The industry development of medical AI big data has been progressing rapidly in recent years.
In terms of technology, data explosion, algorithm upgrading, and computing power improvement continuously ignite the momentum of industry development; At the policy level, the “new infrastructure” policy first proposed by the country as early as 2018, and the “Opinions on Building a More Perfect System and Mechanism for Marketable Allocation of Elements” proposed in April this year, all of which directly “name” and strengthen the construction of emerging technologies such as big data and artificial intelligence; In terms of capital, the integration of healthcare with AI and big data has become one of the hottest tracks in the capital market.
Before 2020, the market has given enterprises sufficient time to grow; Now, it is time for enterprises to “blossom and bear fruit.”.
Own “medical gene” and occupy “home advantage”
As one of the first batch of domestic medical AI big data companies to enter the market, Haisen Health, which occupies the “home advantage” of data resources, has been attracting considerable attention from the industry.
Haisen Health is a member enterprise of Jiahe Meikang Group, one of its predecessors being the medical big data business unit of Beijing Jiahe Meikang Information Technology Co., Ltd. (hereinafter referred to as Jiahe Meikang) in 2016. In 2019, Haisen Health, which had been cultivated for three years, was independently established and listed. At the end of the same year, it quickly completed a round A financing of 200 million yuan.https://www.slw-ele.com/
As is well known, Jiahe Meikang is the largest supplier of electronic medical record systems in China. According to IDC China data in 2020, Jiahe Meikang’s market share of electronic medical records ranked first in the country for the sixth consecutive year, with products covering more than 1300 hospitals in the country. Heisen Health inherits the unique data advantages, technical accumulation, and mature and stable practical experience of Jiahe Meikang. It can be described as carrying its own “medical gene”, and is truly born with a “golden spoon” in it.
At the beginning of its establishment, Haisen Health has set a clear goal: not to make exploratory applications, with the goal of creating high-value medical AI big data applications, to carry out in-depth cooperation with domestic head medical institutions with top comprehensive or specialized strength and leading medical informatization level, to exert its own industry advantages and technical strength, and combine the advantages and strengths of top medical institution partners, in order to polish a comprehensive, comprehensive Full ecological medical AI big data application with independent intellectual property rights.
Up to now, Haisen Health has been highly recognized by a large number of top hospital customers, with cooperative customers covering dozens of provinces and more than 100 Class III hospitals across the country. The top 100 hospitals in Fudan University have a market share of over 20%, including Beijing Union Medical College Hospital, General Hospital of the People’s Liberation Army, Xijing Hospital, Peking University Third Hospital Top medical institutions such as the First Affiliated Hospital of Zhejiang University Medical School, which have absolute advantages in comprehensive or specialized fields.
Consolidate the underlying core technology and innovate the three core achievements of research and development
In recent years, a large number of enterprises have flocked to the medical AI big data circuit, which has ignited the industry. At the same time, it has also caused the industry to fall into an awkward dilemma of blindly pursuing concepts. Few enterprises have calmed down to create underlying technologies. In fact, relying excessively on original medical record data and being subject to unstructured data, resulting in low recall and coarse search granularity in relevant search and retrieval engines, is a common problem faced by medical AI big data manufacturers in the market.
“The underlying core technology determines the value of the application product. If medical AI products want to achieve substantive breakthroughs, consolidating the underlying technology is the only way to break the industry’s dilemma,” said Hesen Health. Based on this sober understanding, Haisen Health is not only not limited to the fickle market environment, but also has “plunged” into the field of medical AI big data, making little appearance outside.
After four years of polishing, Haisen Health has come to the forefront with three major research achievements: an independently developed intelligent medical decision-making engine, a real-time clinical big data search engine, and an AI-based medical big data governance platform.
According to the special attributes of medical care, Hesen Health has innovatively proposed an intelligent decision-making model based on deep learning, and a “dual engine” intelligent medical decision-making architecture based on rule precipitation based on expert experience. This not only relies on the information from the deep learning model and authoritative knowledge base, but also enables timely iteration of models and rules based on medical big data, fundamentally ensuring the continuous evolution of medical AI decision-making power.
Unlike the underlying logic of Keyword commonly used in the industry, Haisen Health, based on its deep understanding of the medical industry, pioneered the choice of a structured full electronic medical record data model based on natural language processing technology, and terminology normalization technology based on knowledge maps to build a real-time clinical big data search engine. The engine not only enables efficient storage of ultra-fine grained underlying diagnosis and treatment data, but also supports tens of thousands of structured and semi structured dimensions, maximizing the retention and restoration of real medical record information, providing stronger support for innovative medical applications.
The Haisen Healthcare Big Data Governance Platform can be described as a “magic weapon” for data governance. Based on the business data of fully electronic medical records, the platform can quickly access arbitrary medical data, perform automated cleaning, processing, transformation, entity and relationship extraction, terminology normalization, patient master index, and obtain standardized data after governance. It has obvious advantages such as less labor involved, high processing efficiency, high data quality, and high accuracy.
Focusing on clinical and scientific research scenarios, create a closed loop of ecological medical AI products
On the basis of the three core achievements mentioned above, Haisen Health has conducted in-depth cooperation with domestic head medical institutions with top comprehensive or specialized strength and leading medical informatization level. By leveraging its own industry advantages and technical strength, and combining the advantages and strengths of top medical institution partners, Haisen Health has carefully polished and laid out a low-key layout with the intelligent medical data center as the core, covering clinical decision-making, scientific research support, medical record quality control The full ecological medical AI closed-loop for different application scenarios such as patient interaction services forms a full ecological product application matrix including a clinical decision support system (CDSS), a big data scientific research and analysis platform, an intelligent specialized disease database, an AI follow-up system, an AI medical record connotation quality control system, a single disease process quality management platform, a VTE intelligent prevention and treatment system, an intelligent triage system, and an intelligent pre consultation system, Realize the full process closed-loop application of “pre diagnosis, during diagnosis, after diagnosis” data, and implement the enterprise goal of “cognitive data, enabling medical treatment”.
Clinical and scientific research are the two major entry points selected by Haisen Health’s “cognitive data, enabling medical treatment”.
China’s medical resources are generally scarce and unevenly distributed. In clinical practice, medical AI can effectively improve the overall level and efficiency of medical diagnosis and treatment in China; In scientific research, it can further strengthen the bottom core competitiveness of large hospitals and enhance the overall strength of medical institutions. The combination of the two can maximize the overall service level and capacity of China’s medical system from the supply side.
In terms of clinical application, Haisen Health pioneered the “dual engine” driven CDSS in China, creating an evolutionary new generation of CDSS from knowledge assistance to decision support, and achieving the entire process of assisted decision-making support from consultation to treatment.
Take the application of CDSS in the Third Hospital of Peking University (hereinafter referred to as the Third Hospital of Peking University) as an example. Haisen Health CDSS was launched in the Third Hospital of Northern Medicine in 2018 and has been running steadily for two years now. The system is based on more than 30 million real diagnosis and treatment data accumulated in the hospital’s electronic medical record system over the past 10 years, integrating BMJ Best Practice clinical practices as evidence-based medicine evidence, forming a dual engine decision-making architecture of “clinical best practices in our hospital”+”evidence-based medicine best practices”; On this basis, artificial intelligence technology is used to build a high-precision decision support model covering the entire diagnosis and treatment process, fully realizing the clinical decision support application function. According to calculations, the CDSS launched by some departments of the Third Hospital of Northern Medicine has a high decision-making support ability, with a recommended clinical diagnosis accuracy rate of up to 91.7%, an average diagnosis time reduction of 0.98 days, and an average hospital stay reduction of 2.02 days.
In terms of scientific research applications, in response to data problems commonly encountered in clinical research, including high repetition rates in data collection, low utilization rates, inconsistent data storage structures, lack of unified monitoring of data quality and integrity, and complex data query logic, Hisen Health has built a big data scientific research and analysis platform driven by artificial intelligence technology and supported by real diagnosis and treatment data both inside and outside the hospital. This platform can automatically integrate massive medical data inside and outside the hospital, and use corresponding algorithm models to conduct in-depth mining and multidimensional analysis of data, greatly optimizing the scientific research process. At the same time, it can effectively improve the efficiency and quality of scientific research, promote the transformation of scientific research achievements, support the validation and optimization of scientific research achievements based on the real world, and achieve the development of “clinical scientific research integration”, thereby boosting the diagnosis and treatment level of medical institutions “Double improvement” of scientific research ability.
From the full ecological medical AI product matrix to the full process data closed-loop application, medical institutions are only one of the application scenarios. Haisen’s vision of health is not limited to this. With the help of medical AI and big data technology, Haisen Health has actively tried and explored in areas such as medical insurance management, drug research and development, business insurance forecasting, and public health decision-making. Haisen Health stated that by focusing on consolidating the underlying core technology and defining the value height of medical AI big data, Haisen Health aims to bring tangible help to the people and bring more value to society.
The industry development of medical AI big data has been progressing rapidly in recent years.