A Systematic Literature Review of Automated ICD Coding and Classification Systems using Discharge Summaries

Kaur, Rajvir; Ginige, Jeewani Anupama; Obst, Oliver

doi:10.48550/arxiv.2107.10652

Cited by 6 publications

(13 citation statements)

References 79 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since bagof-word features were fed to XGBoost, and XGBoost basically has feature selection within itself, it is more similar to our best performing model, TF-IDF with dynamic feature space. Few studies report the efficacy of Doc2vec in coding automation processes [3].Li et al [9] proposed an ICD-9 coding method using a deep learning framework called DeepLabeler, which combines a convoluted neural network (CNN) with Doc2vec to assign ICD codes. They concluded that Doc2vec was critical to their prediction accuracy, and their DeepLabeler outperformed both hierarchy-based and flat-SVM approaches.…”

Section: Discussionmentioning

confidence: 99%

“…Failure to code correctly can result in inadequate patient care and may lead to an increase in expenses or delays in the reimbursement process. Recent studies regarding the automation of the coding process have tried an array of techniques, ranging from traditional text matching to deep learning-based approaches, to categorize clinical notes [3]– [11].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Systematic Evaluation of Common Natural Language Processing Techniques to Codify Clinical Notes

Tavabi

Singh

Pruneski

et al. 2022

Preprint

View full text Add to dashboard Cite

Proper codification of medical diagnoses and procedures is essential for optimized health care management, quality improvement, research, and reimbursement tasks within large healthcare systems. Assignment of diagnostic or procedure codes is a tedious manual process, often prone to human error. Natural Language Processing (NLP) have been suggested to facilitate these manual codification process. Yet, little is known on best practices to utilize NLP for such applications. Here we comprehensively assessed the performance of common NLP techniques to predict current procedural terminology (CPT) from operative notes. CPT codes are commonly used to track surgical procedures and interventions and are the primary means for reimbursement. The direct links between operative notes and CPT codes makes them a perfect vehicle to test the feasibility and performance of NLP for clinical codification. Our analysis of 100 most common musculoskeletal CPT codes suggest that traditional approaches (i.e., TF-IDF) can outperform resource intensive approaches like BERT, in addition to providing interpretability which can be very helpful and even crucial in the clinical domain. We also proposed a complexity measure to quantify the complexity of a classification task and how this measure could influence the effect of dataset size on model's performance. Finally, we provide preliminary evidence that NLP can help minimize the codification error, including mislabeling due to human error.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Systematic Evaluation of Common Natural Language Processing Techniques to Codify Clinical Notes

Tavabi

Singh

Pruneski

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Emphasis to process this kind of data had also been given within several task-specific academic challenges 1 (Huang and Lu, 2015;Dörendahl et al, 2019;Kelly et al, 2019). A recent systematic literature review regarding ICD coding systems was published by Kaur et al (2021). Gehrmann et al (2017) applied convolutional neural networks (CNNs) to the recognition of ten disease phenotypes using 1,610 manually annotated discharge summaries from the MIMIC-III corpus.…”

Section: Background and Related Workmentioning

confidence: 99%

Secondary Use of Clinical Problem List Entries for Neural Network-Based Disease Code Assignment

Kreuzthaler¹,

Schulz²

2021

Preprint

View full text Add to dashboard Cite

Clinical information systems have become large repositories for semi-structured annotated healthcare data, which have reached a critical mass that makes them interesting for supervised data-driven neural network approaches. We explored automated coding of 50 character long clinical problem list entries using the International Classification of Diseases (ICD-10) and evaluated three different types of network architectures on the top 100 ICD-10 three-digit codes. A fastText baseline reached a macro-averaged F 1 -measure of 0.83, followed by a character-level LSTM with a macro-averaged F 1 -measure of 0.84. Top performing was a downstreamed RoBERTa model using a custom language model with a macro-averaged F 1 -measure of 0.88. A neural network activation analysis together with an investigation of the false positives and false negatives unveiled inconsistent manual coding as a main limiting factor.

show abstract

“…For example, the presence of an ICD code (structured) may be used as a proxy for clinician-confirmed diagnosis (unstructured) for a certain condition. However, the use and accuracy of ICD codes may vary substantially across sites, is not validated, and can present challenges when updating to newer versions [7][8][9] . Additionally, CPT codes are frequently miscoded 10 .…”

Section: Introductionmentioning

confidence: 99%

Selective prediction for extracting unstructured clinical data

Swaminathan

López

Wang

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: Electronic health records represent a large data source for outcomes research, but the majority of EHR data is unstructured (e.g. free text of clinical notes) and not conducive to computational methods. While there are currently approaches to handle unstructured data, such as manual abstraction, structured proxy variables, and model-assisted abstraction, these methods are time-consuming, not scalable, and require clinical domain expertise. This paper aims to determine whether selective prediction, which gives a model the option to abstain from generating a prediction, can improve the accuracy and efficiency of unstructured clinical data abstraction. Methods: We trained selective prediction models to identify the presence of four distinct clinical variables in free-text pathology reports: primary cancer diagnosis of glioblastoma (GBM, n = 659), resection of rectal adenocarcinoma (RRA, n = 601), and two procedures for resection of rectal adenocarcinoma: abdominoperineal resection (APR, n = 601) and low anterior resection (LAR, n = 601). Data were manually abstracted from pathology reports and used to train L1-regularized logistic regression models using term-frequency-inverse-document-frequency features. Data points that the model was unable to predict with high certainty were manually abstracted. Findings: All four selective prediction models achieved a test-set sensitivity, specificity, positive predictive value, and negative predictive value above 0.91. The use of selective prediction led to sizable gains in automation (anywhere from 57% to 95% reduction in manual abstraction of charts across the four outcomes). For our GBM classifier, the selective prediction model saw improvements to sensitivity (0.94 to 0.96), specificity (0.79 to 0.96), PPV (0.89 to 0.98), and NPV (0.88 to 0.91) when compared to a non-selective classifier. Interpretation: Selective prediction using utility-based probability thresholds can facilitate unstructured data extraction by giving "easy" charts to a model and "hard" charts to human abstractors, thus increasing efficiency while maintaining or improving accuracy.

show abstract

A Systematic Literature Review of Automated ICD Coding and Classification Systems using Discharge Summaries

Cited by 6 publications

References 79 publications

Systematic Evaluation of Common Natural Language Processing Techniques to Codify Clinical Notes

Systematic Evaluation of Common Natural Language Processing Techniques to Codify Clinical Notes

Secondary Use of Clinical Problem List Entries for Neural Network-Based Disease Code Assignment

Selective prediction for extracting unstructured clinical data

Contact Info

Product

Resources

About