Intelligent analysis and treatment suggestion method and system based on breathing machine data

Abstract

The invention relates to the technical field of breathing machines, and discloses an intelligent analysis and treatment suggestion method and system based on breathing machine data, and the method comprises the steps: calculating the curve similarity between curve data in curve data, and carrying out the problem aggregation of common breathing problems through the curve similarity; constructing a curve-problem relationship between the curve data and the breathing problem set; when the breathing machine fault exists in the problem set category, performing problem self-detection on the breathing machine, and when the breathing machine fault exists in the self-detection result; reducing the data category of the monitoring data, and constructing a numerical value adjustment sequence of the reduced data category; analyzing the improvement effect of the numerical values in the numerical value adjustment sequence on the reduced set category, and generating a reward model of the spliced data by using the improvement effect; and constructing a state set corresponding to the reduced data category and the sign data, and calculating an accumulated reward corresponding to the numerical value adjustment sequence. According to the invention, the intelligence of data analysis and treatment suggestion of the breathing machine can be improved.

Description

Technical Field

[0001] The invention relates to an intelligent analysis and treatment suggestion method and system based on ventilator data, belonging to the technical field of ventilators. Background Art

[0002] As an effective means of artificially replacing the spontaneous ventilation function, the ventilator has been widely used in respiratory failure caused by various reasons, anesthesia respiratory management during major surgery, respiratory support therapy and emergency resuscitation. It occupies a very important position in the field of modern medicine. The ventilator is a vital medical device that can prevent and treat respiratory failure, reduce complications, save and prolong the patient's life. The basic setting parameter data of the ventilator itself and the characteristic data generated by the patient using the ventilator.

[0003] With the advancement of medical technology, ventilators are being used more and more widely in clinical practice. However, traditional ventilators mainly provide mechanical ventilation support and lack in-depth analysis of the patient's condition and personalized treatment recommendations. Therefore, how to use the data generated by the ventilator for intelligent analysis to generate accurate treatment recommendations has become an urgent problem to be solved in the current medical field. Ventilator users need to manually upload data or the device automatically uploads data to the digital sleep management platform to generate digital charts. Professionals then give treatment recommendations based on the interpretation of the chart information. Among them, treatment recommendations are suggestions for adjusting the instrument parameters of the ventilator. After the patient's treatment data is uploaded, the doctor needs to spend a certain amount of time to carefully analyze the data. Since the amount of data may be very large, the analysis work becomes particularly complicated and time-consuming, so the doctor cannot give treatment recommendations immediately. In addition, as the amount of data increases, the required analysis cost will also increase accordingly, which may affect the efficiency and cost-effectiveness of the entire treatment process.

[0004] Currently, the data analysis and treatment recommendations of ventilators are not intelligent enough. Summary of the invention

[0005] The present invention provides a method and system for intelligent analysis and treatment suggestion based on ventilator data, the main purpose of which is to improve the intelligence of ventilator data analysis and treatment suggestion.

[0006] To achieve the above purpose, the present invention provides an intelligent analysis and treatment suggestion method based on ventilator data, comprising:

[0007] Collecting common respiratory problems of a ventilator, simulating curve data of the common respiratory problems, calculating curve similarity between each curve data in the curve data, and performing problem aggregation on the common respiratory problems using the curve similarity to obtain a set of respiratory problems, wherein the curve data includes a respiratory frequency curve, a tidal volume curve, an airway pressure curve, and a respiratory ratio curve;

[0008] Constructing a curve-question relationship between the curve data and the respiratory question set, collecting monitoring data in the ventilator, inputting the monitoring data into the curve-question relationship, and outputting the question set category corresponding to the monitoring data through the curve-question relationship, wherein the monitoring data includes respiratory rate, tidal volume, airway pressure and respiratory ratio;

[0009] Query whether there is a ventilator failure in the problem set category;

[0010] When there is a ventilator fault in the problem set category, performing a problem self-detection on the ventilator to obtain a self-detection result, and determining whether the self-detection result contains the ventilator fault;

[0011] When the self-test result shows that the ventilator is faulty, after the ventilator is repaired, the current data of the ventilator is collected again, the current problem category corresponding to the current data is identified using the curve-problem relationship, and the process returns to the above step S3;

[0012] When the self-test result does not contain the ventilator fault, the number of problems in the problem set category is reduced to obtain a reduced set category, and based on the reduced set category, the data category of the monitoring data is reduced to obtain a reduced data category, and a numerical adjustment sequence of the reduced data category is constructed;

[0013] Collecting vital sign data of the patient of the ventilator, analyzing the improvement effect of the numerical value in the numerical value adjustment sequence on the reduced set category based on the reduced data category and the vital sign data, and generating a reward model for the spliced ​​data using the improvement effect;

[0014] Construct a state set corresponding to the narrowed data category and the vital sign data, calculate the cumulative reward corresponding to the numerical adjustment sequence according to the reward model and the state set, use the cumulative reward to filter the target adjustment sequence from the numerical adjustment sequence, and determine the treatment recommendation of the ventilator through the target adjustment sequence.

[0015] Optionally, the calculating the curve similarity between each curve data in the curve data includes:

[0016] Obtaining a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve in the curve data;

[0017] The first curve similarity corresponding to the respiratory frequency curve is calculated using the following formula:

[0018]

[0019] Among them, DKC(x i ,x j ) represents the first curve similarity, x i Represents the respiratory rate curve for the i-th common respiratory problem, x j represents the breathing frequency curve of the jth common respiratory problem, Φ() represents the kernel function related to the dynamic kernel, Φ(x i ) represents x i The kernel function, Φ(x j ) represents x j The kernel function, R i represents the set of relevant factors of the i-th common respiratory problem, R j represents the set of relevant factors of the jth common respiratory problems, and α represents the preset similarity threshold;

[0020] Calculating a second curve similarity corresponding to the tidal volume curve;

[0021] Calculating a third curve similarity corresponding to the airway pressure curve;

[0022] Calculating a fourth curve similarity corresponding to the breathing ratio curve;

[0023] The first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are used as the curve similarity between each curve data in the curve data.

[0024] Optionally, the using the curve similarity to aggregate the common respiratory problems to obtain a set of respiratory problems includes:

[0025] Obtaining a first curve similarity, a second curve similarity, a third curve similarity, and a fourth curve similarity among the curve similarities;

[0026] Determine whether the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are all not less than a preset similarity threshold;

[0027] When the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are all not less than a preset similarity threshold, obtaining a first target common problem corresponding to the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity from the common breathing problems;

[0028] Aggregating the first target common problems into a respiratory problem set;

[0029] When the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are not all less than a preset similarity threshold, obtaining a second target common problem corresponding to the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity from the common breathing problems;

[0030] The second target common problems are set as a respiratory problem set.

[0031] Optionally, constructing a curve-problem relationship between the curve data and the respiratory problem set includes:

[0032] Obtaining a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve corresponding to the curve data;

[0033] Converting the respiratory frequency curve, the tidal volume curve, the airway pressure curve and the respiratory ratio curve into a vector matrix;

[0034] Extracting the timing features of each row in the vector matrix using a gated recurrent unit;

[0035] Using a convolutional neural network to extract spatial features of each column in the vector matrix;

[0036] Using an attention mechanism to perform spatiotemporal fusion of the temporal features and the spatial features to obtain fused features;

[0037] Analyzing the classification probability value of the fused feature with respect to the set of respiratory problems;

[0038] A curve-problem relationship between the curve data and the set of respiratory problems is identified using the classification probability value.

[0039] Optionally, the performing self-detection of problems on the ventilator to obtain a self-detection result includes:

[0040] Obtaining a ventilator failure of the ventilator;

[0041] Using the ventilator failure as the current state of the ventilator;

[0042] Query the transferable status and transferable conditions of the current status within the historical period;

[0043] Based on the transferable condition, conditionally transfer the ventilator to obtain a transfer condition;

[0044] identifying a transfer state of the ventilator under the transfer condition;

[0045] Determining whether the transfer state is consistent with the transferable state;

[0046] When the transfer state is consistent with the transferable state, the existence of the ventilator fault is taken as the self-detection result;

[0047] When the transfer state is inconsistent with the transferable state, the absence of the ventilator failure is used as the self-detection result.

[0048] Optionally, the narrowing down the data category of the monitoring data based on the narrowed set category to obtain the narrowed data category includes:

[0049] Obtain the ventilator fault corresponding to the reduced set category;

[0050] Based on the ventilator failure and the reduced set of categories, the categories to be removed in the monitoring data are identified in the following manner:

[0051]

[0052] Among them, z represents the category to be removed, R k represents the set of relevant factors of ventilator failure k, R k′ represents the set of relevant factors for narrowing down the set category k′;

[0053] The categories to be removed are removed from the monitoring data to obtain narrowed data categories.

[0054] Optionally, analyzing the improvement effect of the values ​​in the value adjustment sequence on the narrowed set category based on the narrowed data category and the vital sign data includes:

[0055] Performing category assignment on the reduced set category to obtain a category initial value;

[0056] The following formula is used to construct the input data between the numerical value adjustment sequence, the category initial value, the reduced data category and the vital sign data:

[0057]

[0058] Where X represents the input data, Indicates the value in the value adjustment sequence at round t, Indicates the numerical set of the reduced set category in the tth round, when t = 1, represents the initial value of the category, X3 represents the set of reduced data categories, and X4 represents the set of vital sign data;

[0059] Performing feature extraction on the input data to obtain extracted features;

[0060] Identifying an implicit vector corresponding to the extracted feature;

[0061] Outputting the next round of distribution of the category initial value corresponding to the implicit vector, the next round of categories of the narrowed data categories, and the next round of vital signs of the vital sign data;

[0062] Determine whether there is a next round of values ​​in the value adjustment sequence;

[0063] When there are values ​​for the next round in the value adjustment sequence, selecting values ​​for the next round from the value adjustment sequence;

[0064] Constructing the next round of data between the next round of values, the next round of distribution, the next round of categories and the next round of vital signs, and returning to the above step of extracting features from the input data to obtain extracted features;

[0065] When there is no next round of values ​​in the value adjustment sequence, weighted summing the next round of distribution is performed to obtain a weighted sum value;

[0066] The weighted sum value is taken as the improvement effect.

[0067] Optionally, the step of constructing a state set corresponding to the narrowed data category and the vital sign data includes:

[0068] The state set corresponding to the narrowed data category and the vital sign data is constructed using the following formula:

[0069] S={s1,s2,...,s M}

[0070] s m ={s u ,s v}

[0071] Among them, S represents the state set, s1,s2,...,s M represents the M states corresponding to the narrowed data categories and the vital sign data, s m represents s1,s2,...,s M The mth state in s u Represents any set of values ​​corresponding to the narrowed data category, s v Represents any set of numerical values ​​corresponding to vital sign data.

[0072] Optionally, calculating the cumulative reward corresponding to the numerical value adjustment sequence according to the reward model and the state set includes:

[0073] Calculating the action value corresponding to the numerical adjustment sequence according to the reward model and the state set;

[0074] Calculating the state value corresponding to the numerical value adjustment sequence according to the reward model and the state set;

[0075] The action value and the state value are used to determine the cumulative reward corresponding to the numerical adjustment sequence.

[0076] In order to solve the above problems, the present invention also provides an intelligent analysis and treatment suggestion system based on ventilator data, the system comprising:

[0077] A problem aggregation module is used to collect common respiratory problems of the ventilator, simulate curve data of the common respiratory problems, calculate the curve similarity between each curve data in the curve data, and use the curve similarity to aggregate the common respiratory problems to obtain a set of respiratory problems, wherein the curve data includes a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve;

[0078] A category output module, used to construct a curve-question relationship between the curve data and the respiratory problem set, collect monitoring data in the ventilator, input the monitoring data into the curve-question relationship, and output the problem set category corresponding to the monitoring data through the curve-question relationship, wherein the monitoring data includes respiratory rate, tidal volume, airway pressure and respiratory ratio;

[0079] A fault query module, used to query whether there is a ventilator fault in the problem set category;

[0080] A fault judgment module is used to perform a self-detection of the ventilator when there is a ventilator fault in the problem set category, obtain a self-detection result, and determine whether the self-detection result contains the ventilator fault;

[0081] A category identification module, used for collecting the current data of the ventilator again after repairing the ventilator fault when the self-test result shows that the ventilator fault exists, identifying the current problem category corresponding to the current data by using the curve-problem relationship, and returning to the above step S3;

[0082] A sequence construction module, used for reducing the number of problems in the problem set category to obtain a reduced set category when the self-test result does not contain the ventilator fault, and reducing the data category of the monitoring data based on the reduced set category to obtain a reduced data category, and constructing a numerical value adjustment sequence of the reduced data category;

[0083] A reward generation module, used for collecting vital sign data of the patient of the ventilator, analyzing the improvement effect of the numerical value in the numerical value adjustment sequence on the reduced set category based on the reduced data category and the vital sign data, and generating a reward model for the spliced ​​data using the improvement effect;

[0084] A treatment recommendation module is used to construct a state set corresponding to the narrowed data category and the vital sign data, calculate the cumulative reward corresponding to the numerical adjustment sequence according to the reward model and the state set, use the cumulative reward to screen the target adjustment sequence from the numerical adjustment sequence, and determine the treatment recommendation of the ventilator through the target adjustment sequence.

[0085] Compared with the problems described in the background technology, the embodiment of the present invention calculates the curve similarity between each curve data in the curve data to calculate the similarity between each curve based on the relevant factors of common respiratory problems. The embodiment of the present invention constructs a curve-problem relationship between the curve data and the respiratory problem set to identify the input-output relationship between the curve data and the respiratory problem set by using intelligent technologies such as neural networks and deep learning, thereby improving the intelligence of ventilator data analysis. The embodiment of the present invention performs self-detection of problems on the ventilator to monitor whether the state of the ventilator will change accordingly when the conditions change. If it changes accordingly, it indicates that It indicates that an instrument fault in the problem set category exists. If it does not change, it means that the instrument fault in the problem set category has been repaired. Furthermore, the embodiment of the present invention analyzes the improvement effect of the numerical value in the numerical adjustment sequence on the reduced set category based on the reduced data category and the vital sign data, so as to analyze whether the curve data monitored by the ventilator will still be abnormal after the numerical value in the numerical adjustment sequence of the instrument data is adjusted. It should be noted that the intelligent analysis performed by the present invention is based on the situation that the curve data will be abnormal, and does not consider the situation that the curve data is normal. That is to say, when the curve data is normal, there is no need to adjust the instrument parameters of the ventilator. Therefore, the intelligent analysis and treatment suggestion method based on ventilator data proposed by the present invention can improve the intelligence of data analysis and treatment suggestions of the ventilator. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 A flowchart of a method for intelligent analysis and treatment suggestion based on ventilator data provided by an embodiment of the present invention;

[0087] Figure 2 A schematic diagram of a module for implementing the intelligent analysis and treatment suggestion method based on ventilator data provided by an embodiment of the present invention.

[0088] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0089] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0090] The embodiment of the present application provides a method for intelligent analysis and treatment suggestion based on ventilator data. The execution subject of the method for intelligent analysis and treatment suggestion based on ventilator data includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided in the embodiment of the present application. In other words, the method for intelligent analysis and treatment suggestion based on ventilator data can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0091] Embodiment 1:

[0092] Reference Figure 1 FIG. 1 is a flow chart of a method for intelligent analysis and treatment suggestion based on ventilator data provided by an embodiment of the present invention. In this embodiment, the method for intelligent analysis and treatment suggestion based on ventilator data includes:

[0093] S1. Collect common respiratory problems of ventilators, simulate curve data of the common respiratory problems, calculate the curve similarity between each curve data in the curve data, and use the curve similarity to aggregate the common respiratory problems to obtain a set of respiratory problems, wherein the curve data includes a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve.

[0094] In an embodiment of the present invention, the common respiratory problems refer to the reasons why abnormalities appear in the patient data monitored by the ventilator, such as physical reasons of the patient and instrument failure reasons. The instrument failure reasons include failure reasons of instruments that monitor the respiratory rate curve, tidal volume curve, airway pressure curve and respiratory ratio curve. For example, air leakage in the respiratory circuit causes abnormalities in the respiratory rate curve, and airway obstruction causes abnormalities in the airway pressure curve. Physical reasons of the patient include incomplete exhalation of the patient, increased residual air volume in the lungs, and thus abnormal airway pressure curve.

[0095] Optionally, the process of simulating the curve data of the common respiratory problems refers to using a constructed physical patient model and a ventilator to simulate the conditions under which common respiratory problems occur, thereby measuring the curve data when common respiratory problems occur, wherein the curve data refers to continuous values ​​within a fixed interval period.

[0096] Furthermore, the embodiment of the present invention calculates the curve similarity between each curve data in the curve data, so as to calculate the similarity between each curve based on the relevant factors of common respiratory problems.

[0097] In one embodiment of the present invention, the calculating of the curve similarity between each curve data in the curve data includes: obtaining a respiratory rate curve, a tidal volume curve, an airway pressure curve, and a respiratory ratio curve in the curve data; and calculating a first curve similarity corresponding to the respiratory rate curve using the following formula:

[0098]

[0099] Among them, DKC(x i ,x j ) represents the first curve similarity, x i Represents the respiratory rate curve for the i-th common respiratory problem, x j represents the breathing frequency curve of the jth common respiratory problem, Φ() represents the kernel function related to the dynamic kernel, Φ(x i ) represents x i The kernel function, Φ(x j ) represents x j The kernel function, R i represents the set of relevant factors of the i-th common respiratory problem, R j represents the set of relevant factors of the jth common respiratory problems, and α represents the preset similarity threshold;

[0100] Calculate the similarity of the second curve corresponding to the tidal volume curve; calculate the similarity of the third curve corresponding to the airway pressure curve; calculate the similarity of the fourth curve corresponding to the respiratory ratio curve; use the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity as the curve similarity between each curve data in the curve data.

[0101] Among them, the set of related factors refers to the factors that are correlated with the common respiratory problems based on the grey correlation analysis method, such as respiratory rate, tidal volume, airway pressure, respiratory ratio, etc. The preset similarity threshold refers to the numerical value set according to the actual scenario for screening similarity, which can be set to 0.9 or set according to the actual scenario.

[0102] It should be noted that the principles of calculating the similarity of the second curve corresponding to the tidal volume curve; calculating the similarity of the third curve corresponding to the airway pressure curve; and calculating the similarity of the fourth curve corresponding to the respiratory ratio curve are similar to those of calculating the similarity of the first curve corresponding to the respiratory frequency curve, and are not further elaborated here. Further, when , it means that x is an irrelevant factor for the i-th common respiratory problem and the j-th common respiratory problem. In order to ensure the smooth progress of the subsequent process of aggregating the common respiratory problems using the curve similarity, DKC(x i ,x j)=α.

[0103] In one embodiment of the present invention, the method of aggregating the common respiratory problems using the curve similarity to obtain a set of respiratory problems includes: obtaining a first curve similarity, a second curve similarity, a third curve similarity, and a fourth curve similarity in the curve similarity; determining whether the first curve similarity, the second curve similarity, the third curve similarity, and the fourth curve similarity are all not less than a preset similarity threshold; when the first curve similarity, the second curve similarity, the third curve similarity, and the fourth curve similarity are all not less than the preset similarity threshold, obtaining a first target common respiratory problem corresponding to the first curve similarity, the second curve similarity, the third curve similarity, and the fourth curve similarity from the common respiratory problems; aggregating the first target common respiratory problems into a set of respiratory problems; when the first curve similarity, the second curve similarity, the third curve similarity, and the fourth curve similarity are not all not less than the preset similarity threshold, obtaining a second target common respiratory problem corresponding to the first curve similarity, the second curve similarity, the third curve similarity, and the fourth curve similarity from the common respiratory problems; and using the second target common respiratory problems as a set of respiratory problems.

[0104] Optionally, the process of aggregating the first target common questions into a respiratory problem set refers to aggregating multiple first target common questions into one respiratory problem set, and the process of treating the second target common questions as a respiratory problem set refers to treating multiple second target common questions as multiple respiratory problem sets respectively.

[0105] S2. Construct a curve-problem relationship between the curve data and the respiratory problem set, collect monitoring data from the ventilator, input the monitoring data into the curve-problem relationship, and output the problem set category corresponding to the monitoring data through the curve-problem relationship, wherein the monitoring data includes respiratory rate, tidal volume, airway pressure and respiratory ratio.

[0106] The embodiment of the present invention constructs a curve-problem relationship between the curve data and the set of respiratory problems, so as to utilize intelligent technologies such as neural networks and deep learning to identify the input-output relationship between the curve data and the set of respiratory problems, thereby improving the intelligence of ventilator data analysis.

[0107] The curve-problem relationship refers to a neural network model composed of a gated recurrent unit, a convolutional neural network (CNN) and an attention mechanism.

[0108] In one embodiment of the present invention, the construction of the curve-problem relationship between the curve data and the set of respiratory problems includes: obtaining the respiratory frequency curve, tidal volume curve, airway pressure curve and respiratory ratio curve corresponding to the curve data; converting the respiratory frequency curve, the tidal volume curve, the airway pressure curve and the respiratory ratio curve into a vector matrix; using a gated recurrent unit to extract the temporal features of each row in the vector matrix; using a convolutional neural network to extract the spatial features of each column in the vector matrix; using an attention mechanism to perform spatiotemporal fusion of the temporal features and the spatial features to obtain fused features; analyzing the classification probability value of the fused feature with respect to the set of respiratory problems; and using the classification probability value to identify the curve-problem relationship between the curve data and the set of respiratory problems.

[0109] Each row in the vector matrix is ​​continuous data in a curve of one category, such as continuous data in a breathing frequency curve, and the attention mechanism is a self-attention layer.

[0110] Optionally, the analysis of the classification probability value of the fusion feature regarding the set of respiratory problems refers to the process of using an activation function to output the probability of classification, and the process of using the classification probability value to identify the curve-problem relationship between the curve data and the set of respiratory problems refers to the process of obtaining a trained neural network model when the set corresponding to the classification probability value is consistent with the actual set, wherein each classification probability value corresponds to each set of respiratory problems, rather than the problems in the set.

[0111] S3. Check whether there is a ventilator failure in the problem set category.

[0112] The ventilator failure refers to the cause of the ventilator instrument failure.

[0113] Optionally, the process of querying whether there is a ventilator failure in the problem set category refers to querying whether this set (problem set category) contains patient physical reasons or ventilator instrument failure reasons.

[0114] S4. When there is a ventilator fault in the problem set category, perform a self-detection of the problem on the ventilator to obtain a self-detection result, and determine whether the self-detection result contains the ventilator fault.

[0115] The embodiment of the present invention performs self-detection of problems on the ventilator to monitor whether the state of the ventilator changes when conditions change. If it changes, it indicates that an instrument fault in the problem set category exists; if it does not change, it indicates that the instrument fault in the problem set category has been repaired.

[0116] In one embodiment of the present invention, the self-detection of problems on the ventilator to obtain self-detection results includes: obtaining a ventilator fault of the ventilator; taking the ventilator fault as the current state of the ventilator; querying the transferable state and transferable condition of the current state within a historical period; based on the transferable condition, performing conditional transfer on the ventilator to obtain the transfer condition; identifying the transfer state of the ventilator under the transfer condition; determining whether the transfer state is consistent with the transferable state; when the transfer state is consistent with the transferable state, taking the existence of the ventilator fault as the self-detection result; when the transfer state is inconsistent with the transferable state, taking the non-existence of the ventilator fault as the self-detection result.

[0117] Among them, the transferable state refers to other states that the current state can change to, such as other faults or normal states, and the transferable condition refers to the conditions that can occur from the current state to the transferable state, such as ventilator instrument parameters.

[0118] S5. When the self-test result shows that there is a ventilator fault, after repairing the ventilator fault, collect the current data of the ventilator again, use the curve-problem relationship to identify the current problem category corresponding to the current data, and return to the above step S3.

[0119] Optionally, the process of using the curve-problem relationship to identify the current problem category corresponding to the current data refers to the process of reusing the data collected by the repaired ventilator to analyze whether there is a ventilator failure, and step S6 is not performed until there is no ventilator failure.

[0120] S6. When the self-test result shows that the ventilator fault does not exist, the number of problems in the problem set category is reduced to obtain a reduced set category, and based on the reduced set category, the data category of the monitoring data is reduced to obtain a reduced data category, and a numerical adjustment sequence of the reduced data category is constructed.

[0121] In the embodiment of the present invention, the reduced set category refers to the problem set category after removing the ventilator failure.

[0122] In one embodiment of the present invention, the reducing the data category of the monitoring data based on the reduced set category to obtain the reduced data category includes: obtaining a ventilator fault corresponding to the reduced set category; based on the ventilator fault and the reduced set category, identifying the category to be removed in the monitoring data in the following manner:

[0123]

[0124] Among them, z represents the category to be removed, Rk represents the set of relevant factors of ventilator failure k, R k′ represents the set of relevant factors for narrowing down the set category k′;

[0125] The categories to be removed are removed from the monitoring data to obtain narrowed data categories.

[0126] Among them, the numerical adjustment order refers to the sequence of randomly set ventilator instrument parameters, for example, the sequence of ventilator instrument parameters includes A sequence, B sequence, and C sequence, A represents ventilator instrument Class A parameters, B represents ventilator instrument Class A parameters, C represents ventilator instrument Class C parameters, and A sequence is 1, 2, 3, 4, indicating that the Class A parameters are adjusted from the current value to 1, then adjusted to 2, and then adjusted to 3, until the adjustment is completed to 4.

[0127] S7. Collect the vital sign data of the patient on the ventilator, analyze the improvement effect of the numerical value in the numerical adjustment sequence on the narrowed set category based on the narrowed data category and the vital sign data, and use the improvement effect to generate a reward model for the spliced ​​data.

[0128] The vital sign data refers to the physical fitness data of the patient undergoing ventilator treatment, such as blood pressure, body temperature and blood oxygen saturation.

[0129] Furthermore, the embodiment of the present invention analyzes the improvement effect of the numerical values ​​in the numerical adjustment sequence on the narrowed set category based on the narrowed data category and the vital sign data, so as to analyze whether the curve data monitored by the ventilator will still be abnormal after the numerical values ​​in the numerical adjustment sequence of the instrument data are adjusted. It should be noted that the intelligent analysis performed by the present invention is based on the situation where the curve data may be abnormal, and does not consider the situation where the curve data is normal. That is to say, when the curve data is normal, there is no need to adjust the instrument parameters of the ventilator.

[0130] In one embodiment of the present invention, the analyzing the improvement effect of the values ​​in the numerical value adjustment sequence on the reduced set category based on the reduced data category and the vital sign data includes: performing category assignment on the reduced set category to obtain a category initial value; and constructing input data between the numerical value adjustment sequence, the category initial value, the reduced data category, and the vital sign data using the following formula:

[0131]

[0132] Where X represents the input data, Indicates the value in the value adjustment sequence at round t, Indicates the numerical set of the reduced set category in the tth round, when t = 1, represents the initial value of the category, X3 represents the set of reduced data categories, and X4 represents the set of vital sign data;

[0133] Perform feature extraction on the input data to obtain extracted features; identify the implicit vector corresponding to the extracted features; output the next round distribution of the category initial value corresponding to the implicit vector, the next round category of the narrowed data category and the next round of vital signs of the vital sign data; determine whether the next round of values ​​exists in the numerical adjustment sequence; when the next round of values ​​exists in the numerical adjustment sequence, select the next round of values ​​from the numerical adjustment sequence; construct the next round of data between the next round of values, the next round of distribution, the next round of categories and the next round of vital signs, and return to the above step of performing feature extraction on the input data to obtain the extracted features; when the next round of values ​​does not exist in the numerical adjustment sequence, perform weighted summation on the next round of distribution to obtain a weighted summation value; use the weighted summation value as the improvement effect.

[0134] Among them, the initial value of the category is 0. For example, the value of each problem in the reduced set category is set to 0. For example, if there are 3 problems in the reduced set category, 0, 0, 0 are set as the initial value of the category, because the reduced set category is originally the cause of the curve abnormality, that is, it has no value.

[0135] Optionally, the process of performing feature extraction on the input data to obtain extracted features; identifying implicit vectors corresponding to the extracted features; and outputting the next round distribution of the category initial values ​​corresponding to the implicit vectors, the next round categories of the narrowed data categories, and the next round of signs of the sign data refers to the process of classification using multiple classifiers in the input layer, hidden layer, and output layer of the BP neural network, and the next round distribution of the output is, for example, 1, -2, 2, etc., a positive number indicates that the improvement effect is increased by 1, 2, and a negative number indicates that the improvement effect is reduced by 2, and the next round of values ​​refers to the value adjustment sequence that is consistent with The next consecutive value.

[0136] Wherein, the reward model refers to the improvement effect.

[0137] S8. Construct a state set corresponding to the narrowed data category and the vital sign data, calculate the cumulative reward corresponding to the numerical adjustment sequence according to the reward model and the state set, use the cumulative reward to filter the target adjustment sequence from the numerical adjustment sequence, and determine the treatment recommendation of the ventilator through the target adjustment sequence.

[0138] In one embodiment of the present invention, the step of constructing the state set corresponding to the narrowed data category and the vital sign data includes: constructing the state set corresponding to the narrowed data category and the vital sign data using the following formula:

[0139] S={s1,s2,...,s M}

[0140] s m ={s u ,s v}

[0141] Among them, S represents the state set, s1,s2,...,s M represents the M states corresponding to the narrowed data categories and the vital sign data, s m represents s1,s2,...,s M The mth state in s u Represents any set of values ​​corresponding to the narrowed data category, s v Represents any set of numerical values ​​corresponding to vital sign data.

[0142] Among them, s u It indicates any set of values ​​corresponding to the narrowed data category. It does not refer to the set in the narrowed data category, but any set of all possible sets of values ​​for each narrowed data category. For example, the narrowed data category includes category A and category B, and s u It represents the set of any numerical distribution of category A and any numerical distribution of category B, where the numerical distribution refers to continuous data within a fixed time interval, that is, the numerical value on the curve that changes with time.

[0143] In one embodiment of the present invention, calculating the cumulative reward corresponding to the numerical adjustment sequence based on the reward model and the state set includes: calculating the action value corresponding to the numerical adjustment sequence based on the reward model and the state set; calculating the state value corresponding to the numerical adjustment sequence based on the reward model and the state set; and determining the cumulative reward corresponding to the numerical adjustment sequence using the action value and the state value.

[0144] Optionally, the process of determining the cumulative reward corresponding to the numerical adjustment order using the action value and the state value refers to the process of calculating the cumulative reward using the state value function and the action value function, and the reward model acts as the reward function when calculating the state value function and the action value function.

[0145] Optionally, the process of using the cumulative reward to screen the target adjustment sequence from the numerical adjustment sequence and determining the treatment recommendation of the ventilator through the target adjustment sequence refers to selecting the maximum cumulative reward after calculating the cumulative reward corresponding to each numerical adjustment sequence, and using the numerical adjustment sequence corresponding to the maximum cumulative reward as the target adjustment sequence.

[0146] The treatment recommendation refers to the order of adjusting instrument parameters of the ventilator.

[0147] Compared with the problems described in the background technology, the embodiment of the present invention calculates the curve similarity between each curve data in the curve data to calculate the similarity between each curve based on the relevant factors of common respiratory problems. The embodiment of the present invention constructs a curve-problem relationship between the curve data and the respiratory problem set to identify the input-output relationship between the curve data and the respiratory problem set by using intelligent technologies such as neural networks and deep learning, thereby improving the intelligence of ventilator data analysis. The embodiment of the present invention performs self-detection of problems on the ventilator to monitor whether the state of the ventilator will change accordingly when the conditions change. If it changes accordingly, it indicates that It indicates that an instrument fault in the problem set category exists. If it does not change, it means that the instrument fault in the problem set category has been repaired. Furthermore, the embodiment of the present invention analyzes the improvement effect of the numerical value in the numerical adjustment sequence on the reduced set category based on the reduced data category and the vital sign data, so as to analyze whether the curve data monitored by the ventilator will still be abnormal after the numerical value in the numerical adjustment sequence of the instrument data is adjusted. It should be noted that the intelligent analysis performed by the present invention is based on the situation that the curve data will be abnormal, and does not consider the situation that the curve data is normal. That is to say, when the curve data is normal, there is no need to adjust the instrument parameters of the ventilator. Therefore, the intelligent analysis and treatment suggestion method based on ventilator data proposed by the present invention can improve the intelligence of data analysis and treatment suggestions of the ventilator.

[0148] Embodiment 2:

[0149] like Figure 2 The figure shows a functional module diagram of an intelligent analysis and treatment suggestion system based on ventilator data of the present invention.

[0150] The intelligent analysis and treatment suggestion system 200 based on ventilator data described in the present invention can be installed in an electronic device. According to the functions implemented, the intelligent analysis and treatment suggestion system based on ventilator data can include a problem aggregation module 201, a category output module 202, a fault query module 203, a fault judgment module 204, a category identification module 205, a sequence construction module 206, a reward generation module 207 and a treatment suggestion module 208. The module described in the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0151] In the embodiment of the present invention, the functions of each module / unit are as follows:

[0152] The problem aggregation module 201 is used to collect common respiratory problems of the ventilator, simulate curve data of the common respiratory problems, calculate the curve similarity between each curve data in the curve data, and use the curve similarity to aggregate the common respiratory problems to obtain a set of respiratory problems, wherein the curve data includes a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve;

[0153] The category output module 202 is used to construct a curve-question relationship between the curve data and the respiratory problem set, collect monitoring data in the ventilator, input the monitoring data into the curve-question relationship, and output the problem set category corresponding to the monitoring data through the curve-question relationship, wherein the monitoring data includes respiratory rate, tidal volume, airway pressure and respiratory ratio;

[0154] The fault query module 203 is used to query whether there is a ventilator fault in the problem set category;

[0155] The fault judgment module 204 is used to perform a self-test on the ventilator when there is a ventilator fault in the problem set category, obtain a self-test result, and judge whether the self-test result contains the ventilator fault;

[0156] The category identification module 205 is used for collecting the current data of the ventilator again after repairing the ventilator fault when the self-test result shows that the ventilator fault exists, identifying the current problem category corresponding to the current data by using the curve-problem relationship, and returning to the above step S3;

[0157] The sequence construction module 206 is used to reduce the number of problems in the problem set category to obtain a reduced set category when the self-test result does not contain the ventilator fault, and based on the reduced set category, reduce the data category of the monitoring data to obtain a reduced data category, and construct a numerical value adjustment sequence of the reduced data category;

[0158] The reward generation module 207 is used to collect the vital sign data of the patient of the ventilator, analyze the improvement effect of the numerical value in the numerical value adjustment sequence on the reduced set category based on the reduced data category and the vital sign data, and generate the reward model of the spliced ​​data using the improvement effect;

[0159] The treatment recommendation module 208 is used to construct a state set corresponding to the narrowed data category and the vital sign data, calculate the cumulative reward corresponding to the numerical adjustment sequence based on the reward model and the state set, use the cumulative reward to filter the target adjustment sequence from the numerical adjustment sequence, and determine the treatment recommendation of the ventilator through the target adjustment sequence.

[0160] In detail, each module in the intelligent analysis and treatment suggestion system 200 based on ventilator data in the embodiment of the present invention is used in the same manner as described above. Figure 1 The intelligent analysis based on ventilator data and treatment recommendation methods described in the text are the same technical means and can produce the same technical effects, so they will not be repeated here.

[0161] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for intelligent analysis and treatment suggestion based on ventilator data, characterized in that: The method comprises: S1. Collect common respiratory problems of ventilators, simulate curve data of the common respiratory problems, calculate the curve similarity between each curve data in the curve data, and use the curve similarity to aggregate the common respiratory problems to obtain a set of respiratory problems, wherein the curve data includes a respiratory frequency curve, a tidal volume curve, an airway pressure curve, and a respiratory ratio curve; S2, constructing a curve-problem relationship between the curve data and the respiratory problem set, collecting monitoring data in the ventilator, inputting the monitoring data into the curve-problem relationship, and outputting the problem set category corresponding to the monitoring data through the curve-problem relationship, wherein the monitoring data includes respiratory rate, tidal volume, airway pressure and respiratory ratio; S3, querying whether there is a ventilator failure in the problem set category; S4, when there is a ventilator fault in the problem set category, performing a problem self-detection on the ventilator to obtain a self-detection result, and determining whether the self-detection result contains the ventilator fault; S5, when the self-test result shows that the ventilator is faulty, after the ventilator is repaired, the current data of the ventilator is collected again, the current problem category corresponding to the current data is identified by using the curve-problem relationship, and the process returns to the above step S3; S6. When the self-test result shows that there is no ventilator fault, the number of problems in the problem set category is reduced to obtain a reduced set category, and based on the reduced set category, the data category of the monitoring data is reduced to obtain a reduced data category, and a numerical adjustment sequence of the reduced data category is constructed; S7, collecting the vital sign data of the patient on the ventilator, analyzing the improvement effect of the numerical value in the numerical value adjustment sequence on the reduced set category based on the reduced data category and the vital sign data, and generating a reward model for the spliced ​​data using the improvement effect; S8. Construct a state set corresponding to the narrowed data category and the vital sign data, calculate the cumulative reward corresponding to the numerical adjustment sequence according to the reward model and the state set, use the cumulative reward to filter the target adjustment sequence from the numerical adjustment sequence, and determine the treatment recommendation of the ventilator through the target adjustment sequence.

2. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The calculating the curve similarity between each curve data in the curve data comprises: Obtaining a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve in the curve data; The first curve similarity corresponding to the respiratory frequency curve is calculated using the following formula: Among them, DKC(x i ,x j ) represents the first curve similarity, x i Represents the respiratory rate curve for the i-th common respiratory problem, x j represents the breathing frequency curve of the jth common respiratory problem, Φ() represents the kernel function related to the dynamic kernel, Φ(x i ) represents x i The kernel function, Φ(x j ) represents x j The kernel function, R i represents the set of relevant factors of the i-th common respiratory problem, R j represents the set of relevant factors of the jth common respiratory problems, and α represents the preset similarity threshold; Calculating a second curve similarity corresponding to the tidal volume curve; Calculating a third curve similarity corresponding to the airway pressure curve; Calculating a fourth curve similarity corresponding to the breathing ratio curve; The first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are used as the curve similarity between each curve data in the curve data.

3. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The method of aggregating the common respiratory problems by using the curve similarity to obtain a set of respiratory problems includes: Obtaining a first curve similarity, a second curve similarity, a third curve similarity, and a fourth curve similarity among the curve similarities; Determine whether the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are all not less than a preset similarity threshold; When the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are all not less than a preset similarity threshold, obtaining a first target common problem corresponding to the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity from the common breathing problems; Aggregating the first target common problems into a respiratory problem set; When the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity are not all less than a preset similarity threshold, obtaining a second target common problem corresponding to the first curve similarity, the second curve similarity, the third curve similarity and the fourth curve similarity from the common breathing problems; The second target common problems are set as a respiratory problem set.

4. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The constructing of a curve-problem relationship between the curve data and the respiratory problem set includes: Obtaining a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve corresponding to the curve data; Converting the respiratory frequency curve, the tidal volume curve, the airway pressure curve and the respiratory ratio curve into a vector matrix; Extracting the timing features of each row in the vector matrix using a gated recurrent unit; Using a convolutional neural network to extract spatial features of each column in the vector matrix; Using an attention mechanism to perform spatiotemporal fusion of the temporal features and the spatial features to obtain fused features; Analyzing the classification probability value of the fused feature with respect to the set of respiratory problems; A curve-problem relationship between the curve data and the set of respiratory problems is identified using the classification probability value.

5. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The self-detection of the ventilator to obtain the self-detection result includes: Obtaining a ventilator failure of the ventilator; Using the ventilator failure as the current state of the ventilator; Query the transferable status and transferable conditions of the current status within the historical period; Based on the transferable condition, conditionally transfer the ventilator to obtain a transfer condition; identifying a transfer state of the ventilator under the transfer condition; Determining whether the transfer state is consistent with the transferable state; When the transfer state is consistent with the transferable state, the existence of the ventilator fault is taken as the self-detection result; When the transfer state is inconsistent with the transferable state, the absence of the ventilator failure is used as the self-detection result.

6. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The step of narrowing down the data category of the monitoring data based on the narrowed set category to obtain the narrowed data category includes: Obtain the ventilator fault corresponding to the reduced set category; Based on the ventilator failure and the reduced set of categories, the categories to be removed in the monitoring data are identified in the following manner: Among them, z represents the category to be removed, R k represents the set of relevant factors of ventilator failure k, R k′ represents the set of relevant factors for narrowing down the set category k′; The categories to be removed are removed from the monitoring data to obtain narrowed data categories.

7. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The analyzing, based on the narrowed data category and the vital sign data, the improvement effect of the values ​​in the value adjustment sequence on the narrowed set category includes: Performing category assignment on the reduced set category to obtain a category initial value; The following formula is used to construct the input data between the numerical value adjustment sequence, the category initial value, the reduced data category and the vital sign data: Where X represents the input data, Indicates the value in the value adjustment sequence at round t, Indicates the numerical set of the reduced set category in the tth round, when t = 1, represents the initial value of the category, X3 represents the set of reduced data categories, and X4 represents the set of vital sign data; Performing feature extraction on the input data to obtain extracted features; Identifying an implicit vector corresponding to the extracted feature; Outputting the next round of distribution of the category initial value corresponding to the implicit vector, the next round of categories of the narrowed data categories, and the next round of vital signs of the vital sign data; Determine whether there is a next round of values ​​in the value adjustment sequence; When there are values ​​for the next round in the value adjustment sequence, selecting values ​​for the next round from the value adjustment sequence; Constructing the next round of data between the next round of values, the next round of distribution, the next round of categories and the next round of vital signs, and returning to the above step of extracting features from the input data to obtain extracted features; When there is no next round of values ​​in the value adjustment sequence, weighted summing the next round of distribution is performed to obtain a weighted sum value; The weighted sum value is taken as the improvement effect.

8. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The step of constructing a state set corresponding to the narrowed data category and the vital sign data includes: The state set corresponding to the narrowed data category and the vital sign data is constructed using the following formula: S={s1,s2,...,s M } s m ={s u ,s v } Among them, S represents the state set, s1,s2,...,s M represents the M states corresponding to the narrowed data categories and the vital sign data, s m represents s1,s2,...,s M The mth state in s u Represents any set of values ​​corresponding to the narrowed data category, s v Represents any set of numerical values ​​corresponding to vital sign data.

9. The intelligent analysis and treatment suggestion method based on ventilator data according to claim 1, characterized in that: The step of calculating the cumulative reward corresponding to the numerical value adjustment sequence according to the reward model and the state set includes: Calculating the action value corresponding to the numerical adjustment sequence according to the reward model and the state set; Calculating the state value corresponding to the numerical value adjustment sequence according to the reward model and the state set; The action value and the state value are used to determine the cumulative reward corresponding to the numerical adjustment sequence.

10. An intelligent analysis and treatment suggestion system based on ventilator data, characterized in that: The system comprises: A problem aggregation module is used to collect common respiratory problems of the ventilator, simulate curve data of the common respiratory problems, calculate the curve similarity between each curve data in the curve data, and use the curve similarity to aggregate the common respiratory problems to obtain a set of respiratory problems, wherein the curve data includes a respiratory frequency curve, a tidal volume curve, an airway pressure curve and a respiratory ratio curve; A category output module, used to construct a curve-question relationship between the curve data and the respiratory problem set, collect monitoring data in the ventilator, input the monitoring data into the curve-question relationship, and output the problem set category corresponding to the monitoring data through the curve-question relationship, wherein the monitoring data includes respiratory rate, tidal volume, airway pressure and respiratory ratio; A fault query module, used to query whether there is a ventilator fault in the problem set category; A fault judgment module is used to perform a self-detection of the ventilator when there is a ventilator fault in the problem set category, obtain a self-detection result, and determine whether the self-detection result contains the ventilator fault; A category identification module, used for collecting the current data of the ventilator again after repairing the ventilator fault when the self-test result shows that the ventilator fault exists, identifying the current problem category corresponding to the current data by using the curve-problem relationship, and returning to the above step S3; A sequence construction module, used for reducing the number of problems in the problem set category to obtain a reduced set category when the self-test result does not contain the ventilator fault, and reducing the data category of the monitoring data based on the reduced set category to obtain a reduced data category, and constructing a numerical value adjustment sequence of the reduced data category; A reward generation module, used for collecting vital sign data of the patient of the ventilator, analyzing the improvement effect of the numerical value in the numerical value adjustment sequence on the reduced set category based on the reduced data category and the vital sign data, and generating a reward model for the spliced ​​data using the improvement effect; A treatment recommendation module is used to construct a state set corresponding to the narrowed data category and the vital sign data, calculate the cumulative reward corresponding to the numerical adjustment sequence according to the reward model and the state set, use the cumulative reward to screen the target adjustment sequence from the numerical adjustment sequence, and determine the treatment recommendation of the ventilator through the target adjustment sequence.

Images