Medical AI assisted VTE prevention and treatment, Huiyi technology tried to solve the problem of prevention and treatment of pulmonary embolism with new technology

Since the late 1980s, surgeons such as caprini have been committed to the research and design of a very detailed individual risk assessment model for VTE, and constantly optimize its accuracy and clinical availability However, more than 40 years later, pulmonary embolism is still an important cause of unexpected death in hospitalized patients The high mortality rate comes from the low proportion of patients receiving VTE prevention and treatment, and doctors ignore the subjective and objective risk assessment model of VTE Under normal circumstances, if the medical staff can accurately assess the risk of VTE, they can find the high-risk patients of VTE in time and give appropriate preventive treatment But in practice, there are many problems in the implementation of this scoring system "Too much time." When asked why all kinds of VTE evaluation scales did not play their due prevention and treatment effect, a doctor from a top three hospital in Beijing replied without hesitation, "it is a miracle that the ward of a large hospital can not add beds Even if doctors and nurses have the intention to prevent, they can not seriously evaluate the situation of each patient Moreover, the problem of VTE is not only the negligence or irregularity of medical staff, but also the lack of familiarity and evidence-based judgment of doctors, especially young doctors The problem of VTE prevention and treatment is not only in large hospitals, in fact, the problems faced by primary medical institutions may be more serious Many doctors in primary medical institutions do not have the prevention awareness and discrimination ability of VTE There are many problems, but the good thing is that those who have the heart are not afraid to challenge and actively seek solutions What are the key issues in VTE prevention and control? VTE is a preventable disease We can't wait for thrombosis to solve the problem If we can carry out early screening and prevention of VTE, the difficulties of prevention and treatment of VTE will be solved Among them, the key is to integrate enough patient information, realize automatic capture and give reasonable judgment, and simplify the clinical workflow In fact, the automatic Hospital of patient information capture has already been equipped, but the primary information system can only capture the basic information of the patient's name, age, gender and so on For more important and diversified descriptive information, due to different doctors' writing habits and formatting degrees, most of them need to be filled in by doctors manually This can not solve the practical problems in clinical practice, not much help in reducing the workload of doctors, nor can it solve the problems of wrong and missing evaluation that may exist in manually filling in the evaluation form The automatic analysis of information in the process of diagnosis and treatment is very challenging "For example, to let the computer judge the VTE risk factor information of" has cerebral infarction occurred within one month ", it is not only necessary to help the machine understand the time limit of cerebral infarction - the event should occur in the near future, such as a given month, without counting the cerebral infarction events three months or half a year ago At the same time, to judge the disease situation of patients, such as classification and classification, it is often necessary to synthesize the previous medical history, MRI reports and other information to reach a conclusion The traditional information system can not complete such a complex work " Zhang Qi, CEO of Beijing huievery cloud Technology Co., Ltd (hereinafter referred to as huievery Technology), explained to the arterial network reporter The emergence of deep learning brings a turning point for big data processing The emerging NLP technology can realize the semantic understanding of medical record data, structure and standardize descriptive natural language, in-depth learning algorithm can also integrate personalized medical record data into rule base, build VTE knowledge map with clinical guidelines and various scales, improve the accuracy of final machine model, and solve the "number" faced by doctors It is difficult to input data and analyze data But NLP also has its limitations If developers are not familiar with VTE diagnosis process, do not have good and complete data of patients' past medical history, medication history, or do not establish a reasonable knowledge map The accuracy of the results will be greatly affected Establishing an effective NLP is actually not that simple Based on CDSs, it is not easy for Huiyi technology to develop NLP battlefield and apply NLP to VTE prevention and control Taking Huiwei technology, which has VTE quality control products, as an example, the whole product building process is full of twists and hardships With four years of algorithm research and development, knowledge map construction foundation, clinical experts' intervention, and policy promotion, Huiwei technology's VTE quality control system development is advancing rapidly Zhang Qi told arteria network: "the whole system was built in May 2019 In just four months, Huiqi technology has created a set of VTE intelligent prevention and control system with an accuracy rate of more than 97% This success is inseparable from the hard work of Huizhu technology and the full support of the cooperative hospital " When talking about the four month development project, Zhang Qi could not help but feel that: "when we made the first generation of products in August, its accuracy rate was less than 80% At that time, we also faced many problems " The combination of CDSS system and clinical data is the first difficulty that every enterprise will encounter Fortunately, Huiwei technology has been very skilled in handling the access steps of CDSS The real trouble is to standardize the medical record information Because the medical record data is mostly unstructured text description, different doctors have their own writing habits, the terms used are standardized, the degree of identity is low, and the multi-source heterogeneity is prominent Huiwei technology uses a common form to "explain" these different information to the machine, so that the computer can understand its semantics accurately "Even if the team has been rooted in clinical practice for several years, and has incorporated hundreds of thousands of real medical records training AI models of top three hospitals and achieved good recognition results, we iterated and verified the model again for the real patient medical records of the cooperative hospital, and the whole process took up to three months to make the recognition accuracy stable above 97%," Zhang Qi said It is worth mentioning that Huiyi technology platform has a strong generalization ability Even if the doctor's report does not strictly follow the standard, Huiyi AI can accurately capture patient information and generate accurate results It is practical to remind doctors to perform early screening and prevention of VTE Even with the help of VTE quality control system, doctors will inevitably Miss VTE risk assessment of some patients in busy work In view of this situation, the intelligent reminder of each VTE quality control system will come into use "Our goal now is to advance the screening time When the patient is just admitted to the hospital, we will help the doctor to solve the relevant problems and move the quality control gate forward The first 24 hours of admission is an important evaluation period If the doctor misses, we will use the CDSS in the electronic medical record to remind the doctor within the 24 hours to tell the doctor that the patient has not completed his VTE risk assessment " In order to ensure timely data statistics and feedback, Huiyi technology has also set up a statistical management platform in the hospital to record the VTE prevention and control behaviors of each patient, so that patient information can be clearly presented to the responsible doctor, and the missing information can be seen at a glance Taking China Japan Friendship Hospital as an example, Huiyi technology's afterwards big data statistics and analysis platform is projected on the large screen of the nurse station, so that the nurse can intuitively learn the patient's bed information, VTE evaluation and prevention implementation at a glance In addition, Huiyi technology is also trying its best to make VTE prevention and treatment land in primary hospitals Different from the third grade a hospital, the level of hospital informatization in primary medical institutions is not high, and doctors also lack the ability to distinguish VTE In this case, the VTE quality control system of Huiwei technology mainly plays the role of standardization and doctor education In this process, the prevention and treatment ability of doctors will gradually improve The promotion of VTE prevention and treatment comes from the combination of various forces On October 13, 2018, the kick-off meeting of "national capacity building project for prevention and treatment of pulmonary embolism and deep vein thrombosis" and the press conference of "China's guidelines for prevention and treatment of thrombotic diseases" were held in Beijing Sino Japanese Friendship Hospital VTE prevention and treatment has a strong push This is also an important opportunity for Huiyi technology and other enterprises to develop VTE quality control system Many hospitals have also responded to the requirements of the policy by setting up prevention and control management team, establishing VTE prevention and control management supervision mechanism, carrying out discipline construction and health education, etc to establish the VTE prevention and control system in the hospital However, it will be a long battle whether it is the promotion of hospital informatization or the improvement of doctors' awareness of VTE prevention and treatment In this process, the government, enterprises and hospitals need to work together Jeni turtle