Multi-injury patient deep venous thrombosis and pulmonary embolism risk prediction system

Abstract

The invention relates to the technical field of data processing, in particular to a multi-injury patient deep venous thrombosis and pulmonary embolism risk prediction system which comprises a sequence acquisition module, a body recovery index acquisition module, a body injury weight index acquisition module and a risk acquisition module. Acquiring a blood coagulation function index change trend sequence of a current patient and a patient with similar medication after multiple times of medication; obtaining a body recovery index of the current patient based on the similar condition of the blood coagulation function index change trend sequences of the current patient and the patients with similar medication, and obtaining a body injury weight index of the current patient based on the difference between the medication dosage and the medication time interval of each time of the current patient and the patients with similar medication. Finally, the deep venous thrombosis and pulmonary embolism risk of the current patient is obtained by combining the similar body index conditions of the current patient and the patients with the similar medication and whether the patients with the similar medication have thrombus or not, and therefore the accuracy of deep venous thrombosis and pulmonary embolism risk prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries. Background Art

[0002] Polytrauma, a serious medical emergency, refers to the simultaneous or sequential occurrence of severe trauma to two or more anatomical tissues or organs in the human body (e.g., a car accident, a fall from height), caused by the same underlying cause. One or more of these injuries can be life-threatening. Its complexity and high risk require meticulous management throughout the entire process, from first aid to rehabilitation care. Deep vein thrombosis (DVT) and pulmonary embolism (PE) are two manifestations of venous thromboembolism (VTE). Their pathological mechanisms are closely related, and patients with polytrauma are susceptible to these complications due to various factors. DVT is the primary source of PE, with approximately 90% of PE cases arising from the dislodgment of thrombi from the deep veins of the lower extremities or pelvis. Once dislodged, the thrombi travel through the right heart to the pulmonary arteries, causing vascular obstruction and triggering PE. DVT and PE are often considered different stages of the same disease.

[0003] During rehabilitation care for patients with multiple injuries, it is necessary to monitor their vital signs based on their various physical indicators. By analyzing the similarities between the current patient's physical indicators and those of historical patients, the probability of the current patient developing venous thrombosis can be determined. This probability can then be used to provide a risk warning. However, existing risk warnings are only based on the current patient's physical indicators, ignoring the role of nursing care. The probability of a patient developing venous thrombosis after surgery is not only related to their physical indicators, but also to the severity of their injuries and their recovery. Therefore, the accuracy of existing deep vein thrombosis and pulmonary embolism risk predictions is low. Summary of the Invention

[0004] In order to solve the technical problem of low accuracy in existing deep vein thrombosis and pulmonary embolism risk prediction, the present invention aims to provide a deep vein thrombosis and pulmonary embolism risk prediction system for polytrauma patients. The technical solutions adopted are as follows: In a first aspect of the present invention, a system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries is provided, comprising: A sequence acquisition module is used to obtain the change trend of blood coagulation function indicators of the current patient and patients with similar medications after each medication, thereby obtaining a blood coagulation function indicator change trend sequence based on multiple medications; the patients with similar medications are historical patients; A physical recovery index acquisition module is used to obtain the physical recovery index of the current patient based on the similarity between the blood coagulation function index change trend sequence of the current patient and patients taking similar medications; A physical injury severity index acquisition module is used to obtain the physical injury severity index of the current patient based on the difference between the dosage of each medication and the medication time interval of the current patient and other patients with similar medications; The risk acquisition module is used to combine the current patient's physical recovery indicators and physical injury indicators, the similarity of the current patient's physical indicators with patients with similar medications, and whether patients with similar medications have thrombosis, to obtain the current patient's deep vein thrombosis and pulmonary embolism risks.

[0005] In an exemplary embodiment, the process of obtaining the change trend of the blood coagulation function index includes: Obtaining a reduction in the blood coagulation function index based on the blood coagulation function index of the target patient at multiple time points after the target medication; the target patient is any one of the current patient and a patient taking similar medication; the target medication is any medication; Obtain the difference between the blood coagulation function index and the normal index at each time point, thereby obtaining the overall difference between the blood coagulation function index and the normal index of the target patient after the target number of medications; According to the reduction amplitude and the overall difference, the change trend of the blood coagulation function index of the target patient after the target number of medications is obtained; the change trend of the blood coagulation function index is proportional to the reduction amplitude and inversely proportional to the overall difference.

[0006] In an exemplary embodiment, the process of obtaining the physical recovery index includes: Obtaining the overall similarity corresponding to the current patient, wherein the overall similarity is obtained by fusing the similarity of the blood coagulation function indicator change trend sequences of the current patient and patients with similar medications; Based on the overall similarity and the maximum blood coagulation function index change trend, the current patient's physical recovery index is obtained; the physical recovery index is proportional to the overall similarity and the maximum blood coagulation function index change trend, and the maximum blood coagulation function index change trend is the maximum value of the blood coagulation function index change trend of the current patient after each medication.

[0007] In an exemplary embodiment, the process of obtaining the physical injury severity indicator includes: Obtaining the dosage and medication time interval of each medication of the current patient each time, and obtaining a first medication characteristic, where the first medication characteristic is proportional to the dosage and inversely proportional to the medication time interval; Obtaining the average dose and average medication time interval of each drug used each time by all patients with similar medications to obtain a second medication feature, where the second medication feature is directly proportional to the average dose and inversely proportional to the average medication time interval; Obtaining a difference in medication characteristics of each drug used in each medication according to the first medication characteristics and the second medication characteristics; The differences in medication characteristics of all drugs used at all times are integrated to obtain the current patient's physical injury severity index.

[0008] In an exemplary embodiment, the process of obtaining the risk of deep vein thrombosis and pulmonary embolism includes: Obtaining a risk impact weight based on the current patient's physical recovery index and physical injury severity index, wherein the risk impact weight represents the degree of impact of the current patient's physical recovery index and physical injury severity index on the risk; Obtain the similarity of the physical indicators of the current patient and other patients taking similar medications, and determine the risk level based on whether these patients have thrombosis. The risk impact weight and the risk degree are multiplied to obtain the current patient's deep vein thrombosis and pulmonary embolism risks.

[0009] In an exemplary embodiment, the process of obtaining the risk impact weight includes: taking the physical injury severity index as a weight, performing weighted summation on the physical recovery index and the physical injury severity index to obtain the risk impact weight.

[0010] In an exemplary embodiment, the process of obtaining the risk level includes: The similarity between the physical indicators of the current patient and the candidate patient with similar medications and the thrombosis presence result of the candidate patient with similar medications are integrated to obtain the risk sub-level of the candidate patient with similar medications; wherein, when the candidate patient with similar medications does not have thrombosis, the thrombosis presence result is 0, and when the candidate patient with similar medications has thrombosis, the thrombosis presence result is a preset positive number; the candidate patient with similar medications is any patient with similar medications; The risk level is obtained by combining the risk sub-levels of all patients with similar medications.

[0011] In an exemplary embodiment, the system further includes a patient screening module for similar medication use, which is configured to screen patients with similar medication use from historical patients based on similar medication use between the current patient and historical patients.

[0012] In an exemplary embodiment, the screening process for patients with similar medications includes: Obtaining the similarity between each medication of the current patient and each medication of a candidate historical patient; the candidate historical patient is any historical patient; Perform DTW matching based on the similarity between each medication of the current patient and each medication of the candidate historical patients to obtain several matching pairs; The similarity of medication conditions of each matching pair is integrated to obtain the similarity between the current patient and the candidate historical patients; Based on the degree of similarity, determine whether the candidate historical patients are patients with similar medication.

[0013] In an exemplary embodiment, the process of obtaining the similarity of medication usage includes: Get the type of medication used by the current patient for the i-th time, and the type of medication used by the candidate historical patient for the j-th time; the i-th medication is any medication used by the current patient, and the j-th medication is any medication used by the candidate historical patient; Obtain the union of the drug types of the current patient's i-th medication and the drug types of the candidate historical patients' j-th medication; The ratio of the number of drug types in the union to the total number of drug types is calculated as the similarity of the medication between the i-th medication of the current patient and the j-th medication of the candidate historical patient; the total number of drug types is equal to the sum of the number of drug types in the i-th medication of the current patient and the number of drug types in the j-th medication of the candidate historical patient.

[0014] The present invention has the following beneficial effects: after the patient's surgery, historical patients with similar medication have similar medication situations with the current patient, and the historical patients with similar medication can be used as a reference. Then, by comparing the similarity between the changing trends of the blood coagulation function indicators after each medication of the current patient and the historical patients with similar medication, the current patient's physical recovery index is obtained, and since the patient's medication dosage and medication time interval are related to the patient's physical injury, the current patient's physical injury index is obtained based on the difference between the current patient's medication dosage and medication time interval with each similar medication patient, and then, combined with the current patient's physical recovery index and physical injury index, as well as the similarity between the current patient's physical indicators and each similar medication patient and whether each similar medication patient has thrombosis, the current patient's deep vein thrombosis and pulmonary embolism risk is obtained, thereby improving the accuracy of deep vein thrombosis and pulmonary embolism risk prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic structural diagram of a system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries, provided by one embodiment of the present invention; Figure 2 This is a flowchart of the steps corresponding to various modules of a system for predicting the risk of deep vein thrombosis and pulmonary embolism in polytrauma patients provided by one embodiment of the present invention; Figure 3 This is a flowchart for screening patients with similar medications provided by one embodiment of the present invention; Figure 4 This is a flowchart for obtaining similarity of medication situations provided by one embodiment of the present invention; Figure 5This is a flow chart for obtaining a change trend of a blood coagulation function indicator provided by one embodiment of the present invention; Figure 6 This is a flow chart for obtaining body recovery indicators provided by one embodiment of the present invention; Figure 7 This is a flow chart for obtaining a body injury severity index provided by one embodiment of the present invention; Figure 8 This is a flowchart for obtaining the risk of deep vein thrombosis and pulmonary embolism of a current patient provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] To further illustrate the technical means and effects employed by the present invention to achieve its intended objectives, the following detailed description of the specific embodiments, structures, features, and effects of the present invention is provided in conjunction with the accompanying drawings and preferred embodiments. In the following description, references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which this invention pertains. The data and information collected in this application were obtained with full consent and authorization.

[0018] This embodiment provides a system for predicting the risk of deep vein thrombosis and pulmonary embolism in polytrauma patients. This system is applicable to scenarios where, during postoperative rehabilitation care, the system predicts the risk of deep vein thrombosis and pulmonary embolism in polytrauma patients based on their medication use and related physical indicators, combined with the relevant information of historical patients with similar medication use. The historical patients are also postoperative polytrauma patients.

[0019] This embodiment provides a system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries. Figure 1 As shown, it includes: a sequence acquisition module, a physical recovery index acquisition module, a physical injury index acquisition module and a risk acquisition module. Each module can be a software module, which is essentially a corresponding method step; it can also be a hardware module, and the executed method steps are configured in the hardware module so that the hardware module realizes the corresponding function. Accordingly, the risk prediction system for deep vein thrombosis and pulmonary embolism in patients with multiple injuries can be a software system, which is configured in relevant processors, computer hosts, and related medical platforms; it can also be a hardware system, such as a server, computer host, etc. This embodiment does not limit the specific configuration of each module and the risk prediction system for deep vein thrombosis and pulmonary embolism in patients with multiple injuries.

[0020] like Figure 2As shown, the method steps corresponding to each module are as follows: The sequence acquisition module is used to obtain the blood coagulation function index change trend of the current patient and patients with similar medications after each medication, thereby obtaining the blood coagulation function index change trend sequence based on multiple medications.

[0021] The physical recovery index acquisition module is used to obtain the physical recovery index of the current patient based on the similarity between the blood coagulation function index change trend sequence of the current patient and patients with similar medications.

[0022] The physical injury severity index acquisition module is used to obtain the physical injury severity index of the current patient based on the difference between the dosage of each medication and the medication time interval of the current patient and other patients with similar medications.

[0023] The risk acquisition module is used to combine the current patient's physical recovery indicators and physical injury indicators, the similarity of the current patient's physical indicators with patients with similar medications, and whether patients with similar medications have thrombosis, to obtain the current patient's deep vein thrombosis and pulmonary embolism risks.

[0024] The specific implementation process of each module is described below with reference to the accompanying drawings.

[0025] The sequence acquisition module is used to obtain the blood coagulation function index change trend of the current patient and patients with similar medications after each medication, thereby obtaining the blood coagulation function index change trend sequence based on multiple medications.

[0026] Patients with multiple injuries who present with the same pathology often receive the same medications. Medication use is directly related to injury severity, such as antibiotics, anticoagulants, and analgesics, and often reflects the patient's injury or treatment needs. The dosage, type, and duration of medication use can help infer the severity of the injury. High-dose analgesics may indicate a more severe injury, increasing the risk of deep vein thrombosis and pulmonary embolism. Severe injuries can lead to prolonged bed rest and slowed blood flow. Medication use can be used to analyze a patient's risk of complications.

[0027] First, similar medication patients with similar medication usage as the current patient are identified. Similar medication patients are screened from historical patients. In an exemplary embodiment, the system further includes a similar medication patient screening module for screening similar medication patients from historical patients based on similar medication usage between the current patient and the historical patients.

[0028] In an exemplary embodiment, Figure 3 As shown, a specific screening process for patients with similar medications is given below: Step 1-1: Obtain the medication similarity between each medication of the current patient and each medication of the candidate historical patients.

[0029] Acquire multiple historical patients, and the historical patients are also polytrauma patients. Preferably, the historical patients have the same type of polytrauma as the current patient. Moreover, in order to improve data accuracy, when acquiring historical patients, it is necessary to satisfy the following: the number of medications used by the historical patients is the same as the number of medications used by the current patient. For ease of explanation, the candidate historical patient is set to be any historical patient. During the postoperative recovery process, the patient needs to take medication multiple times. Therefore, it is necessary to obtain the similarity between each medication use of the current patient and each medication use of the candidate historical patient based on the similarity between each medication use of the current patient and each medication use of the candidate historical patient. In an exemplary embodiment, as Figure 4 As shown, a specific process of obtaining the similarity of medication conditions is given as follows: Step 1-1-1: Obtain the type of medication used by the current patient for the i-th time, and the type of medication used by the candidate historical patient for the j-th time.

[0030] For ease of explanation, let's assume that the i-th medication usage is any medication usage of the current patient, and the j-th medication usage is any medication usage of the candidate historical patient. It should be understood that the i-th medication usage of the current patient and the j-th medication usage of the candidate historical patient are independent of each other. They can be the same medication usage, such as the second medication usage, or different medication usage, such as the i-th medication usage of the current patient is the second medication usage, and the j-th medication usage of the candidate historical patient is the third medication usage.

[0031] To obtain the number of medication types used by each patient at each time, it is necessary to classify the medications used by each patient by medication type. This classification method is a conventional method, for example, classifying based on drug efficacy. For example, drugs such as dabigatran, rivaroxaban, and apixaban all act as anticoagulants and belong to the same category. Therefore, drugs with the same efficacy are classified into one category, thereby classifying the medications used by each patient at each time by medication type.

[0032] This method obtains the drug type of each patient's medication at each time, that is, the drug type of the current patient's i-th medication, and the drug type of the candidate historical patient's j-th medication. At the same time, the number of drug types used by each patient at each time is obtained, that is, the number of drug types used by the current patient's i-th medication, and the number of drug types used by the candidate historical patient's j-th medication.

[0033] Step 1-1-2: Obtain the union of the drug types of the current patient's i-th medication and the drug types of the candidate historical patients' j-th medication.

[0034] Obtain the union of the current patient's i-th medication type and the candidate patient's j-th medication type. The greater the number of medication types included in the union, the more common medication types are between the current patient's i-th medication type and the candidate patient's j-th medication type, and the higher the similarity between the current patient's i-th medication type and the candidate patient's j-th medication type.

[0035] Step 1-1-3: Calculate the ratio of the number of drug types in the union to the total number of drug types as the similarity between the i-th medication of the current patient and the j-th medication of the candidate historical patient.

[0036] The sum of the number of drug types used by the current patient for the i-th time and the number of drug types used by the candidate historical patients for the j-th time is obtained as the total number of drug types.

[0037] The ratio of the union of the current patient's i-th medication types and the candidate patient's j-th medication types to the total number of medication types is calculated as the medication similarity between the current patient's i-th medication and the candidate patient's j-th medication. The sum of the number of medication types used by the current patient's i-th medication and the number of medication types used by the candidate patient's j-th medication is included in the calculation to avoid the situation where one patient's use of many medication types completely includes another patient's use of fewer medication types, resulting in an overly high similarity between the two patients' medication usage, which deviates from the actual situation.

[0038] By using the above method, the similarity between each medication of the current patient and each medication of the candidate historical patients is obtained. The similarity of medication usage of each medication can be used to infer whether the conditions of the two patients are similar, such as the severity of the patient's injury.

[0039] Step 1-2: Perform DTW matching based on the similarity between each medication of the current patient and each medication of the candidate historical patients to obtain several matching pairs.

[0040] The above calculation calculates the similarity between each medication use between two patients. However, patients typically take medication multiple times during treatment. Therefore, we perform DTW (Dynamic Time Warping) matching on the similarity between each medication use of the current patient and each medication use of candidate historical patients. This yields several matching pairs, each corresponding to a medication use similarity. The DTW matching distance is the obtained medication use similarity between the two medication uses.

[0041] Step 1-3: Fusion of the medication similarities of each matching pair to obtain the similarity between the current patient and the candidate historical patients.

[0042] The similarity of medication usage of each matching pair is integrated. Specifically, the average of the similarity of medication usage of each matching pair is calculated. The result is the similarity between the current patient and the candidate historical patients. Thus, the similarity between the current patient and each historical patient is obtained.

[0043] Step 1-4: Based on the degree of similarity, determine whether the candidate historical patients are patients with similar medication.

[0044] This embodiment presets a similarity threshold value, which ranges from 0 to 1. The specific value of the preset similarity threshold value is set based on actual conditions. This preset similarity threshold value is used to determine the degree of similarity between the current patient and historical patients. If the similarity threshold value is greater than the preset similarity threshold value, the current patient and the historical patient are highly similar; otherwise, the current patient and the historical patient are less similar. The specific value of the preset similarity threshold value is determined based on the judgment needs. In this embodiment, 0.6 is used as an example.

[0045] The degree of similarity between the current patient and the candidate historical patient is compared with the preset similarity threshold. If the similarity is greater than the preset similarity threshold, the candidate historical patient is determined to be a patient with similar medication to the current patient. If the similarity is less than or equal to the preset similarity threshold, the candidate historical patient is determined not to be a patient with similar medication to the current patient. Each historical patient is judged in this way to screen out patients with similar medication from the historical patients.

[0046] Then, the blood coagulation index change trends after each medication use are obtained for the current patient and other patients with similar medications, thereby obtaining a sequence of blood coagulation index change trends based on multiple medication use. The blood coagulation index change trends represent the changing trends of the blood coagulation index in the patient's blood after each medication use. For any patient, a blood coagulation index change trend is obtained after each medication use. Therefore, for this patient's multiple medication uses, multiple blood coagulation index change trends are obtained. Based on the time series, these multiple blood coagulation index change trends can be combined to form a sequence of blood coagulation index change trends for this patient based on multiple medication uses.

[0047] The blood coagulation function index is used to characterize the patient's blood coagulation function. In this embodiment, the blood coagulation function index is specifically the D-dimer level. Under normal circumstances, the D-dimer level cannot be too high and needs to be below a set value. Moreover, one of the purposes of the medication in this embodiment is to reduce the blood coagulation function index, i.e., the D-dimer level, in the blood. Therefore, the blood coagulation function index will decrease after each medication.

[0048] The patient's physical indicators will change after each medication. When analyzing the patient's risk, it is necessary to consider the patient's own recovery after injury. For example, if two patients have the same injury, and one patient has a better effect after taking the medication, then the change trend of their physical indicators will be more normal, and their risk level will be lower when making corresponding risk judgments. At the same time, because the D-dimer level is related to the thrombosis itself, when the blood coagulation function index obtained is too large, it will cause blood coagulation and thrombosis.

[0049] In an exemplary embodiment, Figure 5 As shown, a process for obtaining the changing trend of blood coagulation function indicators is given below, including: Step 1-5: Obtain the reduction range of the blood coagulation function index based on the blood coagulation function index of the target patient at multiple time points after the target number of medications.

[0050] For ease of explanation, the current patient and any one of the patients with similar medications are set as the target patient. Any medication of the target patient is set as the target medication. Multiple monitoring time points are set after the target patient's target medication, and the blood coagulation function index in the target patient's blood is collected at each monitoring time point, so that the blood coagulation function index corresponding to each monitoring time point is formed into a sequence in time sequence. Then, under normal circumstances, the blood coagulation function index in the sequence shows a downward trend. Then, the least squares method is used to perform a straight-line fitting on the sequence to obtain the slope of the fitted straight line. Then, the slope is a negative value. The larger the absolute value of the slope, the greater the reduction in the blood coagulation function index. Then, the absolute value of the slope of the fitted straight line is used as the reduction in the blood coagulation function index of the target patient after the target medication.

[0051] Step 1-6: Obtain the difference between the blood coagulation function index and the normal index at each time point, thereby obtaining the overall difference between the blood coagulation function index and the normal index of the target patient after the target number of medications.

[0052] Under normal circumstances, the blood coagulation function index in the blood has a normal level. Therefore, the value of this normal level is obtained as the normal index. Since the purpose of the target patient's medication is to lower the blood coagulation function index in the blood to ultimately reach a normal level, the target patient's blood coagulation function index at each time point after the target number of medication administrations is greater than or equal to the normal index.

[0053] Obtain the difference between the target patient's blood coagulation function index and the normal index at each time point after the target dose. Specifically, calculate the difference between the target patient's blood coagulation function index and the normal index at each time point after the target dose. Then, calculate the average of the difference between the target patient's blood coagulation function index and the normal index at each time point after the target dose as the overall difference between the target patient's blood coagulation function index and the normal index after the target dose.

[0054] Step 1-7: Based on the reduction amplitude and the overall difference, obtain the change trend of the blood coagulation function index of the target patient after the target number of medications.

[0055] The greater the decrease in the target patient's blood coagulation function index change trend after the target dose, the more significant the decrease in the target patient's blood coagulation function index after the target dose, the better the drug efficacy, and the more significant the change trend of the target patient's blood coagulation function index after the target dose. Therefore, the change trend of the target patient's blood coagulation function index after the target dose is directly proportional to the decrease. The greater the overall difference between the target patient's blood coagulation function index after the target dose and the normal index, the more abnormal the target patient's blood coagulation function index after the target dose, and the less significant the change trend of the target patient's blood coagulation function index after the target dose. Therefore, the change trend of the target patient's blood coagulation function index after the target dose is inversely proportional to the overall difference.

[0056] Based on the above logic, a specific quantitative method for the change trend of the target patient's blood coagulation function indicators after the target number of medications is given below: ; in, is the changing trend of the target patient's blood coagulation function index after the i-th medication, is the reduction in the change trend of the target patient's blood coagulation function index after the i-th medication, is the overall difference between the target patient's blood coagulation function index after the i-th medication and the normal index, is an exponential function with the natural constant e as the base, used to Negative correlation normalization.

[0057] Here For Normalization, this embodiment takes the sigmoid function as an example.

[0058] Using the above process, the change trend of the blood coagulation function index after each medication administration for the current patient and for each patient with similar medications is obtained. In an exemplary embodiment, the maximum and minimum values ​​of the change trend of the blood coagulation function index after each medication administration for the current patient and each patient with similar medications are obtained. Then, the change trend of the blood coagulation function index after each medication administration for each patient is normalized using the maximum-minimum normalization method, thereby unifying the change trend of the blood coagulation function index after each medication administration for each patient, facilitating subsequent data processing. The change trend of the blood coagulation function index used subsequently is the result of the maximum-minimum normalization method.

[0059] Taking the target patient as an example, after obtaining the target patient's blood coagulation function index change trend after each medication use, the blood coagulation function index change trend after each medication use is sorted according to the time sequence, that is, according to the order of each medication use, to obtain the target patient's blood coagulation function index change trend sequence based on multiple medication uses. In this way, the blood coagulation function index change trend sequence corresponding to the current patient and each patient with similar medication use is obtained.

[0060] This embodiment can construct a two-dimensional coordinate system, where the horizontal axis represents each medication and the vertical axis represents the change trend of the blood coagulation function index for each medication, thereby mapping the change trend sequence of the blood coagulation function index of each patient to the two-dimensional coordinate system.

[0061] The physical recovery index acquisition module is used to obtain the physical recovery index of the current patient based on the similarity between the blood coagulation function index change trend sequence of the current patient and patients with similar medications.

[0062] This embodiment quantifies the patient's recovery based on the similarity between the change trend sequences of the current patient's blood coagulation function indicators and those of patients taking similar medications. This similarity indicates the degree of similarity between the overall changes in the current patient's blood coagulation function indicators and those of patients taking similar medications. The greater the similarity, the better the patient's recovery.

[0063] In an exemplary embodiment, Figure 6 As shown, a specific process for obtaining the body recovery index is given below: Step 2-1: Obtain the overall similarity corresponding to the current patient. The overall similarity is obtained by fusing the similarity of the change trend sequence of the blood coagulation function index of the current patient and patients with similar medications.

[0064] Obtain the similarity between the change trend series of the blood coagulation function index of the current patient and each patient taking similar medication. The specific method of similarity is set according to actual needs, such as cosine similarity, Pearson correlation coefficient, etc., and DTW distance can also be used. Then, the DTW distance is negatively correlated to obtain similarity. In this embodiment, using DTW distance as an example, the negative correlation after obtaining the DTW distance is used as the similarity.

[0065] Then, the similarity of the change trend sequence of the blood coagulation function index of the current patient and that of the patients with similar medications is fused. The specific fusion method is: calculate the average value of the similarity of the change trend sequence of the blood coagulation function index of the current patient and that of the patients with similar medications as the overall similarity corresponding to the current patient.

[0066] Step 2-2: Obtain the current patient's physical recovery index based on the overall similarity and the change trend of the maximum blood coagulation function index.

[0067] The maximum value of the blood coagulation function index change trend of the current patient after each medication is obtained as the maximum blood coagulation function index change trend corresponding to the current patient.

[0068] The higher the overall similarity for the current patient, the more similar the current patient's blood coagulation index change trend sequence is to that of patients taking similar medications, and the better the current patient's physical recovery is, that is, the higher the current patient's physical recovery index is, and the physical recovery index is proportional to the overall similarity. Furthermore, the higher the change trend of the current patient's maximum blood coagulation index, the more significant the decline in the current patient's blood coagulation index is, the better the medication is, and the better the current patient's physical recovery is, that is, the higher the current patient's physical recovery index is, and the physical recovery index is proportional to the change trend of the maximum blood coagulation index.

[0069] Based on the above logic, a specific quantitative method for the current patient's physical recovery indicators is given below: ; in, It is the current patient's physical recovery index, is the change trend of the patient's maximum blood coagulation function index. is the DTW distance between the blood coagulation function index change trend sequence of the current patient and the bth patient with similar medication, It represents the similarity of the change trend sequence of blood coagulation function indicators between the current patient and the bth patient with similar medication, and B is the number of patients with similar medication.

[0070] The physical injury severity index acquisition module is used to obtain the physical injury severity index of the current patient based on the difference between the dosage of each medication and the medication time interval of the current patient and other patients with similar medications.

[0071] Patients with multiple injuries require strict care after surgery to reduce the occurrence of venous thrombosis complications. When venous thrombosis occurs, the patient's physical indicators and medication status will change to a certain extent, so the risk of venous thrombosis can be predicted based on physical indicators and medication information.

[0072] The dosage of each medication and the time interval between each medication can reflect the severity of the patient's injury to a certain extent. For example, when the patient's injury is more serious, the dosage of the medication will increase significantly and the time interval between medications will shorten. It should be understood that only the dosage of each medication and the time interval between medications are adjusted, and the number of medications and the amount of medication each time do not change. Moreover, in obtaining the physical injury index, each medication taken at each time is the same medication taken by the current patient and other patients with similar medications. That is, in the calculation of the physical injury index, each medication taken at each time is the same medication taken by the current patient and other patients with similar medications.

[0073] The greater the difference between the dosage of each medication and the time interval between medications of the current patient and those of other patients taking similar medications, the more serious the physical injury of the current patient is. Therefore, based on the difference between the dosage of each medication and the time interval between medications of the current patient and those of other patients taking similar medications, the physical injury index of the current patient is obtained.

[0074] In an exemplary embodiment, Figure 7 As shown, a specific process for obtaining the body injury severity index is given below: Step 3-1: Obtain the dosage and medication time interval of each drug used by the current patient each time to obtain a first medication feature.

[0075] For target patients, they may take multiple medications per dose. Therefore, the dosage of each medication is obtained. The time interval between each dose and the previous dose is then calculated as the medication interval for each dose. Since there are no other doses before the first dose, the time interval between the second dose and the first dose is also used as the medication interval for the first dose.

[0076] Based on the dosage and dosing interval of each medication taken by the patient, a first medication use characteristic is obtained. The first medication use characteristic is related to the severity of the physical injury. Therefore, the greater the dosage of the medication and the shorter the dosing interval, the more pronounced the first medication use characteristic is, and the more severe the physical injury. Therefore, the first medication use characteristic is directly proportional to the dosage and inversely proportional to the dosing interval. In one exemplary embodiment, the first medication use characteristic is equal to the ratio of the dosage to the dosing interval.

[0077] Step 3-2: Obtain the average dose of each drug and the average medication time interval for each medication used by all patients with similar medications to obtain the second medication feature.

[0078] Obtain the dosage of each drug for each patient with similar medications, then calculate the average to obtain the average dosage of each drug for each medication for all patients with similar medications. Obtain the medication interval for each drug for each patient with similar medications, then calculate the average to obtain the average medication interval for each drug for each medication for all patients with similar medications. In this way, the average dosage of each drug for each medication and the average medication interval for each drug are obtained for all patients with similar medications.

[0079] Then, based on the average dose of each drug and the average dosing interval for each drug, a second medication characteristic for each medication is obtained. Similarly, the second medication characteristic is related to the severity of the injury. Therefore, the greater the dose and the shorter the dosing interval, the more pronounced the second medication characteristic and the greater the severity of the injury. Therefore, the second medication characteristic is directly proportional to the average dose and inversely proportional to the average dosing interval. In one exemplary embodiment, the second medication characteristic is equal to the ratio of the average dose of the drug to the average dosing interval.

[0080] Step 3-3: Obtain the medication characteristic difference of each drug for each medication according to the first medication characteristic and the second medication characteristic.

[0081] Each medication used in each session has a corresponding first and second medication characteristics. The first medication characteristic is associated with the current patient, while the second medication characteristic is associated with all patients with similar medications. Therefore, the difference between the first and second medication characteristics corresponding to each medication used in each session is obtained to obtain the medication characteristic difference for each medication used in each session.

[0082] Step 3-4: Combine the differences in medication characteristics of all drugs used at all times to obtain the current patient's physical injury severity index.

[0083] After obtaining the medication characteristic differences for each medication at each dose, the medication characteristic differences for all medications across all doses are combined to obtain the patient's current injury severity index. The patient's medication dosage and medication intervals are analyzed to determine differences among patients with similar medications. The greater the difference, the more severe the patient's injury and the greater the patient's current injury severity index.

[0084] In an exemplary embodiment, based on the above logic, a specific quantification method of the current patient's physical injury severity index is given as follows: ; in, It is the current patient's physical injury indicator. is the dose of the nth drug corresponding to the mth medication of the current patient, is the time interval between the mth and nth medications of the current patient, It is the first medication feature corresponding to the current patient. is the average dose of the nth drug for the mth medication for all patients with similar medications, is the average medication time interval of the nth drug for the mth medication of all patients with similar medications, is the second medication feature corresponding to all patients with similar medications, Represents the absolute value function. Indicates the difference in medication characteristics corresponding to the nth drug used in the mth time.

[0085] The function represented is: calculating the average value of the medication characteristic differences corresponding to all drugs of all medications used at all times and performing normalization processing. The normalization method here can be a sigmoid function.

[0086] The risk acquisition module is used to combine the current patient's physical recovery indicators and physical injury indicators, the similarity of the current patient's physical indicators with patients with similar medications, and whether patients with similar medications have thrombosis, to obtain the current patient's deep vein thrombosis and pulmonary embolism risks.

[0087] The current patient's physical recovery index and physical injury index play an auxiliary role in predicting the current patient's deep vein thrombosis and pulmonary embolism risk. Combined with the similarity of the current patient's physical indicators with patients with similar medications and whether patients with similar medications have thrombosis, the current patient's deep vein thrombosis and pulmonary embolism risk is obtained.

[0088] Physical indicators of the current patient and patients taking similar medications are obtained. The specific physical indicators are set based on actual needs. In one exemplary embodiment, these indicators include heart rate, blood pressure, platelet count, blood lipids, fibrinogen, and other indicators. In addition, the location of the injury may be included. To facilitate data processing, the various physical indicators are normalized to eliminate dimension. Based on these normalized physical indicators, a physical indicator sequence for the current patient and a physical indicator sequence for patients taking similar medications are obtained.

[0089] In an exemplary embodiment, Figure 8 As shown, a specific process for obtaining the current patient's deep vein thrombosis and pulmonary embolism risks is given below: Step 4-1: Obtain the risk impact weight based on the current patient's physical recovery index and physical injury index.

[0090] According to the current patient's physical recovery index and physical injury index, a risk impact weight is obtained, which characterizes the degree of influence of the current patient's physical recovery index and physical injury index on the risk. In an exemplary embodiment, the physical injury index is used as a weight, and the physical recovery index and the physical injury index are weighted and summed to obtain the risk impact weight. Among them, the higher the current patient's physical injury index, that is, the higher the injury level, the greater the impact of the physical recovery index; the lower the current patient's physical injury index, that is, the lower the injury level, the more attention should be paid to the impact of the physical injury. Based on the above logic, a specific quantification method is given as follows: ; Where W is the risk impact weight. Indicates the weight of the impact of the severity of the physical injury on the physical recovery. When the physical injury level is high, the weight of the physical recovery is greater. This means that when the level of physical injury is low, one needs to pay more attention to the impact of physical injury.

[0091] Step 4-2: Obtain the similarity of the physical indicators of the current patient and other patients taking similar medications, and determine the risk level based on whether the patients taking similar medications have thrombosis.

[0092] Obtain the similarity of the physical indicators of the current patient and each patient with similar medications. Specifically, obtain the similarity of the physical indicator sequences of the current patient and each patient with similar medications. The similarity can be cosine similarity, Pearson correlation coefficient, etc., or DTW distance, and then negatively correlate the DTW distance to obtain the similarity. In this embodiment, taking DTW distance as an example, the DTW distance between the current patient and each patient with similar medications is obtained and then negatively correlated as the similarity. Specifically: A quantitative method for the similarity of the physical indicators of the current patient and each patient with similar medications is as follows: ; in, is the similarity of the physical indicators between the current patient and the bth patient with similar medication, is the DTW distance between the current patient and the bth patient with similar medication. It represents the negative correlation normalization of the DTW distance between the current patient and the bth patient with similar medication. Indicates weight normalization of the negative correlation normalized DTW distance. Specifically, obtain the sum of the negative correlation normalized results of the DTW distance between the current patient and the patient with similar medications, and then calculate The ratio of the sum to achieve weight normalization.

[0093] Then, the detector is used to determine whether there are thrombi in the blood of patients taking similar medications. When thrombi are present, and the more similar they are to the current patient's physical indicators, the higher the risk of deep vein thrombosis and pulmonary embolism in the current patient.

[0094] The risk level is determined based on the similarity between the current patient's physical indicators and those of patients taking similar medications, as well as whether each patient taking similar medications has thrombosis. In an exemplary embodiment, a specific process for obtaining the risk level is as follows: For ease of explanation, the candidate patient with similar medication is set to be any patient with similar medication. The thrombosis presence result of the candidate patient with similar medication is obtained. When the candidate patient with similar medication does not have thrombosis, the thrombosis presence result is 0. When the candidate patient with similar medication has thrombosis, the thrombosis presence result is a preset positive number. In an exemplary embodiment, the preset positive number is 1 as an example. The similarity of the physical indicators between the current patient and the candidate patient with similar medication is combined with the thrombosis presence result of the candidate patient with similar medication to obtain the risk sub-level of the candidate patient with similar medication. Then, the risk sub-levels of all patients with similar medication are combined to obtain the risk level of the current patient.

[0095] Based on the above logic, a quantitative method for the current patient's risk level is given below: ; Among them, CD is the current risk level of the patient, The result of the presence of thrombus in the blood of the bth patient with similar medication is: Equal to 1, when there is no thrombus Equal to 0. is the risk sub-level of the bth patient taking similar medication.

[0096] Step 4-3: Multiply the risk impact weight and the risk level to obtain the current patient's risk of deep vein thrombosis and pulmonary embolism.

[0097] The current patient's risk of deep vein thrombosis and pulmonary embolism is calculated using the following formula: ; Among them, FX is the current patient's risk of deep vein thrombosis and pulmonary embolism, thereby completing the current patient's risk prediction of deep vein thrombosis and pulmonary embolism.

[0098] In the follow-up, a subsequent risk warning is performed based on the obtained deep vein thrombosis and pulmonary embolism risk of the current patient. In an exemplary embodiment, two thresholds are preset, namely a first threshold and a second threshold. The first threshold is greater than the second threshold. These two thresholds are used to determine the risk level of the current patient's deep vein thrombosis and pulmonary embolism risk. The specific values ​​of the two thresholds are set according to actual needs, such as the first threshold is 0.7 and the second threshold is 0.3. When the current patient's deep vein thrombosis and pulmonary embolism risk is less than or equal to the second threshold, it means that the current patient's deep vein thrombosis and pulmonary embolism risk is low and belongs to low risk, and no warning measures are taken; when the current patient's deep vein thrombosis and pulmonary embolism risk is greater than the second threshold and less than or equal to the first threshold, it means that the current patient's deep vein thrombosis and pulmonary embolism risk is medium risk and needs close attention; when the current patient's deep vein thrombosis and pulmonary embolism risk is greater than the first threshold, it means that the current patient's deep vein thrombosis and pulmonary embolism risk is high risk, and a warning signal is output, requiring the doctor to intervene as soon as possible.

[0099] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0100] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries, characterized by: include: A sequence acquisition module is used to obtain the change trend of blood coagulation function indicators of the current patient and patients with similar medications after each medication, thereby obtaining a blood coagulation function indicator change trend sequence based on multiple medications; The patients with similar medications are historical patients; A physical recovery index acquisition module is used to obtain the physical recovery index of the current patient based on the similarity between the blood coagulation function index change trend sequence of the current patient and patients taking similar medications; A physical injury severity index acquisition module is used to obtain the physical injury severity index of the current patient based on the difference between the dosage of each medication and the medication time interval of the current patient and other patients with similar medications; The risk acquisition module is used to combine the current patient's physical recovery indicators and physical injury indicators, the similarity of the current patient's physical indicators with patients with similar medications, and whether patients with similar medications have thrombosis, to obtain the current patient's deep vein thrombosis and pulmonary embolism risks.

2. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 1, characterized in that: The process of obtaining the change trend of the blood coagulation function index includes: Obtaining a reduction in the blood coagulation function index based on the blood coagulation function index of the target patient at multiple time points after the target medication; the target patient is any one of the current patient and a patient taking similar medication; the target medication is any medication; Obtain the difference between the blood coagulation function index and the normal index at each time point, thereby obtaining the overall difference between the blood coagulation function index and the normal index of the target patient after the target number of medications; According to the reduction amplitude and the overall difference, the change trend of the blood coagulation function index of the target patient after the target number of medications is obtained; the change trend of the blood coagulation function index is proportional to the reduction amplitude and inversely proportional to the overall difference.

3. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 1, characterized in that: The process of obtaining the physical recovery index includes: Obtaining the overall similarity corresponding to the current patient, wherein the overall similarity is obtained by fusing the similarity of the blood coagulation function indicator change trend sequences of the current patient and patients with similar medications; Based on the overall similarity and the maximum blood coagulation function index change trend, the current patient's physical recovery index is obtained; the physical recovery index is proportional to the overall similarity and the maximum blood coagulation function index change trend, and the maximum blood coagulation function index change trend is the maximum value of the blood coagulation function index change trend of the current patient after each medication.

4. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 1, wherein: The process of obtaining the physical injury severity index includes: Obtaining the dosage and medication time interval of each medication of the current patient each time, and obtaining a first medication characteristic, where the first medication characteristic is proportional to the dosage and inversely proportional to the medication time interval; Obtaining the average dose and average medication time interval of each drug used each time by all patients with similar medications to obtain a second medication feature, where the second medication feature is directly proportional to the average dose and inversely proportional to the average medication time interval; Obtaining a difference in medication characteristics of each drug used in each medication according to the first medication characteristics and the second medication characteristics; The differences in medication characteristics of all drugs used at all times are integrated to obtain the current patient's physical injury severity index.

5. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 1, characterized in that: The process of obtaining the risk of deep vein thrombosis and pulmonary embolism includes: Obtaining a risk impact weight based on the current patient's physical recovery index and physical injury severity index, wherein the risk impact weight represents the degree of impact of the current patient's physical recovery index and physical injury severity index on the risk; Obtain the similarity of the physical indicators of the current patient and other patients taking similar medications, and determine the risk level based on whether these patients have thrombosis. The risk impact weight and the risk degree are multiplied to obtain the current patient's deep vein thrombosis and pulmonary embolism risks.

6. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 5, characterized in that: The process of obtaining the risk impact weight includes: taking the physical injury severity index as a weight, performing weighted summation on the physical recovery index and the physical injury severity index to obtain the risk impact weight.

7. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 5, characterized in that: The process of obtaining the risk level includes: The similarity between the physical indicators of the current patient and the candidate patient with similar medications and the thrombosis presence result of the candidate patient with similar medications are integrated to obtain the risk sub-level of the candidate patient with similar medications; wherein, when the candidate patient with similar medications does not have thrombosis, the thrombosis presence result is 0, and when the candidate patient with similar medications has thrombosis, the thrombosis presence result is a preset positive number; the candidate patient with similar medications is any patient with similar medications; The risk level is obtained by combining the risk sub-levels of all patients with similar medications.

8. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 1, characterized in that: The system also includes a patient screening module for similar medication use, which is used to screen patients with similar medication use from historical patients based on the similarity between the medication use of the current patient and the historical patients.

9. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 8, characterized in that: The screening process for patients with similar medications includes: Obtaining the similarity between each medication of the current patient and each medication of a candidate historical patient; the candidate historical patient is any historical patient; Perform DTW matching based on the similarity between each medication of the current patient and each medication of the candidate historical patients to obtain several matching pairs; The similarity of medication conditions of each matching pair is integrated to obtain the similarity between the current patient and the candidate historical patients; Based on the degree of similarity, determine whether the candidate historical patients are patients with similar medication.

10. The system for predicting the risk of deep vein thrombosis and pulmonary embolism in patients with multiple injuries as claimed in claim 9, characterized in that: The process of obtaining the similarity of medication situations includes: Get the type of medication used by the current patient for the i-th time, and the type of medication used by the candidate historical patient for the j-th time; the i-th medication is any medication used by the current patient, and the j-th medication is any medication used by the candidate historical patient; Obtain the union of the drug types of the current patient's i-th medication and the drug types of the candidate historical patients' j-th medication; The ratio of the number of drug types in the union to the total number of drug types is calculated as the similarity of the medication between the i-th medication of the current patient and the j-th medication of the candidate historical patient; the total number of drug types is equal to the sum of the number of drug types in the i-th medication of the current patient and the number of drug types in the j-th medication of the candidate historical patient.

Images