Treatment audit system

Abstract

To provide capability for auditing whether medical cosmetic treatment is being appropriately provided or not.SOLUTION: Disclosed herein is a medical audit system for generating models regarding treatment procedures from a captured video image of treatment and analyzing the newly acquired video image to audit whether or not the treatment in the video image is being executed according to the treatment procedures. The present invention enables one to audit whether a series of treatment procedures is being appropriately executed, and can ensure treatment quality and patient safety. In the event of a problem, the invention can identify its generation source and swiftly implement remedial measures.SELECTED DRAWING: None

Description

[Technical field]

[0001] The present invention relates to a treatment audit system. [Background technology]

[0002] In recent years, the need for aesthetic medical treatment has increased, and various hospitals and clinics are offering a wide variety of aesthetic medical treatments. In response to this growing need, it is becoming increasingly important to provide aesthetic medical treatment more efficiently, accurately, and safely. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Japan Society of Aesthetic Plastic Surgery, "Japan Society of Aesthetic Plastic Surgery Newsletter Aesthetic Medical Practice Guidelines (Revised Edition for 2021)," published on October 10, 2022 Summary of the Invention [Problem to be solved by the invention]

[0004] The present invention provides a method for evaluating whether cosmetic medical treatment is being performed appropriately in order to meet the need for performing cosmetic medical treatment more efficiently, more accurately, and more safely. [Means for solving the problem]

[0005] According to the present invention, there are provided an evaluation method, evaluation device, evaluation system, evaluation program, analysis method, analysis device, analysis system, analysis program, medical accident prevention method, medical accident prevention device, medical accident prevention system, medical accident prevention program, medical safety assurance method, medical safety assurance device, medical safety assurance system, and medical safety assurance program. Effect of the Invention

[0006] According to the present invention, it is possible to improve the safety and efficiency of treatment, and to ensure the support and safety of treatment by using a terminal. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0007] Embodiments of the present invention can be utilized as an evaluation method, evaluation device, evaluation system, evaluation program, analysis method, analysis device, analysis system, analysis program, medical accident prevention method, medical accident prevention device, medical accident prevention system, medical accident prevention program, medical safety assurance method, medical safety assurance device, medical safety assurance system, and medical safety assurance program, as to whether cosmetic medical care is being performed more efficiently, more accurately, and safely.

[0008] [Technology 1] This invention is a system for monitoring compliance with manuals for procedures at beauty clinics. Specifically, it captures information on the actions of medical staff during procedures and the tools they use in real time, and uses AI to determine whether the procedures are being performed according to the manual. If proper procedures are not performed, the system immediately issues an alert. This improves patient safety and reduces the possibility of medical errors.

[0009] Specifically, the present invention is a system for monitoring compliance with a manual for treatment actions in a beauty clinic. The problem that the present invention aims to solve is the deterioration of patient safety due to human error during surgery or examination or deviation from the manual. Specifically, in treatment at a beauty clinic where compliance with procedures is required, human error and non-compliance with routines are serious problems. The specific content of the present invention is a system that captures information such as the movements of medical personnel and the tools they use in real time and uses artificial intelligence (AI) to determine whether those actions conform to the manual. If appropriate actions are not taken, an alert is issued immediately, allowing appropriate responses. As an effect of the present invention, first of all, patient safety is significantly improved. By detecting deviations from the manual in real time, it is possible to prevent medical errors. In addition, the reaction of medical personnel to the generated alert and their subsequent actions are also recorded, making it possible to collect information that contributes to subsequent improvement activities. Furthermore, the present invention promotes training and ability improvement of medical personnel. Based on the generated alert, it is possible to clarify specific points for improvement, such as in what situations medical personnel tend to deviate from the manual or what procedures are difficult.

[0010] The system of the present invention includes the following elements. First, a sensor device, such as a high-resolution camera or motion sensor, is used to capture the movements of medical staff. This allows detailed recording of the doctor's movements during surgery and the operation of surgical tools. Second, an RFID (Radio Frequency Identification) tag or barcode scanner is installed to identify the tools and medicines used.

[0011] Motion sensors and high-resolution cameras capture the movements of medical professionals in real time and send the data to the AI ​​engine. For example, the moment a doctor picks up a surgical tool, his or her hand movements are captured and information is generated. Specific movement patterns during surgery (e.g., the direction and depth of the incision) are also captured as data.

[0012] RFID tags and barcode scanners are used to identify the instruments and drugs used. For example, RFID tags or barcodes attached to syringes, scalpels, and even medicines can be scanned to record which instruments or drugs were used and when. This ensures that the right instruments and drugs are used at the right time.

[0013] This data is sent in real time to the AI ​​engine, which uses deep learning algorithms to determine whether the data is in line with the manual. This determination covers a wide range of areas, including the timing of use of various instruments, surgical procedures, and the amount and timing of drug use.

[0014] If any behavior that does not comply with the manual is detected, the AI ​​engine generates an alert and warns medical personnel. This alert is communicated in the form of a message displayed on the screen, a voice notification, or a vibration on a specific device (e.g., a smartwatch). At the same time, detailed information about the alert and appropriate responses are also provided. In this way, by combining a specific sensor device and AI technology, the present invention monitors compliance with medical procedures in beauty clinics in real time and improves patient safety.

[0015] The audit system according to this embodiment will be described in detail below. In particular, a system using a device called Cellution will be described as an example. "Celution" is a medical device for activating "stem cells" contained in fat collected from one's own thighs, etc. Stem cells have the ability to transform into new fat or blood vessels, and to replicate themselves, and their safety and effectiveness have been confirmed by many medical institutions. They are also widely used in cosmetic surgery for breast augmentation and skin rejuvenation treatments.

[0016] Celution activates the function of these stem cells, and when injected into the breast, it promotes the process by which the fat tissue creates new cells. This makes it possible to significantly improve the rate and durability of fat attachment. However, because stem cells have reduced function when contained in fat tissue and the amount contained is small, if stem cells are injected into the breast without separating them from the fat, it becomes difficult for the fat to attach, and as a result, the size of the breasts often returns to normal. In addition, the effect of preventing hardening is not fully achieved.

[0017] However, by using "Celution", it is possible to efficiently separate fat and stem cells and extract activated stem cells. As a result, it is possible to achieve excellent breast enlargement effects that could not be achieved with conventional methods. Although the amount of fat collected is greater than with the condensed-rich method, it is possible to achieve a higher fat retention rate by extracting stem cells separately and injecting them into the breast. In addition, this method is the only breast enlargement surgery in Japan that is recognized as a fat stem cell injection method (other methods are classified as simple fat injection methods), and the desired size is achieved one month after surgery. Another advantage is that there is almost no risk of the breast size decreasing six months after surgery. In addition, it is possible to increase the size of the breasts by more than two cups, and a guarantee system for a two-cup increase is also provided.

[0018] The present invention is a medical auditing system that generates a model of the procedure from images of the procedure, analyzes newly acquired videos, and audits whether the procedure in the video is performed according to the procedure.Furthermore, the system records the procedure at a beauty clinic, which includes locking the four casters on the main unit, inserting the power cord into the lower right side of the main unit, operating according to the main unit monitor, opening the centrifuge cover and pump cover, and setting the disposable.

[0019] The present invention solves the problem of verifying the exact procedure of a procedure in a beauty clinic and efficiently auditing whether it is being performed properly, thereby improving the quality of the procedure and patient safety.

[0020] The present invention allows for real-time auditing of whether a series of procedures is being performed properly, ensuring the quality of care and patient safety, and, if a problem does occur, identifying its source and quickly implementing corrective measures.

[0021] The medical audit system of the present invention generates a model from images of treatments at beauty clinics. This model includes treatment procedures such as locking the casters on the main unit, inserting the power cord, following the instructions on the main unit monitor, opening the centrifuge cover and pump cover, and setting up the disposables. By comparing and analyzing newly acquired videos with this model, it is possible to audit in real time whether the treatment is being performed according to the procedures. This makes it possible to immediately detect any problems that may occur during the treatment and take the necessary measures.

[0022] The present invention performs an audit using the following models: a first learning model created by machine learning using information indicating whether or not the first situation is occurring in which the four casters on the main body are locked and an image of the same as training data, a second learning model created by machine learning using the image as training data of information indicating whether or not the second situation is occurring in which the power cord is plugged into the lower right side of the main body, the switch next to the cord plug is turned on, and the power is on and an image of the same as training data, information indicating whether or not the third situation is occurring in which the main body is installed in a location with little foot traffic and an image of the same as training data of information indicating whether or not the fourth situation is occurring in which an operation is being performed according to the main body monitor and an image of the same as training data of information indicating whether or not the fifth situation is occurring in which the centrifuge cover and the pump cover are open and an image of the same as training data of information indicating whether or not the sixth situation is occurring in which a disposable is set, the disposable packaging is opened, and the centrifuge processing container is attached to the centrifuge rotation unit ... a sixth learning model created by machine learning using an image as training data; information indicating whether a seventh situation is present, in which a waste liquid bag is hung on the back side of the main body and the centrifuge processing container is pushed in vertically and symmetrically and attached, and a seventh learning model created by machine learning using the image as training data; information indicating whether an eighth situation is present, in which there is no gap between the centrifuge processing container and the attachment base and the centrifuge processing container is set horizontally to the back, and an eighth learning model created by machine learning using the image as training data; information indicating whether a ninth situation is present, in which a disposable tube set is attached in accordance with the mark on the main body, and a ninth learning model created by machine learning using the image as training data; information indicating whether a tenth situation is present, in which a tube and a centrifuge processing container are connected, and a tenth learning model created by machine learning using the image as training data; and information indicating whether an eleventh situation is present, in which the lid of the pump roller part and the lid of the centrifuge are closed, and an eleventh learning model created by machine learning using the image as training data.A 12th learning model created by machine learning using information indicating whether or not the 12th situation is one in which the tube is passed through the outside of the main body handle and attached, and the image as training data.

[0023] The first step of the present invention captures the situation where the four casters on the main body are locked. This is important for maintaining the stability of the device. This information is used as training data, and a first learning model is created through machine learning. In the second step, the power cord is plugged into the lower right side of the main body, the switch next to the cord plug is turned on, and the power is recognized. This is necessary to confirm the proper power supply of the device. Information on this situation is also used as training data, and a second learning model is created through machine learning. In the third step, the situation is confirmed where the main body is installed in a location with little pedestrian traffic. This is important to ensure a proper working environment for the device. This information is also used as training data, and a third learning model is created through machine learning. In the fourth step, the situation is recognized where operations are being performed according to the main body monitor. This confirms that the device is being operated normally. Information on this situation is used as training data for creating a fourth learning model. In the fifth step, the situation is confirmed where the centrifuge cover and the pump cover are open. This is to confirm that the preparation stage of the device is being performed correctly. This information is also used as training data, and a fifth learning model is created through machine learning. In the sixth step, it is confirmed that the disposable container is set, the disposable packaging is opened, and the centrifuge processing container is attached to the centrifuge rotor. This indicates that the device is properly set up to prepare for the procedure. This information is used as training data to create the sixth learning model. In the seventh step, it is confirmed that the waste liquid bag is hung on the back side of the main unit, and the centrifuge processing container is pushed in vertically and symmetrically and attached. This is to ensure proper waste management and accuracy of the procedure. This information is also used as training data to create the seventh learning model. In the eighth step, it is confirmed that the centrifuge processing container and the mounting base are set horizontally and all the way to the back with no gaps. This is important to ensure the safety and efficiency of the centrifuge process. This information is also used as training data to create the eighth learning model.The ninth step recognizes when the disposable tubing set is attached to the machine in line with the markings on the machine. This is to ensure that the machine is properly assembled and functions properly. This information is used as training data to create the ninth learning model. The tenth step recognizes when the tubing and centrifuge container are connected. This is important to ensure proper flow of fluid. This information is also used as training data to create the tenth learning model. The eleventh step recognizes when the pump roller lid and centrifuge lid are closed. This is necessary to ensure the safety of the centrifuge process. This information is used as training data to create the eleventh learning model. Finally, the twelfth step recognizes when the tubing is attached to the outside of the machine handle. This is important to ensure proper use of the machine. This information is also used as training data to create the twelfth learning model.

[0024] The system may acquire video in real time, and the auditing unit may audit the video in real time. The system may further include an alarm unit that issues an alarm when it is detected that the treatment is not being performed correctly.

[0025] The audit model used by the Audit Department is explained below. The audit model mainly uses deep learning techniques. This model judges the face band application process for each frame of video and evaluates whether the correct application process has been followed. Specifically, it uses a Convolutional Neural Network (CNN) to extract image features for each frame and a Long Short-Term Memory (LSTM) to determine temporal continuity.

[0026] Specifically, the model is generated using the following procedure: Data collection: Collect video data of correct and incorrect use of the faceband, preferably with labels for each step. Data preprocessing: Split the video into frames and organize each frame into a dataset along with its label. Model training: CNN is used to extract image features for each frame, and LSTM is used to learn temporal continuity. Model evaluation: The performance of the model is evaluated using a test dataset. Specific evaluation indicators include accuracy, precision, recall, and F1 score. Model application: The model will be applied to videos from actual medical settings to audit whether treatments are being performed properly.

[0027] [Technology 2] The invention is an AR (Augmented Reality) device that allows medical personnel to smoothly refer to manuals when performing procedures. This device is a headset or glasses type, and displays the manual instructions directly within the field of view of the medical personnel performing surgery. This makes it possible to check the manual more efficiently and quickly than conventional paper-based or electronic manual references.

[0028] The problem that this invention aims to solve is that it is difficult to refer to the manuals required for performing medical procedures, which hinders the smooth implementation of examinations and surgeries. In particular, some procedures are complicated, and if the appropriate manuals cannot be instantly checked, the quality and efficiency of the procedures may decrease.

[0029] The specific content of this invention is a system that utilizes AI and AR (Augmented Reality) technology to provide an appropriate manual for each procedure. When a medical professional starts a procedure, the AI ​​selects the manual corresponding to that procedure from a database and displays it in real time through an AR device (e.g., AR glasses or a head-mounted display). At that time, a voice recognition system is also used, and the medical professional can request the next procedure by voice.

[0030] The effect of this invention is that it becomes much easier to refer to the manual, which results in improved quality and efficiency of medical procedures. Specifically, by visually presenting the manual in real time using AR technology, examinations and surgeries can be performed smoothly. In addition, by using a voice recognition system, it becomes possible to refer to the manual and give instructions on the progress even when one's hands are full.

[0031] Additionally, the present invention can be used in education and training, where the system of the present invention can be used to train new medical personnel and potentially increase the speed at which they learn new procedures.

[0032] As described above, the present invention plays an important role in improving the quality and efficiency of medical treatments in beauty clinics.

[0033] [Technology 3] This invention is an AI audit system that integrates the recording of medical examinations and surgeries with the evaluation of compliance with the manual. This system records video of each stage of medical treatment, and then the AI ​​evaluates the degree of compliance with the manual. The AI ​​evaluation results are used as feedback for medical staff, contributing to the provision of higher quality medical services.

[0034] The problem that this invention aims to solve is the difficulty of properly evaluating whether medical procedures are being performed according to manuals and recording and analyzing the results over the long term. If this problem cannot be solved, it may be difficult to gain insight into improving the quality of medical procedures, preventing medical errors, and improving the education of medical professionals.

[0035] The specific content of this invention is a medical procedure recording and evaluation system that utilizes AI and video recording technology. The AI ​​engine analyzes high-resolution video feeds in real time and evaluates whether each step of the procedure or examination is performed according to the manual. This evaluation is performed automatically and generates a detailed report. The report is stored in an electronic medical history system for later reference and analysis.

[0036] The invention also records high-resolution video of surgical and examination procedures, which are then automatically tagged and evaluated by AI, allowing for the recording of small details and anomalies that may be overlooked during the medical procedure to be recorded and analysed later.

[0037] The benefits of this invention include quality assurance and continuous improvement of medical procedures. Automatic evaluation by AI and high-resolution video recording make it possible to check whether examinations and surgeries are being performed according to the manual and provide appropriate feedback. In addition, the records and evaluation results are stored in an electronic medical history system, making it possible to use them for long-term trend analysis and training material.

[0038] As described above, the present invention plays an important role in improving the efficiency of recording and evaluation of medical procedures in beauty clinics and in improving quality.

[0039] [Technology 4] This invention is a system that uses robotics technology to complement or replace medical procedures in beauty clinics. This system inputs manual information into AI, and the robot performs the medical procedure based on that information. For example, robots are used for procedures that require precision and consistency, such as injections, delicate surgery, and even the application of cosmetics. This makes it possible to standardize the quality of medical care and reduce the workload of medical professionals.

[0040] The problem that this invention aims to solve is that there are limitations to the diagnosis and surgery performed by human medical professionals, and it is difficult to ensure quality and consistency, especially when high precision and delicate operations are required.

[0041] The specific content of this invention is a robotic arm that inputs information from a manual and performs medical procedures as instructed, and an AI engine that controls its movements. First, information from the manual (e.g., surgical procedures, selection and use of surgical instruments, etc.) is input into the system. The AI ​​engine then analyzes the information and generates specific operation instructions for the robotic arm. The robotic arm performs medical procedures as instructed in the manual according to the instructions received from the AI ​​engine.

[0042] The effects of the present invention include improved quality and consistency of medical procedures. In particular, in surgeries that require delicate manipulation, the precise movements of the robot arm can achieve greater accuracy and consistency than human medical personnel. In addition, since the movements of the robot arm based on the manual are controlled by an AI engine, it is easy to respond to updates or changes to the manual.

[0043] Furthermore, the present invention contributes to reducing the workload of medical staff, particularly by reducing fatigue and stress caused by long surgeries and continuous medical examinations, thereby reducing the risk of medical errors.

[0044] As described above, the present invention plays an important role in improving the quality and consistency of medical procedures in beauty clinics and reducing the workload of medical staff.

[0045] Each of the above-mentioned functions (units) can be configured by, for example, hardware provided in a server device, a DSP (Digital Signal Processor), or software. For example, when configured by software, each of the above-mentioned function blocks 11 to 16 is actually configured with a CPU, RAM, ROM, etc. of a computer, and is realized by the operation of a program stored in a recording medium such as the RAM, ROM, hard disk, or semiconductor memory.

[0046] The present invention can include the following four configuration groups.

[0047] [Technology 1] <Item 1> A system that monitors whether medical procedures are being performed appropriately according to a pre-set manual and issues an alarm if inappropriate procedures are detected, A step of storing a standard manual for various medical procedures in beauty clinics in a database; A step of analyzing real-time data acquired during medical procedures and evaluating whether the procedures are being performed in accordance with a database manual; The system has a step of issuing an alarm if a deviation from the manual is detected. <Item 2> In the system of claim 2, the alarm is a combination of visual and audio signals. <Item 3> In the system described in item 1 or 2, the acquisition of real-time data includes video data from a head-mounted display, biometric information from a biosensor, and operation information of a surgical instrument. <Item 4> In the system described in any one of items 1 to 3, the analysis of the real-time data is performed by an AI engine. <Item 5> In the system described in any one of items 1 to 4, the manual describes detailed procedures for various cosmetic treatments and surgical procedures. <Item 6> In the system described in any one of items 1 to 5, the alarm is issued immediately when an action that does not conform to the manual is detected. <Item 7> In the system described in any one of items 1 to 6, after the alarm is issued, the contents of the alarm are stored as a log for review. <Item 8> In the system described in any one of items 1 to 7, if an action that does not conform to the manual continues, a continuous alarm is issued. <Item 9> In the system described in any one of items 1 to 8, the alarm issues different levels of warning depending on the type of violation and its importance. <Item 10> In the system according to any one of items 1 to 9, the manual is periodically updated, and the system performs evaluation based on the latest manual.

[0048] [Technology 2] <Item 1> A system that presents manuals as needed to medical professionals so that they can smoothly perform medical procedures, analyzing the healthcare professional's voice input and identifying an appropriate section of the manual; The system includes the step of immediately presenting the identified section of the manual. <Item 2> In the system described in item 1, the voice input analysis is performed using natural language processing techniques. <Item 3> In the system described in item 1 or 2, the voice input is analyzed not only for the spoken words of the medical professional but also for specific voice commands. <Item 4> In the system described in any one of items 1 to 3, the section of the manual is displayed on a device such as a head mounted display, a tablet, or a smartphone. <Item 5> In the system according to any one of items 1 to 4, when the identified section of the manual is displayed, necessary illustrations and related information are also displayed. <Item 6> In the system according to any one of items 1 to 5, sections of the manual are scrolled according to voice commands of a medical professional. <Item 7> In the system according to any one of items 1 to 6, past patient data and treatment results related to the section of the displayed manual can be referenced. <Item 8> In the system according to any one of items 1 to 7, when a section of the manual is displayed, a specific case study of a patient suitable for that section is also presented. <Item 9> In the system described in any one of items 1 to 8, the manual is periodically updated, and the system operates based on the latest manual. <Item 10> In the system described in any one of items 1 to 9, when a section of the manual is displayed, the document is summarized and key points are highlighted to make it easier for medical professionals to understand.

[0049] [Technology 3] <Item 1> A system for automating the recording and evaluation of medical examinations and surgical procedures in a beauty clinic, comprising: obtaining and storing a high resolution video recording of the medical procedure; analyzing the video recording with an AI engine to evaluate whether the actions were performed according to the manual; The system includes storing the results of the assessment in an electronic medical history system. <Item 2> In the system described in item 1, the real-time video and audio are recorded using a head-mounted display, a body camera, or the like. <Item 3> In the system described in item 1 or 2, the video and audio recording covers all medical procedures. <Item 4> In the system described in any one of items 1 to 3, the analysis of the real-time video and audio data is performed using an AI engine. <Item 5> In the system according to any one of items 1 to 4, the analysis result indicates whether or not an action consistent with a specific manual has been performed. <Item 6> In the system according to any one of items 1 to 5, after the medical procedure is completed, the recorded data is stored as a log for evaluation and review. <Item 7> In the system according to any one of items 1 to 6, the evaluation and review are performed by the medical professional himself, a colleague, or a superior. <Item 8> In the system described in any one of items 1 to 7, the results of the evaluation and review are used as feedback for improvement or training material. <Item 9> In the system according to any one of items 1 to 8, the manual is periodically updated, and the system performs evaluation based on the latest manual. <Item 10> In the system described in any one of items 1 to 9, the AI ​​engine analyzes video and audio data using a machine learning algorithm including deep learning.

[0050] [Technology 4] A system for performing medical procedures according to a manual using a robot arm, inputting medical procedure manual information into the system; The AI ​​engine analyzes the manual information and generates operation commands for the robot arm; A system comprising a step in which a robotic arm performs a medical procedure according to the generated motion command. <Item 2> In the system described in item 1, the robot arm is capable of movement with six or more degrees of freedom. <Item 3> In the system described in item 1 or 2, the robot arm has a high-precision position adjustment function for accurately reproducing specific treatment procedures. <Item 4> In the system described in any one of items 1 to 3, the robot arm detects the body movement of the patient and adjusts its position accordingly. <Item 5> In the system described in any one of items 1 to 4, the robot arm utilizes real-time sensor feedback to respond to various situations that arise during treatment. <Item 6> In the system described in any one of items 1 to 5, the robot arm has an advanced control algorithm for performing delicate medical procedures. <Item 7> In the system described in any one of items 1 to 6, the AI ​​engine controls the robot arm by utilizing a machine learning algorithm including deep learning. <Item 8> In the system described in any one of items 1 to 7, the robot arm can be remotely operated and can be monitored and operated in real time by a specialized medical professional. <Item 9> In the system described in any one of items 1 to 8, the manual is periodically updated, and the system controls the robot arm based on the latest manual. <Item 10> In the system described in any one of items 1 to 9, the operation results of the robot arm are recorded and stored for later review and feedback.

[0051] The above-described embodiment is merely an example for facilitating understanding of the present invention, and is not intended to limit the present invention. The present invention can be modified or improved without departing from the spirit of the present invention, and it goes without saying that the present invention includes equivalents thereof.

Claims

1. A medical audit system for recording and auditing treatments at beauty clinics, comprising: In the images taken during the treatment, A first learning model created by machine learning using information indicating whether or not the first situation is in which the four casters on the main body are locked and the image as training data; a second learning model created by machine learning using as training data information indicating whether or not a second situation is occurring, in which the power cord is plugged into the lower right side of the main body, the switch next to the power cord plug is turned on, and the image; and A third learning model created by machine learning using information indicating whether the main body is installed in a location with little pedestrian traffic or not and the image as training data; and a fourth learning model created by machine learning using information indicating whether or not a fourth situation in which an operation is being performed according to a body monitor and the image as training data; a fifth learning model created by machine learning using information indicating whether or not a fifth situation in which the centrifuge cover and the pump cover are open and the image as training data; and a sixth learning model created by machine learning using as training data information indicating whether or not a sixth situation is occurring, in which the disposable container is set, the disposable container packaging is opened, and the centrifuge processing container is attached to the centrifuge rotor, and the image; and a seventh learning model created by machine learning using information indicating whether or not a seventh situation is occurring, in which a waste liquid bag is hung on the rear side of the main body and the centrifuge processing container is pushed in vertically and symmetrically and attached, and the image as training data; and an eighth learning model created by machine learning using information indicating whether or not an eighth situation is present in which there is no gap between the centrifuge processing container and the mounting base and the image is used as training data; and A ninth learning model created by machine learning using information indicating whether or not the disposable tube set is attached in accordance with the mark on the main body and the image as training data; A tenth learning model created by machine learning using information indicating whether or not a tenth situation in which the tube and the centrifuge processing container are connected and the image as training data; An eleventh learning model created by machine learning using information indicating whether or not an eleventh situation in which the lid of the pump roller part and the lid of the centrifuge are closed and the image as training data; A 12th learning model created by machine learning using information indicating whether or not a 12th situation in which a tube is passed through and attached to the outside of the main body handle and the image as training data; and A learning model storage unit that stores the An acquisition unit for acquiring a video of the treatment; An auditing unit that audits whether the treatment is being performed correctly by providing the video to the first to twelfth learning models and determining whether the video is in the first to twelfth situations; A medical audit system that includes:

2. 2. The audit system according to claim 1, The acquisition unit acquires the video moving image in real time, The auditing unit audits the video footage in real time; The medical audit system further comprises an alarm unit that issues an alarm if it is detected that the treatment method is not being performed correctly.