Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Neural machine translation of clinical procedure codes for medical diagnosis and uncertainty quantification

Pei-Hung Chung, Shuhan He, Norawit Kijpaisalratana, Abdel-badih el Ariss, Byung-Jun Yoon

TL;DR

This study introduces the concept of Medical Entropy, quantifying uncertainties in patient outcomes predicted by neural machine translation based on the ICD-9 code of procedures, and demonstrates the promising capacity to model trends of uncertainties during hospitalizations through a data-driven approach.

Abstract

A Clinical Decision Support System (CDSS) is designed to enhance clinician decision-making by combining system-generated recommendations with medical expertise. Given the high costs, intensive labor, and time-sensitive nature of medical treatments, there is a pressing need for efficient decision support, especially in complex emergency scenarios. In these scenarios, where information can be limited, an advanced CDSS framework that leverages AI (artificial intelligence) models to effectively reduce diagnostic uncertainty has utility. Such an AI-enabled CDSS framework with quantified uncertainty promises to be practical and beneficial in the demanding context of real-world medical care. In this study, we introduce the concept of Medical Entropy, quantifying uncertainties in patient outcomes predicted by neural machine translation based on the ICD-9 code of procedures. Our experimental results not only show strong correlations between procedure and diagnosis sequences based on the simple ICD-9 code but also demonstrate the promising capacity to model trends of uncertainties during hospitalizations through a data-driven approach.

Neural machine translation of clinical procedure codes for medical diagnosis and uncertainty quantification

TL;DR

This study introduces the concept of Medical Entropy, quantifying uncertainties in patient outcomes predicted by neural machine translation based on the ICD-9 code of procedures, and demonstrates the promising capacity to model trends of uncertainties during hospitalizations through a data-driven approach.

Abstract

A Clinical Decision Support System (CDSS) is designed to enhance clinician decision-making by combining system-generated recommendations with medical expertise. Given the high costs, intensive labor, and time-sensitive nature of medical treatments, there is a pressing need for efficient decision support, especially in complex emergency scenarios. In these scenarios, where information can be limited, an advanced CDSS framework that leverages AI (artificial intelligence) models to effectively reduce diagnostic uncertainty has utility. Such an AI-enabled CDSS framework with quantified uncertainty promises to be practical and beneficial in the demanding context of real-world medical care. In this study, we introduce the concept of Medical Entropy, quantifying uncertainties in patient outcomes predicted by neural machine translation based on the ICD-9 code of procedures. Our experimental results not only show strong correlations between procedure and diagnosis sequences based on the simple ICD-9 code but also demonstrate the promising capacity to model trends of uncertainties during hospitalizations through a data-driven approach.
Paper Structure (11 sections, 3 equations, 4 figures, 4 tables)

This paper contains 11 sections, 3 equations, 4 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Materials and Methods
  3. Data Source and Experiment Compute Resource
  4. Model Frameworks
  5. Model Overview
  6. Seq2seq diagnoses predictor
  7. Entropy Quantification in Admissions
  8. Model Performance Evaluation
  9. Results
  10. Discussion
  11. Conclusion

Figures (4)

  • Figure 1: The proposed framework of entropy quantification includes Model Pre-Training Module (MPM) and Diagnosis Predictor Module (DPM) by using the procedure and diagnosis ICD-9 codes in the MIMIC-IV database.
  • Figure 2: Illustration of an implementation of the proposed model by adopting the seq2seq model as the pre-trained model and an example of the entropy quantification of a procedure sequence.
  • Figure 3: Trends of average entropy for admission cases with five procedures. The three colors show the entropy trends clustered by their primal diagnosis, which are diagnosis ICD-9 code 41401 (Coronary atherosclerosis of native coronary artery), 78650 (Chest pain), and all admissions with 5 procedures.
  • Figure 4: Entropy trends of three admissions with sepsis diagnosis.