Generally speaking, medical artificial intelligence uses deep learning to process two types of data: image and text. Although the former is hotter in the capital market and more mature in technology, in terms of application scenarios, text data such as electronic medical records and prescriptions are ubiquitous and have been widely used in medical information systems.
By integrating and processing the raw data scattered in the medical information system into standardized structured data for research and clinical use, combined with the valuable knowledge and experience of experts, artificial intelligence companies deposit these elements on algorithms to create disease models, and even further disease database and disease research network which can support medical research, clinical diagnosis and treatment, as well as hospital operation management. It is not difficult to see that the essence of artificial intelligence is a data processing tool that requires a large amount of data to support machine learning. Therefore, medical big data that exactly matches its role has begun to take shape in 2016. However, not all data can be used for machine learning. On the contrary, artificial intelligence learning has higher requirements for data. Due to historical and habitual reasons, the phenomenon of “focusing on clinic and neglecting data” is very common. Medical data shows the characteristics of large quantity and poor quality, lack of uniform standards, etc., to a large extent which has lagged the development of big data in medical field.
Main advantages of Clinbrain
Qin Xiaohong, co-founder of Clinbrain, believes that there are several difficulties in successfully applying existing medical big data to artificial intelligence. First of all, the existing medical data is huge which can reach the PB level, but there is a big problem with data standardization. The first is that the data structure is not standardized due to the lack of corresponding mandatory standards. The second is that the data content is not standardized due to the lack of a unified template, doctors in different hospitals or even the same hospital may have different descriptions of the same disease when writing medical records. Secondly, with the rapid development in medical industry, the phenomenon of data islands among hospitals and departments is serious, making the use of medical data difficult. Although China is solving this problem, it will take time to solve it completely. Finally, medicine is a highly specialized field. Even if they have data governance capabilities, it is almost a fantasy to make use of these data and further empower clinic or research without corresponding medical background.
It is not difficult to see that the successful combination of medical big data and artificial intelligence has a high barrier to entry. It requires companies not only to have data mining capabilities, but also to have deep data analysis and governance capabilities based on the understanding of the characteristics and needs of the medical industry. In this regard, as an explorer of “medical big data + artificial intelligence”, Clinbrain has an advantage. In the long-term governance process of medical big data, Clinbrain conducts standardized mapping processing of non-standard data, post-structural processing of unstructured data, and cleaning processing of dirty data, building a huge medical industry standard library and medical terminology database; thereby solving the problem of data quality.
At present, hospitals including West China Hospital, Ruijin Hospital, Changhai Hospital, Southwest Hospital of AMU, Fudan University Shanghai Cancer Center, etc have all chosen Clinbrain to build medical big data governance platform and application platform.
In order to break through the data islands inside the hospital, Clinbrain has created ClinData product through continuous research and development. Even if the corresponding HIS and EMR information systems do not open interfaces, Clinbrain can seamlessly integrate the data of hundreds of systems from dozens of data into a unified data center without any interface modification, realizing the hospital connectivity of internal “data islands”. In addition, Clinbrain has many years of experience in the medical industry and has established a large-scale professional medical team. These medical professionals cooperate well with each other to complete the extraction and processing of clinical data.
Clinbrain empowers hospitals
Clinbrain has made corresponding layouts in subdivisions such as natural language processing, knowledge graphs, text recognition, automated machine learning, and clinical decision-making systems, and has been involved in rare disease clinical decision-making and VTE intelligent prevention and control management.
Beginning in 2014, Clinbrain began to deploy in the field of medical big data and AI. Over the past 8 years, Clinbrain has polished the self-developed natural language processing technology NLP, which includes post-structured processing systems for several types of text data such as medical records, CT reports, MR reports, and pathology reports, etc.
Clinbrain also optimizes its products to better meet the unique needs of the medical field. In addition to providing general models, Clinbrain provides a labeling platform for hospitals, so that doctors can mark their own text, and then automatically train a personalized model.
In addition, a large number of medical literature and medical guides are displayed in PDF format, and how to extract these data and use them has always been an urgent need. With the help of natural language processing technology and the accumulation of related models, the company has realized the content recognition and extraction from these PDF files.
Moreover, Clinbrain’s medical knowledge graph is developed around knowledge in the medical field. It aims to systematically organize the knowledge in the text by establishing the association relationship between medical entities, so that the knowledge is easier to be understood and processed by the machine. Also, Clinbrain has also built a huge medical industry standard library and medical term library, with an accumulation of more than 1,600,000 medical terms.
Based on the construction of the above-mentioned artificial intelligence underlying technology, Clinbrain has introduced “big data + artificial intelligence” solutions in a number of specialized disease fields; among them, the VTE intelligent prevention and control management platform and the rare disease intelligent decision-making system are particularly worth mentioning. This platform uses the hospital’s big data center and AI model to provide high-quality decision-making basis; through the provision of standards + customized evaluation scales, while combining AI automatic decision-making engine to evaluate the condition and recommend diagnosis and treatment; through the medical auxiliary diagnosis and treatment system, quality control management system and the patient education follow-up system to manage the entire process of VTE prevention and treatment.
The rare disease clinical decision support system is the latest exploration of the comprehensive application of the underlying technology of Clinbrain, including the rare disease decision-making interactive system, the disease phenotype analysis system and the rare disease decision engine. It provides opportunities for the diagnosis and treatment of rare diseases, and is used for the comprehensive evaluation of clinical phenotypes, disease knowledge and other information, and provides a list of potential rare diseases. Its main function is to synthesize the patient’s disease phenotype, evaluate and score against the existing 7,000+ rare diseases, and assist clinicians in accurate diagnosis.
Through deep cultivation in the field of medical big data, Clinbrain has formed a competitive advantage in “big data + artificial intelligence”. In the future, Clinbrain will explore and practice more medical application scenarios, improve scientific research and clinical quality, help the development of medical data applications, provide intelligent data support for the medical industry, and ultimately create a “ClinBrain ” that serves all medical scenarios.