Improve the quality and efficiency of the hospital sense. Why is this clinical risk system praised by the top three hospitals?

The hospital that saves the wounded and injured is actually at great risk. Health care workers are engaged in high-complexity, high-risk health services such as disease diagnosis, treatment and care every day. These service activities cause patients to face a range of clinical risks once they enter the hospital.

According to a special report published by the Institute of Medicine of the National Academy of Sciences in 2000, medical errors in the United States have become the fifth leading cause of death. The cost per year in the United States is between $17 billion and $29 billion due to patient injuries caused by medical errors. According to data released by the World Health Organization, in developed countries, about 10% of hospitalized patients suffer from various clinical errors or adverse events.

For this reason, reducing medical risks and improving patient safety can bring huge economic and social benefits. In this context, the necessity and urgency of establishing a systematic, specialized, and long-term clinical risk management mechanism in hospitals is beyond doubt.

In recent years, the national medical management departments have paid more and more attention to patient safety, requiring medical institutions to fully understand the importance of patient safety management, and to ensure safety as an important part of medical management, in accordance with "prevention, system optimization, full participation". The principle of “continuous improvement” is vigorously promoted, and the level of patient safety management in medical institutions is continuously improved.

However, subject to management concepts and technical conditions, the current clinical risk management model of medical institutions is mainly the reporting and statistical analysis after the occurrence of risk events. There is a series of problems in this kind of post-event management, and there is an urgent need for improvement.

The current stage of reporting is often directed to the risk event after the occurrence, so the result can only be managed as a management object, which is a lagging management. Risk events are usually reported by doctors, so it is inevitable that they will miss or cover up the problem. In addition, the hospital's risk management channels are scattered, and the risk of hospital infection, patient fall, adverse drug reaction, deep vein thrombosis and other adverse events are reported separately, which has led to the development of multiple information systems for card reporting. The clinical workload burden makes the actual risk management results less than ideal.

In response to this situation, Shanghai Lilian Information Technology Co., Ltd. (hereinafter referred to as Lilian Cognitive) based on medical big data analysis, cognitive computing and other technologies, cooperated with famous domestic hospitals to establish clinical risk monitoring and early warning based on high-quality clinical data. The model provides comprehensive solutions for clinical risk management, helping medical staff to detect clinical risks in a timely manner, and even predict clinical risks based on patient specific conditions, thereby driving more efficient and accurate clinical decisions, reducing clinical risks and reducing clinical risk events. Life and economic loss.

It is reported that Lilian Cognitive founder Niu Yaojun is the former general manager of the IBM Greater China Medical Health Solutions Laboratory team. He has participated in the research and development of IBM Waston Cancer Assistant and its promotion in China. The core members of the company are from the world's top 500 technology companies such as IBM, HP, Teradata, Baidu and other companies in the chief medical industry consultants, top big data analysis experts and technical experts.

Hospital infection refers to the infections acquired by doctors and patients in hospitals. It is an important clinical risk, and the risk events account for a high proportion, resulting in huge losses. A 2014 study by the Department of Infectious Diseases Prevention and Control of Peking University First Hospital showed that there were 4 million hospital infections per year in China. The average economic loss per patient from hospital infection was 29,846 yuan, and the hospitalization time was extended by 13 days. .

In general, ICU wards, neonatal wards, and related departments such as hematology and oncology are the most risky departments for nosocomial infections. Because of the low immunity of these patients, and the application of invasive and open measures such as catheters and ventilators, the incidence of nosocomial infections is also high.

Sun Shumei, director of the Department of Sensory Control of the Southern Hospital affiliated to Southern Medical University, once made an image metaphor. She believes that high-risk departments are like road sections that are prone to traffic accidents. Even if doctors operate the rules, they are prone to accidents; while ordinary departments are like normal road sections, and accidents are caused by medical staff violations.

Therefore, improving the ability of medical staff to sense the risk of infection, and timely and proactively conduct standardized infection prevention and infection response has become the key to hospital risk management.

It is the first step to ensure the quality of hospital risk management by ensuring timely and accurate identification of hospital risk events. Objectively speaking, the current hospital infection event discrimination does not have a unified gold standard, and it is often based on the doctor's experience.

Because some doctors do not have accurate recognition ability or are unwilling to report bad data, the infection is underreported. Such self-reporting can not reflect the true infection status of the hospital, so the corresponding management department, such as the sensory control department, needs to judge according to experience, which brings a huge workload to the sensory department. Sensing and control departments often only use spot checks to supervise the status of the card, making it difficult to conduct comprehensive and effective management.

Under this circumstance, some hospitals use sensory control software to identify and judge the sense of the hospital according to the rules of expert design, which helps the sensory department to improve the quality and efficiency of work. However, due to the methodological limitations of the expert development model, the characteristics of the rules are very limited, often only a dozen to dozens, resulting in a low degree of accuracy and accuracy of the risk judgment model. The effect of the event to identify and determine is not satisfactory.

In response to this problem, Lilian Cognizant has launched a clinical risk management system through intelligent and digital solutions. It has adopted the PDCA concept design for the first time in China, and fully closed-loop control of clinical risks. This system is first applied to the yard risk management scenario, where the AI ​​risk monitoring model is the core of the system.

Lilian AI risk monitoring model is based on a large number of clinical data, using inductive method for supervised learning, based on integrated learning algorithm for hospital risk-assisted diagnosis model. Taking the hospital sense as an example, the rules included in the model mainly include two parts, one is based on the “rules” learned from the hospital's massive infection data, and the other is based on the “rules” summarized by the experience of many hospitals.

The model has sustainable learning capabilities and is able to optimize models based on the specific conditions of different hospitals. One of the major advantages of the AI ​​risk monitoring model with integrated learning algorithms is interpretability, which allows for “evidence-based” and “retrospective research” of results.

At the same time, the “new evidence” and “new rules” obtained from the analysis and research of the results can be added to the model for further optimization, and finally the sustainable learning and optimization of the AI ​​risk monitoring model can be realized. The model has better auxiliary functions in hospital infection monitoring, and can present patients with possible infections to doctors in real time, so that doctors can timely intervene and treat patients, and ultimately reduce the workload of doctors and prevent the occurrence and development of infections.

Compared with the expert development model, the Lilian AI risk monitoring model is based on a large number of real electronic medical record data. The feature engineering is used to explore the characteristics of infection-related features, and many new prognostic variables are identified. The eigenvalues ​​involved are up to three. More than a hundred, which greatly improved the fineness and accuracy of the model, making the model more sensitive and specific for infection discrimination than artificial rules.

Cai Meiping, chief medical consultant of Lilian Cognition, said: "In the future, in addition to continuous optimization of the AI ​​infection risk monitoring model to improve the accuracy of the system monitoring of infection events, Lilian Cognition will also expand horizontally in the field of clinical risk monitoring. A risk monitoring model for areas such as deep vein thrombosis and sepsis.

There is a concept that is often mentioned in hospital clinical risk management, called the gateway forward. As the name implies, a risk warning is given before a clinical risk event occurs. But at home, few hospitals are currently able to do this. First, the hospital lacks corresponding management data, and secondly, it lacks corresponding analysis models and tools.

Cai Meiping believes that risk control should not be a lagging job, but should be identified and predicted from the beginning of the doctor's understanding of the patient's initial illness. Using the risk prediction and early warning model to timely discover the signs of clinical risk events, together with standardized and effective preventive clinical recommendations, can help doctors take timely preventive measures to minimize the probability of occurrence of risk events.

It is reported that Lilian Cognition is conducting research and development work on clinical risk prediction and early warning models. The AI ​​clinical risk management assistant developed based on this model can predict the possibility of occurrence of risk events through the data collection and analysis of patient dynamics before the occurrence of clinical risk events in hospitals, and provide early warning for high-risk events to provide specific doctors. Prevention and control recommendations to help them take effective preventive measures to reduce the occurrence of risk events.

Cai Meiping said: "This AI assistant system will be used as a supplement to the clinical risk management system, bringing new expansion to the company's product line and being widely used in more clinical departments."

The Department of Infection Control of Zhongshan Hospital affiliated to Fudan University in Shanghai introduced Lilian Cognitive Products this year. Gao Xiaodong, director of the sensory department, told the arterial network: "Purchasing this system is mainly because the hospital doctor wants to know the hospital's hospitality rate and whether the patient is infected."

According to reports, in the past, the sensory control department of Zhongshan Hospital was informed by a full-time staff about the situation of each discharged patient. But as more and more patients in the hospital, doctors can't understand all the patients. Through Lilian's cognitive system, doctors can quickly and timely understand the patient's main infection indicators and the possibility of infection, and screen out high-risk groups.

In addition, the hospital should understand whether the clinician has made various diagnoses or interventions according to the regulations, and whether there is any inspection during the infection. Through the doctor's investigation, the current inspection situation is obtained, and then the supervision is carried out. In this kind of investigation, many doctors or nurses used to go directly to the patient to ask about the situation. With the Lilian system, you can use the clinical information of all aspects of the system integration to quickly select the target population.

In the aspect of prevention and control early warning, the sensory control department can inform the clinician to do relevant protective measures in advance according to the risk situation indicated by the system before the hospital infection occurs, so as to avoid the infection.

Take Zhongshan Hospital affiliated to Shanghai Fudan University as an example. At present, the number of patients discharged from hospitals is about 8,000. With an early warning system, the sensory doctor can save 60%-70% of the workload. In this regard, Director Gao said: "In the past, we went to investigate after the patient was discharged from the hospital. It is a kind of lagging management. But now through the Lilian cognitive system, we put the sensory management in front, as long as the clinical detection results, the system Can be early warning."

Hospital infection outbreak warning is a larger application of the system. The sensory department of Zhongshan Hospital set some target values ​​through the system of Lilian Cognition, and under what circumstances, there may be signs of a large-scale infection outbreak in the hospital, thus avoiding some large-scale malignant events. "This is based on a single patient warning, a large system warning for the entire department or hospital." Gao said.

The Department of Sensory Control of the Southern Hospital affiliated to Southern Medical University is also one of the clients of Lilian Cognitive Products. For Lilian's cognitive system, Sun Shumei, director of the sensory department, commented: "Lili's system is a monitoring and intervention tool that can reflect hospital infections and high-risk factors in real time. The patient has a little bit of turmoil and the doctor can promptly Found that the occurrence of group incidents is eliminated."

In addition, Lilian's system is able to dynamically discover some rules. When the patient has several risk factors, the doctor can timely intervene through the timely monitoring and early warning of big data to avoid the development of infection.

To truly achieve the refined management of the hospital, we need to quantify through big data. On the one hand, there is no additional burden on the clinician. The doctor's job is still to treat patients, so the sensory department should not wait for the doctor to report, passive notification, but should use the means of information, take the initiative to collect data, and actively serve the clinic.

According to Director Sun, the current state stipulates that 250 beds are equipped with one sensory control staff. The Southern Hospital affiliated to Southern Medical University has a bed of 2,250 and a total of 12 sensory management doctors. "In the past, we can only conduct random checks and no comprehensive monitoring. Now with this system, we can achieve comprehensive monitoring of hospital risks through accurate data." Sun said.

"If the score is 100, I will give 90 points!" Director Sun gave a high evaluation of Lilian's cognitive system. On the one hand, she hopes that the intelligent warning of the system can be made more precise, and the probability of false negative or false positive can be reduced to a lower level. On the other hand, she hopes that the output interface of the system will be more concise, and the complicated things will be presented simply, so that doctors can operate more conveniently.

Octopus Products

Cut Poulp Squid,Frozen Baby Octopus,Frozen Poulp Squid,Frozen Cut Poulp Squid

Zhoushan Shiji Aquatic Products Co., Ltd. , https://www.shiji-aquatics.com