Electrocardio diagnosis system and device, storage medium and program product

Abstract

The invention provides an electrocardio diagnosis system and device, a storage medium and a program product. The system comprises an electrocardio data acquisition device; the diagnosis terminal equipment is used for realizing the following functions: an AI diagnosis module used for performing AI diagnosis on a to-be-diagnosed person according to the acquired electrocardiogram data and outputting an acquired quality analysis conclusion, a disease diagnosis conclusion and corresponding two-dimensional basic confidence; the dynamic confidence coefficient calculation module is used for determining the dynamic confidence coefficient matched with the current scene according to the output of the AI diagnosis module; and the flow control module is used for comparing the dynamic confidence coefficient with a confidence coefficient threshold value preset for the current scene by each processing flow, and triggering the corresponding processing flow when the dynamic confidence coefficient is greater than or equal to the confidence coefficient threshold value preset for the current scene. By means of the technical scheme, the dynamic confidence coefficient and the confidence coefficient threshold value based on clinical scene adaptation are achieved, the accuracy of the AI diagnosis result and the processing timeliness and efficiency are improved, and the false positive rate is reduced.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, and in particular to an electrocardiogram (ECG) diagnostic system, device, storage medium, and program product. Background Technology

[0002] With the gradual penetration of artificial intelligence technology into the medical field, AI-powered electrocardiogram (ECG) diagnostic systems have begun to be applied in clinical ECG examination procedures. These systems can quickly analyze ECG waveform data and output automatic diagnostic conclusions, providing auxiliary support for doctors' diagnoses and playing a role in improving diagnostic efficiency. However, current AI-powered ECG diagnostic systems still suffer from inaccurate diagnostic results due to issues such as fixed confidence thresholds, neglect of clinical scenario differences, and a disconnect between data acquisition quality requirements and the specific scenarios. These limitations hinder the full realization of their clinical application value. Summary of the Invention

[0003] To address the aforementioned problems in the prior art, embodiments of the present invention provide an electrocardiogram (ECG) diagnostic system, device, storage medium, and program product.

[0004] To achieve the above objectives, on the one hand, an electrocardiogram (ECG) diagnostic system is provided, comprising: Electrocardiogram (ECG) data acquisition equipment is used to collect ECG data from individuals undergoing diagnosis. A diagnostic terminal device, connected to the electrocardiogram (ECG) data acquisition device, receives the ECG data, and includes a memory and a processor. The memory stores at least one program, which is executed by the processor to implement the following modules: The AI ​​diagnostic module is used to perform AI diagnosis on the person to be diagnosed based on the collected electrocardiogram data, and output the disease diagnosis conclusion and the corresponding basic confidence level of the disease diagnosis, as well as the collection quality analysis conclusion and the corresponding basic confidence level of the collection quality. The dynamic confidence calculation module is used to determine the dynamic confidence level adapted to the current scenario based on the output of the AI ​​diagnosis module. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season. Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the data acquisition quality baseline confidence level is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the disease diagnosis baseline confidence level is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; The process control module is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario. The corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

[0005] Preferably, the electrocardiogram diagnostic system pre-sets weighting coefficients through the following steps: Based on the age of the individuals to be diagnosed, they are divided into multiple age groups, and pre-set parameters according to the age group they belong to. The value of corresponds to the age range; the higher the age, the better. The larger the value of ; Pre-set based on the patient's medical history The values ​​for are as follows, among which those with a history of atrial fibrillation or myocardial infarction are considered. Value > when there is an underlying disease Value > when there is no relevant medical history value; Preset according to the current season The values ​​of , among which, the seasons with high incidence rates The value is higher than that of the season with a low incidence rate. value.

[0006] Preferably, the electrocardiogram diagnostic system pre-sets weighting coefficients through the following steps: When the age is within the predetermined childhood and adolescence stage, When the age is within the predetermined youth stage, When the age is within the predetermined middle age, When the age is in the predetermined old age stage, When the age is in the predetermined advanced age stage, ; When there is no relevant medical history =1.0; when there are underlying medical conditions =1.1; in cases with a history of atrial fibrillation or myocardial infarction, =1.3; The underlying conditions include: hypertension or diabetes; Spring γ=1.0; Summer γ=0.95; Autumn γ=1.0; Winter γ=1.15.

[0007] Preferably, in the electrocardiogram diagnostic system, the dynamic confidence calculation module determines whether the current process is in the quality inspection process or a diagnostic process other than the quality inspection process based on the conclusion type currently output by the AI ​​diagnostic module.

[0008] Preferably, the ECG diagnostic system uses different confidence thresholds for different processes when determining whether to trigger the examination quality reminder process, clinical information improvement process, critical value warning process, or normal report grouping process.

[0009] Preferably, the ECG diagnostic system, for each of the examination quality reminder process, clinical information improvement process, critical value warning process, or normal report grouping process, the step of obtaining the preset confidence threshold for the current scenario includes: the confidence threshold pre-adjusted according to the respective preset basic confidence threshold.

[0010] Preferably, the electrocardiogram diagnostic system acquires quality analysis conclusions including one or more of the following: electromyographic interference, power frequency interference, baseline drift, good or qualified electrode contact; for the examination quality process, when the acquisition quality analysis conclusion is electromyographic interference or power frequency interference and the corresponding dynamic confidence level is greater than or equal to the preset acquisition quality reminder confidence level threshold for the current scenario, the examination quality reminder process is triggered.

[0011] Preferably, the electrocardiogram diagnostic system provides the following disease diagnosis conclusions: normal, belonging to a predetermined non-critical value category, and belonging to a predetermined critical value abnormality category. The predetermined non-critical value category includes: arrhythmia; the predetermined critical value abnormality category includes: malignant arrhythmia, acute myocardial infarction, and severe hyperkalemia or hypokalemia. Acute myocardial infarction is classified as a myocardial infarction-related critical value abnormality. When the disease diagnosis conclusion is normal and the corresponding dynamic confidence level is greater than the preset confidence level threshold for normal report grouping for the current scenario, the normal report grouping process is triggered. When the disease diagnosis conclusion belongs to the predetermined non-critical value diagnosis conclusion and the corresponding dynamic confidence level is greater than the preset clinical information improvement confidence level threshold for the current scenario, the clinical information improvement process is triggered. When the disease diagnosis conclusion belongs to the predetermined critical value abnormality diagnosis conclusion and the corresponding dynamic confidence level is greater than the critical value warning confidence level threshold preset for the current scenario, the critical value warning process is triggered.

[0012] Preferably, in the electrocardiogram diagnostic system, when the age of the person to be diagnosed is 0-14 years old, if the disease diagnosis conclusion is an abnormality of critical value for myocardial infarction, the corresponding dynamic confidence level is directly set to 0, without triggering the critical value warning process.

[0013] Preferably, the ECG diagnostic system, in obtaining the preset confidence threshold for the current scenario, includes one or more of the following steps: The baseline confidence threshold for data collection quality alerts is 70%; when the patient to be diagnosed is a child or the current season is winter, the baseline confidence threshold for data collection quality alerts is lowered by a first predetermined percentage; the first predetermined percentage is less than or equal to 5%. The baseline confidence threshold for complete clinical information is 80%. When the patient to be diagnosed has a medical history, the baseline confidence threshold for complete clinical information is increased by a second predetermined percentage; the second predetermined percentage is less than or equal to 5%. The baseline confidence threshold for critical value warning is 60%. When the person to be diagnosed is in a predetermined stage of old age or the current season is winter, the baseline confidence threshold for critical value warning is lowered by a third predetermined percentage; the third predetermined percentage is less than or equal to 5%. The baseline confidence threshold for the normal reporting group is 90%. When the person to be diagnosed is in the predetermined youth stage, the baseline confidence threshold for the normal reporting group is lowered by a fourth predetermined percentage, which is less than or equal to 2%. When the person to be diagnosed is in the predetermined old age stage, the baseline confidence threshold for the normal reporting group is raised by a fifth predetermined percentage, which is less than or equal to 2%.

[0014] On the other hand, an electrocardiogram (ECG) diagnostic device is provided, including a memory and a processor, the memory storing at least one program, the at least one program being executed by the processor to perform the following steps: Based on the obtained electrocardiogram (ECG) data, AI diagnosis is performed on the individuals to be diagnosed, and the disease diagnosis conclusion and corresponding basic confidence level of the disease diagnosis are output, as well as the acquisition quality analysis conclusion and corresponding basic confidence level of the acquisition quality. Dynamic confidence calculation is used to determine a dynamic confidence level that fits the current scenario based on the AI ​​diagnosis output. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season, wherein: Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the data acquisition quality baseline confidence level is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the disease diagnosis baseline confidence level is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; Process control is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario; the corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

[0015] In practice, the electrocardiogram (ECG) diagnostic device is the aforementioned diagnostic terminal device. This ECG diagnostic device is used to execute the steps of the aforementioned diagnostic terminal device.

[0016] In another aspect, a computer-readable storage medium is also provided, wherein at least one program is stored therein, the at least one program being executed by a processor to perform the following steps: Based on the obtained electrocardiogram (ECG) data, AI diagnosis is performed on the individuals to be diagnosed, and the disease diagnosis conclusion and corresponding basic confidence level of the disease diagnosis are output, as well as the acquisition quality analysis conclusion and corresponding basic confidence level of the acquisition quality. Dynamic confidence calculation is used to determine a dynamic confidence level that fits the current scenario based on the AI ​​diagnosis output. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season, wherein: Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the data acquisition quality baseline confidence level is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the disease diagnosis baseline confidence level is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; Process control is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario; the corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

[0017] Preferably, the storage medium also performs the other steps described above performed by the diagnostic terminal device.

[0018] In another aspect, a computer program product is also provided, comprising a computer program that, when executed by a processor, performs the following steps: Based on the obtained electrocardiogram (ECG) data, AI diagnosis is performed on the individuals to be diagnosed, and the disease diagnosis conclusion and corresponding basic confidence level of the disease diagnosis are output, as well as the acquisition quality analysis conclusion and corresponding basic confidence level of the acquisition quality. Dynamic confidence calculation is used to determine a dynamic confidence level that fits the current scenario based on the AI ​​diagnosis output. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season, wherein: Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the data acquisition quality baseline confidence level is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the disease diagnosis baseline confidence level is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; Process control is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario; the corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

[0019] Preferably, the program product also performs the other steps described above performed by the diagnostic terminal device.

[0020] The above technical solution has the following technical effects: The technical solution of this invention quantifies the impact of age, medical history, and season on disease incidence and data collection quality, formulates differentiated data collection quality requirements and process triggering rules, and achieves dynamic confidence based on clinical scenarios. This enables precise linkage between "scenario-quality-confidence-process," improving the accuracy and clinical applicability of AI diagnosis. It also solves the problem of insufficient accuracy in triggering AI diagnosis processes caused by existing solutions relying on a single data dimension and lacking scenario-based algorithm support, which prevents the scientific and dynamic adjustment of confidence.

[0021] In a further embodiment, a dynamic confidence threshold based on clinical scenarios is implemented, which further improves the accuracy of AI diagnostic results, the timeliness of critical value abnormalities, the efficiency of doctor processing, and reduces the false positive rate. It is applicable to ECG examination departments and regional ECG diagnostic centers in medical institutions at all levels, and can achieve precise ECG diagnostic process management for different age, medical history and seasonal scenarios. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the structure of an electrocardiogram diagnostic system according to an embodiment of the present invention; Figure 2 A schematic diagram illustrating the workflow for diagnosis using the system of this embodiment; Figure 3 The following is an example of the implementation process of the inspection quality reminder process using the system of this embodiment of the invention; Figure 4 An example interface for providing quality inspection alerts; Figure 5 This is a schematic diagram illustrating the implementation process of the system for improving clinical information using an embodiment of the present invention. Figure 6 This is a schematic diagram of an exemplary clinical information completion interface; Figure 7 This is a schematic diagram illustrating the implementation process of the critical value early warning procedure using the system of this embodiment of the invention; Figure 8 This is a schematic diagram illustrating the implementation process of a normal report grouping process using a system based on an embodiment of the present invention. Detailed Implementation

[0023] To further illustrate the various embodiments, the present invention provides accompanying drawings. These drawings are part of the disclosure of the present invention, primarily used to illustrate the embodiments and to explain the operating principles of the embodiments in conjunction with the relevant descriptions in the specification. With reference to these drawings, those skilled in the art should be able to understand other possible implementations and the advantages of the present invention. Components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0024] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments.

[0025] In the process of implementing the electrocardiogram AI diagnosis system business, the inventors of this application found that the existing technology uses a unified and fixed AI confidence threshold to trigger the diagnosis process, without considering the core clinical influencing factors: in terms of age, the incidence of diseases such as myocardial infarction and atrial fibrillation increases significantly with age (the incidence of myocardial infarction in people over 65 years old is 8 times that of people under 45 years old), and children basically have no risk of myocardial infarction; in terms of medical history, patients with a history of hypertension and diabetes have a higher demand for detecting arrhythmias; in terms of season, the incidence of myocardial infarction in winter is 37% higher than that in summer, and low temperature is likely to increase the interference in electrocardiogram signal acquisition. The fixed threshold leads to a high risk of missed diagnosis in elderly high-risk populations and a high false positive rate in low-risk populations such as children. In addition, the data acquisition quality requirements of the existing technology are disconnected from the scenarios: the quality of electrocardiogram signal acquisition in different scenarios is affected by different interference factors. For example, children are prone to movement, resulting in electromyogram interference, and thick winter clothes lead to poor electrode contact. The existing technology does not formulate differentiated acquisition quality standards for scenarios and only relies on AI post-processing, resulting in the proportion of low-quality data reaching 15%-20%, seriously affecting the diagnostic accuracy.

[0026] Embodiment 1: Figure 1 It is a schematic structural diagram of the electrocardiogram diagnosis system according to an embodiment of the present invention. As Figure 1 , the electrocardiogram diagnosis system of this embodiment includes: An electrocardiogram data acquisition device, which is used to acquire the electrocardiogram data of the person to be diagnosed; A diagnostic terminal device, connected to the electrocardiogram data acquisition device, receives the electrocardiogram data, and includes: a memory and a processor. The memory stores at least one program, and at least one program is executed by the processor to implement the following modules: An AI diagnosis module, which is used to perform AI diagnosis on the person to be diagnosed according to the acquired electrocardiogram data, and output a disease diagnosis conclusion and the corresponding disease diagnosis basic confidence level, and a collection quality analysis conclusion and the corresponding collection quality basic confidence level; in a specific implementation, this module is implemented through the existing Cardio AI module (Medical Device Registration Certificate Number: National Medical Device Registration Approval 20223211426); the confidence level ranges from 0-100%, and the higher the value, the stronger the reliability of the AI analysis result; among them, the collection quality analysis conclusion is the AI analysis result of the quality of the currently acquired electrocardiogram data; the disease diagnosis conclusion is the AI analysis result of disease diagnosis for the currently acquired electrocardiogram data; A dynamic confidence level calculation module, which is used to determine the dynamic confidence level adapted to the current scenario according to the output of the AI diagnosis module. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season, where: Dynamic confidence level = basic confidence level × × × ; Specifically, for the inspection quality process, the basic confidence level of data acquisition quality is used to calculate the dynamic confidence level; for the pre-set diagnostic process other than the inspection quality process, the basic confidence level of disease diagnosis is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; The process control module compares the dynamic confidence level with the preset confidence threshold for each processing step in the current scenario. When the dynamic confidence level is greater than or equal to the preset confidence threshold for the current scenario, the corresponding processing step is triggered. The corresponding processing steps include one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process. These processing steps are the corresponding processing steps that may be triggered subsequently. Preferably, the confidence thresholds for each processing step are different, i.e., a differentiated trigger confidence threshold setting is adopted. In one specific implementation, the base confidence thresholds for each processing step are different. The preset confidence threshold for each processing step in the current scenario is obtained by adjusting its respective base confidence threshold according to the current scenario.

[0027] In a specific implementation, the diagnostic terminal device includes a display for displaying an electrocardiogram and providing an interface for operation or control by an operator, such as a user interface or control buttons within the interface.

[0028] In one specific implementation, the weighting coefficients are pre-set through the following steps: Based on the age of the individuals to be diagnosed, they are divided into multiple age groups, and pre-set parameters according to the age group they belong to. The value of corresponds to the age range; the higher the age, the better. The larger the value of ; Pre-set based on the patient's medical history The values ​​for are as follows, among which those with a history of atrial fibrillation or myocardial infarction are considered. Value > when there is an underlying disease Value > when there is no relevant medical history value; Preset according to the current season The values ​​of , among which, the seasons with high incidence rates The value is higher than that of the season with a low incidence rate. value.

[0029] In one specific implementation, the dynamic confidence calculation module determines whether the current AI diagnosis process is a quality control process or a diagnosis process other than a quality control process based on the type of conclusion currently output by the AI ​​diagnosis module. In another specific implementation, the AI ​​diagnosis process is determined based on whether the conclusion was received after ECG data acquisition and before disease diagnosis. If so, the current AI diagnosis process is a quality control process; otherwise, it is a disease diagnosis process other than a quality control process.

[0030] Figure 2 This is a schematic diagram illustrating the workflow for diagnosis using the system of this embodiment of the invention. Figure 2 The steps for diagnosis using the system of this embodiment include: Based on the current scenario, input the preset information of the person to be diagnosed, such as the patient's age, medical history, seasonal weighting coefficient, and the preset basic confidence threshold for each diagnostic trigger process. Perform electrocardiogram (ECG) data acquisition and upload the acquired ECG data to the diagnostic terminal device; The AI ​​diagnostic module outputs corresponding data acquisition quality analysis conclusions and disease diagnosis conclusions, as well as corresponding two-dimensional basic confidence scores. The two-dimensional basic confidence scores include: data acquisition quality basic confidence score Cbasic-quality and disease diagnosis basic confidence score Cbasic-disease. Using the dynamic confidence calculation model described above, select the corresponding baseline confidence level according to the diagnostic process, and calculate the corresponding dynamic confidence level Cdynamic. The dynamic value C is compared with the preset scenario-based threshold, which is the confidence threshold preset for each process to be triggered for the current scenario. When the dynamic value C is greater than or equal to the aforementioned confidence threshold preset for the current scenario, the corresponding processing process is triggered. The system executes the triggered process and obtains clinical feedback, iteratively optimizing the weighting coefficients for dynamic confidence calculation. In one specific implementation, the doctor executes the early warning process and issues a diagnostic conclusion. The system records the associated data of "scenario parameters - C - diagnostic results," where C represents dynamic confidence. The weighting coefficients are iteratively optimized based on historical data according to a predetermined cycle, such as quarterly. For example, if the incidence of myocardial infarction is higher in a certain region during winter, γ can be adjusted to 1.2.

[0031] Example 2: Based on Example 1, this embodiment sets different confidence thresholds (i.e., trigger thresholds) for different processes when determining whether to trigger the examination quality reminder process, clinical information improvement process, critical value warning process, or normal report grouping process.

[0032] In one specific implementation, for each of the inspection quality reminder process, clinical information improvement process, critical value warning process, or normal report grouping process, the step of obtaining the preset confidence threshold for the current scenario includes: the confidence threshold pre-adjusted according to the respective preset basic confidence threshold.

[0033] In one specific implementation, the acquisition quality analysis conclusion includes one or more of the following: electromyography interference, power frequency interference, baseline drift, good or qualified electrode contact; for the inspection quality process, when the acquisition quality analysis conclusion is electromyography interference or power frequency interference and the corresponding dynamic confidence level is greater than or equal to the preset acquisition quality reminder confidence level threshold for the current scenario, the inspection quality reminder process is triggered.

[0034] In one specific implementation, the disease diagnosis conclusion includes: normal, a conclusion belonging to the predetermined non-critical value category, and a conclusion belonging to the predetermined critical value abnormality category; the predetermined non-critical value category includes: arrhythmia; the predetermined critical value abnormality category includes: malignant arrhythmia, acute myocardial infarction, severe hyperkalemia or hypokalemia, and other critical value abnormalities; among which, acute myocardial infarction belongs to the myocardial infarction category of critical value abnormalities; wherein: When the disease diagnosis conclusion is normal and the corresponding dynamic confidence level is greater than the preset normal report grouping confidence level threshold for the current scenario, the normal report grouping process is triggered; in one specific implementation, the normal report grouping process includes: assigning the patient to the primary physician group; When the disease diagnosis conclusion belongs to the predetermined non-critical value diagnosis conclusion and the corresponding dynamic confidence level is greater than the preset clinical information improvement confidence level threshold for the current scenario, the clinical information improvement process is triggered. In one specific implementation, the clinical information improvement process includes: pop-up reminder; for example, a corresponding clinical information improvement window pops up for the doctor to supplement the relevant information of the patient to be diagnosed; for example, it may include: patient's chief complaint, past medical history, contact information, address, etc. When the disease diagnosis conclusion belongs to the predetermined critical value abnormality diagnosis conclusion and the corresponding dynamic confidence level is greater than the critical value warning confidence level threshold preset for the current scenario, the critical value warning process is triggered. In one specific implementation, the critical value warning process includes warnings in multiple ways, such as pop-up windows, text messages, and pinning to the top, to prompt doctors to handle or treat the disease in a timely manner.

[0035] In one specific implementation, the weighting coefficients in the dynamic confidence calculation are iteratively optimized based on clinical feedback data.

[0036] In one specific implementation, when the age of the person to be diagnosed is 0-14 years old, and the disease diagnosis conclusion is an abnormal critical value for myocardial infarction, the corresponding dynamic confidence is directly set to 0, and the critical value warning process is not triggered.

[0037] In one specific implementation, the steps for obtaining the preset confidence threshold for the current scene include one or more of the following: The baseline confidence threshold for data collection quality alerts is 70%; when the patient to be diagnosed is a child or the current season is winter, the baseline confidence threshold for data collection quality alerts is lowered from 70% by a first predetermined percentage; the first predetermined percentage is less than or equal to 5%. The baseline confidence threshold for complete clinical information is 80%. When the patient to be diagnosed has a medical history, the baseline confidence threshold for complete clinical information is increased by a second predetermined percentage; the second predetermined percentage is less than or equal to 5%. The baseline confidence threshold for critical value warning is 60%. When the patient to be diagnosed is in the predetermined old age stage or the current season is winter, the baseline confidence threshold for critical value warning is lowered by a third predetermined percentage; the third predetermined percentage is less than or equal to 5%. The baseline confidence threshold for the normal reporting group is 90%. When the patient to be diagnosed is in the predetermined youth stage, the baseline confidence threshold for the normal reporting group is lowered by a fourth predetermined percentage, which is less than or equal to 2%. When the patient to be diagnosed is in the predetermined old age stage, the baseline confidence threshold for the normal reporting group is raised by a fifth predetermined percentage, which is less than or equal to 2%. Example 3: Based on Example 1 or Example 2, this embodiment establishes the following quantitative indicator system for weighting coefficients according to three core scenario factors: age, medical history, and season: Age-stratified quantification: Based on the disease incidence pattern, the disease is divided into 5 groups, each with an age-weighted coefficient. For children and adolescents aged 0-18, the risk of myocardial infarction is approximately 0. =0.8; 18-44 years old, youth. =0.9; 45-64 years old, middle-aged. =1.0; 65-79 years old, elderly. =1.2; Over 80 years old, advanced age, =1.3; Medical history grading and quantification: Based on the correlation between underlying diseases and electrocardiographic disorders, medical history is divided into three levels, and a weighting coefficient is assigned to each level. No relevant medical history =1.0; has underlying conditions such as hypertension or diabetes. =1.1; history of atrial fibrillation or myocardial infarction, =1.3; Seasonal adaptation quantification: Based on the collection quality and incidence rate, the data is divided into 4 seasons, and a seasonal weighting coefficient γ is assigned: Spring, γ=1.0, Summer, γ=0.95, Autumn, γ=1.0, Winter, γ=1.15, with a high incidence of myocardial infarction and significant collection interference.

[0038] The above coefficients are the system default configurations and can be adjusted according to the seasonality and incidence rate characteristics of different regions.

[0039] In one specific implementation, when using the dynamic confidence calculation model described above to calculate the dynamic confidence, the model constraint rules include: if the age group is 0-14 years old, the dynamic confidence of myocardial infarction diagnosis is directly set to 0, and the critical value process is not triggered; atrial fibrillation and other diseases that children may develop are calculated normally.

[0040] Table 1 shows the basic trigger thresholds, i.e., basic confidence thresholds, and corresponding scenario-based adjustment instructions set for the four types of processes to be triggered in this embodiment of the present invention.

[0041] Table 1:

[0042] The above groups are categorized based on the age, medical history, and current season of the person or patient to be diagnosed.

[0043] The specific percentage for adjusting the confidence threshold based on the current scenario mentioned above is an example. In other implementations, the corresponding percentage for adjustment can be determined based on local conditions such as age, medical history, and the relationship between season and incidence rate.

[0044] The following provides specific examples of using the system of this invention for diagnosis for each of the four types of triggering processes.

[0045] Example 1: Inspection Quality Reminder - Child's Winter Medical Visit Scenario Specific scenario: 6-year-old child - hyperactivity leading to electromyography interference, seeking medical attention in winter - heavy clothing increases interference, no underlying medical history, routine electrocardiogram examination.

[0046] Figure 3 This example demonstrates the implementation process of a quality inspection reminder procedure using the system of an embodiment of the present invention. Figure 3 The implementation process includes: Configure the weight coefficients on the corresponding interface of the device. Since it's for a child, the configuration... =0.8; Due to the lack of medical history, configuration =1.0; Due to the winter season, the basic threshold for inspection quality is set to 65%; Collect electrocardiogram (ECG) data and upload it to the AI ​​diagnostic module; The AI ​​diagnostic module outputs an examination quality analysis conclusion of "borderline electromyography interference," with a corresponding baseline confidence level of acquisition quality Cbaseline - quality = 75%; the disease diagnosis conclusion is "sinus rhythm," with a corresponding baseline confidence level of disease diagnosis Cbaseline - disease = 90%. Calculate the dynamic confidence level Cdynamic for the inspection quality alert process: 75% × 0.8 × 1.0 × 1.15 = 69%; 69% 65%, triggering a quality alert process: a corresponding preset warning icon and / or text pops up; for example, such as Figure 4 As shown, the ECG interface outputs "Warning: The current ECG is suspected of having artifacts, which may affect the diagnostic results" and provides a corresponding "Reacquire" button. In other implementations, users can also select "Cancel Acquisition" or "Save Still" through the corresponding buttons. Based on the above warning, the ECG signal was reacquired. This time, the output acquisition quality analysis result was qualified and Cbaseline - quality = 90%. Calculate C dynamics: 90% × 0.8 × 1.0 × 1.15 = 82.8%; To proceed with the subsequent disease diagnosis process.

[0047] The effects of using the system of this invention include: combining confidence calculation with children and high-interference winter scenarios to identify critical quality data in advance; improving data accuracy by 25% after resampling; and avoiding misdiagnosis caused by interference.

[0048] Example 2: Clinical Information Completion Process - Mid-aged Patient's Basic Medical History Scenario Specific scenario: A 55-year-old middle-aged male with a history of hypertension (β=1.1) visited the clinic in the spring (γ=1.0), complaining of occasional palpitations. The initial AI diagnosis suggested myocardial infarction.

[0049] Figure 5 This example demonstrates the implementation process of a clinical information improvement workflow using the system of an embodiment of the present invention. Figure 5 The implementation process includes: Configure the weight coefficients on the corresponding interface of the device. Since it's for a middle-aged person, the configuration... =1.0; Due to a history of hypertension, configuration =1.1; Due to the spring season, the baseline confidence threshold for complete clinical information configuration is 85%; Collect electrocardiogram (ECG) data and upload it to the AI ​​diagnostic module; The AI ​​diagnostic module outputs a quality analysis conclusion of "qualified," with a baseline confidence level of 90% (Cbaseline - quality) for the data acquisition quality; the disease diagnosis conclusion is "myocardial infarction," with a baseline confidence level of 82% (Cbaseline - disease). Calculate the dynamic confidence level Cdynamic for the disease diagnosis process: 82% × 1.0 × 1.1 × 1.0 = 90.2%; 90.2% 85%, triggering the clinical information completion process: a corresponding interface pops up for the doctor to complete the clinical information; for example, Figure 6 The example shown is a clinical information completion interface for doctors to complete: the patient's chief complaint and past medical history; After reviewing the doctor's supplementary information, apply for a diagnosis; Based on the supplementary information, the diagnosing doctor determined it to be a myocardial infarction, with an accuracy rate of 95%.

[0050] The effects of using the system of this invention include: raising the threshold for patients with underlying medical history, accurately triggering the supplementation of key information, improving the diagnostic accuracy by 13% after combining the information, and avoiding missed diagnosis of the cause of the disease.

[0051] Example 3: Critical Value Early Warning Procedure - High-Risk Scenarios for the Elderly in Winter Specific scenario: An 82-year-old male with a 10-year history of hypertension (β=1.1) experiences sudden chest pain in the early morning of winter (γ=1.15), requiring rapid identification of critical values ​​for myocardial infarction.

[0052] Figure 7 This example demonstrates the implementation process of a critical value early warning procedure using a system from an embodiment of the present invention. For instance... Figure 7 The implementation process includes: Configure the weight coefficients in the corresponding interface of the device. Due to its advanced age, the configuration... =1.3; Due to a history of hypertension, the configuration =1.1; Due to the winter season, the basic confidence threshold for critical value anomalies is set at 55%; Collect electrocardiogram (ECG) data and upload it to the AI ​​diagnostic module; The AI ​​diagnostic module outputs a quality analysis conclusion of "qualified," with a baseline confidence level of 85% (Cbaseline - quality) for the data acquisition quality. The disease diagnosis conclusion is "suspected myocardial infarction," with a baseline confidence level of 62% (Cbaseline - disease). Calculate the dynamic confidence level Cdynamic for the disease diagnosis process: 62% × 1.3 × 1.1 × 1.15 = 97.1%; 97.1% 55% triggers the critical value exception process; in one specific implementation, triggering the critical value exception process includes: triggering a triple warning: pop-up, pinned message, and SMS notification; Based on the warning, the diagnosing physician prioritizes treatment, and in this case, a report can be issued within 15 minutes. Initiating thrombolytic therapy shortens the procedure time, reducing it by 65% ​​in this case.

[0053] The effects of using the system of this invention include: adjusting the threshold to adapt to high-risk scenarios for the elderly in winter, quickly triggering early warnings, buying critical time for treatment, and reducing the risk of missed diagnosis by 40%.

[0054] Example 4: Normal Reporting Grouping Process - Youth Physical Examination Scenario Specific scenario: A 30-year-old young woman with no underlying medical history (β=1.0), undergoing a summer physical examination (γ=0.95), experiencing no discomfort, and requiring efficient processing of routine reports.

[0055] Figure 8 This example demonstrates the implementation process of a normal report grouping workflow using a system based on an embodiment of the present invention. Figure 8 The implementation process includes: Configure the weight coefficients on the corresponding interface of the device. Since the user is young, the configuration... =0.9; Due to the lack of medical history, configuration =1.0; Due to the summer season, the basic confidence threshold for configuring normal report grouping is 80%; Collect electrocardiogram (ECG) data and upload it to the AI ​​diagnostic module; The AI ​​diagnostic module outputs a quality analysis conclusion of "qualified," with a baseline confidence level of 90% (Cbaseline - quality) for the acquisition quality; the disease diagnosis conclusion is "normal sinus rhythm," with a baseline confidence level of 90% (Cbaseline - disease). Calculate the dynamic confidence level Cdynamic for the disease diagnosis process: 95% × 0.9 × 1.0 × 0.95 = 81.23%; 81.23% 80%, meaning the patient is normal, triggers the normal reporting grouping process, such as assigning them to the primary care physician group; In this example, a junior physician reviewed 40 reports in batches and submitted them within 1 hour. Feedback: Processing efficiency improved by 50%, with no missed diagnoses.

[0056] The effects of using the system of this invention include: lowering the threshold in low-risk scenarios for young people, improving efficiency in batch processing, increasing the average daily processing volume of primary physicians from 50 to 120 cases, and reducing the false positive rate to 2%.

[0057] In the system of this invention embodiment, the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the computer unit, connecting various parts of the entire computer unit via various interfaces and lines.

[0058] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the computer unit by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0059] Example 4: The present invention also provides an electrocardiogram (ECG) diagnostic device, including a memory and a processor. The memory stores at least one program, which is executed by the processor as described in the above system embodiment using the steps performed by the diagnostic terminal device. In one specific implementation, the ECG diagnostic device is the diagnostic terminal device as described above.

[0060] Example 5: The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps performed by the diagnostic terminal device as described in the system embodiment above.

[0061] Example 6: The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps performed by the diagnostic terminal device as described in the system embodiment above.

[0062] Although the invention has been specifically shown and described in conjunction with preferred embodiments, those skilled in the art should understand that various changes in form and detail may be made to the invention without departing from the spirit and scope of the invention as defined in the appended claims, all of which shall be within the scope of protection of the invention.

Claims

1. An electrocardiogram (ECG) diagnostic system, characterized in that, include: Electrocardiogram (ECG) data acquisition equipment is used to collect ECG data from individuals undergoing diagnosis. A diagnostic terminal device, connected to the electrocardiogram (ECG) data acquisition device, receives the ECG data, and includes a memory and a processor. The memory stores at least one program, which is executed by the processor to implement the following modules: The AI ​​diagnostic module is used to perform AI diagnosis on the person to be diagnosed based on the collected electrocardiogram data, and output the disease diagnosis conclusion and the corresponding basic confidence level of the disease diagnosis, as well as the collection quality analysis conclusion and the corresponding basic confidence level of the collection quality. The dynamic confidence calculation module is used to determine the dynamic confidence level adapted to the current scenario based on the output of the AI ​​diagnosis module. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season. Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the basic confidence level of data collection quality is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the basic confidence level of disease diagnosis is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; The process control module is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario. The corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

2. The electrocardiogram diagnostic system according to claim 1, characterized in that, Pre-set the weighting coefficients using the following steps: Based on the age of the individuals to be diagnosed, they are divided into multiple age groups, and pre-set parameters according to the age group they belong to. The value of corresponds to the age range; the higher the age, the better. The larger the value of ; Pre-set based on the patient's medical history The values ​​for are as follows, among which those with a history of atrial fibrillation or myocardial infarction are considered. Value > when there is an underlying disease Value > when there is no relevant medical history value; Preset according to the current season The values ​​of , among which, the seasons with high incidence rates The value is higher than that of the season with a low incidence rate. value.

3. The electrocardiogram diagnostic system according to claim 1, characterized in that, When determining whether to trigger the examination quality reminder process, clinical information improvement process, critical value warning process, or normal report grouping process, the preset confidence thresholds for different processes are different. Specifically, for each of the examination quality reminder process, clinical information improvement process, critical value warning process, or normal report grouping process, the steps to obtain the preset confidence threshold for the current scenario include: adjusting the confidence threshold in advance based on the respective preset basic confidence threshold.

4. The electrocardiogram diagnostic system according to claim 1, characterized in that, The acquisition quality analysis conclusion includes one or more of the following: electromyography interference, power frequency interference, baseline drift, good or qualified electrode contact; for the inspection quality process, when the acquisition quality analysis conclusion is electromyography interference or power frequency interference and the corresponding dynamic confidence level is greater than or equal to the preset acquisition quality reminder confidence level threshold for the current scenario, the inspection quality reminder process is triggered.

5. The electrocardiogram diagnostic system according to claim 3, characterized in that, Disease diagnostic conclusions include: normal, conclusions belonging to the predetermined non-critical value category, and conclusions belonging to the predetermined critical value abnormality category; predetermined non-critical value category conclusions include: arrhythmia; predetermined critical value abnormality category conclusions include: malignant arrhythmia, acute myocardial infarction, severe hyperkalemia or hypokalemia; among which: When the disease diagnosis conclusion is normal and the corresponding dynamic confidence level is greater than the preset confidence level threshold for normal report grouping for the current scenario, the normal report grouping process is triggered. When the disease diagnosis conclusion belongs to the predetermined non-critical value diagnosis conclusion and the corresponding dynamic confidence level is greater than the preset clinical information improvement confidence level threshold for the current scenario, the clinical information improvement process is triggered. When the disease diagnosis conclusion belongs to the predetermined critical value abnormality diagnosis conclusion and the corresponding dynamic confidence level is greater than the critical value warning confidence level threshold preset for the current scenario, the critical value warning process is triggered.

6. The electrocardiogram diagnostic system according to claim 5, characterized in that, When the age of the person to be diagnosed is 0-14 years old, if the disease diagnosis conclusion is abnormal critical value for myocardial infarction, the corresponding dynamic confidence level is directly set to 0, and the critical value warning process is not triggered.

7. The electrocardiogram diagnostic system according to claim 5, characterized in that, The steps to obtain the preset confidence threshold for the current scene include one or more of the following: The baseline confidence threshold for data collection quality alerts is 70%; when the patient to be diagnosed is a child or the current season is winter, the baseline confidence threshold for data collection quality alerts is lowered by a first predetermined percentage; the first predetermined percentage is less than or equal to 5%. The baseline confidence threshold for complete clinical information is 80%. When the patient to be diagnosed has a medical history, the baseline confidence threshold for complete clinical information is increased by a second predetermined percentage; the second predetermined percentage is less than or equal to 5%. The baseline confidence threshold for critical value warning is 60%. When the person to be diagnosed is in a predetermined stage of old age or the current season is winter, the baseline confidence threshold for critical value warning is lowered by a third predetermined percentage; the third predetermined percentage is less than or equal to 5%. The baseline confidence threshold for the normal reporting group is 90%. When the person to be diagnosed is in the predetermined youth stage, the baseline confidence threshold for the normal reporting group is lowered by a fourth predetermined percentage, which is less than or equal to 2%. When the person to be diagnosed is in the predetermined old age stage, the baseline confidence threshold for the normal reporting group is raised by a fifth predetermined percentage, which is less than or equal to 2%.

8. An electrocardiogram (ECG) diagnostic device, characterized in that, Includes a memory and a processor, wherein the memory stores at least one program, which is executed by the processor to perform the following steps: Based on the obtained electrocardiogram (ECG) data, AI diagnosis is performed on the individuals to be diagnosed, and the disease diagnosis conclusion and corresponding basic confidence level of the disease diagnosis are output, as well as the acquisition quality analysis conclusion and corresponding basic confidence level of the acquisition quality. Dynamic confidence calculation is used to determine a dynamic confidence level that fits the current scenario based on the AI ​​diagnosis output. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season, wherein: Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the basic confidence level of data collection quality is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the basic confidence level of disease diagnosis is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; Process control is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario; the corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

9. A computer-readable storage medium, characterized in that, The storage medium stores at least one program, which is executed by a processor to perform the following steps: Based on the obtained electrocardiogram (ECG) data, AI diagnosis is performed on the individuals to be diagnosed, and the disease diagnosis conclusion and corresponding basic confidence level of the disease diagnosis are output, as well as the acquisition quality analysis conclusion and corresponding basic confidence level of the acquisition quality. Dynamic confidence calculation is used to determine a dynamic confidence level that fits the current scenario based on the AI ​​diagnosis output. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season, wherein: Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the basic confidence level of data collection quality is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the basic confidence level of disease diagnosis is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; Process control is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario; the corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it performs the following steps: Based on the obtained electrocardiogram (ECG) data, AI diagnosis is performed on the individuals to be diagnosed, and the disease diagnosis conclusion and corresponding basic confidence level of the disease diagnosis are output, as well as the acquisition quality analysis conclusion and corresponding basic confidence level of the acquisition quality. Dynamic confidence calculation is used to determine a dynamic confidence level that fits the current scenario based on the AI ​​diagnosis output. The current scenario includes: the age of the person to be diagnosed, the medical history of the person to be diagnosed, and the current season, wherein: Dynamic confidence level = Baseline confidence level × × × ; Specifically, for the inspection quality process, the basic confidence level of data collection quality is used to calculate the dynamic confidence level; for the diagnostic process other than the inspection quality process, the basic confidence level of disease diagnosis is used to calculate the dynamic confidence level. The predetermined age weighting coefficient, The predetermined medical history weighting coefficient, The predetermined seasonal weighting coefficient; Process control is used to compare the dynamic confidence level with the confidence threshold preset for the current scenario for each processing flow, and to trigger the corresponding processing flow when the dynamic confidence level is greater than or equal to the confidence threshold preset for the current scenario; the corresponding processing flow includes one of the following: inspection quality reminder process, clinical information improvement process, critical value warning process, and normal report grouping process.

Images