A data quality verification method based on medical treatment

Through ETL technology, the hospital business system is connected to the hospital, and medical data is automatically collected and processed, which solves the problem of low data quality in the hospital, realizes automatic verification and display of data quality, and improves data application and analysis results.

CN118152385BActive Publication Date: 2024-11-05XIAN GEOMETRY DIGITAL INFORMATION TECH CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202410373704.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-29
Publication Date
2024-11-05
Estimated Expiration
2044-03-29

AI Technical Summary

Technical Problem

There are low quality of filling out medical records in the hospital, poor connectivity of business systems, lack of key information and inconsistent statistical caliber of multi-system data, resulting in low data quality and affecting the sharing and analysis of data within the hospital.

Method used

Use ETL technology to connect to various business systems, automatically collect medical data, perform data normalization and unified format processing, design data quality verification methods, including verification of data sources and data warehouses, and generate verification through SQL statements and data warehouse queries. Verify the results and generate a self-increment primary key in the data warehouse to realize the automatic checksum display of data quality.

Benefits of technology

It realizes automated verification of hospital data quality, saves labor costs, provides intuitive data quality verification results, facilitates problem rectification, and forms a closed-loop management process that continuously improves.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118152385B_ABST
    Figure CN118152385B_ABST
Patent Text Reader

Abstract

The present invention relates to a data quality verification method based on medical treatment, and the steps are as follows: collect medical data in a business system according to the business scope of a hospital; store the collected medical data into a data warehouse; design a business analysis theme for the data in the data warehouse, divide the business analysis theme into several subject domains, and sort out the key indicators in each subject domain, determine the verification type of each subject domain, use the SQL statement in the business system as the verification caliber according to the verification type, and convert the SQL statement into a data source, and use ETL to verify the data quality of the data warehouse; construct a medical data model table of the verification result, and store the verification result in the model table, use the spring boot architecture, develop a data quality result display function, and output the verification result. Through this method, the statistical caliber data of various systems in the hospital can be automatically verified, and the quality of data sharing and analysis within the hospital can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of medical data processing, and in particular relates to a medical-based data quality verification method. Background Art

[0002] Hospital data plays a pivotal role in the medical field and has a direct impact on hospital management and medical quality. In the era of informatization and intelligence, the effective use of hospital data is the key to achieving rational resource allocation and high-quality development of hospitals. Modern hospital management needs to take appropriate measures to effectively use data to promote the development of hospitals and improve the quality of medical services.

[0003] In the daily operation of hospitals, in order to achieve refined management, data must be used to speak, but there are also multiple challenges. At present, hospitals generally have problems such as low quality of filling out medical record homepages, poor connectivity of various business systems, missing key information, and inconsistent statistical caliber of data in multiple systems, which leads to problems with problematic data and low data quality when data is shared and analyzed within the hospital. Summary of the invention

[0004] In view of the shortcomings of the existing technology, the present invention provides a medical-based data quality verification method, which realizes the automated verification of the data quality of various systems in the hospital and unifies the statistical caliber of indicators of multiple systems, thereby improving the application and analysis effect of data in the hospital.

[0005] In order to achieve the above object, the technical solution of the present invention is:

[0006] A data quality verification method based on medical treatment specifically comprises the following steps:

[0007] Step 1) Sort out the business scope of hospital operation and management, connect various business systems in the hospital through ETL according to the business scope, and automatically collect medical data in each business system;

[0008] Step 2) In ETL, the collected medical data is imported into the data landing area; then the data content is normalized according to the unified data dictionary and format, and the loading of the data preparation area is completed; the processed data is loaded into the data storage area, and the basic data layer, processing summary layer and application market layer data are loaded in the data storage area. After the processing is completed, the data is stored in the data warehouse;

[0009] Step 3) Sort out the dimensions, metrics and indicators of the data in the data warehouse, design corresponding business analysis topics, divide the business analysis topics into several subject domains, sort out the key indicators in each subject domain, and determine the verification type of each subject domain;

[0010] Step 4) According to the verification type, ETL is used to verify the data quality of the data warehouse using the SQL statements in the business system as the verification caliber; the data quality verification method is as follows:

[0011] 41) Connect to the business system and use the SQL statements in the business system as the data source;

[0012] 42) Sort, summarize and convert rows to columns for key fields of data sources;

[0013] 43) Connect to the data warehouse and write SQL statements in the data warehouse to query the same content as the data source;

[0014] 44) The results of the data source and the data warehouse are associated and merged by time and verification content as the data quality verification result;

[0015] 45) Add a sequence, a data warehouse table, and a calculator to the data warehouse, obtain the maximum primary key value of the data warehouse table, associate the sequence with the maximum primary key value of the data warehouse table, and generate an auto-increment number through the calculator as the primary key of the data warehouse table to be landed;

[0016] 46) The processed and summarized data is used as the data quality verification result;

[0017] Step 5) Build a medical data model table of data quality verification results, store the data quality verification results in the medical data model table, that is, implement the data warehouse, and use the spring boot architecture to develop a data quality result display function to output and display the data quality verification results.

[0018] Preferably, the specific method of automatically collecting medical data in each business system is:

[0019] Step 11) According to the business scope, the process of collecting data from each business system is divided into several ordered ETL sub-process task sequences, and the construction of the data flow is started according to the ETL sub-process task sequence; a process scheduling database related to the ETL sub-process scheduling is established to save the information of the ETL sub-process scheduling server, task data, task data triggering rules, task groups, scheduling plans and scheduling logs;

[0020] Step 12) Monitor the task data according to the task data triggering rules and the scheduling plan. When the business system data meets the task data triggering rules, the task group automatically executes the corresponding task;

[0021] Step 13) Record the execution logs of each server component and each task of the ETL scheduling management, and save the logs on a daily basis.

[0022] Preferably, an input table is configured in each ordered ETL sub-process task sequence to input the data source, database and table name, and configure the data source connection method, data processing operation method, output table name and data update method.

[0023] Preferably, the method of loading data of the basic data layer, processing summary layer and application mart layer in the data storage area is: clean the data in the data storage area and perform dimensional degradation processing at the basic data layer, generate dimension table data and detailed data, and use common dimensions to perform light aggregation on the detailed data to construct a summary indicator table of public granularity.

[0024] Preferably, the method of generating dimension table data is: taking dimension as modeling driver, completing attribute definition by adding dimension attributes and definition calculation logic of associated dimensions according to the business meaning of each dimension, and establishing a consistent data analysis dimension table using a snowflake model;

[0025] The method of generating detailed data is: using the business process as the modeling driver, and building the most fine-grained detailed fact table based on the characteristics of each specific business process.

[0026] Preferably, the several subject domains include a management cockpit subject domain, an outpatient business analysis subject domain, an inpatient business analysis subject domain, a medical quality analysis subject domain, a surgical business analysis subject domain, a nursing business analysis subject domain, a medical record management analysis subject domain and a medical technology business analysis subject domain.

[0027] Preferably, the key indicators of the management cockpit theme domain include outpatient visits, emergency visits, hospital admissions, discharges, outpatient income, inpatient income, drug share, antibiotic use rate, mortality rate, and basic drug share;

[0028] The key indicators of the outpatient business analysis subject domain include outpatient flow, number of outpatient surgeries, and number of observation patients;

[0029] The key indicators of the inpatient business analysis subject domain include average cost per visit, discharge income, and inpatient income;

[0030] Key indicators in the subject domain of medical quality analysis include three-day diagnosis rate, readmission rate, and mortality rate;

[0031] The key indicators of the surgical business analysis subject domain include the number of surgeries, distribution of surgeries at each level, and the number of anesthesia patients;

[0032] Key indicators in the nursing business analysis subject domain include the number of nursing staff and duration at all levels;

[0033] Key indicators of the medical record management analysis subject domain include the number of admissions and discharges, actual bed occupancy, and bed utilization rate;

[0034] The key indicators of the medical technology business analysis subject domain include medical technology income, equipment income, and technician income.

[0035] Preferably, the data quality verification type of each subject domain in the data warehouse includes one or more of consistency, completeness, repeatability, logic, validity and comparability;

[0036] Among them, consistency verification is a consistency check for the operational indicators, pharmaceutical indicators, and medical quality indicators that are of concern to the hospital, ensuring that the statistical data of the data warehouse and the business system are consistent; logical verification is a logical verification rule formulated based on the application scenarios of the medical industry and national key assessment indicators;

[0037] Integrity verification is to check the integrity of key information;

[0038] Repeatability check is to check the repeatability of data for some subject information that cannot be repeated;

[0039] Validity check is to check whether the content filled in the business system exceeds the value range for some value domains;

[0040] Comparability verification is a consistency check of data in multiple tables in a business system.

[0041] Preferably, ETL is used in the medical data model table to implement the output of data quality verification data, and the execution status and execution log are monitored daily.

[0042] Technical effects and advantages of the present invention:

[0043] 1. The present invention provides a medical-based data quality verification method, which automatically verifies data quality by using ETL technology to achieve automated data quality verification. The hospital does not need to obtain the indicators of the business system and the indicators of the data warehouse separately and then perform manual verification. The ETL technology is used for automatic verification, saving labor costs.

[0044] 2. The present invention provides a medical-based data quality verification method, which has developed a rich verification indicator system that fits the medical business scenarios and is of concern to hospital managers. By using ETL technology to automatically verify data quality, the verification scope can be flexibly set and adjusted, and the consistency, repeatability, completeness, validity, and comparability of data verification can be increased, deleted, or modified according to the needs of the hospital, thus realizing custom configuration.

[0045] 3. The medical-based data quality verification method provided by the present invention can display the results of data quality verification, thereby making the data quality verification results more intuitive. Users can intuitively see the data quality verification results without manual participation.

[0046] 4. The present invention provides a medical-based data quality verification method, which automatically verifies data quality by using ETL technology. It can display daily data quality verification results and output problem data to facilitate the rectification of data problems and review after data rectification, thereby forming a closed-loop management process for continuous improvement of data quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flowchart of the data quality verification method of the present invention;

[0048] Figure 2 It is a flow chart of the data quality verification logic implemented by the present invention using ETL;

[0049] Figure 3 It is a schematic diagram of the structure of the medical data model table of the data quality verification result of the present invention;

[0050] Figure 4 It is a display effect diagram of the data quality verification result of the present invention. DETAILED DESCRIPTION

[0051] The present invention is further described in detail below with reference to the embodiments given in the accompanying drawings.

[0052] See also Figure 1 As shown, a medical data quality verification method specifically includes the following steps:

[0053] Step 1) Sort out the business scope of hospital operation and management, connect various business systems in the hospital through ETL according to the business scope, and automatically collect medical data in each business system;

[0054] Step 2) In ETL, the collected medical data is imported into the data landing area; then the data content is normalized according to the unified data dictionary and format, and the loading of the data preparation area is completed; the processed data is loaded into the data storage area, and the basic data layer, processing summary layer and application market layer data are loaded in the data storage area. After the processing is completed, the data is stored in the data warehouse;

[0055] Step 3) Sort out the dimensions, metrics and indicators of the data in the data warehouse, design corresponding business analysis topics, divide the business analysis topics into several subject domains, sort out the key indicators in each subject domain, and determine the verification type of each subject domain;

[0056] Step 4) Based on the verification type, ETL is used to verify the data quality of the data warehouse using the SQL statements in the business system as the verification caliber;

[0057] Step 5) Build a medical data model table of data quality verification results, store the data quality verification results in the medical data model table, that is, implement the data warehouse, and use the spring boot architecture to develop a data quality result display function to output and display the data quality verification results.

[0058] The following is a detailed description of each step:

[0059] Step 1) Sort out the business scope of hospital operation and management, connect various business systems within the hospital through ETL according to the business scope, and automatically collect medical data in each business system.

[0060] The specific methods for automatically collecting medical data in various business systems are:

[0061] Step 11) According to the business scope, the process of collecting data from each business system is divided into several ordered ETL sub-process task sequences. Each ordered ETL sub-process task sequence is configured with an input table to enter the data source, database and table name, and configure the connection method of the data source, the data processing operation method, the output table name and the data update method. Start the construction of the data flow according to the ETL sub-process task sequence; establish a process scheduling database related to the ETL sub-process scheduling to save the information of the ETL sub-process scheduling server, task data, task data triggering rules, task groups, scheduling plans and scheduling logs.

[0062] Among them, the connection method of the data source can be set to extraction or direct connection, the configuration of data processing operations can use operations such as association, merger, collection, row to column, etc., and the output table name and data update method can be set to full coverage, append, etc.

[0063] Step 12) Monitor the task data according to the task data triggering rules and the scheduling plan. When the business system data meets the task data triggering rules, the task group automatically executes the corresponding task;

[0064] Step 13) Record the execution log of each task in the scheduling management and save the log on a daily basis.

[0065] Step 2) In ETL, the collected medical data is imported into the data landing area; then the data content is normalized according to the unified data dictionary and format to complete the loading of the data preparation area; the processed data is loaded into the data storage area, and the basic data layer, processing summary layer and application market layer data are loaded in the data storage area. After the processing is completed, the data is dropped into the data warehouse.

[0066] Among them, the method of loading data of the basic data layer, processing summary layer and application market layer in the data storage area is: in the basic data layer, the data in the data storage area is cleaned and dimensionally degraded to generate dimension table data and detailed data, and the detailed data is lightly summarized using common dimensions to construct a summary indicator table of public granularity.

[0067] Furthermore, when cleaning the data of the data location area at the basic data layer, methods such as removing null values ​​and dirty data are used for cleaning; the commonly used dimensions include time, organizational level, and patients.

[0068] The method of generating dimension table data is to use dimension as modeling driver, complete attribute definition by adding dimension attributes and definition calculation logic of associated dimensions according to the business meaning of each dimension, and use snowflake model to establish a consistent data analysis dimension table.

[0069] The way to generate detailed data is to use the business process as the modeling driver and build the most fine-grained detailed fact table based on the characteristics of each specific business process.

[0070] Step 3) Sort out the dimensions, metrics and indicators of the data in the data warehouse, design corresponding business analysis topics, divide the business analysis topics into several subject domains, sort out the key indicators in each subject domain, and determine the verification type for each subject domain.

[0071] Among them, several subject domains include the management cockpit subject domain, outpatient business analysis subject domain, inpatient business analysis subject domain, medical quality analysis subject domain, surgical business analysis subject domain, nursing business analysis subject domain, medical record management analysis subject domain and medical technology business analysis subject domain.

[0072] The key indicators of the management cockpit theme domain include outpatient visits, emergency visits, hospital admissions, discharges, outpatient revenue, inpatient revenue, drug share, antibiotic use rate, mortality rate, and basic drug share;

[0073] The key indicators of the outpatient business analysis subject domain include outpatient flow, number of outpatient surgeries, and number of observation patients;

[0074] The key indicators of the inpatient business analysis subject domain include average cost per visit, discharge income, and inpatient income;

[0075] Key indicators in the subject domain of medical quality analysis include three-day diagnosis rate, readmission rate, and mortality rate;

[0076] The key indicators of the surgical business analysis subject domain include the number of surgeries, distribution of surgeries at each level, and the number of anesthesia patients;

[0077] Key indicators in the nursing business analysis subject domain include the number of nursing staff and duration at all levels;

[0078] Key indicators of the medical record management analysis subject domain include the number of admissions and discharges, actual bed occupancy, and bed utilization rate;

[0079] The key indicators of the medical technology business analysis subject domain include medical technology income, equipment income, and technician income.

[0080] The data quality verification types for each subject domain in the data warehouse include one or more of consistency, completeness, duplication, logic, validity and comparability;

[0081] Among them, consistency verification is to check the consistency of operational indicators, pharmaceutical indicators, and medical quality indicators that are of concern to the hospital, to ensure that the statistical data of the data warehouse and the business system are consistent;

[0082] Logical verification is a logical verification rule formulated based on the application scenarios of the medical industry and national key assessment indicators;

[0083] Integrity verification is to check the integrity of key information;

[0084] Repeatability check is to check the repeatability of data for some subject information that cannot be repeated;

[0085] Validity check is to check whether the content filled in the business system exceeds the value range for some value domains;

[0086] Comparability verification is a consistency check of data in multiple tables in a business system.

[0087] Step 4) According to the verification type, ETL is used to verify the data quality of the data warehouse using the SQL statements in the business system as the verification caliber; the data quality verification method is as follows:

[0088] 41) Connect to the business system and use the SQL statements in the business system as the data source;

[0089] 42) Sort, summarize and convert rows to columns for key fields of data sources;

[0090] 43) Connect to the data warehouse and write SQL statements in the data warehouse to query the same content as the data source;

[0091] 44) The results of the data source and the data warehouse are associated and merged by time and verification content as the data quality verification result;

[0092] 45) Add a sequence, a data warehouse table, and a calculator to the data warehouse, obtain the maximum primary key value of the data warehouse table, associate the sequence with the maximum primary key value of the data warehouse table, and generate an auto-increment number through the calculator as the primary key of the data warehouse table to be landed;

[0093] 46) The processed and summarized data is used as the data quality verification result.

[0094] During specific implementation, the above-mentioned verification methods can be used to verify the data quality of each subject domain in the data warehouse. The verification types include consistency verification, integrity verification, repeatability verification, logic verification, validity verification and comparison verification.

[0095] The data verification type of each subject domain is different. For example, consistency verification is a consistency check for the operational indicators, pharmaceutical indicators, and medical quality indicators that the hospital is concerned about, ensuring that the statistical data of the data warehouse and the business system are consistent, and outputting the consistency verification results. For example: outpatient visits, emergency visits, outpatient income, inpatient income, average cost per visit, rescue visits, critically ill visits, basic drug ratio, antibiotic use rate, number of surgeries, discharges, admissions, number of people at the beginning of the period, number of people at the end of the period, etc.

[0096] Logical verification is a logical verification rule formulated based on the application scenarios of the medical industry and the national key assessment indicators, and data that does not meet the rules will be output. For example: the number of patient rescues < the number of successful rescues; total cost < the sum of sub-item costs; total hospitalization costs ≥ total hospitalization costs including self-paid amount; surgical treatment costs < anesthesia costs + surgical costs; Western medicine costs < antibacterial drug costs; admission date > diagnosis date after admission; admission date > surgical date; non-surgical treatment project costs < clinical physical therapy costs; there is a surgical record but the surgical cost < 0, etc.

[0097] The integrity check verifies the integrity of key information and outputs data that does not conform to the rules. For example: patients whose discharge method is death do not have the time of death, cause of death, diagnosis code of cause of death, etc.; the record of hand numbness does not have the surgical incision, surgeon, anesthesia method, and surgical level; the patient is ≤28 days old but has no admission weight; there is a plan to be re-hospitalized within 31 days of discharge, but there is no purpose of re-hospitalization within 31 days; the patient's diagnosis code is Z37, but there is no birth weight; the patient's diagnosis is maternity, but there is no delivery outcome; the billing department and billing doctor of the fee are empty; the patient's diagnosis number is empty, etc.

[0098] Repeatability check is to check the repeatability of some main information that cannot be repeated, and output the repeated data. For example: there are multiple patient information for the same ID number; there are multiple hospitalization records for the same patient at the same time; there are multiple medical record covers for the same patient at the same time; there are multiple records for the same disease code and they are valid, etc.

[0099] Validity verification is to check for some value domains, where the content filled in the business system exceeds the value range. For example: the patient's discharge method exceeds the value range; the patient's cause of death does not fall within the ICD-10 coding range; the coding of the admission and discharge diagnosis compliance on the medical record homepage does not fall within the diagnosis compliance range; the medical record quality on the medical record homepage does not fall within the medical record quality range.

[0100] Comparative verification is a consistency check of multi-table data in the business system. For example: a patient has an operation record on the medical record homepage, but the hand numbness system does not exist, or the hand numbness system exists but the medical record homepage does not exist; a patient's anesthesia type, surgeon, surgery name, first assistant, second assistant, surgery level, etc. on the medical record homepage are inconsistent with the hand numbness system; a patient's discharge time on the medical record homepage and the change record are not equal.

[0101] Step 5) Build a medical data model table of data quality verification results, store the data quality verification results in the medical data model table, that is, implement the data warehouse, and use the spring boot architecture to develop a data quality result display function to output and display the data quality verification results.

[0102] In specific implementation, the method of outputting data quality verification results is to use ETL in the medical data model table to output data quality verification data, and monitor the execution status and execution log on a daily basis.

[0103] This application realizes the automated verification of data quality of various systems in the hospital, and unifies the statistical caliber of indicators of multiple systems, thereby improving the application and analysis effect of data in the hospital. At the same time, it outputs the problems existing in data quality to facilitate the hospital to carry out targeted rectification work and further improve data quality management.

[0104] Next, we will conduct a data quality check on the consistency of the outpatient visit data in the business system and the data warehouse. The steps are as follows:

[0105] Step 1) Sort out and investigate the business scope of hospital operation and management, understand the key indicators that managers are concerned about, and determine that the number of outpatient visits is one of the most basic and most critical indicators in the hospital;

[0106] Step 2) Because the number of outpatient visits needs to be counted through the patient registration record table of the HIS system, the patient registration records of the HIS system are collected and put into the data warehouse to count the number of outpatient visits. According to the unified data dictionary, such as registration category, execution department, doctor who issues the order, registration time, etc., the data content is processed, and the calculation and loading of the data of the basic layer and the application market layer are completed, and the data is stored in the data warehouse;

[0107] Step 3) Sort out the dimensions, measurements and display methods of the outpatient visit indicators, and design corresponding business analysis topics, such as hourly outpatient flow analysis, outpatient volume analysis of each department, registration volume analysis of each doctor, monthly outpatient volume analysis of this year, etc., all of which are analyzed and displayed around the key indicator of outpatient visits;

[0108] Step 4) Communicate with the hospital, use the SQL statements in the business system as the verification criteria, and the hospital recognizes the SQL statistical results, and uses ETL technology to implement the verification logic; the way to use ETL technology to implement the verification logic is as follows Figure 2 As shown, the following steps are included:

[0109] 41) Connect to the HIS system (HIS_TBI) and write the SQL for outpatient statistics of the HIS system;

[0110] 42) Connect to the data warehouse (DW_TBI) and write SQL for outpatient statistics in the data warehouse;

[0111] 43) Sort the above two parts of SQL execution results (ST_01, ST_02);

[0112] 44) Associating and merging the sorted results of the above two parts, and setting a name for the merged result column (record set connection);

[0113] 45) After merging, add a sequence (SEQ_01) and a calculator (CAL_01) to the data warehouse, obtain the maximum primary key value of the data warehouse table (TBI_10), associate the sequence with the maximum primary key value of the data warehouse table (JNC_01), and generate an auto-increment number through the calculator as the primary key of the data warehouse table;

[0114] 46) The result of merging and generating primary keys is used as the data quality verification result and landed in the medical data model table of the data quality verification result (insert / update). The medical data model table is referred to Figure 3 As shown, the medical data model table of the data quality verification result is based on the final display requirements, for example Figure 3 The corresponding ① part represents the column name of the data table, from top to bottom: table primary key, verification content, business system verification result, data warehouse verification result, verification time, constant content recorded in the ETL processing process, business department. Virtual corresponds to the virtual column of the database, which is not used yet. The ② part represents the data type of the ① part. Each column will correspond to a fixed data type. For example, if the verification content is a character type, it corresponds to the varchar type of the database table. The ③ part represents whether the column can be empty. Currently, it can be empty except for the primary key.

[0115] Step 5) Use the spring boot architecture to develop a data quality verification result analysis function and display the data quality verification results. For verification result display, see Figure 4 As shown, after selecting the start time, end time, and verification result (all / consistent / inconsistent) and querying, the verification results that meet the conditions will be displayed. The data verification name represents the key indicator of the verification, the system business data represents the result of the indicator in the business system, the DW data represents the result of the indicator in the data warehouse, and the difference represents the difference between the business system and the data warehouse.

[0116] The above is only a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several modifications and improvements without departing from the inventive concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A medical data quality verification method, characterized in that: The specific steps include: Step 1) Sort out the business scope of hospital operation and management, connect various business systems in the hospital through ETL according to the business scope, and automatically collect medical data in each business system; Step 2) Import the collected medical data into the data landing area in ETL; then normalize the data content according to the unified data dictionary and format to complete the loading of the data preparation area; The processed data is loaded into the data storage area, and the data of the basic data layer, processing summary layer and application market layer are loaded into the data storage area. The method of loading the data of the basic data layer, processing summary layer and application market layer into the data storage area is as follows: the data of the data storage area is cleaned and dimensionally degraded at the basic data layer to generate dimension table data and detailed data, and the detailed data is lightly summarized using common dimensions to construct a summary indicator table of public granularity; The method of generating dimension table data is: using dimension as modeling driver, completing attribute definition by adding dimension attributes and definition calculation logic of associated dimensions according to the business meaning of each dimension, and establishing a consistent data analysis dimension table using a snowflake model; The method of generating detailed data is: using the business process as the modeling driver, and building the most fine-grained detailed fact table according to the characteristics of each specific business process; After processing, the data is stored in the data warehouse; Step 3) Sort out the dimensions, metrics and indicators of the data in the data warehouse, design corresponding business analysis topics, divide the business analysis topics into several subject domains, sort out the key indicators in each subject domain, and determine the verification type of each subject domain; Step 4) According to the verification type, ETL is used to verify the data quality of the data warehouse using the SQL statements in the business system as the verification caliber; the data quality verification method is as follows: 41) Connect to the business system and use the SQL statements in the business system as the data source; 42) Sort, summarize and convert rows to columns for key fields of data sources; 43) Connect to the data warehouse and write SQL statements in the data warehouse to query the same content as the data source; 44) The results of the data source and the data warehouse are associated and merged by time and verification content as the data quality verification result; 45) Add a sequence, a data warehouse table, and a calculator to the data warehouse, obtain the maximum primary key value of the data warehouse table, associate the sequence with the maximum primary key value of the data warehouse table, and generate an auto-increment number through the calculator as the primary key of the data warehouse table to be landed; 46) The processed and summarized data is used as the data quality verification result; Step 5) Build a medical data model table of data quality verification results, store the data quality verification results in the medical data model table, and use the spring boot architecture to develop a data quality result display function to output and display the data quality verification results.

2. A medical-based data quality verification method according to claim 1, characterized in that: The specific method of automatically collecting medical data in each business system is: Step 11) According to the business scope, the process of collecting data from each business system is divided into several ordered ETL sub-process task sequences, and the construction of the data flow is started according to the ETL sub-process task sequence; a process scheduling database related to the ETL sub-process scheduling is established to save the information of the ETL sub-process scheduling server, task data, task data triggering rules, task groups, scheduling plans and scheduling logs; Step 12) Monitor the task data according to the task data triggering rules and the scheduling plan. When the business system data meets the task data triggering rules, the task group automatically executes the corresponding task; Step 13) Record the execution log of each task in the scheduling management and save the log on a daily basis.

3. A medical-based data quality verification method according to claim 2, characterized in that: An input table is configured in each ordered ETL sub-process task sequence to enter the data source, database and table name, and configure the data source connection method, data processing operation method, output table name and data update method.

4. The medical data quality verification method according to claim 1, characterized in that: The several subject domains include a management cockpit subject domain, an outpatient business analysis subject domain, an inpatient business analysis subject domain, a medical quality analysis subject domain, a surgical business analysis subject domain, a nursing business analysis subject domain, a medical record management analysis subject domain and a medical technology business analysis subject domain.

5. A medical-based data quality verification method according to claim 4, characterized in that: The key indicators of the management cockpit theme domain include outpatient visits, emergency visits, hospital admissions, discharges, outpatient revenue, inpatient revenue, drug share, antibiotic use rate, mortality rate, and basic drug share; The key indicators of the outpatient business analysis subject domain include outpatient flow, number of outpatient surgeries, and number of observation patients; The key indicators of the inpatient business analysis subject domain include average cost per visit, discharge income, and inpatient income; Key indicators in the subject domain of medical quality analysis include three-day diagnosis rate, readmission rate, and mortality rate; The key indicators of the surgical business analysis subject domain include the number of surgeries, distribution of surgeries at each level, and the number of anesthesia patients; Key indicators in the nursing business analysis subject domain include the number of nursing staff and duration at all levels; Key indicators of the medical record management analysis subject domain include the number of admissions and discharges, actual bed occupancy, and bed utilization rate; The key indicators of the medical technology business analysis subject domain include medical technology income, equipment income, and technician income.

6. A medical-based data quality verification method according to claim 5, characterized in that: The data quality verification types for each subject domain in the data warehouse include one or more of consistency, completeness, duplication, logic, validity and comparability; Among them, consistency verification is to check the consistency of operational indicators, pharmaceutical indicators, and medical quality indicators that are of concern to the hospital, to ensure that the statistical data of the data warehouse and the business system are consistent; Logical verification is a logical verification rule formulated based on the application scenarios of the medical industry and national key assessment indicators; Integrity verification is to check the integrity of key information; Repeatability check is to check the repeatability of data for some subject information that cannot be repeated; Validity check is to check whether the content filled in the business system exceeds the value range for some value domains; Comparability verification is a consistency check of data in multiple tables in a business system.

7. The medical data quality verification method according to claim 1, characterized in that: Use ETL in the medical data model table to output data quality verification data, and monitor the execution status and execution logs daily.

Citation Information

Patent Citations

  • Data warehouse system based on data management and blood relationship design

    CN110232098A

  • Data quality improvement method and system based on medical big data

    CN111241086A

  • Modeling method and model for data warehouse construction, computer equipment and storage medium

    CN111460045A

  • Index library generation method and device, computer equipment and storage medium

    CN116978526A

  • System and Method for Producing Performance Reporting and Comparative Analytics for Finance, Clinical Operations, Physician Management, Patient Encounter, and Quality of Patient Care

    US20130054260A1