectional matching technique then refines similarity estimation. Experiments demonstrate the method yields up to 0.91 F1 score in coding accuracy, significantly outperforming a baseline tool. This work enables efficient high-quality ICD-11 coding to support healthcare informatics.

show abstract