Intel Big Data technology helps improve medical services

Drag the sick body in the hospital floors, the shuttle between departments, registration, outpatient, payment, take test results, take medicine such a series of suffering complicated medical treatment process you must have experienced. Some people lamented "the difficulty of seeing a doctor in a big hospital", but did we choose a big hospital with a reputation for being worried about the level of medical care and the level of testing in small hospitals with small size and simple facilities? Where is the convenient, efficient, high quality and inexpensive service? Can you enjoy the service of a big hospital in a small hospital?

Informatization, especially the regional informatization is a key to solve the problem, imagine that in small hospitals, because of the network and information and large hospitals, remote, intelligent, synergistic application of the use of small hospitals can not solve the large hospital doctors directly involved, it is bound to bring about the revolution in medical services.

How to achieve a good vision? Must face the reality of health informatization, the current island-like information system needs to be integrated, but the biggest problem in the integration is that the increase in geometric multiples of medical information data brings unprecedented pressure.

A CT scan image, 150MB, a genome sequence file, 750MB, standard pathological map, 5GB. If you multiply this data by the number of people and the life expectancy, only one community hospital can accumulate as many terabytes or even petabytes as possible, not to mention the data sets compiled by larger medical institutions and even regional medical authorities.

According to McKinsey, medical data will rise sharply to 35ZB by 2020, equivalent to 99 times times the amount of data in 2009. Facing the development trend of interconnection and the rapid growth of data, how to minimize the cost of storage, processing and protection of massive and complex data, and at the same time it can be handled in real time or quasi real time, query demand response of second level and intelligent, in-depth analysis, facing difficulties.

As the practitioner of regional medical information industry, the Information center of Shanghai Health Bureau began to demonstrate the construction of regional health informatization in 2009. But in the face of more than 20 million permanent residents and more than 16 million medical data per day, the Shanghai Health Bureau Information Center faces greater challenges than ever before.

Opening large data platform brings breakthrough to Shanghai Health Bureau

There are more ways than difficulties, under the pressure of huge data service, Shanghai Health Bureau United Intel and Wanda Information Co., Ltd. launched the "Shanghai Health Information Network Cloud computing" project to improve how to achieve three-level hospital, District Hospital, Community Health Service Center information interconnection, data sharing in large data age, We will further strengthen the integration of health resources under the framework of "1+19" in Shanghai, so as to truly establish a patient-centric information system.

To achieve these objectives, in the "Sea Health Information Network Cloud Computing" project, in addition to the public to provide online booking, online inquiries and other basic health services, the Shanghai Municipal Health Bureau also increased the "Smart prompt" function (refers to the doctor in the course of the diagnosis and treatment of repeated testing, repeated inspection, repeated medication for background supervision, In order to avoid unnecessary waste of medical resources, on the other hand, it reduces the risk of medical quality and medical safety. The system will have a prompt box for these three "repetitions", and the final decision is in the doctor's discretion. The main function of the set is to rely on the market mechanism, management mechanism, game mechanism to monitor the doctor's diagnosis and treatment, drug behavior, to avoid the patients may encounter three major "duplication" problem. "For example, if the repeat rate is 30% for most of the city's level three hospitals, and a third-level hospital examination repeat rate of 80%, then this three-level hospital may have an excessive examination of the problem, through the intelligent prompt function in this situation before the warning doctor, can better solve the problem" Shanghai Health Bureau, deputy director of the Information Center Chevy, "Shanghai about 120,000 work stations in use every day, to do three tips to serve each workstation, the background of the enormous pressure can be imagined, huge data pressure, access pressure and integration pressure are coming." and the storage and computing challenges brought by massive concurrency, applications and access require new technologies to solve. ”

Therefore, with the support of Shanghai Science and Technology Commission and Shanghai Development and Reform Bureau, after more than one year's argument and discussion, the Shanghai Municipal Health board eventually abstracted and stripped the actual needs, and worked with partners to translate the abstract requirements into technical languages and organize them into operable design programs. "In the project discussion, we also considered many open source large data solutions, but after testing, found that the stability is not enough to long-term use," Chevy said, "Until 2012, Intel provides us with large data storage, management, processing and query requirements for the development of Apache The Intel distribution of the Hadoop software and the Intel Xeon solution. Considering that Intel's Open data platform has been successfully applied in a wide range of industries, with its high-performance, energy-efficient features, we have, after trial, opted for the Intel Architecture virtual machine as the primary distributed storage and distributed computing node infrastructure, with Intel's dedicated Apache The Intel distribution of the Hadoop software further perfects the access issue. ”

The Hadoop solution results in significant

"Shanghai Health Information Network Cloud computing" project officially run more than a year to date, cleaning data has reached 4.3 billion, and through large data technology, has achieved data I/O layer 5000 concurrency in a second all corresponding, data query speed also accelerated. The result of the project shows that the situation of repeated inspection and inspection has been reduced and the utilization of medical resources is becoming more and more balanced through automatic intelligent reminders of each link. This remarkable result not only witnessed the high concurrent processing ability of the platform, the fast data retrieval ability also means the practice of the "Shanghai Health Information Network Cloud Computing" project, and it also means that our medical profession has taken a step towards "patient-centered" service-oriented medical system.

At present, the "Shanghai Health Information Network Cloud computing" project has now achieved the second stage, the strong combination of soft and hard has brought new opportunities and breakthroughs for the application of medical data in Shanghai, and promoted the standardization and standardization of the diagnosis and treatment process of the major hospitals in Shanghai. In the next stage, the Shanghai Municipal Bureau of Health will use distributed storage and distributed computing to further excavate the value of large data in the key areas of integrated management, decision making, data mining and data exploration, and will share the same mechanism with private hospitals for interoperability and cooperation. After a wider range of shared exchange and integration, even a doctor at a Community Health Service center will monitor his behavior when he is in the hospital and prescribe medicine. Only in this way can the real use of information technology to better mine data value, improve the hospital process, improve the fairness and effectiveness of diagnosis and treatment services, will gradually alleviate the "expensive, difficult to see a doctor" livelihood issues.