Research On Electronic Medical Record Data Quality Detection Based On Semantic Constraints

Electronic medical records are an important part of the medical information field and can be used for disease prediction,diagnosis,treatment and other research.However,the absence of relevant standard constraints in the process of collecting and processing electronic medical record data leads to data quality problems,and these problems can affect the effectiveness of its secondary use.Data quality detection is an important part of guaranteeing data quality,so the research on data quality detection is of great significance.Traditional data quality detection methods have low pervasiveness and poor logical reasoning ability.The existing semantic-based data quality detection methods have limitations such as weak ability to detect the number of attribute values and separation of ontology model and validation rules.In order to solve the above problems,the following research is conducted in this thesis based on the semantic-based data quality detection.First,a data quality detection method based on SHACL is proposed,which establishes an ontology framework with constraints,converts the relational model into an RDF model,enriches the information in the data to be verified by reasoning,and finally compares the data with the ontology model to find the anomalies existing in the data.The method has a simpler way of constraint writing,higher universality,more scalability,and stronger inference capability,which can improve the completeness and accuracy of detection results.In this thesis,experiments are conducted on the method,and the experiments achieve the expected results and prove the feasibility of the method.Then,an electronic medical record ontology construction method is proposed,which uses a seven-step method to construct an electronic medical record ontology,and uses an event ontology to represent the patient’s treatment process,which connects the thing classes in the electronic medical record in series.Five inference rules are proposed to represent the implicit information in electronic medical records.On the basis of this method,the construction of an electronic medical record ontology is completed with the nucleic acid test report as an example.Finally,the above method and model were practically tested using nucleic acid detection data in Chongqing,and the test results proved that the method has engineering feasibility and effectiveness.Based on the above study,a nucleic acid testing data quality management system was designed and programmed to make it easier for nucleic acid testing data managers to complete data quality detection,and to visualize the status of data quality and the presence of abnormal data.In summary,the research on data quality detection based on semantic constraints is completed in this thesis,which solves the problems of poor universality and logical reasoning ability of traditional electronic medical record data quality detection methods and improves the detection effect of electronic medical record data quality.