A picc-dvt risk prediction method and system

Abstract

The application provides a PICC-DVT risk prediction method and system, and relates to the technical field of medical devices and medical data analysis. The method comprises the following steps: screening a target patient; obtaining blood vessel depth data, arm circumference data, blood vessel diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, blood vessel diameter data and blood flow velocity data on the first day after operation of the target patient; obtaining a judgment result of whether the target patient has deep vein thrombosis; training a PICC-DVT risk prediction model; and predicting the deep vein thrombosis risk of the patient. According to the application, the blood vessel depth, diameter, arm circumference and blood flow velocity of the catheterized vein can be monitored in real time through high-frequency ultrasound, relevant blood vessel parameters and hemodynamic data can be obtained, a combined model can be constructed, multiple parameters can be fused, the parameter weight can be optimized, the interaction between the blood vessel parameters and the blood flow velocity can be revealed, and the accuracy and reliability of the DVT risk assessment can be improved.

Description

Technical Field

[0001] This invention relates to the field of medical device and medical data analysis technology, and in particular to a PICC-DVT risk prediction method and system. Background Technology

[0002] In recent years, with the continuous advancement of medical technology, central venous catheters (CVCs) have been increasingly widely used in clinical treatment, especially in areas such as intensive care and cancer treatment. Peripherally inserted central venous catheters (PICCs) are favored due to their ease of operation and low complication rate. However, the use of PICCs also carries the risk of deep vein thrombosis (DVT). DVT not only affects patient recovery but can also lead to serious complications, such as pulmonary embolism. Therefore, effectively predicting and assessing the risk of PICC-DVT has become a hot topic in clinical research. Among related technologies, methods for predicting the risk of PICC-DVT mainly rely on clinical experience and some traditional risk assessment tools, such as the Caprini scoring system. These tools are typically based on basic patient information (such as age, sex, and medical history). However, they often overlook the morphological and hemodynamic characteristics of blood vessels, as well as the important role of parameters such as vessel depth, diameter, and blood flow velocity in the formation of deep vein thrombosis (DVT). They lack a comprehensive analysis of vascular morphology and hemodynamic parameters, and the limitations of traditional assessment tools reduce their applicability to specific patient groups. They cannot fully reflect the true risk of patients and adjust treatment plans in a timely manner. Therefore, they have problems such as low ability to monitor and dynamically assess DVT risk in real time, and low accuracy and reliability of DVT risk prediction. They are difficult to comprehensively consider multiple physiological parameters to assess DVT risk, resulting in some high-risk patients not being identified in time, which affects the effectiveness of clinical intervention.

[0003] The information disclosed in the background section of this application is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0004] This invention provides a PICC-DVT risk prediction method and system, which can solve the technical problem that related technologies are difficult to comprehensively consider multiple physiological parameters to assess DVT risk.

[0005] According to a first aspect of the present invention, a PICC-DVT risk prediction method is provided, comprising:

[0006] Target patients are selected based on preset patient screening rules;

[0007] Acquire vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery from the target patient.

[0008] To obtain the assessment results of whether the target patient has deep vein thrombosis;

[0009] Based on vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery, as well as the judgment results, a PICC-DVT risk prediction model was trained.

[0010] The trained PICC-DVT risk prediction model can be used to predict the risk of deep vein thrombosis in patients.

[0011] According to the present invention, screening target patients includes:

[0012] Determine the screening and inclusion criteria;

[0013] Determine the screening and exclusion criteria;

[0014] The screening rules are determined based on the inclusion and exclusion criteria.

[0015] Target patients are selected based on the screening criteria.

[0016] According to the present invention, the method further includes:

[0017] Based on the assessment results, the target patients were divided into two groups: the thrombosis group and the non-thrombosis group.

[0018] Determine the first trend judgment results regarding vessel depth data in target patients in the thrombosis group and the non-thrombosis group;

[0019] Determine the second trend judgment results regarding arm circumference data for target patients in the thrombosis group and the non-thrombosis group;

[0020] The third trend judgment results were determined for target patients in the thrombosis group and the non-thrombosis group regarding vessel diameter data on the day of catheter placement and on the first day after the procedure.

[0021] The fourth trend judgment results were determined for blood flow velocity data on the day of catheter placement and on the first day after the procedure in target patients in the thrombosis group and the non-thrombosis group.

[0022] According to the present invention, training a PICC-DVT risk prediction model includes:

[0023] Set the target function;

[0024] Based on the vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, blood flow velocity data on the first day after surgery, and the judgment results of multiple target patients, the target coefficients of the objective function are solved to obtain the target coefficient solution values.

[0025] The solution values ​​of the target coefficients are verified using the results of the first trend judgment, the second trend judgment, the third trend judgment, and the fourth trend judgment.

[0026] If the validation is successful, the trained PICC-DVT risk prediction model is obtained based on the objective function and the validated objective coefficients.

[0027] According to the present invention, setting the objective function includes:

[0028] According to the formula

[0029] outp i =a×x 1,i +b×x 2,i +c×x 3,i +d×x 4,i +e×x 5,i +f×x 6,i +g

[0030] Obtain the objective function, where outp i For the judgment result of the i-th target patient, x 1,i For the vascular depth data of the i-th target patient, x 2,i For the arm circumference data of the i-th target patient, x 3,i For the blood vessel diameter data on the day of catheter placement for the i-th target patient, x 4,i For the blood flow velocity data of the i-th target patient on the day of catheter placement, x 5,i For the vessel diameter data on the first day after surgery of the i-th target patient, x 6,i Here are the blood flow velocity data for the i-th target patient on the first postoperative day, where a, b, c, d, e, f, and g are the target coefficients.

[0031] According to the present invention, verifying the solution value of the target coefficient includes:

[0032] The solution value of a is verified by comparing it with the result of the first trend judgment.

[0033] The solution value of b is verified by comparing it with the result of the second trend judgment.

[0034] pass The calculated values ​​of c and e are verified against the results of the third trend judgment. Let c be the solution value. Let be the solution value of e, and n be the number of target patients;

[0035] pass The calculated values ​​of d and f are verified against the results of the fourth trend judgment. Let d be the solution value. Let f be the solution value.

[0036] According to the present invention, the method further includes:

[0037] according to Determine the patient's vascular depth contribution value, among which, For the patient's vascular depth data, Let be the solution value of 'a', and p be the probability that a patient has a risk of deep vein thrombosis obtained by the trained PICC-DVT risk prediction model.

[0038] according to Determine the contribution value of the patient's arm circumference data, among which, For the patient's arm circumference data, Let b be the solution value;

[0039] according to Determine the contribution value of vasodilation in the patient, among which, The patient's blood vessel diameter data on the day of catheter placement. The data is the patient's vessel diameter on the first day after surgery;

[0040] according to Determine the contribution value of the patient's blood flow acceleration, among which, For the patient's blood flow velocity data on the day of catheter placement, This is the blood flow velocity data for the patient on the first postoperative day.

[0041] According to a second aspect of the present invention, a PICC-DVT risk prediction system is provided, comprising:

[0042] The screening module filters target patients according to preset patient screening rules;

[0043] The acquisition module acquires the target patient's vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery.

[0044] The judgment result module obtains the judgment result of whether the target patient has deep vein thrombosis;

[0045] The training module trains a PICC-DVT risk prediction model based on vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, blood flow velocity data on the first day after surgery, and judgment results from multiple target patients.

[0046] The prediction module uses a trained PICC-DVT risk prediction model to predict the risk of deep vein thrombosis in patients.

[0047] According to a third aspect of the present invention, a PICC-DVT risk prediction device is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the PICC-DVT risk prediction method.

[0048] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having computer program instructions stored thereon, which, when executed by a processor, implement the PICC-DVT risk prediction method.

[0049] By adopting the above technical solution, the present invention can achieve the following technical effects:

[0050] According to this invention, target patients can be screened based on preset patient screening rules, and data on vascular depth, arm circumference, vessel diameter on the day of catheterization, blood flow velocity on the day of catheterization, vessel diameter on the first postoperative day, and blood flow velocity on the first postoperative day can be obtained. This allows for the determination of whether the target patient has developed deep vein thrombosis (DVT), and the PICC-DVT risk prediction model can be trained to predict the patient's DVT risk. High-frequency ultrasound can be used to monitor the vessel depth, diameter, arm circumference, and blood flow velocity of the inserted vein in real time, obtaining relevant vascular parameters and hemodynamic data. A combined model can then be constructed, fusing multiple parameters and optimizing parameter weights to reveal the interaction between vascular parameters and blood flow velocity. This comprehensive consideration of multiple physiological parameters in assessing DVT risk improves the accuracy and reliability of DVT risk assessment. Based on inclusion and exclusion criteria, pre-defined patient screening rules can be established to select target patients. Pre-processed data on vascular depth, arm circumference, vessel diameter on the day of catheterization, blood flow velocity on the day of catheterization, and vessel diameter and blood flow velocity on the first postoperative day can be obtained, providing basic data for training the PICC-DVT risk prediction model. During training, the model can be trained based on vascular depth, arm circumference, vessel diameter on the day of catheterization, blood flow velocity on the day of catheterization, vessel diameter and blood flow velocity on the first postoperative day, and judgment results from multiple target patients. The first, second, third, and fourth trend judgment results can be determined, and the target coefficient solution values ​​can be validated. Using particle swarm optimization to train the risk prediction model improves model performance and enhances the accuracy and stability of risk prediction. Furthermore, the trained PICC-DVT risk prediction model can be used to predict the risk of deep vein thrombosis in patients and calculate the contribution values ​​of vessel depth, arm circumference, vascular dilation, and blood flow acceleration. Combining contribution analysis provides an intuitive interpretation of the characteristics influencing clinical decision-making, improving the accuracy and comprehensiveness of DVT risk assessment.

[0051] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Other features and aspects of the invention will become clearer from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0053] Figure 1 An exemplary flowchart of the PICC-DVT risk prediction method according to an embodiment of the present invention is shown.

[0054] Figure 2 An exemplary schematic diagram illustrating the trend changes of body mass index, vascular conditions, and blood flow velocity on the first postoperative day according to embodiments of the present invention with PICC-DVT.

[0055] Figure 3 An exemplary schematic diagram illustrating the relationship between a patient's clinical characteristics and a PICC line according to an embodiment of the present invention is shown.

[0056] Figure 4 Exemplary diagrams of single-factor and multi-factor logistic regression according to embodiments of the present invention are shown;

[0057] Figure 5 A block diagram of the PICC-DVT risk prediction system according to an embodiment of the present invention is shown as an example. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0060] Figure 1 An exemplary flowchart of a PICC-DVT risk prediction method according to an embodiment of the present invention is shown, the method comprising:

[0061] Step S1: Select target patients according to preset patient screening rules;

[0062] Step S2: Obtain the target patient's vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery.

[0063] Step S3: Obtain the assessment result of whether the target patient has deep vein thrombosis;

[0064] Step S4: Based on the vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, blood flow velocity data on the first day after surgery, and the judgment results of multiple target patients, train the PICC-DVT risk prediction model.

[0065] Step S5: Predict the patient's risk of deep vein thrombosis using the trained PICC-DVT risk prediction model.

[0066] The PICC-DVT risk prediction method and system according to embodiments of the present invention can screen target patients according to preset patient screening rules, and acquire vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery. This allows for the determination of whether the target patient has developed deep vein thrombosis, and the PICC-DVT risk prediction model can be trained to predict the patient's deep vein thrombosis risk. High-frequency ultrasound can be used to monitor the vascular depth, diameter, arm circumference, and blood flow velocity of the catheterized vein in real time, acquiring relevant vascular parameters and hemodynamic data. A combined model can then be constructed, fusing multiple parameters and optimizing parameter weights to reveal the interaction between vascular parameters and blood flow velocity. This comprehensive consideration of multiple physiological parameters in assessing DVT risk improves the accuracy and reliability of DVT risk assessment.

[0067] Example 1:

[0068] According to an embodiment of the present invention, in step S1, screening target patients according to preset patient screening rules includes: determining screening inclusion criteria; determining screening exclusion criteria; determining screening rules according to screening inclusion criteria and screening exclusion criteria; and screening target patients according to screening rules.

[0069] According to embodiments of the present invention, inclusion criteria are determined. For example, age ≥18 years; pathological or cytological diagnosis of malignancy; PICC placement at a designated medical institution and planned systemic treatment; being of sound mind, voluntarily joining after understanding the study methods and objectives, and signing an informed consent form; tolerating ultrasound examination; not receiving anticoagulation or antiplatelet therapy prior to placement; and requiring a single-lumen PICC catheter for treatment. Selection rules can be determined based on these inclusion criteria.

[0070] According to embodiments of the present invention, screening and exclusion criteria are determined. For example, a history of deep vein thrombosis (DVT) or thrombosis in the implanted vessel; expected survival of less than six months; vessel diameter too small (<3mm), unsuitable for PICC catheter placement; patients with atrial fibrillation or pacemakers where P waves cannot be observed; blood disorders; pregnant or lactating women; or patients with active infections requiring treatment. Screening rules can be determined based on these screening and exclusion criteria.

[0071] According to an embodiment of the present invention, screening rules are determined based on inclusion criteria and exclusion criteria. The screening criteria, composed of the inclusion criteria and exclusion criteria, constitute the screening rules. Target patients can be screened based on these screening rules.

[0072] According to an embodiment of the present invention, target patients are screened based on screening rules. Patients meeting all inclusion criteria in the screening rules are included as target patients, while patients meeting any one of the exclusion criteria are excluded, thus obtaining the target patients.

[0073] Example 2:

[0074] According to an embodiment of the present invention, in step S2, the following data are acquired for the target patient: vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first postoperative day, and blood flow velocity data on the first postoperative day. The arm circumference of the arm on the side of catheter placement is measured before catheter insertion, which is the arm circumference data. The depth, diameter, and blood flow velocity of the indwelling catheter vein are measured by ultrasound, which are the vascular depth data, vascular diameter data on the day of catheter placement, and blood flow velocity data on the day of catheter placement. The diameter and blood flow velocity of the indwelling catheter vein are measured by ultrasound on the first postoperative day, which are the vascular diameter data and blood flow velocity data on the first postoperative day. Based on the same processing method, vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first postoperative day, and blood flow velocity data on the first postoperative day can be acquired for all target patients. Of course, the above-mentioned data for the target patients can be stored in the electronic medical record system of each target patient and can be directly obtained from the electronic medical record system. Furthermore, by preprocessing the data of all the target patients (e.g., deleting data containing arbitrary missing values, replacing missing labels with NaN, etc.), we can obtain the preprocessed data of the target patients, which can be used to train the PICC-DVT risk prediction model.

[0075] In this way, pre-defined patient screening rules can be determined based on inclusion and exclusion criteria, thereby screening target patients and obtaining pre-processed vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery for target patients, providing basic data for training the PICC-DVT risk prediction model.

[0076] Example 3:

[0077] According to an embodiment of the present invention, in step S3, the determination result of whether the target patient has deep vein thrombosis is obtained. Through DVT-related examinations, it is determined whether each target patient has deep vein thrombosis. The determination result of the target patient with deep vein thrombosis is set to 1, and the determination result of the target patient without deep vein thrombosis is set to 0.

[0078] Example 4:

[0079] Figure 2 An exemplary schematic diagram illustrating the trend changes of body mass index, vascular conditions, and blood flow velocity on the first postoperative day according to an embodiment of the present invention with PICC-DVT.

[0080] According to an embodiment of the present invention, the method further includes: grouping target patients according to the judgment result to obtain target patients in the thrombosis group and target patients in the non-thrombosis group; determining a first trend judgment result of vascular depth data for the target patients in the thrombosis group and the target patients in the non-thrombosis group; determining a second trend judgment result of arm circumference data for the target patients in the thrombosis group and the target patients in the non-thrombosis group; determining a third trend judgment result of vascular diameter data on the day of catheterization and vascular diameter data on the first day after surgery for the target patients in the thrombosis group and the target patients in the non-thrombosis group; and determining a fourth trend judgment result of blood flow velocity data on the day of catheterization and vascular flow velocity data on the first day after surgery for the target patients in the thrombosis group and the target patients in the non-thrombosis group.

[0081] According to an embodiment of the present invention, target patients are grouped based on the judgment result to obtain target patients in the thrombosis group and target patients in the non-thrombosis group. Target patients with a judgment result of 1 are identified as target patients in the thrombosis group, and target patients with a judgment result of 0 are identified as target patients in the non-thrombosis group.

[0082] According to an embodiment of the present invention, a first trend judgment result regarding vascular depth data is determined for target patients in the thrombosis group and the non-thrombosis group. When a certain data of a target patient is positively correlated with the risk of DVT, the corresponding trend judgment result can be determined as 0; when it is negatively correlated with the risk of DVT, the corresponding trend judgment result can be determined as 1. For example... Figure 2As shown, through the study and analysis of the relationship between body mass index and vascular conditions, as well as changes in blood flow velocity on the first postoperative day, and their impact on thrombosis in multiple target patients (338 target patients in the example), it can be concluded that as the vascular depth data of the target patients increases, the risk of DVT will decrease. The vascular depth data of the target patients is negatively correlated with the risk of DVT. Therefore, the first trend judgment result can be determined as 1.

[0083] According to an embodiment of the present invention, a second trend judgment result regarding arm circumference data is determined for target patients in the thrombosis group and the non-thrombosis group. For example... Figure 2 As shown, through the study and analysis of the relationship between body mass index and vascular conditions of multiple target patients, as well as the changes in blood flow velocity on the first day after surgery, and their impact on thrombosis, it can be concluded that as the arm circumference data of target patients increases, the risk of DVT will decrease. The arm circumference data of target patients is negatively correlated with the risk of DVT. Therefore, the second trend judgment result can be determined as 1.

[0084] According to an embodiment of the present invention, a third trend judgment result is determined for the vessel diameter data on the day of catheter placement and the first day post-procedure in target patients in the thrombosis group and the non-thrombosis group. The difference between the vessel diameter data on the first day post-procedure and the vessel diameter data on the day of catheter placement indicates a change in vessel diameter; when the difference is greater than 0, the vessel diameter data increases; when the difference is less than 0, the vessel diameter data decreases. Figure 2 As shown, through the study and analysis of the relationship between body mass index and vascular conditions of multiple target patients, as well as the changes in blood flow velocity on the first day after surgery, and their impact on thrombosis, it can be concluded that as the vascular diameter data of target patients increases, and as the amount of increase in vascular diameter data (i.e., the amount of vascular dilation) increases, the risk of DVT will decrease. Therefore, the third trend judgment result can be determined as 1.

[0085] According to an embodiment of the present invention, a fourth trend judgment result is determined for blood flow velocity data on the day of catheter placement and on the first postoperative day for target patients in the thrombosis group and the non-thrombosis group. The difference between the blood flow velocity data on the first postoperative day and the blood flow velocity data on the day of catheter placement indicates a change in blood flow velocity; when the difference is greater than 0, the blood flow velocity data increases; when the difference is less than 0, the blood flow velocity data decreases. Figure 2 As shown, through the study and analysis of the relationship between body mass index and vascular conditions of multiple target patients, as well as the changes in blood flow velocity on the first day after surgery, and their impact on thrombosis, it can be concluded that as the blood flow velocity data of the target patients increases, and as the amount of increase in blood flow velocity data (i.e., the amount of blood flow acceleration) increases, the risk of DVT will decrease. Therefore, the fourth trend judgment result can be determined as 1.

[0086] According to embodiments of the present invention, such as Figure 2 As shown, as the body mass index of the target patients increases, the data on vascular depth, arm circumference, and vascular diameter all increase. Furthermore, the data on vascular depth is not significantly correlated with the change in blood flow velocity on the first postoperative day. However, the data on arm circumference and vascular diameter are positively correlated with the change in blood flow velocity (i.e., the degree of blood flow acceleration) on the first postoperative day. In other words, as the data on arm circumference and vascular diameter increase, the degree of blood flow acceleration on the first postoperative day will increase.

[0087] Example 5:

[0088] According to an embodiment of the present invention, in step S4, a PICC-DVT risk prediction model is trained based on vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery, as well as the judgment results of multiple target patients. This includes: setting an objective function; solving for the objective coefficients of the objective function based on the vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery, as well as the judgment results of multiple target patients, to obtain the target coefficient solution values; verifying the target coefficient solution values ​​through the first trend judgment results, the second trend judgment results, the third trend judgment results, and the fourth trend judgment results; and obtaining the trained PICC-DVT risk prediction model based on the objective function and the verified target coefficient solution values ​​if the verification is successful.

[0089] According to an embodiment of the present invention, setting the objective function includes: obtaining the objective function according to formula (1).

[0090] outp i =a×x 1,i +b×x 2,i +c×x 3,i +d×x 4,i +e×x 5,i +f×x 6,i +g(1)

[0091] Among them, outp i For the judgment result of the i-th target patient, x 1,i For the vascular depth data of the i-th target patient, x 2,i For the arm circumference data of the i-th target patient, x 3,i For the blood vessel diameter data on the day of catheter placement for the i-th target patient, x 4,i For the blood flow velocity data of the i-th target patient on the day of catheter placement, x 5,i For the vessel diameter data on the first day after surgery of the i-th target patient, x 6,iHere are the blood flow velocity data for the i-th target patient on the first postoperative day, where a, b, c, d, e, f, and g are the target coefficients.

[0092] According to an embodiment of the present invention, the target coefficients of the objective function are solved based on vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery, as well as the judgment results of multiple target patients, to obtain the target coefficient solution values. Using a particle swarm optimization algorithm, the optimization variable dimension is set to 7, corresponding to the seven target coefficients in the objective function. The number of individual particles is set to 1000, and the maximum number of iterations is set to 300. Further, based on the vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery, as well as the judgment results of multiple target patients, the target coefficients of the objective function are solved. The optimal solution for each target coefficient (i.e., the position with the best fitness found by each particle during the iteration process) and the global optimal solution (i.e., the position with the best fitness found by all particles in the entire ion swarm) are found until the termination condition is met (e.g., reaching the maximum number of iterations). The global optimal solution for each target coefficient is then obtained, which is the target coefficient solution value. It can effectively handle complex linear and nonlinear relationships and does not depend on the gradient information of the objective function, thus improving the training efficiency of the PICC-DVT risk prediction model.

[0093] According to an embodiment of the present invention, the solution value of the target coefficient is verified by using the first trend judgment result, the second trend judgment result, the third trend judgment result, and the fourth trend judgment result, including: verifying the solution value of 'a' by comparing it with the first trend judgment result; verifying the solution value of 'b' by comparing it with the second trend judgment result; and verifying the solution value of 'b' by comparing it with the second trend judgment result. The calculated values ​​of c and e are verified against the results of the third trend judgment. Let c be the solution value. Let e ​​be the solution value, and n be the number of target patients; through The calculated values ​​of d and f are verified against the results of the fourth trend judgment. Let d be the solution value. Let f be the solution value.

[0094] According to an embodiment of the present invention, the solution value of 'a' is verified by comparing it with the first trend judgment result. The solution value of 'a' is the solution value of the target coefficient corresponding to the vascular depth data of the target patient. A first trend judgment result of 1 indicates that as the vascular depth data of the target patient increases, the risk of DVT decreases, and the vascular depth data of the target patient is negatively correlated with the risk of DVT. Therefore, the solution value of the target coefficient corresponding to the vascular depth data of the target patient should be less than 0. When the solution value of 'a' is less than 0, it can be considered that the vascular depth data of the target patient is negatively correlated with the risk of DVT, which conforms to the first trend judgment result, and the verification of the solution value of 'a' is successful. Conversely, when other trend judgment results are 0, it indicates that as the corresponding data increases, the risk of DVT increases, and the corresponding data is positively correlated with the risk of DVT. Therefore, the solution value of the corresponding target coefficient should be greater than 0, thereby verifying the solution values ​​of the target coefficients corresponding to other data.

[0095] According to an embodiment of the present invention, the solution value of b is verified by comparing it with the second trend judgment result. Similar to the verification of the solution value of a, a second trend judgment result of 1 indicates that as the target patient's arm circumference data increases, the risk of DVT decreases, and the target patient's arm circumference data is negatively correlated with the risk of DVT. When the solution value of b is less than 0, it can be considered that the target patient's arm circumference data is negatively correlated with the risk of DVT, which is consistent with the second trend judgment result, and the verification of the solution value of b is successful.

[0096] According to an embodiment of the present invention, by The calculated values ​​of c and e are verified against the results of the third trend judgment. Let c be the solution value. Let be the solution value of e, and n be the number of target patients. This represents the product of the sum of the vessel diameter data on the day of catheterization for each target patient and the calculated value of c. This represents the product of the sum of the vessel diameter data on the first postoperative day for each target patient and the calculated value of e. The third trend judgment result is 1, indicating that as the vessel diameter data of the target patients increases, and as the amount of increase in vessel diameter data (i.e., vascular dilation) increases, the risk of DVT decreases. Therefore, the calculated values ​​of c and e should be one positive and one negative, and the sum of the product of the sum of the vessel diameter data on the day of catheterization for each target patient and the calculated value of c, and the sum of the product of the sum of the vessel diameter data on the first postoperative day for each target patient and the calculated value of e, should be less than 0. When the value is less than 0, it can be considered that as the difference between the vessel diameter data on the day of catheterization and the vessel diameter data on the first day after surgery (i.e., the amount of vascular dilation) increases, the risk of DVT decreases. That is, the amount of vascular dilation is negatively correlated with the risk of DVT, which is consistent with the third trend judgment result. The verification of the solution values ​​of c and e is passed.

[0097] According to an embodiment of the present invention, by The calculated values ​​of d and f are verified against the results of the fourth trend judgment. Let d be the solution value. Let f be the solution value. The fourth trend judgment result is determined to be 1, indicating that as the blood flow velocity data of the target patient increases, and as the amount of increase in blood flow velocity data (i.e., the amount of blood flow acceleration) increases, the risk of DVT will decrease. Similar to the verification of the solution values ​​of d and f, when When the value is less than 0, the solution values ​​of d and f can be considered to be positive and negative respectively. As the difference between the blood flow velocity data on the day of catheter placement and the blood flow velocity data on the first day after surgery (i.e., the amount of blood flow acceleration) increases, the risk of DVT decreases. That is, the amount of blood flow acceleration is negatively correlated with the risk of DVT, which is consistent with the fourth trend judgment result. The verification of the solution values ​​of d and f is passed.

[0098] According to an embodiment of the present invention, if the verification passes, the trained PICC-DVT risk prediction model is obtained based on the objective function and the verified target coefficient solutions. If the verification of all target coefficient solutions passes, the target coefficient solutions are considered correct. Then, the verified target coefficient solutions are substituted into the objective function to obtain the trained PICC-DVT risk prediction model. Conversely, if the verification of any one target coefficient solution fails, training continues until the verification of all target coefficient solutions passes.

[0099] In this way, a PICC-DVT risk prediction model can be trained based on vascular depth data, arm circumference data, vessel diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vessel diameter data on the first postoperative day, and blood flow velocity data on the first postoperative day, along with the judgment results from multiple target patients. The model then determines the first, second, third, and fourth trend judgment results, and subsequently validates the target coefficient solution. Using particle swarm optimization to train the risk prediction model improves its performance and enhances the accuracy and stability of risk prediction.

[0100] Example 6:

[0101] According to an embodiment of the present invention, in step S5, the patient's deep vein thrombosis risk is predicted using a trained PICC-DVT risk prediction model. The patient's vessel depth data, arm circumference data, vessel diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vessel diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery are obtained. These data are then input into the trained PICC-DVT risk prediction model. The predicted deep vein thrombosis risk is mapped to the range [0,1] through an activation layer (i.e., a sigmoid activation function), representing the patient's deep vein thrombosis risk probability. When the probability is greater than 0.5, the patient is considered to have a risk of deep vein thrombosis; when the probability is less than 0.5, the patient is considered not to have a risk of deep vein thrombosis.

[0102] Example 7:

[0103] According to an embodiment of the present invention, the method further includes: based on Determine the patient's vascular depth contribution value, among which, For the patient's vascular depth data, Let be the solved value of 'a', and p be the probability that a patient has a risk of deep vein thrombosis, obtained by the trained PICC-DVT risk prediction model; according to Determine the contribution value of the patient's arm circumference data, among which, For the patient's arm circumference data, Let b be the solution value; according to Determine the contribution value of vasodilation in the patient, among which, The patient's blood vessel diameter data on the day of catheter placement. The data for the patient's vessel diameter on the first day after surgery; based on Determine the contribution value of the patient's blood flow acceleration, among which, For the patient's blood flow velocity data on the day of catheter placement, This is the blood flow velocity data for the patient on the first postoperative day.

[0104] According to an embodiment of the present invention, based on Determine the patient's vascular depth contribution value, among which, For the patient's vascular depth data, Let be the solution value of 'a', and p be the probability that a patient has a risk of deep vein thrombosis obtained by the trained PICC-DVT risk prediction model. The product of the patient's vascular depth data and the solved value of 'a' represents the proportion of the probability that the patient has a risk of deep vein thrombosis obtained by the trained PICC-DVT risk prediction model. It can be considered as the patient's vascular depth contribution value, which describes the impact of the patient's vascular depth data on the probability of the patient having a risk of deep vein thrombosis. The larger the vascular depth contribution value, the greater the impact of the patient's vascular depth data on the probability of the patient having a risk of deep vein thrombosis, and the greater the attention should be paid to vascular depth data.

[0105] According to an embodiment of the present invention, based on Determine the contribution value of the patient's arm circumference data, among which, For the patient's arm circumference data, This is the solution value for b. Similar to determining the contribution value of the patient's vascular depth data, The product of the patient's arm circumference data and the calculated value of b represents the proportion of the probability that the patient has a risk of deep vein thrombosis obtained by the trained PICC-DVT risk prediction model. It can be considered as the contribution value of the patient's arm circumference data and can describe the influence of the patient's arm circumference data on the probability that the patient has a risk of deep vein thrombosis.

[0106] According to an embodiment of the present invention, based on Determine the contribution value of vasodilation in the patient, among which, The patient's blood vessel diameter data on the day of catheter placement. The data shows the diameter of the blood vessels on the first day after surgery. The sum of the product of the patient's vessel diameter data on the day of catheter placement and the calculated value of c, and the product of the patient's vessel diameter data on the first day after surgery and the calculated value of e, represents the proportion of the probability of the patient having a risk of deep vein thrombosis obtained by the trained PICC-DVT risk prediction model. It can be used as the patient's vascular dilation contribution value, which describes the impact of the degree of vascular dilation on the probability of the patient having a risk of deep vein thrombosis. The larger the vascular dilation contribution value, the greater the impact of the degree of vascular dilation on the probability of the patient having a risk of deep vein thrombosis, and the greater the attention should be paid to the degree of vascular dilation.

[0107] According to an embodiment of the present invention, based on Determine the contribution value of the patient's blood flow acceleration, among which, For the patient's blood flow velocity data on the day of catheter placement, This data represents the patient's blood flow velocity on the first postoperative day. Similar to obtaining the patient's vasodilation contribution value, It can be used as a contribution value of blood flow acceleration in patients, which can describe the impact of the degree of blood flow acceleration on the probability of patients having the risk of deep vein thrombosis.

[0108] In this way, the trained PICC-DVT risk prediction model can predict a patient's deep vein thrombosis risk and calculate the contribution values ​​of vessel depth, arm circumference, vascular dilation, and blood flow acceleration. Combined with contribution analysis, this provides an intuitive interpretation of the characteristics affecting clinical decision-making, improving the accuracy and comprehensiveness of DVT risk assessment.

[0109] The PICC-DVT risk prediction method and system according to embodiments of the present invention can screen target patients according to preset patient screening rules, and acquire vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery. This allows for the determination of whether the target patient has developed deep vein thrombosis, and the PICC-DVT risk prediction model can be trained to predict the patient's deep vein thrombosis risk. High-frequency ultrasound can be used to monitor the vascular depth, diameter, arm circumference, and blood flow velocity of the catheterized vein in real time, acquiring relevant vascular parameters and hemodynamic data. A combined model can then be constructed, fusing multiple parameters and optimizing parameter weights to reveal the interaction between vascular parameters and blood flow velocity. This comprehensive consideration of multiple physiological parameters in assessing DVT risk improves the accuracy and reliability of DVT risk assessment. Based on inclusion and exclusion criteria, pre-defined patient screening rules can be established to select target patients. Pre-processed data on vascular depth, arm circumference, vessel diameter on the day of catheterization, blood flow velocity on the day of catheterization, and vessel diameter and blood flow velocity on the first postoperative day can be obtained, providing basic data for training the PICC-DVT risk prediction model. During training, the model can be trained based on vascular depth, arm circumference, vessel diameter on the day of catheterization, blood flow velocity on the day of catheterization, vessel diameter and blood flow velocity on the first postoperative day, and judgment results from multiple target patients. The first, second, third, and fourth trend judgment results can be determined, and the target coefficient solution values ​​can be validated. Using particle swarm optimization to train the risk prediction model improves model performance and enhances the accuracy and stability of risk prediction. Furthermore, the trained PICC-DVT risk prediction model can be used to predict the risk of deep vein thrombosis in patients and calculate the contribution values ​​of vessel depth, arm circumference, vascular dilation, and blood flow acceleration. Combining contribution analysis provides an intuitive interpretation of the characteristics influencing clinical decision-making, improving the accuracy and comprehensiveness of DVT risk assessment.

[0110] Example 8:

[0111] Figure 3 An exemplary schematic diagram illustrating the relationship between a patient's clinical characteristics and a PICC line according to an embodiment of the present invention is shown.

[0112] According to embodiments of the present invention, an analysis of the relationship between a patient's clinical characteristics and a PICC line is provided, such as... Figure 3 As shown, with increasing body mass index (BMI), arm circumference, vessel depth, and vessel diameter also increase, while the incidence of thrombosis decreases. Blood flow velocity also increases with increasing BMI, but peaks at a BMI of 28.3 kg / m², decreasing beyond this value. Postoperative changes in blood flow velocity were not related to BMI. Notably, 90.8% of patients in this cohort had a BMI below 28.3 kg / m², indicating a positive correlation between blood flow velocity and BMI within the primary observation range.

[0113] Example 9:

[0114] Figure 4 Exemplary diagrams of single-factor and multi-factor logistic regressions according to embodiments of the present invention are shown.

[0115] According to embodiments of the present invention, univariate and multivariate analyses of thrombosis are provided. For example... Figure 4 As shown, demographic analysis revealed a significant difference in gender distribution between the thrombosis group and the non-thrombosis group (P=0.040), with a higher incidence of thrombosis in male patients (42.6%) than in female patients (31.5%). However, multivariate analysis showed that gender was not an independent risk factor (OR=0.747, P=0.369). There was no statistically significant difference in age between the two groups (51.68±13.32 vs 51.83±12.96, P=0.921), nor were there significant differences in the prevalence of hypertension (35.5% vs 35.5%) and diabetes (40.0% vs 35.1%) (P>0.05).

[0116] According to embodiments of the present invention, such as Figure 4As shown, analysis of vascular morphology parameters revealed that the thrombosis group had significantly shallower vessels (median 0.918 cm vs 1.07 cm, P < 0.001), and multivariate analysis showed a reduced risk of thrombosis (OR = 0.324, 95%, CI: 0.120–0.874, P = 0.026). Smaller arm circumference (median 26.61 cm vs 27.26 cm, P = 0.042) was also observed, but multivariate analysis did not show independence (P = 0.587). Smaller vessel diameters were more prone to thrombosis, but this was not statistically significant (median 0.467 cm vs 0.487 cm, P = 0.072). On the first day after surgery, the diameter of blood vessels in the thrombosis group was lower than that in the non-thrombosis group (0.458 cm vs 0.493 cm, P=0.008), and multivariate analysis also showed that it was a protective factor (OR=0.025, 95%, CI: 0.001-0.550, P=0.019).

[0117] According to embodiments of the present invention, such as Figure 4 As shown, the analysis of hemodynamic parameters indicates that there was no difference in blood flow velocity between the groups on the day of catheter placement (median 14.95 cm / s vs 14.97 cm / s, P = 0.968). However, the blood flow velocity was significantly lower in the thrombosis group on the first day after the procedure (median 13.4 cm / s vs 14.9 cm / s, P = 0.008), but the multivariate analysis did not reach a significant level (OR = 0.956, P = 0.099).

[0118] According to embodiments of the present invention, such as Figure 4 As shown, the analysis of body mass index (BMI) and grouping revealed that the thrombosis group had a lower BMI (median 23.18 vs. 24.19, P = 0.014). Further BMI grouping analysis indicated that the low BMI group (<21.45) had the highest incidence of thrombosis (47.6% vs. 26.7%-38.6%, P = 0.022), and the statistical significance of BMI disappeared in the multivariate analysis (OR = 0.946, P = 0.398), suggesting that its influence may be mediated through vascular parameters.

[0119] According to embodiments of the present invention, such as Figure 4 As shown, the analysis of operational factors suggests that the success rate of a single puncture is 97.6%. Patients with ≥2 punctures have a slightly higher risk of thrombosis (62.5% vs 34.8%, P=0.109), and patients with ≥2 catheter insertions have a slightly higher risk of thrombosis (40.0% vs 34.9%, P=0.527), but none of these results are statistically significant.

[0120] Example 10:

[0121] Figure 5An exemplary block diagram of a PICC-DVT risk prediction system according to an embodiment of the present invention is shown, the system comprising:

[0122] The screening module filters target patients according to preset patient screening rules;

[0123] The acquisition module acquires the target patient's vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery.

[0124] The judgment result module obtains the judgment result of whether the target patient has deep vein thrombosis;

[0125] The training module trains a PICC-DVT risk prediction model based on vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, blood flow velocity data on the first day after surgery, and judgment results from multiple target patients.

[0126] The prediction module uses a trained PICC-DVT risk prediction model to predict the risk of deep vein thrombosis in patients.

[0127] According to an embodiment of the present invention, a PICC-DVT risk prediction device is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the PICC-DVT risk prediction method.

[0128] According to one embodiment of the present invention, a computer-readable storage medium is provided, on which computer program instructions are stored, wherein the computer program instructions, when executed by a processor, implement the PICC-DVT risk prediction method.

[0129] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0130] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functions and structural principles of the present invention have been demonstrated and explained in the embodiments, and any variations or modifications may be made to the implementation of the present invention without departing from the stated principles.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A PICC-DVT risk prediction method, characterized in that, include: Target patients are selected based on preset patient screening rules; Acquire vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery from the target patient. To obtain the assessment results of whether the target patient has deep vein thrombosis; Based on vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery, as well as the judgment results, a PICC-DVT risk prediction model was trained. The risk of deep vein thrombosis in patients is predicted using a trained PICC-DVT risk prediction model. The method further includes: Based on the assessment results, the target patients were divided into two groups: the thrombosis group and the non-thrombosis group. Determine the first trend judgment results regarding vessel depth data in target patients in the thrombosis group and the non-thrombosis group; Determine the second trend judgment results regarding arm circumference data for target patients in the thrombosis group and the non-thrombosis group; The third trend judgment results were determined for target patients in the thrombosis group and the non-thrombosis group regarding vessel diameter data on the day of catheter placement and on the first day after the procedure. The fourth trend judgment results were determined for blood flow velocity data on the day of catheter placement and on the first day post-procedure in target patients in the thrombosis group and the non-thrombosis group. The trend judgment result is either positive or negative correlation; Based on vascular depth data, arm circumference data, vessel diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vessel diameter data on the first postoperative day, and blood flow velocity data on the first postoperative day, along with assessment results from multiple target patients, a PICC-DVT risk prediction model was trained, including: Set the target function; Based on the vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, blood flow velocity data on the first day after surgery, and the judgment results of multiple target patients, the target coefficients of the objective function are solved to obtain the target coefficient solution values. The solution values ​​of the target coefficients are verified using the results of the first trend judgment, the second trend judgment, the third trend judgment, and the fourth trend judgment. If the verification is successful, the trained PICC-DVT risk prediction model is obtained based on the objective function and the verified objective coefficients. If the solution value of any target coefficient fails verification, training continues until the solution values ​​of all target coefficients pass verification.

2. The PICC-DVT risk prediction method according to claim 1, characterized in that, Based on preset patient screening rules, target patients are screened, including: Determine the screening and inclusion criteria; Determine the screening and exclusion criteria; The screening rules are determined based on the inclusion and exclusion criteria. Target patients are selected based on the screening criteria.

3. The PICC-DVT risk prediction method according to claim 1, characterized in that, Set the target function, including: According to the formula outp i =a×x 1,i +b×x 2,i +c×x 3,i +d×x 4,i +e×x 5,i +f×x 6,i +g Obtain the objective function, where outp i For the judgment result of the i-th target patient, x 1,i For the vascular depth data of the i-th target patient, x 2,i For the arm circumference data of the i-th target patient, x 3,i For the blood vessel diameter data on the day of catheter placement for the i-th target patient, x 4,i For the blood flow velocity data of the i-th target patient on the day of catheter placement, x 5,i For the vessel diameter data on the first day after surgery of the i-th target patient, x 6,i Here are the blood flow velocity data for the i-th target patient on the first postoperative day, where a, b, c, d, e, f, and g are the target coefficients.

4. The PICC-DVT risk prediction method according to claim 3, characterized in that, The calculated target coefficients are verified using the results of the first trend assessment, the second trend assessment, the third trend assessment, and the fourth trend assessment, including: The solution value of a is verified by comparing it with the result of the first trend judgment. The solution value of b is verified by comparing it with the result of the second trend judgment. pass The calculated values ​​of c and e are verified against the results of the third trend judgment. Let c be the solution value. Let be the solution value of e, and n be the number of target patients; pass The calculated values ​​of d and f are verified against the results of the fourth trend judgment. Let d be the solution value. Let f be the solution value.

5. The PICC-DVT risk prediction method according to claim 4, characterized in that, The method further includes: according to Determine the patient's vascular depth contribution value, among which, For the patient's vascular depth data, Let be the solution value of 'a', and p be the probability that a patient has a risk of deep vein thrombosis obtained by the trained PICC-DVT risk prediction model. according to Determine the contribution value of the patient's arm circumference data, among which, For the patient's arm circumference data, Let b be the solution value; according to Determine the contribution value of vasodilation in the patient, among which, The patient's blood vessel diameter data on the day of catheter placement. The data is the patient's vessel diameter on the first day after surgery; according to Determine the contribution value of the patient's blood flow acceleration, among which, For the patient's blood flow velocity data on the day of catheter placement, This is the blood flow velocity data for the patient on the first postoperative day.

6. A PICC-DVT risk prediction system, characterized in that, For performing the method as described in any one of claims 1-5, comprising: The screening module filters target patients according to preset patient screening rules; The acquisition module acquires the target patient's vascular depth data, arm circumference data, vascular diameter data on the day of catheter placement, blood flow velocity data on the day of catheter placement, vascular diameter data on the first day after surgery, and blood flow velocity data on the first day after surgery. The judgment result module obtains the judgment result of whether the target patient has deep vein thrombosis; The training module trains a PICC-DVT risk prediction model based on vascular depth data, arm circumference data, vascular diameter data on the day of catheterization, blood flow velocity data on the day of catheterization, vascular diameter data on the first day after surgery, blood flow velocity data on the first day after surgery, and judgment results from multiple target patients. The prediction module uses a trained PICC-DVT risk prediction model to predict the risk of deep vein thrombosis in patients.

7. A PICC-DVT risk prediction device, characterized in that, include: processor; A memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, It stores computer program instructions that, when executed by a processor, implement the method of any one of claims 1-5.

Images