Dialogue with international authorities: listen Dr. Sten Westgard Some opinions on special courses

2022-02-17 1214

picture 12. png


01

Sten Westgard Introduction to Professor's Lecture Content

   

Sten Westgard The professor focuses on the types of patient data quality control methods, Feature comparison, MA The Development of Law, Current Status of Laboratory Applications Abroad, MA Provide a clear and concise explanation of the establishment of the law and the handling process for loss of control, Absolutely brilliant, benefit a lot!

picture 1. png

4 Comparison of the advantages and disadvantages of patient data quality control calculation techniques


picture 2. png

Sten Westgard Professor's presentation MA Quality control chart, MA Quality control setting method and out of control correction process


02

My thoughts

The best method is the compromise moving average method (MA law) Or

Difficult professional skills EWMA law?

only use 3SD control limit?


IFCC Yu 2019 Year and 2020 The guidance document published in the year proposes real-time quality control based on patient data ( patient based realtime quality control , PBRTQC) It is a method of utilizing patient test results in real-time, Quality control methods for analyzing performance in continuous monitoring and testing processes. Numerous studies have shown that, PBRTQC Compared with traditional quality control methods, it has lower cost, No matrix effect, Can be dynamically monitored, Advantages such as sensitivity to identify errors before analysis. With the development of information technology and clinical research, PBRTQC Multiple computational programs have been developed, including BULL law, AON law, MA law, Percentile method, Median method and EWMA Law, etc.


Sten Westgard The professor mentioned in class:


picture 3. png


and PBRTQC Professional software tools -


Senxu Medical Patient Data Real time Quality Control Intelligent Monitoring Platform (AI-MA) Across the country 16 Medical institutions of different scales and levels, 6 A professional field, 13 A brand, 36 model number, 65 Taiwan Inspection System, 110 A testing item, Experience and Results of Multi Center Clinical Practice Research on Patient Big Data Performance Validation, ( "based on AI Real time quality control of real-world patient data PBRTQC Multi center clinical research" , Wen Dongmei, Shanghai Senxu Medical)


right PBRTQC middle MA Law and EWMA The quality control effectiveness of the method in identifying quality risks and analyzing performance changes was compared.


We have selected Professor Wang Xuefeng from Ruijin Hospital affiliated with Shanghai Jiao Tong University School of Medicine and Professor Li Guixia from Hebei Children's Hospital for the intelligent construction team AI-MA Elaborate on two research cases of real-world big data.


picture 4. png

Professor Wang Xuefeng from Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine&Senxu Medical Intelligence Construction Team AI-MA Clinical application site


picture 5. png

Professor Li Guixia from Hebei Children's Hospital&Senxu Medical Intelligence Construction Team AI-MA Clinical application site



03

Two cases


1, MA False negative cases


(1) Choose Senxu Medical AI-MA of EWMA, Z Score chart calculation program and intelligent quality control rules, Regarding Hisen Meikang XN1000 Blood cell analyzer RBC and HGB Patient data for the project (n=31853) Conduct analysis.


The results show, Analyze the performance of the detection system from 5 moon 11 Negative bias begins to appear on the day, And in 6 moon 14 Daily correction, See figure 8, According to the laboratory's on-site quality records and instrument alarm information display, 6 moon 13 day XN1000 The blood cell analysis instrument has triggered an alarm, Insufficient sample size for alarm, And automatically stop detecting samples, Laboratory in 6 moon 13 Replace the sample probe in the afternoon, 14 The instrument has resumed normal testing. AI-MA of EWMA The program canNearly before the fault alarm 1 monthChanges in recognition and warning analysis performance, And the warning elimination time is completely consistent with the time node for replacing the sample probe in the laboratory. Transitioning from traditional instrument alarm correction to real-time monitoring, early warning!


picture 6. png

Senxu Medical AI-MA Intelligent monitoring platform RBC of EWMA Quality control chart, RBC& HGB Z Score chart (EWMA law)


remark: The green dot is RBC Patient testing results; The purple dot is HGB Patient testing results; The yellow dots are the detection results of patients with negative bias and analysis error warnings caused by changes in analysis performance by the system.


(2) Choose Senxu Medical AI-MA of MA operation program, reference Sten Westgard Professor's recommendation 3SD Set control limits, Verify the same batch of patient data (n=31853) , The results show, RBC, HGB Patient data MA Quality control chartNo warning message appeared, Moreover, the trend and time nodes of analysis performance changes caused by corresponding instrument failures are not obvious, namelyUnable to provide early warning of instrument malfunctions and potential quality risks, See figure 9.


picture 7. png

Same group of patients RBC& HGB The test results MA Quality control chart (MA law)


2, MA False positive cases in law

(1) Choose Senxu Medical AI-MA Patient data EWMA quality control material QC, integration Z Score chart calculation program and intelligent quality control rules, Regarding Beckman AU5800ALT project 20-45U/L Patient population data (n=860) Conduct analysis, The results show, EWMA Quality control chart and quality control product QC consistent trend, All displayed ALT Good detection and analysis performance, stable, Actual Accumulation CV Less than the quality target of the laboratory, The reagent blank on the instrument and indoor quality control both reflect stable analytical performance, See figure 8-B.

Choose Senxu Medical AI-MA of MA operation program, reference Sten Westgard Professor's recommendation 3SD Set control limits, verificationthe same batchPatient data (n=860) , The results show, ALT Patient data MA The quality control chart shows that the precision analysis performance is average, appear 8 Next Super 2SD Warning and 2 Next Super 3SD False positive alarm, Regarding quality control products QC The displayed analytical performance does not match. ifpress Sten Westgard Professor's suggestion is super 2SD And super 3SD process flow, It will bring additional manpower to the laboratory, Time and cost investment, See figure 8-A.


picture 8. png

picture 9. png

picture 8-A  Senxu Medical AI-MA Intelligent monitoring platform for patient data ALT of MA Quality control chart

picture 8-B Senxu Medical AI-MA Intelligent monitoring platform for patient data ALT of EWMA Quality control chart&quality control material L-J Graph integration Z Score chart

remark: The blue-green dots represent the patient's test results MA Value and EWMA value; The triangle symbol represents height, middle, Low concentration level quality control product testing results; Yellow dots are warning points, The red dot is the point of loss of control



04

summary


From the above two cases, it can be seen that, although EWMA More complex calculations, But as long as the appropriate quality control parameters are selected, It can demonstrate excellent performance in identifying instrument failures and monitoring potential quality risks "sensitivity ", " specificity "and "immediacy" Waiting for quality control effectiveness.


EWMA It is one of the most famous algorithms used internationally to monitor public health events such as outbreaks of epidemics or changes in influenza virus activity, Research on quality control in other industries internationally has shown that, EWMA The role of control charts in monitoring inaccuracies and Westgard Multi rule control chart is equivalent, Even better, Can detect small variations and trend changes in accuracy or imprecision during the analysis process.


Senxu Medical PBRTQC Professional software tools AI-MA Performance validation results of patient big data in clinical laboratories of different scales, confirmed EWMA The above excellent performance of the algorithm, Simultaneously superior to MA algorithm.



05

Outlook and Reflections

  

Thank you to the organizing committee for giving us the opportunity to listen to and learn from international quality control experts online StenWestgard Professor's class, Broaden one's horizons, Understand foreign countries MAQC Research progress and application methods. StentWestgard The professor mentioned, Only experts understand the value of patient data quality control methods, Only enthusiasts of this technology will appreciate it, feel the same way! !


By listening to lectures from international quality control experts, Comparing patient data from abroad MA Progress in the Study of Law, alsoDeeply proud, Our Forest Xu Medical AI-MA The intelligent development application team adheres to the tradition of inheriting classics, Innovative applications, Together with the spirit of research to effectively address clinical pain points and needsConquered it StenWestgard The professor mentioned EWMA The International Technical Difficulties of Law.


at the same time, very happy, It is an honor to collaborate with many outstanding discipline managers across the country, Excellent team based together AI-MA Conduct research on big data performance validation and accumulate professional experience, Feeling increasingly affectedIFCC The mentioned: PBRTQC It is an exciting innovative quality management tool, also AI-MA Professionalism of the technology platform, Innovation is full of confidence!

With the increasing emphasis on patient risk quality management strategies, The Chinese inspection industry has begun to pay attention to and attach importance to research on patient data quality control methods. Believe in the future as the industry develops and progresses, Promotion and popularization of learning, Strengthening personnel training, The application of big data and artificial intelligence technology in the field of medical testing, Will accelerate PBRTQC The Development and Application Process of Clinical Laboratories in China, Quality risk identification in the laboratory, Plays a significant role in lean management and intelligent construction.


Sincere thanks for participating "China's first based AI Real time quality control of real-world patient data PBRTQC Multi center clinical research" 16 The valuable research platform provided by home medical institutions, For the domestic market PBRTQC Provide rich professional practical experience in application and implementation!

picture 10. png