A method for recommending diagnosis and treatment plans based on patient data analysis

Abstract

The present invention relates to the field of data processing technology, and specifically to a method for recommending a diagnosis and treatment plan based on patient data analysis. The method for recommending a diagnosis and treatment plan based on patient data analysis includes: obtaining the current patient's physical state data. Analyze the patient's current physical state data, and calculate the symptom bias coefficients of each symptom of the patient. Determine whether the current patient's maximum symptom bias coefficient is greater than or equal to a preset standard coefficient. If so, determine that the symptom corresponding to the symptom bias coefficient is the symptom suffered by the patient, and then find a pre-set symptom-treatment plan information table according to the symptom to obtain the treatment plan for the symptom. If not, recommend the inspection of the test items that are required and not detected for the symptom corresponding to the symptom bias coefficient. This method can automatically diagnose and analyze the patient, and determine whether the diagnosis plan can meet the judgment conditions. The recommendation level is good and can be more targeted at the actual situation of the patient.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for recommending diagnosis and treatment plans based on patient data analysis. Background Art

[0002] Patient data analysis refers to the use of statistics and data mining techniques to analyze patient data in order to better understand the development process of the disease, the severity of the disease, the effectiveness of treatment, and other issues. The following are some commonly used patient data analysis methods: 1. Descriptive statistics: This is one of the most basic analysis methods, which is used to describe the general characteristics of the data set, including mean, median, standard deviation, minimum and maximum values. 2. Association analysis: This is a statistical method used to determine the relationship between variables. It can be used to identify which factors may cause the occurrence of a certain disease and the correlation between different factors. For example, it can be analyzed whether factors such as patient age, gender, occupation, marital status, etc. are related to the incidence of a certain disease. 3. Regression analysis: Regression analysis is a method used to determine the relationship between independent variables and dependent variables. This can be used to analyze the extent to which a certain factor affects the development of a disease, as well as the interaction between this factor and other factors.

[0003] Treatment plan recommendation Several common methods are used to recommend treatment plans based on patient data: 1. Rule-based system: This method uses pre-defined rules to match the patient's condition and medication or treatment method. For example, if the patient has high blood pressure, the system will match the type and dosage of medication used. 2. Machine learning algorithm: Use machine learning algorithms to analyze patient data and generate new prediction models. These models are then used to match the best treatment plan for the patient. For example, use techniques such as support vector machines (SVM) or random forests to predict whether a patient needs medication. 3. Cluster analysis: Cluster patient data into different clusters and recommend the best treatment plan based on the characteristics of each cluster. For example, if patients belong to the same cluster group, it means that they have similar conditions and treatment needs, so they may be suitable for the same treatment plan.

[0004] The existing method of recommending diagnosis and treatment plans is mainly based on test reports, which cannot be analyzed based on the patient's actual physical condition and has poor objectivity. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a diagnosis and treatment plan recommendation method based on patient data analysis, which is used to analyze the patient's current status and recommend the best diagnosis and treatment plan. The diagnosis and treatment plan recommendation method based on patient data analysis may include:

[0006] S1. Obtaining the physical status data of the current patient. There are various ways to obtain the physical status data of the current patient, and the specific method can be designed according to actual needs.

[0007] S2. Analyze the patient's current physical condition data and calculate the symptom bias coefficient B of each symptom of the patient. j , where j is the number of each disease. We can count and sort each disease and its corresponding symptoms in advance, and sort each disease. The sorting method can be sorted from more to less patients, and of course other sorting methods are not excluded. The patient's current physical condition data fully analyzes the current information.

[0008] S3, determine the maximum disease bias coefficient Max{B j} is greater than or equal to a preset standard coefficient B 标 If yes, then execute S5, if no, then execute S4. The standard coefficient B 标 It can be determined according to the specific situation, usually between 0.5 and 1. When j is ranked high or the disease is mild, the general standard coefficient B 标 Smaller, when j is ranked lower or the disease is more serious, the general standard coefficient B 标 Larger, most ideal state standard coefficient B 标 The specific value is 1 and will not be described here.

[0009] S4, recommended disease bias coefficient Max{B j} Check the test items that are needed for the corresponding symptoms and have not been tested. In actual test items, various symptoms will correspond to various test items, and the judgment effects of various test items for fixed symptoms are good and bad.

[0010] S5, determine the disease bias coefficient Max{B j The corresponding disease is the disease suffered by the patient, and then a pre-set disease-treatment plan information table is searched according to the disease, so as to obtain the treatment plan of the disease. The disease-treatment plan information table can be prepared according to the existing disease treatment plan.

[0011] Preferred: generally including basic physical data of patients, variable data, consultation data and test data. Of course, other forms of data are not necessarily excluded. They can be optimized according to actual development needs. The difficulty of obtaining these data increases accordingly.

[0012] Preferably: the patient's basic physical data, variable data, and medical consultation data are all transmitted through a voice playback module, a voice input module, and a voice analysis module; the voice playback module is used to play data question items or instruction items according to an inquiry item set, and then obtain the patient's voice information content through a voice input module, and then analyze the obtained voice information content through a voice analysis module to obtain the patient's basic physical data, variable data, and medical consultation data.

[0013] Preferably: the patient's basic physical data, variable data, and medical consultation data can all be obtained through the voice playback module, voice input module, voice analysis module, video input module, and video analysis module. The voice playback module can play the data question items or instruction bars according to the current inquiry item set, and then obtain the patient's voice information content through the voice input module, and then analyze the obtained voice information content through the voice analysis module to obtain the patient's basic physical data, variable data, and medical consultation data. The inquiry item set is composed of many data question items, and these data question items can be arranged according to a preset data, so that they can be played in sequence. In our data acquisition process, multiple data question items of the same type or the same disease will be counted together to form an inquiry item set, and this inquiry item set will be named according to the type or disease. The video input module is used to obtain the video image of the patient's collection part, and the video analysis module is used to analyze the video image of the collection part to obtain the body state data. The video analysis module analyzes the video image to obtain the disease feature points, and then analyzes the disease feature points to obtain the body state data.

[0014] Preferably: the video image acquisition method includes: shooting to obtain multiple video frames, constructing these video frames into a detection image frame set, then calculating the ghosting coefficient F of each video frame in the detection image frame set, then calculating the video frame corresponding to Min{F} as the video image, and then analyzing the video image through a video analysis module.

[0015] Preferably: the method for obtaining the ghosting coefficient F may specifically include: performing detection feature recognition on the video frames in the detection image frame set, and statistically analyzing the video frames containing the detection features to form a feature image frame set. When collecting images of various parts of the human body, one or more detection features will be pre-constructed according to the different parts. When the video frame contains these detection features, the video frame is defined as a feature image frame set. The video frame is placed in a preset plane coordinate system, and then the detection features of the video frame currently to be analyzed in the feature image frame set are compared with the detection features in the next video frame, and the movement vector p of the detection feature is obtained. The movement vector is the movement direction and position of the patient during the detection process. Sampling is performed according to a preset sampling density to obtain sampling points, and then according to the sampling points and the movement vectors, the detection features are analyzed. Obtain the offset point. The method for sampling the coordinates A' of the offset point can be Where A is the coordinate point of the sampling point, ɑ is the offset acquisition ratio, and the gray value G of the sampling point is obtained by analyzing the video frame. i and the gray value G of the offset point i ', the ghosting coefficient of the video frame In the formula, i is the number of the sampling point or offset point, and the sampling point and the offset point are one-to-one corresponding. i Represents the gray value of the sampling point numbered i; G i ' is the gray value of the offset point numbered i, I is the total number of boundary points, i=1, 2, ┈, I.

[0016] Preferably, the calculation method of the disease bias coefficient may include: extracting disease keywords and matching coefficient c from the physical state data n , the disease keywords are pre-set. For example, the values ​​of various indicators, the names of diseases, the types of diseases, etc. When these keywords appear in the patient's physical condition data, the keywords and their corresponding contents are defined as the disease keywords of the physical condition data. Then the disease bias coefficient Where n is the sequence number of the symptom keyword in the current patient's physical state data, N is the total number of symptom keywords in the patient's current physical state data, k is the symptom numbered j and is the sequence number of the symptom keyword in the current patient's physical state data, n = 1, 2, ┈, N; w k is the influence weight of the symptom keyword numbered k on symptom j. k It is the matching coefficient of the symptom keyword when the number k = n. When there is only one symptom keyword in the symptom, the influence weight of the symptom keyword is 1. When there are multiple symptom keywords, it is necessary to check the influence of each symptom keyword on the symptom. The details are not repeated here.

[0017] Preferred: Coefficient of fit c n The acquisition method can generally be calculated based on the disease similarity, numerical overflow ratio, etc. Disease similarity mainly analyzes the biometrics, mainly comparing the current biometrics with the standard biometrics to obtain their similarity. The specific existing technology will not be described here. The numerical overflow ratio is to compare the detected value with the standard value, which can be Among them, C is the disease keyword detection value of patient number n, C' is the standard limit value of the disease keyword detection item of patient number n, and ε is the matching coefficient, which can basically be 0.01-100. It needs to be set according to actual conditions and will not be elaborated here.

[0018] The technical effects and advantages of the present invention: The diagnosis and treatment plan carried out by the present method can automatically diagnose and analyze the patient, determine whether the diagnosis plan can meet the judgment conditions, and recommend the detection plan. The recommendation level is good and more targeted at the actual situation of the patient, avoiding the subjective influence of the doctor and increasing objectivity. The detection items are basically sorted from good to bad according to the effect. Basically, if the data obtained through the consultation cannot be judged, 1-2 more detection data can be obtained. The number of tests can be minimized, the detection cost is reduced, and the judgment ability is strong. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A flowchart of a method for recommending a diagnosis and treatment plan based on patient data analysis proposed in the present invention.

[0020] Figure 2 This is a structural block diagram of Embodiment 1 of the structure for acquiring basic patient physical data, variable data, and consultation data in a method for recommending a diagnosis and treatment plan based on patient data analysis proposed by the present invention.

[0021] Figure 3 This is a structural block diagram of Embodiment 2 of the structure for obtaining basic patient physical data, variable data, and consultation data in a method for recommending a diagnosis and treatment plan based on patient data analysis proposed by the present invention. DETAILED DESCRIPTION

[0022] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are provided for the purpose of illustration and description, and are not intended to be exhaustive or to limit the present invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention, and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

[0023] Example 1

[0024] refer to Figure 1 In this embodiment, a diagnosis and treatment plan recommendation method based on patient data analysis is proposed, which is used to analyze the patient's current status and recommend the best diagnosis and treatment plan. The diagnosis and treatment plan recommendation method based on patient data analysis may include:

[0025] S1. Obtain the physical status data of the current patient. There are various ways to obtain the physical status data of the current patient, which can be specifically designed according to actual needs. Generally, it includes basic physical data of the patient, variable data, consultation data and test data. Of course, other data forms are not necessarily excluded, and can be optimized according to actual development needs. The difficulty of obtaining these data increases successively. The basic physical data can be some data that will not change and will not change with the change of time. The basic physical data generally include name, gender, date of birth, etc. The variable data can include weight, age, address, height, etc. The basic physical data and variable data of the patient can be obtained according to the patient's case. Specifically, these data can be directly entered into the case when the patient builds the case, and then bound with the patient's identity information to obtain the case data information package. The case data information package includes the basic physical data of the patient, variable data, consultation data, test data and the patient's identity information, and then the basic physical data, variable data, consultation data and test data of the patient are obtained by searching the patient's identity information. The patient's identity information can be obtained by scanning the ID card, the QR code on the medical record, the medical card, the electronic diagnosis code, etc. The details are not described here. Of course, it can also be obtained by direct input, and the details are not described here. The medical consultation data may include complexion, tongue coating, pain location, patient symptoms, etc., which can be obtained through the doctor's consultation content. These data require the cooperation of the patient to answer. In general, the doctor inquires and the patient answers according to his or her actual situation, thereby obtaining the medical consultation data. Reference Figure 2, whether it is the patient's basic physical data, variable data or consultation data, it can be obtained through the voice playback module, the voice input module, and the voice analysis module. The voice playback module can play the data question items or instruction bars according to the current inquiry item set, and then obtain the patient's voice information content through the voice input module, and then analyze the obtained voice information content through the voice analysis module to obtain the patient's basic physical data, variable data, and consultation data. The inquiry item set is composed of many data question items, and these data question items can be arranged according to a preset data, so that they can be played in sequence. In our data acquisition process, multiple data question items of the same type or the same disease will be counted together to form an inquiry item set, and this inquiry item set will be named according to the type or disease. For example, in the actual use process, it is necessary to input the patient's variable data, then the data question items in the inquiry item set at this time can be "Please answer your age", "Please accurately say your weight", etc. For example, the data question items in hand, foot and mouth disease can be "whether there is fever, how many degrees Celsius of fever", "whether there is rash or blisters", "whether there are symptoms such as headache / cough / runny nose", etc. Whether it is the order of data question items or inquiry item sets, they are simple first and then complex. Generally, they start with the basic physical data of the patient, or directly ask "Please tell me your symptoms". It is a prior art to obtain physical status data by analyzing the content of the sound information obtained by the voice analysis module, which is a specific prior art and will not be described in detail here. The test data may include laboratory test data and equipment test data. The laboratory test data may include blood tests, molecular biology tests, and urine and feces, etc. The blood test may be a routine blood test, a biochemical test, an immune test, etc., which will not be described in detail here. These data results can directly obtain physical status data through reports. Equipment detection data may include X-ray examinations, CT, nuclear magnetic resonance, blood pressure measuring instruments, etc., which require equipment and instruments to be detected. These data results can directly obtain physical status data through reports or image analysis. It is a specific prior art and will not be described here.

[0026] S2. Analyze the patient's current physical condition data and calculate the symptom bias coefficient B of each symptom of the patient. j , where j is the number of each symptom. We can count and sort each symptom and its corresponding symptoms in advance, and sort each symptom. The sorting method can be sorted from most to few patients, and of course other sorting methods are not excluded. The patient's current physical condition data fully analyzes the current information. The calculation method of the symptom bias coefficient may include: extracting the symptom keywords and the matching coefficient c of the physical condition data n, the disease keywords are pre-set. For example, the values ​​of various indicators, the names of diseases, the types of diseases, etc. When these keywords appear in the patient's physical condition data, the keywords and their corresponding contents are defined as the disease keywords of the physical condition data. Then the disease bias coefficient Where n is the sequence number of the symptom keyword in the current patient's physical state data, N is the total number of symptom keywords in the patient's current physical state data, k is the symptom numbered j and is the sequence number of the symptom keyword in the current patient's physical state data, n = 1, 2, ┈, N; w k is the influence weight of the disease keyword numbered k on disease j, c k It is the matching coefficient of the symptom keyword when the number k=n. When there is only one symptom keyword in the symptom, the influence weight of the symptom keyword is 1. When there are multiple symptom keywords, it is necessary to check the influence of each symptom keyword on the symptom. The details are not repeated here. Matching coefficient c n The acquisition method can generally be calculated based on the disease similarity, numerical overflow ratio, etc. Disease similarity mainly analyzes the biometrics, mainly comparing the current biometrics with the standard biometrics to obtain their similarity. The specific existing technology will not be described here. The numerical overflow ratio is to compare the detected value with the standard value, which can be Among them, C is the disease keyword detection value of the patient numbered n, C' is the standard limit value of the disease keyword detection item of the patient numbered n, and ε is the matching coefficient, and its value can basically be 0.01-100. It needs to be set according to the actual situation, and the details will not be repeated here. For example, when we diagnose and treat the upper respiratory tract, we observe the patient's oral cavity and obtain the patient's symptoms of abnormal stool, runny nose, fever, and sputum. The symptoms of upper respiratory tract infection are runny nose, fever, sputum, sore throat, headache, and tears. We analyzed and obtained a total of four disease keywords for the patient, of which 3 were consistent with the symptoms of upper respiratory tract infection, of which w1, w2, and w3 were all 0.33, and the matching coefficient c n are both 3, then we calculate Of course, this is just a simple example and may not be universal. Other specific situations will not be elaborated here.

[0027] S3, determine the maximum disease bias coefficient Max{B j} is greater than or equal to a preset standard coefficient B 标 If yes, then execute S5, if no, then execute S4. The standard coefficient B 标 It can be determined according to the specific situation, usually between 0.5 and 1. When j is ranked high or the disease is mild, the general standard coefficient B 标 Smaller, when j is ranked lower or the disease is more serious, the general standard coefficient B标 Larger, most ideal state standard coefficient B 标 The specific value is 1 and will not be described here.

[0028] S4, recommended disease bias coefficient Max{B j} Check the test items that are needed for the corresponding symptoms and have not been tested. In the actual test items, various symptoms will correspond to various test items, and the judgment effects of various test items for fixed symptoms are good and bad (generally speaking, special cases can be determined according to actual conditions). The test items are basically sorted from good to bad according to the effect. Basically, if it cannot be judged after obtaining the consultation data, 1-2 test data acquisitions can be performed. The number of tests can be minimized, the test costs are reduced, and the judgment ability is strong. For example, for the detection of pulmonary tuberculosis, the test items are basically imaging examinations, sputum examinations, tuberculosis contact tests, etc., and the ranking is basically the same as above. If the effects are basically the same, they can be sorted from small to large by cost. The details will not be repeated here.

[0029] S5, determine the disease bias coefficient Max{B j The corresponding symptom is the symptom suffered by the patient, and then a pre-set symptom-treatment plan information table is searched according to the symptom to obtain the treatment plan for the symptom. The symptom-treatment plan information table can be made according to the existing symptom treatment plan. For example, if the patient is a child, has yellow sputum, and is diagnosed with a bacterial upper respiratory tract infection, the recommended treatment plan is ceftriaxone, etc., which will not be described in detail here. The diagnosis and treatment plan carried out by this method can automatically diagnose and analyze the patient, determine whether the diagnosis plan can meet the judgment conditions, and recommend a detection plan. The recommendation level is good and can be more targeted at the actual situation of the patient, avoiding the influence of the doctor's subjectivity and increasing objectivity.

[0030] Example 2

[0031] refer to Figure 3, the patient's basic physical data, variable data, and medical consultation data can also be obtained through the voice playback module, voice input module, voice analysis module, video input module, and video analysis module. The voice playback module can play the data question item or instruction bar according to the current inquiry item set, and then obtain the patient's voice information content through the voice input module, and then analyze the obtained voice information content through the voice analysis module to obtain the patient's basic physical data, variable data, and medical consultation data. The inquiry item set is composed of many data question items, and these data question items can be arranged according to a preset data, so that they can be played in sequence. In our data acquisition process, multiple data question items of the same type or the same disease will be counted together to form an inquiry item set, and this inquiry item set will be named after the type or disease. The video input module is used to obtain the video image of the patient's collection part, and the video analysis module is used to analyze the video image of the collection part to obtain the physical state data. For example, when the tongue coating needs to be collected, it is necessary to play the collection instruction bar "Please open your mouth and aim at the collection frame" through the voice playback module, and then irradiate the oral cavity with light, and collect the video inside the oral cavity through the video input module to obtain the video image. The video image is analyzed by the video analysis module to obtain the disease feature points, and then the disease feature points are analyzed to obtain the body state data. At this time, the video input module can be a high-definition camera or a high-frequency camera, wherein the high-definition camera is preferred. The high-definition camera can obtain multiple video frames by shooting, and these video frames are constructed into a detection image frame set, and then the ghosting coefficient F of each video frame in the detection image frame set is calculated, and then the video frame corresponding to Min{F} is calculated as the video image, and then the video image is analyzed by the video analysis module. The ghosting coefficient F acquisition method can specifically include: performing detection feature recognition on the video frames in the detection image frame set, and statistically analyzing the video frames containing the detection features to form a feature image frame set, when collecting images of various parts of the human body, one or more detection features will be pre-constructed according to each different part, and when the video frame contains these detection features, the video frame is defined as a feature image frame set. Recognizing biological features from images is a prior art, and the specific details are not repeated here. For example, we collect images of the tonsils in the patient's mouth. The detection features at this time can be the tooth outline and tongue coating. When the patient extends the tooth outline and tongue coating, the mouth is open wide, which can fully expose the tonsils within the shooting range of the video recording module. Place the video frame in a preset plane coordinate system, then compare the detection features in the current video frame to be analyzed with the next video frame, and obtain the movement vector of the detection feature. The movement vector is the patient's movement direction and position during the test. Sampling is performed according to a preset sampling density to obtain sampling points, and then the patient's movement vector is determined based on the sampling points and the movement vector. Obtain the offset point. The sampling density can be set according to actual needs. For example, one point is taken as a sampling point out of 100 pixels. The method of sampling coordinates A' of the offset point can be Where A is the coordinate point of the sampling point, ɑ is the offset acquisition ratio, and its value is generally 0.01-0.5. The specific details are not repeated here. The gray value G of the sampling point is obtained by analyzing the video frame. i and the gray value G of the offset point i ', the ghosting coefficient of the video frame In the formula, i is the number of the sampling point or offset point, and the sampling point and the offset point are one-to-one corresponding. i Represents the gray value of the sampling point numbered i; G i ' is the gray value of the offset point numbered i, I is the total number of boundary points, i = 1, 2, ┈, I; when the video recording module is performing video acquisition, the patient will move or shake, which will cause the video frame to have ghosting or unclear conditions. The ghosting phenomenon that occurs is very small, but it has a great impact on the observation of detailed features. Therefore, analyzing the video frames with small ghosting coefficients can ensure the accuracy of the analysis. Selection before analysis can greatly reduce the amount of calculation and improve the efficiency of processing and analysis. Calculation can be performed in a directional and quantitative manner by calculating the direction and size of the moving vector, avoiding the occurrence of ghosting deviations.

[0032] Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field and related fields without creative work should fall within the scope of protection of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention are implemented according to the conventional means in the field unless otherwise specified and limited.

Claims

1. A method for recommending a diagnosis and treatment plan based on patient data analysis, characterized in that: The method for recommending a diagnosis and treatment plan based on patient data analysis includes: S1. Obtain the current patient's physical status data; S2. Analyze the patient's current physical condition data and calculate the symptom bias coefficient B of each symptom of the patient. j , where j is the number of each disease; S3, determine the maximum disease bias coefficient Max{B j } is greater than or equal to a preset standard coefficient B 标 , if yes, execute S5, if no, execute S4; S4, recommended disease bias coefficient Max{B j } Check the test items that are required for the corresponding disease but not tested; S5, determine the disease bias coefficient Max{B j The corresponding symptom is the symptom suffered by the patient, and then a pre-set symptom-treatment plan information table is searched according to the symptom to obtain the treatment plan for the symptom; The calculation method of the disease bias coefficient includes: extracting disease keywords and matching coefficient c of physical condition data n , then the disease bias coefficient , where n is the sequence number of the symptom keyword in the current patient's physical status data, N is the total number of symptom keywords in the patient's current physical status data, k is the symptom numbered j and is the sequence number of the symptom keyword in the current patient's physical status data, n=1, 2, ┈, N; w k is the influence weight of the disease keyword numbered k on disease j, c k is the matching coefficient of the disease keyword when number k=n; Coefficient of fit c n The acquisition method is calculated by the disease similarity or the numerical overflow ratio; Symptom similarity includes analyzing the biometrics, comparing the current biometrics with the standard biometrics, and obtaining the symptom similarity; The numerical overflow ratio is obtained by the matching coefficient , where C is the disease keyword detection value of patient number n, C' is the standard limit value of the disease keyword detection item of patient number n, and ε is the matching coefficient.

2. A method for recommending a diagnosis and treatment plan based on patient data analysis according to claim 1, characterized in that: Physical condition data includes the patient's basic physical data, variable data, consultation data and test data.

3. A method for recommending a diagnosis and treatment plan based on patient data analysis according to claim 2, characterized in that: The patient's basic physical data, variable data, and medical consultation data are obtained through the voice playback module, voice input module, and voice analysis module; the voice playback module is used to play data question items or instruction items according to the inquiry item set, and then obtain the patient's voice information content through the voice input module, and then analyze the obtained voice information content through the voice analysis module to obtain the patient's basic physical data, variable data, and medical consultation data.

4. A method for recommending a diagnosis and treatment plan based on patient data analysis according to claim 2, characterized in that: The patient's basic physical data, variable data, and medical consultation data are all obtained through the voice playback module, voice input module, voice analysis module, video input module, and video analysis module; the voice playback module is used to play data question items or instruction items according to the inquiry item set, and then obtain the patient's voice information content through the voice input module, and then analyze the obtained voice information content through the voice analysis module to obtain the patient's basic physical data, variable data, and medical consultation data; the video input module is used to obtain the patient's collection part video image, and the video analysis module is used to analyze the collection part video image to obtain physical condition data.

5. A method for recommending a diagnosis and treatment plan based on patient data analysis according to claim 3 or 4, characterized in that: The video image acquisition method includes: shooting to obtain multiple video frames, constructing these video frames into a detection image frame set, then calculating the ghosting coefficient F of each video frame in the detection image frame set, and then calculating the video frame corresponding to Min{F} as the video image.

6. A method for recommending a diagnosis and treatment plan based on patient data analysis according to claim 2, characterized in that: The test data includes laboratory test data and equipment test data.

Images