CN-122024981-A - Intelligent analysis quality control method and system for integrated cloud medical records

CN122024981ACN 122024981 ACN122024981 ACN 122024981ACN-122024981-A

Abstract

The invention relates to the technical field of informationized medical treatment, in particular to an integrated cloud medical records intelligent analysis quality control method and system. The technical scheme includes that the method comprises the following steps of collecting multi-source medical records data and identifying the multi-source medical records data to obtain a pathological data set. Preprocessing and fusing the pathology data set to generate a standardized medical record data set. And carrying out disease association analysis on the standardized medical records data set to obtain disease risk factors. And aiming at time series data of multiple times of detection of the patient, obtaining trend risk factors through capturing long-term dependency of the data. And performing quality control verification. And outputting the patient condition data which is qualified through quality control test and comprises the disease risk factors and the trend risk factors. The method integrates the multi-source data integrally, breaks the data island, digs the data relevance and improves the analysis comprehensiveness. The analysis efficiency and accuracy are high, the quality control pertinence and comprehensiveness are strong, the expansibility and the safety are high, the clinical practicality is strong, and the accurate support and the boost medical quality improvement are provided.

Inventors

FENG QING
YANG ZHI
SHAO TIANJUN

Assignees

杭州博石科技有限公司
杭州博海云惠医疗科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260123

Claims (10)

1. The intelligent analysis quality control method for the integrated cloud medical records is characterized by comprising the following steps of: S1, collecting multi-source medical record data and identifying the multi-source medical record data to obtain a pathological data set; S2, preprocessing and fusing the pathological data set to generate a standardized medical record data set; s3, carrying out disease association analysis on the standardized medical records data set to obtain disease risk factors; S4, aiming at time sequence data of repeated detection of a patient, acquiring trend risk factors through capturing long-term dependency of the data; S5, performing quality control verification on the whole steps of the steps S1-S4, executing the step S6 if the quality control verification passes, and executing the step S1 if the quality control verification does not pass; And S6, outputting the patient condition data which is qualified through quality control test and comprises disease risk factors and trend risk factors.
2. The method for intelligent analysis and quality control of an integrated cloud medical record according to claim 1, wherein the multi-source medical record data comprises a disease description text, detection data, a detection report and/or a detection image.
3. The intelligent analysis quality control method for the integrated cloud medical records, which is disclosed in claim 1, is characterized in that the preprocessing comprises S2.1, text processing, S2.2, numerical processing, S2.3, report analysis and S2.4, and data fusion.
4. The integrated cloud medical records intelligent analysis quality control method is characterized by comprising the steps of generating a plurality of disease items through medical records data set extraction, prefiltering the disease items based on clinical thresholds, removing clinically insignificant disease items, distributing medical records data in slices to generate the plurality of disease items in parallel, removing disease items with symptoms not clinically relevant to detection indexes at the same time, calculating to obtain weighted confidence, and outputting the weighted confidence which is more than or equal to a preset association rule threshold as the disease risk factor.
5. The integrated cloud medical records intelligent analysis quality control method is characterized by comprising the specific steps of 1, data preprocessing and disease item encoding, converting into a disease item format which can be identified by an algorithm, 2, setting algorithm core parameters, setting minimum support, minimum weighted confidence and clinical weight coefficients by combining medical data characteristics and clinical requirements, 3, mining distributed candidate set generation and associated disease items, and 4, association rule generation and weighted confidence calculation.
6. The method of claim 5, wherein the condition terms include a symptom condition term extracted from a condition description text, a detection index condition term of detection data boolean according to a clinical threshold, and a report result condition term of pathological/image features extracted from a detection report.
7. The integrated cloud medical records intelligent analysis quality control method according to claim 5 is characterized in that the distributed candidate set generation and associated disease item mining comprises the following steps of 1, scanning all single disease items, obtaining undetermined diagnosis disease risk factors Y through the single disease items, counting the support degree of the single disease items, screening out associated single disease items with the support degree not less than a set minimum support degree, 2, carrying out K-order connection on K disease items to generate K-order disease item candidate sets, eliminating candidate sets without clinical association, and finally obtaining all candidate sets meeting the minimum support degree to form K-order disease items.
8. The method for intelligent analysis and quality control of integrated cloud medical records of claim 7, wherein the association rule generation and weighted confidence calculation comprises the steps of calculating the weighted confidence of each k-order disease term Wherein I is the number of each individual disorder item in the k-order disorder item, I is the total number of individual disorder items in the k-order disorder item, w i is the clinical weight coefficient of the I-item numbered in the k-order disorder item, i=xuy is the single item data related to diagnosing the disease risk factor, T X∪Y is the probability that the disorder item X is judged to be diagnosing the disease risk factor Y, and T X is the probability that the disorder item X appears.
9. The intelligent analysis quality control method for the integrated cloud medical records, which is disclosed in claim 1, is characterized by comprising the specific steps of S41, time sequence data extraction and preprocessing, S42, mapping detection index values to a [0,1] interval, then constructing a detection index curve, S43, obtaining abnormal parameters according to the detection index curve, S44, and outputting detection indexes with the abnormal parameters D larger than a preset abnormal parameter threshold as trend dangerous factors.
10. The utility model provides an integration cloud medical records wisdom analysis quality control system, its characterized in that, integration cloud medical records wisdom analysis quality control system includes: The data acquisition module is used for acquiring multi-source medical record data and identifying to obtain a pathological data set; the cloud storage module is used for classifying and storing the acquired multi-source medical record data and pathological data sets; the data preprocessing module is used for cleaning and standardizing the multi-source medical record data in the cloud storage module; the abnormal value analysis module is used for carrying out disease association analysis on the standardized medical records data set to obtain disease risk factors; The abnormal trend analysis module is used for carrying out trend prediction on the multiple detection data of the patient and identifying trend risk factors of potential disease deterioration or abnormal fluctuation trend; The quality control module is used for constructing a full-flow quality control system and carrying out quality control aiming at all links of data acquisition, preprocessing, abnormal constant value analysis and abnormal trend analysis; And the result output module is used for outputting the patient condition data qualified through quality control.

Description

Intelligent analysis quality control method and system for integrated cloud medical records Technical Field The invention relates to the technical field of informationized medical treatment, in particular to an integrated cloud medical records intelligent analysis quality control method and system. Background Along with the advanced advancement of medical informatization, multi-source heterogeneous systems such as hospital electronic medical records, laboratory information systems, image archiving and communication systems and the like accumulate massive medical records data, and cover various types such as illness state description texts, structural detection indexes, semi-structural reports and the like. The data is not only a core basis of clinical diagnosis and treatment and quality control management, but also a potential association rule of occurrence, development and complication risks of diseases, especially for chronic diseases such as diabetes, hypertension and the like, and can provide important support for early warning and accurate intervention by mining association relations between time sequence change trend of detection data and clinical characteristics. At present, the analysis of medical records is basically manual analysis, the traditional method relies on manual arrangement, analysis and quality control of medical record data, when facing massive medical record data, the processing period is long, the labor cost is high, the requirement of large-scale medical record analysis is difficult to meet, multi-source data such as disease description, detection data, detection reports and the like are stored in different systems in a scattered manner, an effective integrated integration mechanism is lacked, data relevance is difficult to mine, and analysis results are one-sided. Disclosure of Invention In order to solve the technical problems, the invention provides an integrated cloud medical records intelligent analysis quality control method, which realizes the integrated integration of multi-source medical records data, improves the medical records analysis efficiency and accuracy and enhances the quality control pertinence by intelligent analysis and full-flow quality control, and meanwhile realizes the elastic expansion of resources and the cooperation of a cross mechanism by depending on a cloud architecture. The intelligent analysis quality control method for the integrated cloud medical records comprises the following steps: s1, collecting multi-source medical record data and identifying the multi-source medical record data to obtain a pathological data set. S2, preprocessing and fusing the pathological data set to generate a standardized medical record data set. S3, carrying out disease association analysis on the standardized medical records data set to obtain disease risk factors. And S4, aiming at time sequence data of multiple times of detection of the patient, obtaining trend risk factors through capturing long-term dependency of the data. And S5, performing quality control verification on the whole steps of the steps S1-S4, executing S6 if the quality control verification passes, and executing S1 if the quality control verification does not pass. And S6, outputting the patient condition data which is qualified through quality control test and comprises disease risk factors and trend risk factors. Preferably, the multi-source medical records data comprise illness state description text, detection data, detection reports and/or detection images. The preprocessing preferably comprises S2.1 of text processing, S2.2 of numerical processing, S2.3 of report analysis and S2.4 of data fusion. Preferably, the text processing comprises word segmentation, entity recognition and noise reduction of the illness state description text, removal of messy codes, repeated sections and the like, and extraction of keywords. Preferably, the numerical processing includes normalizing the test data based on a medical unit conversion library. Preferably, the data fusion includes generating a standardized medical records data set based on patient hospitalization number or identification card number association fusion. The method for obtaining the disease risk factors comprises the steps of extracting and generating a plurality of disease items through a medical record data set, prefiltering the disease items based on clinical thresholds to remove disease items without clinical significance, distributing medical record data in slices to generate the plurality of disease items in parallel, removing disease items with symptoms not clinically related to detection indexes at the same time, calculating to obtain weighted confidence, and outputting the weighted confidence which is more than or equal to a preset association rule threshold as the disease risk factors. The method for calculating the weighted confidence coefficient comprises the specific steps of 1, encoding data preprocessing and disorder items, conve