Big data platform
Product Description
In order to realize the intelligent management, sharing, transformation and utilization of samples, we will establish a big data platform with the characteristics of a third-party sample library, and then realize: based on the advanced data acquisition module, connect with the HIS, LIS, PACS or EMR system of the source of samples, and form an accurate, efficient and timely standard data acquisition tool; According to the construction requirements of the sample database and the relevant information of the sample database submitted by different sources, and in combination with the characteristics of the collected samples, determine the fields to obtain the required data from different systems of the source to establish a comprehensive big data system; Combined with the clinical information provided by the sample source and the bioinformatics data provided by the sample bank and the bioinformatics data provided by the multi-organization center, and the biological sample information in the big data platform functional modules (data source management system, biological sample resource management system, clinical research management system, follow-up system, intelligent medical data system, disease resource management system, etc.), Form a triad of hospital clinical data, sample information and bioinformatics data (see Figure 10 for details).
Finally, based on the big data platform under the structure of the sample information storage and processing platform of the information center, a standardized and easy-to-operate data access interface will be established to provide available scientific research data for scientific researchers, technicians, managers, medical staff and other users, which will be applied to data pathology, scientific research transformation, clinical medicine, drug research and development, precision medicine and other fields.
(1) Build a basic information system to realize the complete preservation of sample information
The basic information system of the big data platform is mainly composed of seven subsystems, including data element management system, biological sample resource management system, clinical research management system, follow-up system, disease resource management system, intelligent medical data system, and big data scientific research analysis system, which can realize the link record of the whole process from the sample leaving the body to the completion of warehousing and outbound; Management of the whole process/life cycle of samples, complete quality control, whole-process monitoring and paperless operation; The smooth connection with the hospital health care big data system; Meet the needs of cohort research projects carried out in cooperation with medical institutions in the later stage, support the unified management of all follow-up projects in the sample database, and be able to define and implement different follow-up plans according to the follow-up needs of different disciplines.
Finally, based on the big data platform under the structure of the sample information storage and processing platform of the information center, a standardized and easy-to-operate data access interface will be established to provide available scientific research data for scientific researchers, technicians, managers, medical staff and other users, which will be applied to data pathology, scientific research transformation, clinical medicine, drug research and development, precision medicine and other fields.
(1) Build a basic information system to realize the complete preservation of sample information
The basic information system of the big data platform is mainly composed of seven subsystems, including data element management system, biological sample resource management system, clinical research management system, follow-up system, disease resource management system, intelligent medical data system, and big data scientific research analysis system, which can realize the link record of the whole process from the sample leaving the body to the completion of warehousing and outbound; Management of the whole process/life cycle of samples, complete quality control, whole-process monitoring and paperless operation; The smooth connection with the hospital health care big data system; Meet the needs of cohort research projects carried out in cooperation with medical institutions in the later stage, support the unified management of all follow-up projects in the sample database, and be able to define and implement different follow-up plans according to the follow-up needs of different disciplines.
Figure 10 Overall architecture of sample database informatization construction
① Data element management system
The establishment of the data element management system will help to establish the medical semantic network of the sample database and realize the medical semantic retrieval; Realize the mining of biomedical knowledge; Establishing the subject gateway of medical quality control; Obtain medical core metadata set; It can integrate heterogeneous sample information resources and provide distributed sample information resource sharing.
The data element management system will use the internationally recognized data element standard specification ISO11179 to standardize the data items, and refer to the health information data element model, attributes, naming, definition, classification of health information data elements, and the health information data element content standard of the Ministry of Health WST 303-306.
The data element management system mainly includes but is not limited to the following functional modules: data element model construction, data element management, data group management, data set management, threshold management, etc.
② Biological sample resource management system
The biological sample resource management system will be able to realize the management of the whole process/life cycle of the sample, meet the requirements of recording the whole process from the sample leaving the human body to the completion of warehousing and outbound, and have a special report module to support it.
The system mainly includes but is not limited to the following functional modules: platform management, project research management, system management, storage space management, research object management, warehousing management, outbound management, quality control management, data query management, statistics management, equipment management, reagent consumption management, software and hardware interface management, data security management, etc.
③ Clinical research management system
The establishment of the clinical research management system can enable researchers from different institutions to conduct online data entry and collection through the standards formulated by the unified "data standard management system", and understand the project progress in real time.
The clinical research management system will comply with the Good Clinical Practice for Drugs (GCP), and meet the researchers' self-defined research plan, such as cohort study, drug clinical trial, etc. The system supports multi-project and multi-center platform management and can be deployed on the Internet.
The clinical research system mainly includes but is not limited to the following functional modules: multi-center project management, organization management, e-CRF form management, object recruitment management, external data access, data acquisition function, data quality control, data export, system management, etc.
④ Follow-up system
The establishment of follow-up system will be able to provide necessary support for the research project of large natural population cohort and disease-specific cohort in the sample bank.
The follow-up system will support the unified management of all follow-up items in the sample database, define and implement different follow-up plans according to the follow-up needs of different disciplines, and provide the project PI with the management and progress view function of follow-up items within the scope of authority.
The follow-up system mainly includes but is not limited to the following functional modules: follow-up scheme editing, follow-up template editing, follow-up personnel management, follow-up population management, multi-terminal and multi-form follow-up, external data access, data verification management, data export, etc.
⑤ Disease resource management system
The disease resource management system is an application system based on the scientific research data center and based on the underlying model of "multidimensional weight confirmation".
The system will realize the sharing and utilization of heterogeneous and non-heterogeneous data in various business systems, and establish complete full-course patient information for each disease database to provide detailed data support for the viewing and use of legally authorized users within their rights and responsibilities.
The disease resource management system mainly includes, but is not limited to, the following functional modules: disease resource overview, full-course patient information (including case dimension, data variable dimension, etc.), data quality, data ownership and exchange, etc.
⑥ Intelligent medical data system
The intelligent medical data system will collect and gather multi-source and heterogeneous medical data of clinical samples from different hospitals, use big data technology to manage the data, realize the standardization and unification of the data, and provide data services for the business system on the basis of sufficient data mining analysis.
Intelligent medical data system mainly includes but is not limited to the following functional modules: standardized medical terminology system, standardized terminology ontology library (including data extraction processing, natural language processing, medical data standardization processing, medical document standardization processing, medical data fusion processing, etc.), data quality control management system (including data quality detection and analysis, data processing monitoring platform, etc.), Medical data governance system, data resource management and display platform, etc.
⑦ Big data scientific research and analysis system
The big data scientific research and analysis system has six major applications, including population analysis, research modeling, data export, data quality control, data mining and patient case view, and can provide big data scientific research specialist services for researchers in different fields, including: research scheme design services, special disease database construction services, statistical analysis services, and academic achievement services.
The big data scientific research analysis system mainly includes but is not limited to the following functional modules: research population screening, research population management, research population characteristic analysis, research population data quality control, research population statistical analysis, data export, authority control, etc.
(2) Improve the basic information system and realize the standardization and unification of data
Based on the established basic information system, establish the medical semantic network of sample database, and realize the integration of heterogeneous sample information resources of different medical institutions; Realize the multi-project and multi-center platform-based management of the sample library, and realize the online data entry, collection, analysis and use of the users of each center through the standards specified by the unified data standard management system; Use big data technology to manage the collected medical data with diverse and heterogeneous data, realize the standardization and unification of data, and fully mine and analyze the data.
(3) Develop big data mining analysis methods and platforms
On the basis of the basic information system built in the early stage, the efficient integration and analysis of sample information, clinical data and scientific data can be realized through the development of big data mining analysis methods and platforms: with the help of big data processing, natural language segmentation, machine learning, knowledge mapping and other technologies, the integration and mining includes electronic medical records, laboratory examination, medical imaging, genomics, transcriptomics Massive medical and scientific research data including proteomics and metabolomics form a complete time series research resource library centered on samples, and use data mining algorithms to achieve in-depth analysis and visualization of clinical data and scientific research data, and assist scientific researchers in carrying out scientific research ideas.
The establishment of the data element management system will help to establish the medical semantic network of the sample database and realize the medical semantic retrieval; Realize the mining of biomedical knowledge; Establishing the subject gateway of medical quality control; Obtain medical core metadata set; It can integrate heterogeneous sample information resources and provide distributed sample information resource sharing.
The data element management system will use the internationally recognized data element standard specification ISO11179 to standardize the data items, and refer to the health information data element model, attributes, naming, definition, classification of health information data elements, and the health information data element content standard of the Ministry of Health WST 303-306.
The data element management system mainly includes but is not limited to the following functional modules: data element model construction, data element management, data group management, data set management, threshold management, etc.
② Biological sample resource management system
The biological sample resource management system will be able to realize the management of the whole process/life cycle of the sample, meet the requirements of recording the whole process from the sample leaving the human body to the completion of warehousing and outbound, and have a special report module to support it.
The system mainly includes but is not limited to the following functional modules: platform management, project research management, system management, storage space management, research object management, warehousing management, outbound management, quality control management, data query management, statistics management, equipment management, reagent consumption management, software and hardware interface management, data security management, etc.
③ Clinical research management system
The establishment of the clinical research management system can enable researchers from different institutions to conduct online data entry and collection through the standards formulated by the unified "data standard management system", and understand the project progress in real time.
The clinical research management system will comply with the Good Clinical Practice for Drugs (GCP), and meet the researchers' self-defined research plan, such as cohort study, drug clinical trial, etc. The system supports multi-project and multi-center platform management and can be deployed on the Internet.
The clinical research system mainly includes but is not limited to the following functional modules: multi-center project management, organization management, e-CRF form management, object recruitment management, external data access, data acquisition function, data quality control, data export, system management, etc.
④ Follow-up system
The establishment of follow-up system will be able to provide necessary support for the research project of large natural population cohort and disease-specific cohort in the sample bank.
The follow-up system will support the unified management of all follow-up items in the sample database, define and implement different follow-up plans according to the follow-up needs of different disciplines, and provide the project PI with the management and progress view function of follow-up items within the scope of authority.
The follow-up system mainly includes but is not limited to the following functional modules: follow-up scheme editing, follow-up template editing, follow-up personnel management, follow-up population management, multi-terminal and multi-form follow-up, external data access, data verification management, data export, etc.
⑤ Disease resource management system
The disease resource management system is an application system based on the scientific research data center and based on the underlying model of "multidimensional weight confirmation".
The system will realize the sharing and utilization of heterogeneous and non-heterogeneous data in various business systems, and establish complete full-course patient information for each disease database to provide detailed data support for the viewing and use of legally authorized users within their rights and responsibilities.
The disease resource management system mainly includes, but is not limited to, the following functional modules: disease resource overview, full-course patient information (including case dimension, data variable dimension, etc.), data quality, data ownership and exchange, etc.
⑥ Intelligent medical data system
The intelligent medical data system will collect and gather multi-source and heterogeneous medical data of clinical samples from different hospitals, use big data technology to manage the data, realize the standardization and unification of the data, and provide data services for the business system on the basis of sufficient data mining analysis.
Intelligent medical data system mainly includes but is not limited to the following functional modules: standardized medical terminology system, standardized terminology ontology library (including data extraction processing, natural language processing, medical data standardization processing, medical document standardization processing, medical data fusion processing, etc.), data quality control management system (including data quality detection and analysis, data processing monitoring platform, etc.), Medical data governance system, data resource management and display platform, etc.
⑦ Big data scientific research and analysis system
The big data scientific research and analysis system has six major applications, including population analysis, research modeling, data export, data quality control, data mining and patient case view, and can provide big data scientific research specialist services for researchers in different fields, including: research scheme design services, special disease database construction services, statistical analysis services, and academic achievement services.
The big data scientific research analysis system mainly includes but is not limited to the following functional modules: research population screening, research population management, research population characteristic analysis, research population data quality control, research population statistical analysis, data export, authority control, etc.
(2) Improve the basic information system and realize the standardization and unification of data
Based on the established basic information system, establish the medical semantic network of sample database, and realize the integration of heterogeneous sample information resources of different medical institutions; Realize the multi-project and multi-center platform-based management of the sample library, and realize the online data entry, collection, analysis and use of the users of each center through the standards specified by the unified data standard management system; Use big data technology to manage the collected medical data with diverse and heterogeneous data, realize the standardization and unification of data, and fully mine and analyze the data.
(3) Develop big data mining analysis methods and platforms
On the basis of the basic information system built in the early stage, the efficient integration and analysis of sample information, clinical data and scientific data can be realized through the development of big data mining analysis methods and platforms: with the help of big data processing, natural language segmentation, machine learning, knowledge mapping and other technologies, the integration and mining includes electronic medical records, laboratory examination, medical imaging, genomics, transcriptomics Massive medical and scientific research data including proteomics and metabolomics form a complete time series research resource library centered on samples, and use data mining algorithms to achieve in-depth analysis and visualization of clinical data and scientific research data, and assist scientific researchers in carrying out scientific research ideas.
