Real-time intelligent auxiliary ICD encoding system and method based on medical record

Abstract

The invention discloses a real-time intelligent auxiliary ICD encoding system and method based on a medical record. The method comprises: acquiring medical record information from a medical record database of a hospital; performing preprocessing and feature vector construction on the medical record information to obtain two types of training sample sets; based on the training sample sets, traininga main coding model to obtain a seq2seq + action model; training a missed diagnosis detection model based on the training sample sets; and respectively obtaining ICD codes and confidence coefficientsrecommended by the main coding model and the missed diagnosis detection model, and recommending the ICD codes and confidence coefficients to a doctor for confirmation. According to the invention, thetraining sample sets are constructed through an automatic disease name alignment technology so as to reduce the cost of manually constructing the sample set; through a deep learning technology, diagnosis names written by doctors in different styles are accurately coded into standard ICD disease names and codes, so that the ICD coding accuracy is improved; and the medical record text is analyzed in real time through the deep learning model, and doctors are assisted in discovering possible miss-writing and wrong-writing disease diagnosis, so that the recall ratio of the system is increased.

Description

technical field [0001] The invention relates to the technical field of ICD coding, in particular to a medical record-based real-time intelligently assisted ICD coding system and method. Background technique [0002] Disease Diagnosis Related Groups (DRGs) is a "patient-centered" case combination system, and it is also an effective tool for controlling medical payment, hospital fine management and medical evaluation. Its main data comes from the front page of medical records. The International Classification of Diseases (ICD) is an international standard for coding and categorizing different types of diseases and health-related problems. The first threshold for DRGs to join the group is the ICD-10 disease code, so the ICD disease code directly affects the smooth development of DRGs. [0003] In current practical applications, the coding process of ICD generated by hospitals is difficult to meet various management needs including hospitals. In my country, the coding of disea...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the problems of low accuracy and timeliness of ICD coding in the prior art, the present invention proposes a real-time intelligent auxiliary ICD coding system based on medical records, which assists doctors in the process of writing medical records by writing diagnoses and medical record texts in real time. Analyze and recommend the ICD disease code closest to the case to the doctor

Images