After more than 20 years of digitizing medical information, China already has a huge medical data base. According to IDC, the global medical data volume has reached 153 EB in 2013, exceeded 600 EB in 2017, and is expected to reach 2.314 PB by 2020.
It is precisely because of the deep accumulation of medical big data that in recent years, after the rise of AI technologies such as deep learning, many startups have tried to use natural language understanding, image recognition and other technologies to clean clinical data. In the process, it was found that the quantity of medical data is enough, but the quality of medical data is worrying.
At the 2018 Shenzhen International BT Leaders Summit not long ago, Hu Shengshou, President of Fuwai Hospital, Chinese Academy of Medical Sciences, mentioned in his keynote speech that if the data pass rate of medical institutions at this stage can reach 50% to 60%, it is quite good. .
At this stage, most artificial intelligence companies still need to invite a large number of experienced doctors and experts to perform manual annotation and machine assistance in the cleaning process of medical data, in order to model.
How on earth can you generate high-quality medical data from the start? Recently, at the 80th China International Medical Equipment Expo, Lv Xudong, professor and doctoral supervisor of the School of Biomedical Engineering and Instrument Science of Zhejiang University, Zhao Dongsheng, researcher of the Academy of Military Medical Sciences, and Zhang Siyu, general manager of the medical product line of Shenzhen Zhongxing Netcom Technology Co., Ltd., etc. People, delivered a keynote speech on "How to use the open international standard openEHR to build high-quality and standardized medical big data".
The "wave" set off by medical big data
Whether it is a country, an enterprise or a university, in recent years, there has been a great "reaction" to medical big data.
In 2015, the State Council issued the "Outline of Action for Promoting the Development of Big Data", which clarified the general requirements for data use. At the end of June 2016, the State Council issued the "Guiding Opinions on Promoting and Regulating the Application and Development of Health and Medical Big Data", which formally incorporated medical big data into national development, and opened up the integration and sharing of medical big data. , medical insurance and other aspects of the application, as well as the use of security and other aspects of a comprehensive standard. In 2017, national key enterprises led the establishment of three health big data companies: China Health Care Big Data Industry Development Group Corporation, China Health Care Big Data Technology Development Group Corporation, and China Health Care Big Data Co., Ltd.
On the enterprise side, with the continuous expansion of the "Al+Medical" cake, the importance of medical big data is also becoming increasingly prominent. Whether it is pharmaceutical companies, medical device manufacturers, life science companies and other parties, they all want to take a share of the pie. . The market size of medical big data is also expanding. According to McKinsey's forecast, the market size of medical big data in the United States is 300 billion to 450 billion US dollars per year, and China also has a market size of hundreds of billions of yuan in the field of medical big data. In this regard, investors also smell business opportunities. The report released by Zhiyan Consulting also shows that in the first quarter of 2018, there were 35 investments in the field of medical and health big data, accounting for 22.2% of the big health field.
In terms of colleges and universities, the combination of production and research has always been strongly advocated by the state. In August this year, with the approval of the China Society of Health Information and Health and Medical Big Data, Xiamen University established the "Xiamen University National Institute of Health and Medical Big Data". In October, Wuhan University announced the establishment of the "Wuhan University National Institute of Health and Medical Big Data" to promote and standardize the development of health and medical big data applications.
Using openEHR to build high-quality medical big data
The application fields of medical big data can be described as wide, including intelligent auxiliary diagnosis, new drug research and development, etc. However, many companies have found that the low quality of medical data has become a "stumbling block" in the process of "fastening the horse". Taking clinical medical data as an example, the main reasons for low quality are:
First, when doctors use the clinical data collection system, the standards for writing medical records are inconsistent and incomplete. Especially in the top three hospitals, doctors have a large daily workload, and it is easy to fill in electronic medical records hastily.
Second, in the data processing of electronic medical records in hospitals, although the medical industry has a high degree of informatization, the degree of dataization is very low. Most hospitals have achieved full coverage of the HIS system, and a lot of patient data can be collected through the HIS system. However, due to the unclear underlying logic of patient information, most of this type of patient data is unstructured document data, which cannot be directly analyzed and applied.
Third, in the data quality control analysis link, the quality control team was not serious enough to verify the data. This makes it easy for junk data to pass the audit and enter into medical big data.
At the meeting, Lv Xudong of Zhejiang University proposed to use openEHR to create high-quality medical data from the source. But at present, most people are relatively unfamiliar with openEHR.
It is shown that openEHR is an open electronic health record specification proposed by the international openEHR organization in 1999. The core of the openEHR specification is to separate the medical domain knowledge from the specific clinical information, and establish a two-layer model - a reference model and a prototype model. The reference model models the stable and unchanging concepts in the information system, and defines the basic data types and data structures required for information expression. The prototype model includes prototypes and templates. The prototype defines clinical content and expresses domain knowledge by adding constraints to the reference model. The template meets practical application requirements by constraining and customizing the prototype.
The open medical data platform driven by the openEHR model can solve the problem that the data requirements of different roles change rapidly, but the response and modification of each business system are slow. In addition, it can also solve the problem of the continuous increase of various medical business systems and the continuous growth of data sources but the inability to integrate them in a timely and effective manner, resulting in data silos.
In fact, OpenEHR has been widely popularized in Europe, Australia, Japan and other countries and regions, and was accepted by the International Standards Organization in 2008 and developed into the ISO 13606-2 standard. So far, the national electronic health record data center in many European countries has adopted this standard, and Japan's newly launched national electronic health record data center project in 2015 also plans to adopt the standard.
Industrial Router,Industrial 3G Router,Industrial Cellular Router,Industrial Router Din Rail
Shenzhen MovingComm Technology Co., Ltd. , https://www.movingcommiot.com