A disease-aided differential diagnosis system based on a causal medical knowledge graph

Abstract

The invention discloses an auxiliary disease differential diagnosis system based on a causal medical knowledge graph. Extract causal knowledge related to disease diagnosis and treatment from medical knowledge sources, and construct a causal medical knowledge graph containing key diagnostic information such as medical history, symptoms, signs, laboratory test results, medication, and demographic information. The corresponding knowledge extraction rules are designed in order to reduce the noise of the causal medical knowledge map and improve the accuracy and operating efficiency of the disease differential diagnosis model. Based on the causal medical knowledge map, the present invention extracts all the personalized diagnosis data of the patient, makes full use of the negative test results and negative symptoms and other negative data in the patient data, uses the disease differential diagnosis model to carry out explicit reasoning, and gives a clear The result of inference on which the diagnosis is based.

Description

technical field [0001] The invention belongs to the technical field of medical and health information, and in particular relates to an auxiliary disease differential diagnosis system based on a causal medical knowledge graph. Background technique [0002] Differential diagnosis of diseases is an important part of the process of disease diagnosis and treatment. Through consultation, clinicians based on existing medical knowledge and experience in diagnosis and treatment, comprehensive analysis and reasoning based on the patient's current medical history, laboratory tests, signs and other information, give the patient's suspected symptoms. list of diseases. However, due to the complex and varied disease conditions of patients, many patients suffer from multiple diseases, and the incidence of some diseases is low. Doctors lack corresponding experience in diagnosis and treatment. Under realistic high-intensity working conditions, missed diagnosis and misdiagnosis are prone to oc...

AI Technical Summary

Problems solved by technology

[0005] 1. Most of the knowledge graphs built based on electronic medical records only use EHR data and simple relationships, and do not build deeper relationships between diseases, such as the relationship between diseases and diagnosis and treatment methods, diseases and drugs, diseases and living habits, etc. At the same time, EHR data is very noisy. Large, the quality of the established knowledge map is not good;

[0006] 2. Most of the models for differential diagnosis of diseases are based on pre-selected variables for prediction, and do not make full use of all the personalized data of patients;

[0007] 3. Most of the research on algorithms for differential diagnosis of diseases only utilizes positive test results and positive symptoms, while ignoring the value of negative test results and negative symptoms;

[0008] 4. The knowledge map reasoning algorithm based solely on graph embedding lacks interpretability, while the rule reasoning algorithm with strong interpretability is too simple, requiring manual formulation and updating of rules, and it is difficult to cope with complex and changeable clinical environments

Images