In the 19th century, British epidemiologist and anesthesiologist John Snow used early modern data science to record the number of deaths and injuries each day and put the addresses of those who died on a map to create a "cluster" map of the cholera outbreak in London, which had been widely believed to have been caused by noxious air. By aggregating the data from his investigations, Snow identified contaminated public wells as the culprits of the cholera outbreak and laid the foundation for the germ theory of the disease, which was one of the early uses of Big Data.
Snow probably would not have thought that nearly two hundred years later, the use of big data is no longer accidental, with the rapid development of health care information technology, its penetration through the combination of AI in biomedical research and development, disease management, public **** sanitation, and health management has been gradually normalized, but the problem has also been highlighted accordingly.
Information silos still exist
In the past two years, policies on medical big health data have emerged frequently, with relevant guidance from top-level design, specific planning guidance, data privacy and security, data management, and many other aspects.
In June 2016, the General Office of the State Council issued the "Guiding Opinions on Promoting and Standardizing the Development of Healthcare Big Data Applications", which pointed out that various types of medical and healthcare institutions are encouraged to promote the collection and storage of healthcare big data, strengthen the support of application and operation and maintenance technology, open up the channel of data resource **** enjoyment, and accelerate the construction and improvement of basic databases centered on residents' e-health files, e-morbidity, e-prescription and so on. The basic database centered on residents' electronic health records, electronic medical records, electronic prescriptions, and so on.
In September 2018, the National Health Commission issued the Measures for the Management of National Healthcare Big Data Standards, Security and Services (for Trial Implementation) to regulate the medical and healthcare big data industry from the perspective of standardized management and development and utilization. The Measures provide guidance on four aspects: medical big data standards, medical big data security, medical big data services, and medical big data supervision, which directly addresses the current pain points in the field of medical big data, and is of great significance in the future for the integrated standard management of data, the implementation of security responsibilities, and the standardization of data services and management.
However, even with the support of special policies, which are limited to the macro level, the laws and regulations in the field of healthcare big data are still lagging behind, and the lack of comprehensive, detailed, and clear guidelines and rules severely restricts its development compared to other mature fields. Although at this stage, there are many enterprises in the field of medical big data layout, but subject to market access and industrial policy uncertainty, is still groping the stones to cross the river, the market enthusiasm and vitality has not been fully and effectively released.
Liu Lei, a professor at the Institute of Biomedical Research, Shanghai Medical College of Fudan University, believes that it is the uncertainty of the medical big data policy, the standard is not uniform, but also directly leads to the difficulty of data exchange and information **** between the various systems, resulting in a large number of "information islands". To give a simple example, patients in A hospital film to the B hospital is not recognized, B hospital doctors want to understand the patient's information needs to start from scratch, the patient was in A hospital to do the examination needs to be in the B hospital to come back to another round, "want to open up the clinical big data resources between the medical institutions *** enjoyment of the channel, at least at this stage is a very difficult thing. " Liu Lei said.
Similar troubles also occurred in the United States more than 10,000 kilometers away, Philip Paynes, director of the Institute for Information Studies at the University of Washington School of Medicine, said in an interview with the medical valley: clinical big data between each other "isolation" to the national health care institutions, patients and hospitals have brought a burden to achieve big data. The realization of big data interoperability and interoperability is an issue that is being addressed worldwide.
As renowned researchers from two top universities, Liu Lei and Paynes want to do something about clinical big data.
The idea of the two men*** quickly gained strong support at the university level, and the first workshop on Applied Clinical Informatics and Data Analytics, jointly taught by Fudan University School of Medicine and Washington University School of Medicine in St. Louis, was held on July 26-29, 2019, with the first session.
Liu Lei, a professor at the Institute of Biomedical Research at Fudan University and director of the Institute of Medical Information and Intelligent Diagnosis of Medical Imaging at the Institute of Big Data at Fudan University, gave the lecture
According to Liu Lei, the workshop has received a positive response from the industry, with one-third of the first class of students coming from hospitals, healthcare companies, and colleges and universities, and it's a pure attempt to bring the knowledge of clinical big data analytics into the classroom. I want to bring together the industry players who are interested in clinical big data analysis and, through *** there efforts, can push the effective use of clinical big data even further."
Lecture by Philip Paynes, Director of the Informatics Institute at Washington University School of Medicine in St. Louis
"Hopefully, through this kind of internationalized collaboration, there will be one more possibility for clinical big data to really 'run' across healthcare organizations and even across borders. ." Paynes said.
Explorations
And both Liu Lei and Paynes have done a lot of work at their respective research organizations before this was possible.
It is reported that Liu Lei's biomedical research institute at Shanghai Medical College of Fudan University, which is committed to creating "China's first, world-class biomedical cross-disciplinary academic research institution," has formed a "molecular and cellular biology of metabolism and tumors," "medical cell biology," "medical cell biology," "medical cell biology," "medical cell biology," and "medical cell biology. We have already formed three advantageous directions in biomedical interdisciplinary fields: "molecular cell biology of metabolism and tumor", "medical epigenetics", and "systemic biomedicine", and are now working on expanding translational and precision medicine research, including geriatrics, oncology and cardiovascular diseases, and birth defects, We are also working to expand translational medicine research and precision medicine research, including geriatrics, oncology and cardiovascular diseases, birth defects, target structure and active small molecules, histology and big data, and biotherapeutics and interventions, which will lead to new cross-disciplinary growth and downstream technologies.
It is also noted that, at present, the Shanghai Medical College of Fudan University biomedical research is also applying for a supercomputing center construction project to support the research of biological big data, "Fudan University has 17 affiliated teaching hospitals, including Sun Yat-sen Hospital, Huashan Hospital, Renji Hospital and so on, which has some hospitals are also doing their own clinical big data center, from the Institute level, we hope to provide them with some strong support for talent training and technical research." Liu Lei said.
The Institute for Information Studies at the University of Washington School of Medicine, where Paynes is based, is at the center of all of the University of Washington's big data initiatives, "We have the best genome institute and the most productive and influential basic science research enterprise in the world," and is very strong in medical information technology, but in big data integration has yet to be strengthened." And that became the focus of Paynes' work after he became the first director of the Institute for Information Studies at the University of Washington School of Medicine.
Since Paynes took office, he has first linked the Institute with its 15 affiliated teaching hospitals, paving a full chain of clinical big data from collection to integration to mining and finally to application.
In Paynes' view: the Institute's 15 teaching hospitals are simply a treasure trove of big data sources, and these 15 hospitals ranked among the top medical institutions in the United States generate a large amount of clinical data every day, relying on the retrospective study of these existing clinical data is one of the most basic and important research methods to analyze and study diseases, and through the statistical analysis of these massive clinical data, the results of the analysis will in turn be used to analyze and analyze the clinical data. Through the statistical analysis of these massive clinical data, the results of the analysis will, in turn, provide doctors with information on the overall situation of the onset and treatment of the disease ****heeded during the entire process of clinical diagnosis and treatment, and help doctors make scientific decisions to achieve precision medicine.
"Our dream is not just to use clinical big data to help patients, but to have these clinical big data penetrate into their lives and work, and even leisure and entertainment, and to minimize the probability of their illnesses through the analysis of big data, so that people can always maintain a healthy state." Paynes looks forward to the Medical Valley.
Future Development Ideas
In the large amount of clinical data integration work done by Liu Lei, Paynes and his team, due to their own strong teaching hospitals, the source of data is no longer a problem, however, there are two more realistic problems in front of them, the first is to solve the problem of multi-modal clinical data selection. Selection Issues. Clinical big data comes from a variety of sources and is a kind of multimodal data, which includes well-structured data, such as laboratory tests and prescriptions; some semi-structured data, such as hospitalization summaries and discharge summaries; completely unstructured data, such as medical images; there are also genomics data like gene sequencing; as well as time-series data, such as in the ICU, you will see patients inserting a variety of instruments to measure the blood pressure, heart rate, pulse, and other kinds of data. blood pressure, heart rate, pulse rate, and other streaming data.
How to select the needed data from these different modal data, Liu Lei said they, they need more structured clinical data, in order to get this part of the data, will be through a certain technology platform will be certain data cleaning, from which to select high-quality effective data.
After this problem is solved, there is a clinical big data has been a controversy that can not be bypassed - security and privacy issues.
In this regard, Liu Lei said, relying on the existing technology, the current collection of clinical big data can basically do "no hospital", which to a certain extent to ensure the security of the data. Paynes also pointed out that the United States has a very strict protection of medical big data regulations, the patient's key private data, such as name, address, phone number, telephone number, and so on. Such as name, address, phone number, ID number, etc. into the data management must be mosaic, at the same time, strong encryption of data, data even if leaked is not decryptable, all the data access (who what time can access what) to have a set of strict access control, through this way to ensure data security.
When technical issues are no longer a problem, this means that the combination of clinical big data and AI will become even more seamless, so Liu Lei and Paynes hope that regulators will do more to assess the effectiveness and safety of AI based on big data training in the future, which means that the approval of access should be strict, and that the public's knowledge of medical AI should be strengthened, no matter how advanced the AI is. AI development to how advanced degree, there are always certain limitations, it can never replace the doctor, can only be a doctor's auxiliary diagnostic tool.
Although there is still a way to go, but for clinical big data and AI with, Liu Lei and Paynes are full of confidence, at least in their existing work in the planning, "Applied Clinical Informatics and Data Analysis Workshop" can be gradually developed into a master's degree training program for clinical big data and artificial intelligence to train more professionals. Artificial Intelligence. At the same time, based on the work carried out by the two research institutes at this stage, someday the convergence and unification of cross-border can be realized, and all the clinical big data can be unified in the same model, to establish a consortium similar to the Union data, which will become easy for the integration and application of data.
Where this website indicates that the source is not a big health Pai works, are reproduced from other media, the purpose is to convey more information, and does not mean that this website agrees with its views and is responsible for its authenticity.