Medical data analysis platform, analysis method, computer and storage medium

Abstract

The present invention discloses a medical data analysis platform, an analysis method, a computer and a storage medium, and belongs to the field of medical data analysis technology. The present invention includes a data collection management module, a data governance module, a data quality control module, a medical data knowledge base module, a background data processing module, a data analysis model module, a data analysis engine module, a foreground data analysis module and a data deduction module. The present invention is a complete and unified medical data analysis and statistics platform, from data collection, transcoding, cleaning, quality control, layered processing application, display, and deduction; the whole link solves the problems of poor hospital data quality, difficult data application, and large data dispersion; the medical data analysis platform is based on a unified and configurable underlying architecture processing, which can meet the needs of various business scenarios of the hospital; based on the data structure of layered processing, the custom configuration processing can quickly generate various indicator set applications; it can expand business and applications with the highest efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical data analysis, and specifically relates to a medical data analysis platform, an analysis method, a computer and a storage medium. Background Art

[0002] With the rapid development of medical information technology, the amount of medical data is growing exponentially, and these data contain huge value. However, most medical institutions currently lack effective methods and technologies to deeply analyze and utilize data, and are unable to transform data into strategic decision-making basis. Therefore, building an efficient and reliable medical data analysis platform to realize the mining and analysis of massive medical data has become a problem that needs to be solved urgently. The present invention aims to solve the construction method and technology of the medical data analysis platform, in order to improve the utilization efficiency of medical data and the value conversion ability.

[0003] At present, all medical data analysis platforms on the market can only make simple result analysis on the data. It is impossible to deeply integrate the mined data with business intelligence and provide professional guidance. It is more necessary to rely on experts from various aspects of the industry to combine their own understanding of the business to analyze on the spot to give opinions and reports on hospital operations and development. Not only is the speed slow, inefficient, and infrequent, but it is also unable to provide real-time professional guidance for hospital operations and development.

[0004] For hospitals located in third-, fourth- and fifth-tier cities, there are very few industry experts. How to maximize the value of hospital data, how to expose problems in hospital operations and development, and how to find the optimal solution for future development directions are all difficult problems. Summary of the invention

[0005] The problem to be solved by the present invention is as follows: With respect to the existing information technology and medical services, policies, regulations, and industry requirements are not matched or related; data cannot be directly matched with services to meet the hospital's needs for deep mining and deep application of massive data, and operations managers cannot arbitrarily slice and dice medical data for all-round three-dimensional display. The present invention proposes a medical data analysis platform, a medical data analysis method based on the medical data analysis platform, a computer, and a storage medium. The entire link solves the problems of poor hospital data quality, difficult data application, and large data dispersion, and can meet the needs of various business scenarios of hospitals.

[0006] Technical solution: In order to solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0007] A medical data analysis platform includes a data collection management module, a data governance module, a data quality control module, a medical data knowledge base module, a background data processing module, a data analysis model module, a data analysis engine module, a front-end data analysis module and a data deduction module.

[0008] Data collection management module: Automatically collect data from various decentralized systems in medical institutions. During the data collection process, the data will be split and relevant fields will be extracted according to the metadata of the analysis, and then stored and indexed for subsequent big data analysis.

[0009] Data governance module: provides various types of front-end plug-ins embedded in the business system that generates data to implement governance before and during data generation. The platform calls the analysis engine, executes the associated dynamic embedded scripts to perform analysis and generate reports;

[0010] Data quality control module: Quality control is performed before the business system data is stored in the database. The platform calls the quality control engine, executes the associated dynamic embedded quality control scripts to analyze and generate reports, including rule quality control, conditional quality control and logical verification quality control of the data;

[0011] Medical data knowledge base module: mainly includes indicator knowledge, clause knowledge, medical industry dictionary library, and medical coding library;

[0012] Backend data processing module: divides indicator data into atomic data items, calculation data items, and analysis data items according to the depth of analysis;

[0013] Data analysis model module: an analysis model built based on indicators and clauses constructed based on data. Indicators are composed of data items. Different data items are configured according to rules to produce arbitrary indicators. Different indicators are customized according to clause rules to form a new data analysis model.

[0014] Data analysis engine module: When the data analysis engine module is executed, it calls the drawn kettle script to obtain the configuration indicator information in the analysis model;

[0015] Front-end data analysis module: The front-end application is an application presentation of the data set processed by the back-end data analysis module;

[0016] Data deduction module: Use the original data as a benchmark to simulate and deduce future business development. Through the analysis model, import relevant indicator data and simulate the adjustment of the value of each indicator data item. Based on the execution results of the analysis model, help users adjust their decision-making goals.

[0017] A medical data analysis method based on a medical data analysis platform comprises the following steps:

[0018] S1: Start the data collection system, configure the collection system execution plan, and obtain data from various business systems within the medical institution; determine whether to configure data standard coding for the acquired data, and if not, configure data industry standard coding information;

[0019] S2: Transcode the data encoded by the data standard configured in the previous step and perform data quality control verification;

[0020] S3: The data that have passed the quality control are entered into the database of the medical data analysis platform, and the data after entry are subjected to connotation quality control verification;

[0021] S4: Perform layered processing and application operations on the stored data and supplementary data;

[0022] S5: The processed data is displayed on the front-end page according to the rules of the indicators;

[0023] S6: Analysis and deduction of medical data after processing in S5.

[0024] Preferably, in S4, the stored data and the supplementary data are subjected to layered processing and application operations, specifically:

[0025] S41: Set up basic data and related operations on the medical data analysis platform, specifically:

[0026] Configure organization information, indicator information, clause information and model information;

[0027] Configure basic role information, user information, account information, API interface configuration and data collection cycle information;

[0028] Configure the indicator data analysis model. The indicator data analysis model is used by the analysis engine to call the indicator template for index processing. Users can dynamically configure the model according to business scenarios and execute the model on demand.

[0029] Configure the analysis engine execution rules, support automation support for different cycle types, and support user-defined editing and configuration operations;

[0030] S42: The data that have passed the connotation quality control verification and the supplementary data are processed and calculated using the kettle script.

[0031] Preferably, in S42, the data that have passed the connotation quality control verification and the supplementary data are processed and calculated using a kettle script, specifically:

[0032] S421: Use the kettle script to process and calculate the original data items and output the atomic data items;

[0033] S422: using the kettle script to output the calculated data items according to the calculation rules for the atomic data items;

[0034] S423: Use the kettle script to add calculation rules to the calculation data items and output the analysis data items.

[0035] Preferably, in S5, the processed data is displayed on the front-end page according to the rules of the indicator, specifically:

[0036] S51: configure the underlying basic data of the system;

[0037] S52: Click the intelligent analysis module on the platform to perform intelligent analysis on the incoming data;

[0038] S53: Click the indicator overview module to enter the indicator overview summary analysis page;

[0039] S54: The data analysis platform performs annual data trend analysis or monthly data trend analysis on the indicator data in a single year based on the indicator attributes.

[0040] Preferably, in S54, the data analysis platform performs an annual data trend analysis or a monthly data trend analysis of the indicator data in a single year according to the indicator attributes, specifically:

[0041] S541: Click on the annual data of any indicator to jump to the annual data details, and display the detailed values ​​of the numerator / denominator data items respectively;

[0042] S542: Click "360 View" in the detailed data operation column, and according to the configured 360 jump link, cross-platform index data original system page to trace back data;

[0043] S543: Click any indicator name to jump to the indicator analysis page, and drill down to display the indicator by department, year, month, disease type, and ward according to the indicator attributes;

[0044] S544: Click the numerator and denominator values ​​in the indicator analysis display page list to display the data details. Click the "360 View" cross-platform index data original system page in the operation column to trace the data back.

[0045] Preferably, in S6, the medical data processed in S5 is analyzed and deduced, specifically:

[0046] S61: Customize the display, query, drill-down and classification of the data processed by S5;

[0047] S62: Click the data deduction module on the medical data analysis platform to enter the function page, specifically:

[0048] S621: Select any annual indicator data value that has been processed into atomic data items, calculation data items, and analysis data items as the benchmark data;

[0049] S622: Configure and select the indicators to be deduced, and set a calculation data item threshold for each indicator;

[0050] S623: deduce a value for any atomic data item in the selected indicator;

[0051] S624: Start the data deduction model, call the analysis engine to perform deduction calculations, output the deduction data results, and output new deduction analysis data item values.

[0052] As a preference, perform the following operations on the data that cannot be collected by the system:

[0053] Supplemental data recording for data that cannot be collected by the system;

[0054] Assign responsibilities to the indicators according to departments, and after assigning departments, reassign the reporting, verification, and review personnel for the indicators that need to be supplemented;

[0055] Perform data entry operations according to the assigned filler information and responsibility authority;

[0056] Verify the entered data, and conduct data review on the index data that have passed the verification; return the index data that have failed the verification to the data entry page, revise the data and submit it again;

[0057] The data that has passed the review will be reviewed. The index data that has passed the review will be directly stored in the database. The index data that has failed the review will be returned to the data entry page for revision and resubmission and verification.

[0058] The approved data will be stored in the database, and the stored data will be processed by script calculation.

[0059] A computer comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the above-mentioned medical data analysis method based on the medical data analysis platform when executing the computer program.

[0060] A storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned medical data analysis method based on the medical data analysis platform.

[0061] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0062] (1) The present invention is an application method based on medical data analysis. Its advantage is that it is a complete and unified medical data analysis and statistical platform, from data collection, transcoding, cleaning, quality control, layered processing application, display, and deduction. The whole link solves the problems of poor data quality, difficult data application, and large data dispersion in hospitals. The medical data analysis platform is based on a unified and configurable underlying architecture processing, which can meet the needs of various business scenarios in hospitals. Based on the data structure of layered processing, custom configuration processing can quickly generate various indicator set applications, flexible and configurable analysis models, and the front-end controllable analysis engine can expand business and applications with the highest efficiency.

[0063] (2) The hierarchical data processing of the present invention solves the performance problem of real-time call of billions of data in hospitals. It can quickly index data according to needs and make real-time calls, which greatly improves the feedback speed of data display.

[0064] (3) The data deduction function of the present invention solves the practical problems of the lack of information technology experts in the medical field and the difficulty in guiding business development. The data of the whole hospital is simulated and deduced, and the problems of business development are located through variables and quantitative relationships. The data is used to support the business and simulate the business, providing an operational decision-making basis and method for the self-circulating development of the hospital. It has greatly improved the progress of the hospital's high-quality development. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 This is a structural framework diagram of the medical data analysis platform of the present invention;

[0066] Figure 2 A schematic diagram illustrating the process of data collection and quality control processing of the present invention;

[0067] Figure 3 This is a schematic diagram illustrating the data display process of the medical data analysis platform of the present invention;

[0068] Figure 4 This is a schematic diagram illustrating the data deduction process of the medical data analysis platform of the present invention;

[0069] Figure 5 It is a schematic diagram of the hierarchical calculation process of atomic data items, calculation data items, and analysis data items of the present invention;

[0070] Figure 6 This is the indicator analysis flow chart of the present invention. DETAILED DESCRIPTION

[0071] The present invention is further illustrated below in conjunction with specific examples. The examples are implemented based on the technical solutions of the present invention. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0072] The medical data analysis platform provided in this embodiment includes a data collection management module, a data governance module, a data quality control module, a medical data knowledge base module, a background data processing module, a data analysis model module, a data analysis engine module, a front-end data analysis module and a data deduction module.

[0073] Data collection management module: used to automatically collect data from various decentralized systems in medical institutions, so as to build the underlying data foundation of the medical data analysis platform.

[0074] The data acquisition module of the present invention realizes metadata extraction and storage of data during the acquisition process, making it more suitable for the analysis of big data. Most source data are provided in a wide table format. When the amount of data reaches a massive amount of data, the efficiency will drop drastically. During the acquisition process, the present invention will split and extract relevant fields (key-values ​​format) according to the metadata to be analyzed (i.e. the requirements of atomic data items), and then store and create indexes for subsequent analysis, which optimizes the speed of column splitting in relational database row storage and greatly improves the analysis efficiency.

[0075] Data governance module: It is used to use artificial intelligence technologies such as natural language processing (NLP) to process the collected data, identify missing items, abnormal items and standardize the data to achieve structured processing of different types of data, and improve data availability and processing efficiency.

[0076] The data governance of the present invention covers the monitoring of the entire life cycle before, during and after data generation. The present invention provides various types of front-end plug-ins that can be embedded in the business system that generates data to implement governance of the links before and during data generation, and assist the business system in producing high-quality data. The present invention implements post-management before the business system data is stored in the warehouse. The platform calls the analysis engine and executes the associated dynamic embedded scripts (including SQL, JS, PYTHON, JAVA, XML) to perform analysis and give a report. Because it is a data analysis flow chart composed of dynamic scripts interpreted by the engine, the difficulty, dimension and depth of analysis it can perform far exceeds the analysis method based solely on SQL statements.

[0077] Data quality control module: used to perform rule quality control, condition quality control and logical verification quality control on data. This verifies and checks the correctness of the data and provides guarantee for the accurate display of data.

[0078] The present invention performs quality control before the business system data is entered into the database. The platform calls the quality control engine, executes the associated dynamic embedded quality control scripts (including sql, js, python, java, xml, ktr) to perform analysis and give a report. Because it is a data quality control analysis flow chart composed of dynamic scripts interpreted by the engine, the difficulty, dimension and depth of the analysis it can perform far exceeds the analysis method based solely on sql statements.

[0079] The rule-based quality control, conditional quality control and logical verification quality control involved in the present invention can be dynamically maintained in the quality control rule library and then executed by the quality control engine.

[0080] Step 1: Define quality control points.

[0081] Step 2: Specify the quality control script type (the construction script type of the analysis flow chart of the quality control point, such as: sql, js, python, java, xml, ktr).

[0082] Step 3: Write the quality control analysis script or reference the analysis script path.

[0083] Step 4: After completing the analysis flow chart settings for all quality control points, build a quality control model analysis flow chart for a certain subject.

[0084] Step 5: Load the data, perform quality control analysis and output a report through the quality control engine.

[0085] Example of quality control analysis script:

[0086] 1. Mobile phone number verification:

[0087] function check(var1){ / / non-empty judgment

[0088] if(var1===undefined||var1===null||var1===”){

[0089] return false; / / If the phone number is empty, return false

[0090] }

[0091] / / Regular expression: starts with 1, followed by 9 digits

[0092] var regex= / ^1[3-9]\d{9}$ / ;

[0093] / / Check if the phone number is valid

[0094] return regex.test(var1); / / Return true or false

[0095] }

[0096] 2. Age is not equal to date of admission minus date of birth:

[0097] function check(var1,var2,var3){ / / Convert the date string to a Date object

[0098] var admissionDate=new Date(var2);

[0099] var birthDate=new Date(var3);

[0100] / / Check if the date is valid

[0101] if(isNaN(admissionDate.getTime())||isNaN(birthDate.getTime())){

[0102] return false;

[0103] }

[0104] / / Calculate the difference in years

[0105] var ageInYears=admissionDate.getFullYear()-birthDate.getFullYear();

[0106] / / Adjust if the admission date is before the birth date

[0107] var birthMonthDay=birthDate.getMonth()*100+birthDate.getDate();

[0108] var admissionMonthDay=admissionDate.getMonth()*100+

[0109] admissionDate.getDate();

[0110] if(admissionMonthDay <birthMonthDay){

[0111] ageInYears--;

[0112] }

[0113] / / Check if age is equal to the calculated difference in years

[0114] if(var1!=ageInYears){

[0115] return false;

[0116] }

[0117] return true;

[0118] }

[0119] 3. Professional title verification for medical and health professionals:

[0120] function check(var1){ / / check if empty

[0121] if(var1==null||var1====”){

[0122] return false;

[0123] }

[0124] / / Type conversion

[0125] var str = String(var1);

[0126] / / Array of valid values

[0127] var validValues ​​= [

[0128] '1', / / Chief Physician

[0129] '2', / / Deputy Chief Physician

[0130] '3', / / Attending physician

[0131] '4', / / Doctor

[0132] '5' / / Healer ];

[0134] / / Check if it is within the range

[0135] return validValues.indexOf(str)! ==-1;

[0136] }

[0137] Through the above examples, it can be seen that the quality control module of the present invention relies on the powerful interpreter of its quality control engine to interpret and execute various rule scripts. It can perform direct quality control, multi-level progressive quality control, and logical compliance quality control across business tables. It is more powerful and advanced than traditional analysis based solely on SQL statements.

[0138] Medical data knowledge base module: It is to instill medical industry knowledge into data, organize and maintain knowledge according to data types, and form a knowledge base module. The knowledge base module of data is mainly divided into: indicator knowledge, clause knowledge, medical industry dictionary library, and medical coding library. The basic knowledge framework system is built based on the bottom layer to index knowledge for data application.

[0139] Background data processing module: Data analysis and processing can be carried out in layers and in depth. According to the depth of analysis, indicator data can be divided into atomic data items, calculation data items, and analysis data items.

[0140] Define atomic data items as the smallest particles of application data, obtain analysis model information and value time range information, and add script operation rules to calculate and output the value of atomic data items after extracting the original detailed data; the atomic data items are reconfigured according to the rules of numerator and denominator, and the output results of the business-based fine-tuning parameter rules are defined as calculated data items.

[0141] Calculate data item configuration analysis rules and output analysis data item results.

[0142] The analytical data item is a direct evaluation feedback of a business result in a certain medical field.

[0143] For the generation of the values ​​of the above data items, except for those with high real-time requirements and simple data acquisition, which uses SQL to obtain data from the real-time data center, other data items are obtained by executing the analysis script through the built-in analysis engine. The pre-data analysis method used in the present invention abandons the real-time data pulling of SQL, greatly reduces the operating load on the main data of the business center, and also improves the efficiency of presenting the analysis results. The present invention solves the problem of sharing data items in multiple indicators through the definition of data items, thereby unifying the data caliber and efficiently realizing data consistency and integrity.

[0144] Step 1: Build a list of data items based on the analysis model requirements.

[0145] Step 2: Analyze the list of data items and classify them according to the analysis level of the acquired data (categories include atomic items, calculation items, and analysis items).

[0146] Step 3: Maintain the data retrieval script or data retrieval script reference path for each atomic data item.

[0147] Step 4: Maintain an analysis script or an analysis script reference path for each calculated data item, where the analysis unit can be an atomic data item.

[0148] Step 5: Maintain an analysis script or an analysis script reference path for each analysis data item, where the analysis unit can be an atomic data item or a calculation data item.

[0149] Step 6: Define analysis indicators and maintain the analysis script of the indicator analysis flow chart. The operation units in the script can be atomic data items, calculation data items, and analysis data items.

[0150] Step 7: Build an analysis model and maintain the analysis script of the analysis model analysis flow chart. The operation unit in the script is the indicator.

[0151] Step 8: Call the analysis engine, execute the analysis model analysis script, and output the analysis results.

[0152] like Figure 6 As shown, an example of the indicator analysis flow chart is as follows:

[0153] The data analysis flow chart of the indicator "Number of clinical-related patents obtained per 100 health technicians per year" is displayed through a visual graphic. The chart describes the data flow and processing logic, and finally saves it as a ktr script for the analysis engine to call and parse. This chart is a visual graphic display of the entire data analysis process, including data acquisition definition, data analysis steps, analysis rules, algorithms and analysis result output.

[0154] Data analysis model module: The data analysis model module is an analysis model built based on indicators and clauses constructed based on data. Indicators are composed of data items. Different data items can produce any indicators according to the configuration rules. Different indicators can be customized according to the clause rules to form a new data analysis model.

[0155] In the present invention, whether the indicators are originally obtained or calculated, they are all determined according to the standards issued by the national or provincial competent authorities. There are basically no formulas for atomic data items, and there are corresponding formulas for calculation data items and analysis data items. The analysis model is composed of indicators. If there is a formula for the indicator or the data item in the indicator, there is a formula. All formulas are reflected in the corresponding analysis flow.

[0156] After the data analysis model is built, the data analysis engine can be used to process and analyze the indicator data and output different result values. Configuring different attribute models according to different business scenarios can effectively solve the complex data and diverse business needs of hospitals. It completely solves the single, fixed and unchangeable model in traditional data applications.

[0157] Data analysis engine module: When the data analysis engine module is executed, it calls the drawn kettle script to obtain the configuration indicator information in the analysis model.

[0158] The indicator-related data items are obtained through the indicator information. The indicator-related data items are the atomic data items that have been calculated. Multiple atomic data items are calculated data items according to the calculation rules, and the calculated data items are output as analysis data items according to the rules.

[0159] From atomic data items to calculation data items to analysis data items, three-level analysis results are output, providing data display for the front-end data analysis module.

[0160] Front-end data analysis module: The front-end application is an application display of the data set processed by the back-end data analysis module. The data is split into three levels, with horizontal and vertical drilling down, and multi-dimensional slice and block display. The front-end application is associated with the search engine, and the search engine completes the index query of the data and builds the index process between the data item and the search engine. The index result includes the data rules, definitions, types, clause numbers, data trend results, and detailed results. The front-end data analysis module builds a data query mechanism with other systems to realize cross-system raw data traceability indexing.

[0161] Data deduction module: This module is the embodiment of the deep integration of data and business, and uses the original data as a benchmark to simulate and deduce future business development. Data is processed in layers and divided into atomic data items, calculation data items, and analysis data items. Indicators are composed of multiple atomic data items. Based on this, the indicators composed of data items form a relationship tree structure, in which any data item is a variable and the rest are quantitative. If any variable is modified, it will cause changes in the value of the calculation data item of the related indicator in the entire relationship tree. Specific indicators are the embodiment of actual business. By fine-tuning the value of the indicator data item, the entire model ecosystem will change. By continuously adjusting the data calculation, users are provided with a sandbox deduction tool based on business adjustments.

[0162] The present invention can use the analysis model to import relevant indicator data, then simulate and adjust the value of each indicator data item, and then quickly execute the analysis model to help users adjust decision-making goals and achieve the optimal solution as much as possible through quick conclusions. Here, the platform also has a built-in EDA analysis model to help users perform performance analysis.

[0163] The medical data analysis platform of this embodiment is based on the underlying data-driven indicator application for indexing, analysis, and display. By grading the data, building atomic data items, calculating data items, and analyzing data items, the data is associated with each other, providing a way to expand any application for the indicator application analysis of the data. When the data remains unchanged and the policy business changes, a response mechanism can be quickly provided for the data application by quickly modifying the rule script.

[0164] like Figure 2-Figure 4 As shown, this embodiment also provides a medical data analysis method based on a medical data analysis platform, comprising the following steps:

[0165] S1: Start the data collection system, configure the collection system execution plan, and obtain data from various business systems within the medical institution; determine whether to configure data standard coding for the acquired data, and if not, configure data industry standard coding information;

[0166] Users can customize cycle configuration and status switching operations, and automatically pull data from the intermediate table view at a fixed time, weekly or monthly time period.

[0167] The medical data analysis platform configures industry standard codes and industry dictionary data within the system. Users are supported to customize additions and revisions.

[0168] When the system cannot collect data, it is necessary to support data re-entry operations to ensure data integrity; assign responsibilities to indicators according to departments, and after assigning departments, re-assign the reporters, verifiers, and auditors for the indicators that need to be re-entered. Complete data reporting cycle.

[0169] According to the assigned filler information, data entry operations are performed according to the responsibility and authority. Batch data entry and direct data entry operations are supported. The entered data can be supported by supporting material information to judge the rationality of the data.

[0170] Verify the entered data, and conduct data review on the index data that have passed the verification; return the index data that have failed the verification to the data entry page, revise the data and submit it again;

[0171] The data that has passed the review will be reviewed. The index data that has passed the review will be directly stored in the database. The index data that has failed the review will be returned to the data entry page for revision and resubmission and verification.

[0172] The approved data will be stored in the database, and the stored data will be processed by script calculation.

[0173] S2: Transcode the data encoded by the data standard configured in the previous step and perform data quality control verification;

[0174] The data quality control rules are configured for the transcoded data. If no data quality control rules are configured, data quality control rules need to be configured; then the original data quality control verification is performed on the data with configured data quality control rules;

[0175] Configure the field quality control rule data in the system on the medical data analysis platform. Users can customize additions and revisions.

[0176] S3: The data that have passed the quality control are entered into the database of the medical data analysis platform, and the data after entry are subjected to connotation quality control verification;

[0177] The data collection management system collects and stores data and performs pre-quality control, and batch data with a quality control compliance rate of ≥90% is entered into the medical data analysis platform; before the connotation quality control verification, it is necessary to determine whether the connotation quality control rules are maintained. If the connotation quality control rules are not maintained, the connotation quality control rules need to be maintained. The connotation quality control verification rules are configured on the medical data analysis platform, and the connotation quality control rules are maintained according to the rule relationship between the quality control model, quality control field, and quality control point. Users can customize additions and revisions.

[0178] Configure the connotation quality control execution plan, support regular execution of the plan at a fixed time, weekly or monthly time period, and support the activation and modification of the execution status.

[0179] Draw the script execution engine of the connotation quality control kettle (ETL tool) to call rules for data quality control;

[0180] The system performs connotation quality control verification on the data in the warehouse according to the maintained connotation quality control rules and the drawn connotation quality control engine. After the connotation quality control is completed, the connotation quality control report and the quality control abnormality detailed table data are output. It determines whether the logic, relevance and conditionality of the data in the warehouse are reasonable.

[0181] S4: Perform layered processing and application operations on the stored data and supplementary data;

[0182] S41: Set up basic data and related operations on the medical data analysis platform, specifically:

[0183] Configure basic information such as organization information, indicator information, clause information, model information, etc.; support users to customize additions and modifications based on actual conditions;

[0184] Configure basic role information, user information, account information, API interface configuration, data collection cycle and other basic setting information; support users to customize additions and modifications based on actual conditions;

[0185] Configure the indicator data analysis model. The indicator data analysis model is used by the analysis engine to call the indicator template for index processing; it supports user-defined configuration processing; users can dynamically configure according to business scenarios and execute the model on demand.

[0186] Configure the analysis engine execution rules, support automation support for different cycle types, and support user-defined editing and configuration operations.

[0187] S42: The kettle script is used to process and calculate the data that have passed the connotation quality control verification and the supplementary data. The system periodically executes the engine to perform script calculation processing on the analysis model data according to the configuration rules. Specifically:

[0188] S421: Use the kettle script to process and calculate the original data items and output the atomic data items;

[0189] Atomic data items are the first step in layering the original data. Atomic data item data can be directly called according to the application direction, reducing the efficiency of the front-end application display calling the back-end original data.

[0190] In the present invention, whether the index data items are originally obtained or calculated, they are all determined according to the standards issued by the national or provincial competent authorities.

[0191] Example:

[0192] Data item: Annual number of national invention patents, which can be directly obtained by statistics.

[0193] S422: using the kettle script to output the calculated data items according to the calculation rules for the atomic data items;

[0194] Calculating data items is the second step of layering and reprocessing the original data. The calculated data item data can be directly called according to the application direction, reducing the efficiency of the front-end application display calling the back-end original data.

[0195] In the present invention, whether the index data items are originally obtained or calculated, they are all determined according to the standards issued by the national or provincial competent authorities, and the calculation data items have corresponding formulas.

[0196] Example:

[0197] The number of clinical-related national patents obtained per 100 health technicians per year = the number of national invention patents per year / the number of health technicians per year × 100.

[0198] Indicator: Number of clinical-related national patents obtained per 100 health technicians per year.

[0199] Calculation data item: Number of clinical-related national patents obtained per 100 health technicians per year.

[0200] Atomic data items: annual number of national invention patents, annual number of health technicians.

[0201] S423: using the kettle script to add calculation rules to the calculation data items and output analysis data items;

[0202] Analyzing data items is the third step of layering and reprocessing the original data. The analyzed data items can be directly called according to the application direction, reducing the efficiency of the front-end application display calling the back-end original data.

[0203] In the present invention, whether the index data items are originally obtained or calculated is determined according to the standards issued by the national or provincial competent authorities, and the analysis data items have corresponding formulas.

[0204] Example:

[0205] Medical income from fixed assets of 100 yuan (excluding income from medicines and sanitary materials) = medical income (excluding income from medicines and sanitary materials) / [(fixed assets at the beginning of the period + fixed assets at the end of the period) / 2]×100.

[0206] Indicator: Medical income per 100 yuan of fixed assets (excluding income from medicines and sanitary materials).

[0207] Analytical data items: Medical income per 100 yuan of fixed assets (excluding income from medicines and sanitary materials).

[0208] Calculation data items: Medical income = total medical income - drug income - health materials income; Fixed assets = (fixed assets at the beginning of the period + fixed assets at the end of the period) / 2.

[0209] Atomic data items: total medical income, total medical income, total medical income, beginning fixed assets, and ending fixed assets.

[0210] S5: The processed data is displayed on the front-end page according to the rules of the indicators (supporting annual, monthly and detailed data display, query, drill-down, source data backtracking and other operations), specifically:

[0211] S51: Configure the underlying basic data of the system, specifically:

[0212] Configure basic information such as medical institutions, personnel, roles, accounts, etc.;

[0213] Maintain data items, indicators, and clause information; maintain analysis models, industry dictionary libraries, business dictionary libraries, and standard coding data; complete the underlying parameter data of the system;

[0214] Configure indicator parameter warning rules and 360 access jump links, configure system analysis models, and analysis engine execution rules;

[0215] S52: Click the intelligent analysis module on the platform to perform intelligent analysis on the incoming data;

[0216] Output atomic data items, calculation data items, analysis data items, and output analysis result reports and execution log information. Click on the analysis report to support custom configuration of indicator scoring methods, support expert review input and summarize the detailed summary results of experts based on indicator evaluation, click on the indicator report html format to output the system definition summary results, support abnormal prompts for abnormal indicator score results, and provide a global overview of all abnormal indicators. Load and update data for the entire system.

[0217] S53: Click the indicator overview module to enter the indicator overview summary analysis page;

[0218] The analysis page processes and calculates the atomic data items, calculation data items, and analysis data items of each year's data according to the script; the page supports conditional screening of indicator data sources, supports screening of data collection frequency, and supports fuzzy search queries of indicator names;

[0219] S54: The data analysis platform can perform annual data trend analysis or monthly data trend analysis on the indicator data in a single year according to the indicator attributes;

[0220] S541: Click on the annual data of any indicator to jump to the annual data details, and display the detailed values ​​of the numerator / denominator data items respectively;

[0221] On the indicator overview page, click on any annual calculated data item value to trace back the calculated data item value; click on the value to jump to the page, which displays the business data information corresponding to each data item. According to the indicator attributes, the data header can be dynamically displayed to display the data item related information.

[0222] Data item information can be traced back to doctor information, nurse information, and patient information based on indicators. Customized dynamic filtering condition queries can be performed for the information of the three groups of people, reducing the efficiency problems of fixed filtering condition settings on the dynamic header page, while meeting the needs of any user's query direction; when querying conditions, the system provides raw data SQL query statements to meet users' needs for quick filtering of raw data.

[0223] S542: Click "360 View" in the detailed data operation column, and according to the configured 360 jump link, the original system page of the data can be indexed across platforms to trace back the data;

[0224] S543: Click any indicator name to jump to the indicator analysis page, and drill down to display the indicator by department, year, month, disease type, and ward according to the indicator attributes;

[0225] Click the indicator name on the data overview page to jump to the single indicator analysis and display page; customize the analysis and display of each indicator based on the indicator data, and slice and dice the data according to the indicator's data collection caliber, cycle, department, ward, and disease type; the indicator data analysis dimensions include annual and monthly trend analysis of atomic data items, and annual and monthly trend analysis of calculated data items. Click any annual and monthly area in the trend chart to switch the display list to numerator, denominator and ratio information.

[0226] Click on specific detailed data in the operation column to perform simple query and advanced query on the original detailed data; the page displays the calculated data item values ​​of the indicator in the last year, the average calculated value of the calculated data item in each year, and the median calculated value; click the indicator definition button to display the indicator data definition details, click on department information to slice the indicator by department, and click on data item details to support advanced filtering query of data item details.

[0227] S544: Click the numerator and denominator values ​​in the indicator analysis display page list to display the data details. Click "360 View" in the operation column to index the original system page of the data across platforms and trace back the data.

[0228] Configure indicator data warning rules to support abnormal data abnormality alerts;

[0229] Configure data deduction model parameters, support deduction of future data models, and output deduction reports and results.

[0230] S6: Analysis and deduction of medical data after S5 processing, specifically:

[0231] S61: Perform customized display, query, drill-down and classification processing on the data processed by S5;

[0232] Users analyze and sort out the generated data to seek the basis for operational development issues, and simulate and deduce based on the positioning issues to predict future development trends;

[0233] S62: Click the data deduction module on the medical data analysis platform to enter the function page;

[0234] S621: Select any annual data that has been processed into atomic data items, calculation data items, and analysis data items as the benchmark data;

[0235] Enter the function page and click the simulation button, select the benchmark data year, use any annual data in the historical period as the benchmark data, and select the scoring method (the scoring methods include median scoring, average scoring, and any year scoring);

[0236] S622: Configure and select the indicators to be deduced, and set a calculation data item threshold for each indicator;

[0237] S623: deduce a value for any atomic data item in the selected indicator;

[0238] After setting the threshold, modify the value of any atomic data item under the indicator (each atomic data item value is data generated and calculated by actual business). The indicator data is generated by multiple atomic data items according to the rules. Modifying the value of an atomic data item will cause the internal calculation conditions of the associated indicator data to change, thereby triggering a global change of the associated analysis indicators in the entire analysis model system.

[0239] S624: Start the data deduction model, call the analysis engine to perform deduction calculations, output deduction data results, and output new deduction analysis data item values;

[0240] Set several atomic data items in the deduction indicator data as variables, and the rest as quantitative. Enable the data deduction model; after the parameters are set, execute the analysis engine, output the deduction results, and output the new deduction analysis data item values. Among them, each atomic data item is the summary calculation value of the actual business generated data.

[0241] The above S621 and S622 processes can be executed repeatedly, and the parameters can be fine-tuned continuously until the entire deduction model ecosystem reaches the expected value. The expected ideal value parameters output can be used to make targeted adjustments to the actual business of the hospital. In this way, the pre-deduction operation of future business development management is completed.

[0242] like Figure 5 As shown, the hierarchical calculation process diagram of atomic data items, calculation data items, and analysis data items in this embodiment. By processing the original data in layers and structures, each level of data can be directly called according to the application direction, which greatly reduces the performance problem of the front-end calling the underlying data; the following will explain the process of building a hierarchical data structure through the script operation process steps, specifically:

[0243] Step 1: Get the analysis model ID information;

[0244] Analyze the indicator data information configured in the model;

[0245] Step 2: Obtain the time range information of data calculation;

[0246] Step 3: Clear the historical data information that has been generated;

[0247] Step 31: Obtaining raw data information within the calculation time range;

[0248] Call the original data information collected, entered and stored in the database;

[0249] Step 311: Add script operation rules;

[0250] Add script calculation summary information to the acquired data. The script calculation rules are customized and drawn according to the data type of each indicator. At this time, all indicators are processed and calculated according to the drawn script information.

[0251] Step 312: Execute the operation rules and output statistical results;

[0252] Output the calculation result of each atomic data item according to the operation of the above steps;

[0253] Step 313: the historical data has been cleared in step 3, and now the calculated value needs to be inserted, that is, the operation result value data is inserted;

[0254] Step 32: Output the atomic data item value;

[0255] Step 321: The atomic data item value can directly call the output application, add calculation rules to multiple atomic data items, and configure different calculation rules for each calculation data item according to the indicator application direction;

[0256] Step 322: Output the calculated data item value according to step 321;

[0257] Step 323: configuring a new operation rule for the calculated data item value according to step 322;

[0258] Step 324: Output the analysis data item value according to the operation rules configured in step 323.

[0259] A computer comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any of the above-mentioned medical data analysis methods based on the medical data analysis platform is implemented.

[0260] A storage medium stores a computer program, which, when executed by a processor, implements any of the above-mentioned medical data analysis methods based on a medical data analysis platform.

[0261] The medical data analysis platform of the present invention is based on the underlying data-driven indicator application for indexing, analysis, and display. By grading the data, building atomic data items, calculating data items, and analyzing data items, the data is associated with each other, providing an arbitrary application expansion method for the indicator application analysis of the data. In the case of unchanged data and changes in policy and business, a response mechanism can be quickly provided for the application of the data by quickly modifying the rule script.

[0262] 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. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A medical data analysis platform, characterized by: It includes data collection management module, data governance module, data quality control module, medical data knowledge base module, background data processing module, data analysis model module, data analysis engine module, front-end data analysis module and data deduction module; Data collection management module: Automatically collect data from various decentralized systems in medical institutions. During the data collection process, the data will be split and relevant fields will be extracted according to the metadata of the analysis, and then stored and indexed for subsequent big data analysis. Data governance module: provides various types of front-end plug-ins embedded in the business system that generates data to implement governance before and during data generation. The platform calls the analysis engine, executes the associated dynamic embedded scripts to perform analysis and generate reports; Data quality control module: Quality control is performed before the business system data is stored in the database. The platform calls the quality control engine, executes the associated dynamic embedded quality control scripts to analyze and generate reports, including rule quality control, conditional quality control and logical verification quality control of the data; Medical data knowledge base module: mainly includes indicator knowledge, clause knowledge, medical industry dictionary library, and medical coding library; Backend data processing module: divides indicator data into atomic data items, calculation data items, and analysis data items according to the depth of analysis; Data analysis model module: an analysis model built based on indicators and clauses constructed based on data. Indicators are composed of data items. Different data items are configured according to rules to produce arbitrary indicators. Different indicators are customized according to clause rules to form a new data analysis model. Data analysis engine module: When the data analysis engine module is executed, it calls the drawn kettle script to obtain the configuration indicator information in the analysis model; Front-end data analysis module: The front-end application is an application presentation of the data set processed by the back-end data analysis module; Data deduction module: Use the original data as a benchmark to simulate and deduce future business development. Through the analysis model, import relevant indicator data and simulate the adjustment of the value of each indicator data item. Based on the execution results of the analysis model, help users adjust their decision-making goals.

2. A medical data analysis method based on a medical data analysis platform, characterized in that: The following steps are involved: S1: Start the data collection system, configure the collection system execution plan, and obtain data from various business systems within the medical institution; determine whether to configure data standard coding for the acquired data, and if not, configure data industry standard coding information; S2: Transcode the data encoded by the data standard configured in the previous step and perform data quality control verification; S3: Enter the qualified data into the database of the medical data analysis platform, and perform connotation quality control verification on the data after entering the database; S4: Perform layered processing and application operations on the stored data and supplementary data; S5: The processed data is displayed on the front-end page according to the rules of the indicators; S6: Analysis and deduction of medical data after processing in S5.

3. The medical data analysis method based on the medical data analysis platform according to claim 2, characterized in that: In S4, the stored data and supplementary data are processed and applied in layers, specifically: S41: Set up basic data and related operations on the medical data analysis platform, specifically: Configure organization information, indicator information, clause information and model information; Configure basic role information, user information, account information, API interface configuration and data collection cycle information; Configure the indicator data analysis model. The indicator data analysis model is used by the analysis engine to call the indicator template for index processing. Users can dynamically configure it according to business scenarios and execute the model on demand. Configure the analysis engine execution rules, support automation support for different cycle types, and support user-defined editing and configuration operations; S42: The data that have passed the connotation quality control verification and the supplementary data are processed and calculated using the kettle script.

4. The medical data analysis method based on the medical data analysis platform according to claim 3, characterized in that: In S42, the data that have passed the connotation quality control verification and the supplementary data are processed and calculated using the kettle script, specifically: S421: Use the kettle script to process and calculate the original data items and output the atomic data items; S422: using the kettle script to output the calculated data items according to the calculation rules for the atomic data items; S423: Use the kettle script to add calculation rules to the calculation data items and output the analysis data items.

5. The medical data analysis method based on the medical data analysis platform according to claim 2, characterized in that: In S5, the processed data is displayed on the front-end page according to the rules of the indicators, specifically: S51: configure the underlying basic data of the system; S52: Click the intelligent analysis module on the platform to perform intelligent analysis on the incoming data; S53: Click the indicator overview module to enter the indicator overview summary analysis page; S54: The data analysis platform performs annual data trend analysis or monthly data trend analysis on the indicator data in a single year based on the indicator attributes.

6. The medical data analysis method based on the medical data analysis platform according to claim 5, characterized in that: In S54, the data analysis platform performs annual data trend analysis or monthly data trend analysis for the indicator data in a single year according to the indicator attributes, specifically: S541: Click on the annual data of any indicator to jump to the annual data details, and display the detailed values ​​of the numerator / denominator data items respectively; S542: Click "360 View" in the detailed data operation column, and according to the configured 360 jump link, cross-platform index data original system page to trace back data; S543: Click any indicator name to jump to the indicator analysis page, and drill down to display the indicator by department, year, month, disease type, and ward according to the indicator attributes; S544: Click the numerator and denominator values ​​in the indicator analysis display page list to display the data details. Click "360 View" in the operation column to access the original system page of the cross-platform index data and trace back the data.

7. The medical data analysis method based on the medical data analysis platform according to claim 2, characterized in that: In S6, the medical data processed in S5 is analyzed and deduced, specifically: S61: Customize the display, query, drill-down and classification of the data processed by S5; S62: Click the data deduction module on the medical data analysis platform to enter the function page, specifically: S621: Select any annual indicator data value that has been processed into atomic data items, calculation data items, and analysis data items as the benchmark data; S622: Configure and select the indicators to be deduced, and set a calculation data item threshold for each indicator; S623: deduce a value for any atomic data item in the selected indicator; S624: Start the data deduction model, call the analysis engine to perform deduction calculations, output the deduction data results, and output new deduction analysis data item values.

8. The medical data analysis method based on the medical data analysis platform according to claim 2, characterized in that: Perform the following operations on data that the system cannot collect: Supplemental data recording for data that cannot be collected by the system; Assign responsibilities to the indicators according to departments, and after assigning departments, reassign the reporting, verification, and review personnel for the indicators that need to be supplemented; Perform data entry operations according to the assigned filler information and responsibility authority; Verify the entered data and conduct data review on the index data that have passed the verification; The indicator data that failed the verification will be returned to the data entry page for revision and resubmission; The data that has passed the review will be reviewed. The index data that has passed the review will be directly stored in the database. The index data that has failed the review will be returned to the data entry page for revision and resubmission and verification. The approved data will be stored in the database, and the stored data will be processed by script calculation.

9. A computer comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the medical data analysis method based on the medical data analysis platform as described in any one of claims 2 to 8 is implemented.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the medical data analysis method based on the medical data analysis platform as described in any one of claims 2 to 8 is implemented.

Images