Method and system for docking sepsis management software and original medical system
Through intelligent interfaces, data from various document formats of the original medical system are obtained and processed. Data mining and machine learning are used to differentiate the condition of septic patients, solving the problem of poor connection between septic management software and the original medical system. Improves the efficiency and accuracy of early warning and diagnosis.
Patent Information
- Application Number
- CN202411116104.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-26
AI Technical Summary
In the prior art, the connection between sepsis management software and the original medical system is poor, resulting in large workloads in early warning and diagnosis of sepsis and easy to miss diagnosis.
Data from various document formats are obtained from the original medical system through the intelligent prompt interface, and data is sent to the sepsis management software using JDBC interface, webservice&HTTP interface, message queue, etc., for XML parsing, plain text word segmentation, key-value pair extraction, and Data mining and machine learning are used to differentiate patient condition, and auxiliary diagnosis and treatment information is generated and written back to the original medical system.
It realizes the rapid docking of sepsis management software with the original medical system, automatically obtains and processes data, reduces the workload of early warning and diagnosis, and improves the accuracy and efficiency of diagnosis.
Smart Images

Figure CN119025121A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical system docking, and in particular to a method and system for docking sepsis management software with an original medical system. Background Art
[0002] In my country, sepsis is characterized by severe illness, high mortality, and rapid disease progression. Early diagnosis and treatment of sepsis is an important treatment measure to effectively avoid septic shock and death. Clinically, it is recommended to use a sofa score greater than or equal to two points plus clear signs of infection to diagnose sepsis. Therefore, the guidelines of the United States, Europe, and China strongly recommend early sofa evaluation of sepsis.
[0003] In the early warning and diagnosis of sepsis, how to conduct a full sample and dynamic risk assessment of in-hospital patients to screen out patients with high risk of confirmed sepsis is the most difficult and labor-intensive link. It is a very difficult task to accurately find high-risk patients. It is necessary to perform sofa scoring and determine whether the patient has infection indications, and the sofa scoring results of the patient need to be repeatedly evaluated every day. The traditional assessment is that medical staff check the corresponding items on the paper version of the assessment scale and then accumulate the scores. According to the score range recommended by the guidelines, the risk level is stratified according to the score. Some hospitals also embed the assessment scale into the hospital information system (HIS) to form an electronic form, but medical staff are still required to check the items in the electronic form like writing electronic medical records, and perform a "modified manual assessment". The above two assessment methods will form a certain burden on the daily work of medical staff, which is easy to cause medical staff to resist sepsis screening and evaluation, resulting in missed diagnosis of sepsis.
[0004] In order to achieve early warning and diagnosis of sepsis, if the data of existing medical systems can be used for rapid warning and diagnosis, the workload of early warning and diagnosis of sepsis will be greatly reduced. In order to achieve rapid warning and diagnosis using the data of existing medical systems, the first thing to do is to connect the sepsis management software with the original medical system. However, since the data in the original medical system is unstructured, there is no better way to connect the sepsis management software with the original medical system in the existing technology. Therefore, we urgently need a method to connect the sepsis management software with the original medical system. Summary of the invention
[0005] In view of the above problems, the purpose of the present invention is to provide a method and system for docking sepsis management software with an original medical system, which can realize rapid docking of sepsis management software with the original medical system, and the sepsis management software automatically obtains the data content in the original medical system, and feeds back the final data processing results of the sepsis management software to the original medical system.
[0006] The above-mentioned object of the present invention is achieved through the following technical solutions:
[0007] A method for connecting sepsis management software with an original medical system comprises the following steps:
[0008] S1: Obtain the original data in various document formats including XML, plain text, key-value pairs from the original medical systems including electronic medical record system EMR, hospital information system HIS, laboratory information system LIS, radiology information system RIS, hospital integration platform, clinical data warehouse CDR, and image archiving and communication system PACS through the intelligent prompt interface, and send it to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, and message queue;
[0009] S2: the sepsis management software processes the acquired raw data in various document formats, including XML parsing, plain text segmentation, and key-value pair direct extraction;
[0010] S3: extracting key information from the raw data for the sepsis management software to perform differential diagnosis on the patient's condition using data mining and machine learning;
[0011] S4: After the sepsis management software performs differential diagnosis of the patient's condition using the key information, the generated auxiliary diagnosis and treatment information is written back to the original medical system.
[0012] Further, in step S1, the original data in various document formats including XML, plain text, and key-value pairs are obtained from the original medical systems including the electronic medical record system EMR, the hospital information system HIS, the laboratory information system LIS, the radiology information system RIS, the hospital integration platform, the clinical data warehouse CDR, and the image archiving and communication system PACS through the intelligent prompt interface, and are sent to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, and message queue, specifically:
[0013] Determine the data type of the original data obtained from each of the original medical systems, the data type including information including basic patient information, laboratory results, imaging reports, examination reports, and medical advice contents;
[0014] When the JDBC interface is used to send the original data, the original medical system opens a database view mode related to the original data, and the sepsis management software obtains the original data incrementally and regularly according to the original data;
[0015] When the webservice&HTTP interface is used to send the original data, the original medical system provides a webservice interface or an HTTP interface, and the sepsis management software obtains the original data according to the interface definition;
[0016] When the original data is sent using the message queue, the sepsis management software obtains the original data in real time by subscribing to the message queue of the original medical system.
[0017] Furthermore, in step S2, the sepsis management software processes the acquired raw data in various document formats including XML parsing, plain text segmentation, and key-value pair direct extraction, specifically:
[0018] When the original data is in XML format, an XML parsing library is used to load the XML document, the original data is extracted by parsing the XML tree structure, and the extracted original data is converted into structured data including a JSON object;
[0019] When the original data is in plain text format, the plain text data is preprocessed including removing stop words, punctuation marks, and standardizing text content, and the text is segmented using natural language processing (NLP). The key entities corresponding to the original data are extracted from the segmented data using named entity recognition (NER) technology, and stored in a structured form;
[0020] When the original data is in a key-value pair format, the key-value pair data format is directly parsed, the corresponding value is searched by the key name, and the extracted key-value pair information is mapped to the internal data structure of the sepsis management software.
[0021] Furthermore, in step S2, after the sepsis management software processes the acquired raw data in various document formats including XML parsing, plain text segmentation, and key-value pair direct extraction, it further includes:
[0022] The structured data extracted from different document formats are integrated to form a complete patient file, which includes basic patient information, laboratory results, imaging reports, examination reports, and medical advice. The integrated data is stored in the sepsis management software to ensure that the data can be used for subsequent sepsis risk assessment, intelligent analysis, and decision support.
[0023] Further, in step S3, data mining and machine learning are used to extract key information for the sepsis management software in the original data for differential diagnosis of the patient's condition, specifically:
[0024] Performing data preprocessing on the original data;
[0025] Extract important features related to sepsis from the raw data, including age, respiratory disease, chronic liver disease, diabetes, hypertension, chronic kidney disease, renal transplant status, portal hypertension, severe malnutrition, acute events, prerenal factors, renal factors, postrenal factors, and laboratory-related tests;
[0026] The pre-trained machine learning evaluation model was used to perform differential diagnosis of sepsis based on the important features.
[0027] Furthermore, data preprocessing is performed on the original data, specifically:
[0028] Process the missing values of the original data using methods including mean filling, median filling, interpolation or deleting records with missing values, and use statistical methods or machine learning algorithms to detect and process abnormal data;
[0029] The data of different dimensions are converted into a standard normal distribution to eliminate the influence of the dimension, and the original data are normalized.
[0030] Further, in step S4, after the sepsis management software performs differential diagnosis of the patient's condition through the key information, the generated auxiliary diagnosis and treatment information is written back to the original medical system, specifically:
[0031] Auxiliary diagnosis and treatment information including sepsis risk assessment results, recommended treatment plans, and diagnostic suggestions generated and integrated from the sepsis management software, wherein the integrated data format meets the requirements of the original medical system;
[0032] The auxiliary diagnosis and treatment information is mapped into the data field of the original medical system.
[0033] A sepsis management software and original medical system docking system for executing the above-mentioned sepsis management software and original medical system docking method comprises:
[0034] The original data acquisition module is used to obtain the original data in various document formats including XML, plain text, key-value pairs, etc. from the original medical systems including the electronic medical record system EMR, hospital information system HIS, laboratory information system LIS, radiology information system RIS, hospital integration platform, clinical data warehouse CDR, and image archiving and communication system PACS through the intelligent prompt interface, and send it to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, and message queue;
[0035] A raw data processing module, used for the sepsis management software to process the raw data in various document formats obtained, including XML parsing, plain text segmentation, and direct extraction of key-value pairs;
[0036] A key information extraction module, for extracting key information from the raw data when the sepsis management software performs differential diagnosis of a patient's condition by using data mining and machine learning;
[0037] The auxiliary information write-back module is used to write the generated auxiliary diagnosis and treatment information back to the original medical system after the sepsis management software performs differential diagnosis of the patient's condition through the key information.
[0038] A computer device comprises a memory and one or more processors, wherein the memory stores computer codes, and when the computer codes are executed by the one or more processors, the one or more processors execute the above method.
[0039] A computer-readable storage medium stores computer codes. When the computer codes are executed, the above method is executed.
[0040] Compared with the prior art, the present invention has the following beneficial effects:
[0041] A method for connecting sepsis management software with an original medical system is provided, comprising: S1: acquiring original data in various document formats including XML, plain text, and key-value pairs from original medical systems including an electronic medical record system EMR, a hospital information system HIS, a laboratory information system LIS, a radiology information system RIS, a hospital integration platform, a clinical data warehouse CDR, and an image archiving and communication system PACS through an intelligent prompt interface, and sending the original data to the sepsis management software in the form of a JDBC interface, a webservice & HTTP interface, and a message queue; S2: the sepsis management software processes the original data in various document formats obtained, including XML parsing, plain text segmentation, and direct extraction of key-value pairs; S3: extracting key information from the original data when the sepsis management software performs differential diagnosis of a patient's condition using data mining and machine learning; and S4: after the sepsis management software performs differential diagnosis of a patient's condition using the key information, writing the generated auxiliary diagnosis and treatment information back to the original medical system. The above technical solution can realize the rapid connection between the sepsis management software and the original medical system. The sepsis management software automatically obtains the data content in the original medical system, automatically predicts sepsis through the acquired data, and feeds back the final data processing results of the sepsis management software to the original medical system. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 It is an overall flow chart of the method for docking the sepsis management software of the present invention with the original medical system;
[0043] Figure 2 It is an overall structural diagram of the sepsis management software of the present invention and the original medical system docking system. DETAILED DESCRIPTION
[0044] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0045] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.
[0046] First embodiment
[0047] like Figure 1 As shown, this embodiment provides a method for connecting sepsis management software with an original medical system, comprising the following steps:
[0048] S1: The original data in various document formats including XML, plain text, and key-value pairs are obtained from the original medical systems including the electronic medical record system EMR, hospital information system HIS, laboratory information system LIS, radiology information system RIS, hospital integration platform, clinical data warehouse CDR, and image archiving and communication system PACS through the intelligent prompt interface, and sent to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, and message queue.
[0049] In this embodiment, it is first necessary to automatically obtain the original data for sepsis diagnosis from the original medical system, specifically through the following steps:
[0050] Determine the data type of the original data obtained from each of the original medical systems, the data type including information including basic patient information, laboratory results, imaging reports, examination reports, and medical advice contents;
[0051] When the JDBC interface is used to send the original data, the original medical system opens a database view mode related to the original data, and the sepsis management software obtains the original data incrementally and regularly according to the original data;
[0052] When the webservice&HTTP interface is used to send the original data, the original medical system provides a webservice interface or an HTTP interface, and the sepsis management software obtains the original data according to the interface definition;
[0053] When the original data is sent using the message queue, the sepsis management software obtains the original data in real time by subscribing to the message queue of the original medical system.
[0054] S2: The sepsis management software processes the acquired raw data in various document formats by performing processes including XML parsing, plain text word segmentation, and direct extraction of key-value pairs.
[0055] When the original data is in XML format, an XML parsing library is used to load the XML document, the original data is extracted by parsing the XML tree structure, and the extracted original data is converted into structured data including a JSON object.
[0056] Common XML parsing libraries are as follows:
[0057] (1) Python
[0058] xml.etree.ElementTree: A module in the Python standard library that provides lightweight XML parsing capabilities.
[0059] lxml: A powerful XML parsing library that supports XPath, XSLT and other functions.
[0060] minidom: A module in the Python standard library that provides parsing functions based on DOM (Document Object Model).
[0061] xmltodict: A library that converts XML to Python dictionaries, suitable for processing simple XML.
[0062] (2) Java
[0063] JAXP (Java API for XML Processing): Java standard API that includes SAX, DOM, and StAX parsers.
[0064] DOM4J: An open source XML framework for the Java platform that supports XPath and XSLT.
[0065] JDOM: A lightweight XML parsing library designed specifically for Java that is easy to use.
[0066] Xerces: A comprehensive XML parsing library provided by Apache that supports SAX, DOM, and XML Schema.
[0067] (3) JavaScript
[0068] DOMParser: Native API in browser environment for parsing XML strings and generating DOM objects.
[0069] xml2js: A library in Node.js that parses XML into JavaScript objects.
[0070] fast-xml-parser: An efficient Node.js XML parsing library that supports converting XML to JSON.
[0071] (4) C# (.NET)
[0072] System.Xml: Namespace in .NET Framework, providing SAX and DOM parsers.
[0073] XmlDocument: DOM parser for loading and manipulating XML documents.
[0074] XDocument: Part of LINQ to XML, providing a modern, object-oriented way to operate XML.
[0075] XmlReader: Provides a stream-based forward-reading parser (SAX style).
[0076] (5) C / C++
[0077] libxml2: A commonly used open source XML parsing library that supports SAX and DOM parsing.
[0078] TinyXML: A lightweight XML parsing library suitable for embedded systems.
[0079] RapidXML: A very fast XML parsing library based on a DOM-style parser.
[0080] (6)PHP
[0081] SimpleXML: A built-in parsing library in PHP for converting XML data into objects.
[0082] DOMDocument: A DOM-based XML parsing library in PHP that provides detailed control over XML documents.
[0083] XMLReader: Provides a forward-only, stream-based parser, suitable for processing large documents.
[0084] (7) Ruby
[0085] REXML: An XML parser in the Ruby standard library that supports SAX and DOM parsing.
[0086] Nokogiri: A powerful XML and HTML parsing library that provides a simple API and powerful features.
[0087] When the original data is in plain text format, preprocess the plain text data, including removing stop words, punctuation marks, and standardizing the text content. Use natural language processing (NLP) to tokenize the text, and use named entity recognition (NER) technology to extract key entities corresponding to the original data from the tokenized data, and store them in a structured form.
[0088] An example of the process of natural language processing (NLP) is as follows:
[0089] (1) Plain text preprocessing
[0090] a. Removing stop words
[0091] Stop words: Stop words refer to words that are frequently used in a specific language but have little significance for text analysis (such as "de", "shi" in Chinese). By using a stop words list, these words can be filtered out.
[0092] Implementation method: Use the stop words list in the NLP library or a custom stop words list to remove stop words from the text.
[0093] b. Removing punctuation marks
[0094] Punctuation marks: Punctuation marks are usually considered irrelevant information in text processing and can be removed to simplify processing.
[0095] Implementation method: Use regular expressions or functions in the NLP library to remove punctuation marks.
[0096] c. Standardizing text content
[0097] Lowercase conversion: Convert all characters in the text to lowercase to eliminate the impact caused by case differences.
[0098] Lemmatization: restoring vocabulary to its original form, for example, restoring the various inflections of a verb (such as "running" and "ran") to the base form "run". This can be done using stemming or lemmatization techniques.
[0099] (2) Word segmentation
[0100] a. Participle
[0101] Definition: Word segmentation is the process of dividing a continuous text string into meaningful words or phrases.
[0102] Implementation method: Use NLP libraries (such as Python's NLTK, spaCy, Java's StanfordNLP, etc.) for word segmentation. Different languages have different word segmentation algorithms and tools.
[0103] (3) Named Entity Recognition (NER)
[0104] a. Named Entity Recognition
[0105] Definition: NER is a technology for identifying entities with specific meanings in text (such as names of people, places, time, drug names, diseases, etc.).
[0106] Implementation method: Use pre-trained NER models or train custom models for entity recognition. Common tools and libraries include spaCy, NLTK, Stanford NER, OpenNLP, etc.
[0107] b. Key entity extraction
[0108] Extraction process: Identify key entities in the text through the NER model, such as the patient's name, age, gender, diagnosis results, drug name, etc.
[0109] Verification and correction: Verify the extracted entities to ensure their accuracy and completeness. Further verification and correction can be performed in combination with domain knowledge base or rules.
[0110] (4) Structured Storage
[0111] a. Structured data model
[0112] Define data model: Define a suitable structured data model for the extracted entities. You can use JSON, XML, or a relational database format to store them.
[0113] Field Definition: Determine the field name and data type for each entity, such as "patient_name" (string), "diagnosis" (string), "age" (integer), etc.
[0114] b. Data storage
[0115] Storage method: Store structured data in a suitable database or data storage system according to the system requirements. You can use relational databases (such as MySQL, PostgreSQL), NoSQL databases (such as MongoDB), file systems (such as JSON files), etc.
[0116] Data writing: Write code to write the extracted structured data into the specified storage system to ensure data consistency and integrity.
[0117] (5) Data verification and quality control
[0118] a. Data Verification
[0119] Verify the accuracy and completeness of structured data to check for missing or incorrect entity information.
[0120] Implement quality control measures such as duplicate data checking, data range verification, etc.
[0121] b. Data quality monitoring
[0122] Set data quality monitoring indicators and regularly evaluate the accuracy and completeness of data.
[0123] Based on the feedback and problems found, we optimized the preprocessing, word segmentation, and NER models to improve the accuracy and stability of data extraction.
[0124] When the original data is in a key-value pair format, the key-value pair data format is directly parsed, the corresponding value is searched by the key name, and the extracted key-value pair information is mapped to the internal data structure of the sepsis management software.
[0125] In addition, in step S2, after the sepsis management software processes the acquired raw data in various document formats including XML parsing, plain text segmentation, and direct extraction of key-value pairs, it also includes: integrating the structured data extracted from different document formats to form a complete patient file, which includes basic patient information, laboratory results, imaging reports, examination reports, and medical advice content, and storing the integrated data in the sepsis management software to ensure that the data can be used for subsequent sepsis risk assessment, intelligent analysis, and decision support.
[0126] S3: Data mining and machine learning are used to extract key information from the raw data for the sepsis management software to perform differential diagnosis on the patient's condition.
[0127] In this embodiment, step S3 is specifically as follows:
[0128] Performing data preprocessing on the original data;
[0129] Extract important features related to sepsis from the raw data, including age, respiratory diseases, chronic liver diseases, diabetes, hypertension, chronic kidney disease, renal transplant status, portal hypertension, severe malnutrition, acute events, prerenal factors (including at least 20 common factors such as severe vomiting, shock, ECMO, etc.), renal factors (including at least 10 items such as the use of nephrotoxic drugs), postrenal factors, and laboratory-related tests (including at least 5 items such as elevated bilirubin);
[0130] The pre-trained machine learning evaluation model was used to perform differential diagnosis of sepsis based on the important features.
[0131] Among them, the original data is preprocessed, specifically: the missing value processing of the original data is performed using methods including mean filling, median filling, interpolation or deletion of records with missing values, and abnormal data is detected and processed using statistical methods or machine learning algorithms; data of different dimensions are converted into a standard normal distribution to eliminate the influence of the dimension, and the original data is normalized.
[0132] An example of the process of pre-training a machine learning evaluation model is as follows:
[0133] (1) Data Collection
[0134] a. Data Source
[0135] Collect patient medical record data from original medical systems such as EMR, HIS, LIS, RIS, etc., including medical records, laboratory results, imaging reports, etc.
[0136] Several important features associated with sepsis were obtained, such as age, respiratory disease, chronic liver disease, diabetes, hypertension, chronic kidney disease, and renal transplant status.
[0137] b. Data annotation
[0138] Collect annotated sepsis case data, including confirmed sepsis and non-sepsis cases.
[0139] Medical experts annotate and verify the data to ensure the accuracy of the annotations.
[0140] 2. Feature extraction
[0141] a. Feature selection
[0142] Clinical characteristics: Clinical characteristics related to sepsis were extracted from the patient's medical records, including age, chronic diseases, acute events, etc.
[0143] Medical history and pathological characteristics: including respiratory diseases, chronic liver disease, diabetes, hypertension, chronic kidney disease, renal transplant status, portal hypertension, severe malnutrition, etc.
[0144] Acute events and risk factors: Extract acute events such as severe vomiting, shock, and prerenal, renal, and postrenal factors.
[0145] Laboratory test results: Extract relevant laboratory test results, such as elevated bilirubin, etc.
[0146] b. Feature encoding
[0147] Encode categorical features, such as disease categories, into numeric or binary form.
[0148] Standardize or normalize numerical features to ensure the scale consistency of features.
[0149] (3) Data preprocessing
[0150] a. Data cleaning
[0151] Remove features or samples with many missing values.
[0152] Fill or impute missing data using mean, median, or interpolation methods.
[0153] b. Data partitioning
[0154] The dataset is divided into training set, validation set and test set, usually in a ratio of 8:1:1.
[0155] (4) Model selection
[0156] a. Model type
[0157] Choose an appropriate machine learning model, such as Logistic Regression, Support Vector Machine (SVM), Random Forest, Gradient Boosting, Neural Network, etc.
[0158] Select the model type based on data characteristics and task requirements.
[0159] b. Hyperparameter Tuning
[0160] Determine the hyperparameters of the model and use cross-validation techniques to tune the hyperparameters to optimize model performance.
[0161] (5) Model training
[0162] a. Model training process
[0163] The model is trained using the training set data, and the model learns the data features by minimizing the loss function.
[0164] The training process includes steps such as forward propagation, back propagation and gradient descent.
[0165] b. Model evaluation
[0166] The model performance was evaluated on the validation set using evaluation indicators such as accuracy, precision, recall, F1 score, area under the receiver operating characteristic curve (AUC-ROC), etc.
[0167] (6) Model verification and testing
[0168] a. Model Validation
[0169] Use the validation set to tune and select the model and choose the model with the best performance.
[0170] Evaluate the performance of the model under different feature combinations to ensure the generalization ability of the model.
[0171] b. Model testing
[0172] Use the test set to evaluate the final performance of the model to ensure that the model performs stably on unseen data.
[0173] Ensure the model's ability to identify samples of different categories and avoid bias.
[0174] (7) Model deployment and monitoring
[0175] a. Model deployment
[0176] Integrate the trained model into sepsis management software for real-time analysis and diagnostic support.
[0177] Configure the model's input and output interfaces to ensure that the model can receive structured data and output diagnostic results.
[0178] b. Model monitoring
[0179] Implement a model monitoring mechanism to regularly evaluate the model’s predictive accuracy and stability.
[0180] Collect and analyze new data, retrain and update the model, and ensure that the model remains accurate as medical data and treatment methods change.
[0181] c. User feedback and optimization
[0182] Collect feedback from doctors and users to optimize the model's prediction results and diagnostic recommendations.
[0183] Adjust the model’s feature selection, hyperparameters, and training process based on the feedback.
[0184] S4: After the sepsis management software performs differential diagnosis of the patient's condition using the key information, the generated auxiliary diagnosis and treatment information is written back to the original medical system.
[0185] In this embodiment, step S4 is specifically as follows:
[0186] The auxiliary diagnosis and treatment information including sepsis risk assessment results, recommended treatment plans, and diagnostic suggestions are integrated and generated from the sepsis management software, and the integrated data format meets the requirements of the original medical system; the auxiliary diagnosis and treatment information is mapped to the data field of the original medical system.
[0187] Second embodiment
[0188] like Figure 2 As shown, this embodiment provides a sepsis management software and original medical system docking system for executing the sepsis management software and original medical system docking method of the first embodiment, including:
[0189] The original data acquisition module 1 is used to obtain the original data in various document formats including XML, plain text, key-value pairs, etc. from the original medical systems including the electronic medical record system EMR, the hospital information system HIS, the laboratory information system LIS, the radiology information system RIS, the hospital integration platform, the clinical data warehouse CDR, and the image archiving and communication system PACS through the intelligent prompt interface, and send it to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, message queue, etc.;
[0190] The raw data processing module 2 is used for the sepsis management software to process the raw data in various document formats obtained, including XML parsing, plain text segmentation, and direct extraction of key-value pairs;
[0191] A key information extraction module 3 is used to extract key information when the sepsis management software performs differential diagnosis of the patient's condition in the original data by using data mining and machine learning;
[0192] The auxiliary information writing back module 4 is used to write the generated auxiliary diagnosis and treatment information back to the original medical system after the sepsis management software performs differential diagnosis of the patient's condition through the key information.
[0193] A computer-readable storage medium stores a computer code. When the computer code is executed, the above method is executed. A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program. The program can be stored in a computer-readable storage medium. The storage medium can include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.
[0194] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.
[0195] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0196] It should be noted that the above embodiments can be freely combined as needed. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention.
Claims
1. A method for docking sepsis management software with an original medical system, characterized in that: The following steps are involved: S1: Obtain the original data in various document formats including XML, plain text, key-value pairs from the original medical systems including electronic medical record system EMR, hospital information system HIS, laboratory information system LIS, radiology information system RIS, hospital integration platform, clinical data warehouse CDR, and image archiving and communication system PACS through the intelligent prompt interface, and send it to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, and message queue; S2: the sepsis management software processes the acquired raw data in various document formats, including XML parsing, plain text segmentation, and key-value pair direct extraction; S3: extracting key information from the raw data for the sepsis management software to perform differential diagnosis on the patient's condition using data mining and machine learning; S4: After the sepsis management software performs differential diagnosis of the patient's condition using the key information, the generated auxiliary diagnosis and treatment information is written back to the original medical system.
2. The method for docking sepsis management software with an original medical system according to claim 1, characterized in that: In step S1, the original data in various document formats including XML, plain text, and key-value pairs are obtained from the original medical systems including the electronic medical record system EMR, the hospital information system HIS, the laboratory information system LIS, the radiology information system RIS, the hospital integration platform, the clinical data warehouse CDR, and the image archiving and communication system PACS through the intelligent prompt interface, and are sent to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, and message queue, specifically: Determine the data type of the original data obtained from each of the original medical systems, the data type including information including basic patient information, laboratory results, imaging reports, examination reports, and medical advice contents; When the JDBC interface is used to send the original data, the original medical system opens a database view mode related to the original data, and the sepsis management software obtains the original data incrementally and regularly according to the original data; When the webservice&HTTP interface is used to send the original data, the original medical system provides a webservice interface or an HTTP interface, and the sepsis management software obtains the original data according to the interface definition; When the original data is sent using the message queue, the sepsis management software obtains the original data in real time by subscribing to the message queue of the original medical system.
3. The method for docking sepsis management software with an original medical system according to claim 1, characterized in that: In step S2, the sepsis management software processes the acquired raw data in various document formats, including XML parsing, plain text segmentation, and key-value pair direct extraction, specifically: When the original data is in XML format, an XML parsing library is used to load the XML document, the original data is extracted by parsing the XML tree structure, and the extracted original data is converted into structured data including a JSON object; When the original data is in plain text format, the plain text data is preprocessed including removing stop words, punctuation marks, and standardizing text content, and the text is segmented using natural language processing (NLP). The key entities corresponding to the original data are extracted from the segmented data using named entity recognition (NER) technology, and stored in a structured form; When the original data is in a key-value pair format, the key-value pair data format is directly parsed, the corresponding value is searched by the key name, and the extracted key-value pair information is mapped to the internal data structure of the sepsis management software.
4. The method for connecting sepsis management software with an original medical system according to claim 3, characterized in that: In step S2, after the sepsis management software processes the acquired raw data in various document formats including XML parsing, plain text segmentation, and key-value pair direct extraction, it also includes: The structured data extracted from different document formats are integrated to form a complete patient file, which includes basic patient information, laboratory results, imaging reports, examination reports, and medical advice. The integrated data is stored in the sepsis management software to ensure that the data can be used for subsequent sepsis risk assessment, intelligent analysis, and decision support.
5. The method for docking sepsis management software with an original medical system according to claim 1, characterized in that: In step S3, data mining and machine learning are used to extract key information from the raw data for the sepsis management software to perform differential diagnosis of the patient's condition, specifically: Performing data preprocessing on the original data; Extract important features related to sepsis from the raw data, including age, respiratory disease, chronic liver disease, diabetes, hypertension, chronic kidney disease, renal transplant status, portal hypertension, severe malnutrition, acute events, prerenal factors, renal factors, postrenal factors, and laboratory-related tests; The pre-trained machine learning evaluation model was used to perform differential diagnosis of sepsis based on the important features.
6. The method for connecting sepsis management software with an original medical system according to claim 5, characterized in that: The raw data is preprocessed, specifically: Process the missing values of the original data using methods including mean filling, median filling, interpolation or deleting records with missing values, and use statistical methods or machine learning algorithms to detect and process abnormal data; The data of different dimensions are converted into a standard normal distribution to eliminate the influence of the dimension, and the original data are normalized.
7. The method for docking sepsis management software with an original medical system according to claim 1, characterized in that: In step S4, after the sepsis management software performs differential diagnosis of the patient's condition through the key information, the generated auxiliary diagnosis and treatment information is written back to the original medical system, specifically: Auxiliary diagnosis and treatment information including sepsis risk assessment results, recommended treatment plans, and diagnostic suggestions generated and integrated from the sepsis management software, wherein the integrated data format meets the requirements of the original medical system; The auxiliary diagnosis and treatment information is mapped into the data field of the original medical system.
8. A sepsis management software and original medical system docking system for executing the sepsis management software and original medical system docking method according to any one of claims 1 to 7, characterized in that: include: The original data acquisition module is used to obtain the original data in various document formats including XML, plain text, key-value pairs, etc. from the original medical systems including the electronic medical record system EMR, hospital information system HIS, laboratory information system LIS, radiology information system RIS, hospital integration platform, clinical data warehouse CDR, and image archiving and communication system PACS through the intelligent prompt interface, and send it to the sepsis management software in the form of JDBC interface, webservice&HTTP interface, and message queue; A raw data processing module, used for the sepsis management software to process the raw data in various document formats obtained, including XML parsing, plain text segmentation, and direct extraction of key-value pairs; A key information extraction module, for extracting key information from the raw data when the sepsis management software performs differential diagnosis of a patient's condition by using data mining and machine learning; The auxiliary information write-back module is used to write the generated auxiliary diagnosis and treatment information back to the original medical system after the sepsis management software performs differential diagnosis of the patient's condition through the key information.
9. A computer device comprising a memory and one or more processors, wherein the memory stores computer codes, and when the computer codes are executed by the one or more processors, the one or more processors execute the method according to any one of claims 1 to 7. 10 . A computer-readable storage medium storing a computer code. When the computer code is executed, the method according to claim 1 is executed.
