A multi-modal data fusion diabetic foot high-risk assessment and nursing system

Abstract

The application relates to the technical field of diabetic foot analysis, in particular to a multi-modal data fusion diabetic foot high-risk assessment and nursing system. The system obtains blood glucose data and foot pressure data of a diabetic foot patient; according to the fluctuation of the blood glucose data in the current time period and the change of abnormal blood glucose data exceeding the preset normal blood glucose range every day, the current blood glucose risk degree of the diabetic foot patient is obtained; according to the synchronization of the abnormal blood glucose data and the foot pressure data, the current hospital examination doctor assessment severity level and the current blood glucose risk degree, the current overall abnormality degree of the diabetic foot patient is obtained, and the diabetic foot high-risk assessment and nursing are carried out. The current overall abnormality degree is obtained, which is beneficial to timely and accurate assessment of the diabetic foot risk of the diabetic foot patient, and then clinical nursing measures are formulated for the diabetic foot patient in a timely and accurate manner, and the aggravation of the diabetic foot is effectively avoided.

Description

Technical Field

[0001] This invention relates to the field of diabetic foot analysis technology, specifically to a multimodal data fusion system for assessing and caring for high-risk diabetic foot. Background Technology

[0002] Diabetic foot is one of the most serious and costly complications of diabetes. It is characterized by external stress acting on a foundation of neuropathy and vascular disease, leading to ulcers, infections, and even gangrene in the foot. It is characterized by high rates of disability and amputation. Timely and accurate high-risk assessment of diabetic foot is crucial for effective intervention and preventing disease progression.

[0003] Currently, clinical assessment of diabetic foot risk largely relies on data from a single clinical examination. However, in reality, the occurrence and development of diabetic foot is a continuous and dynamic process, resulting from the combined effects of multiple pathological mechanisms. Relying solely on data from a single clinical examination cannot effectively integrate dynamic indicators reflecting disease progression and multi-source dynamic information. This leads to inaccurate risk assessment of diabetic foot, hinders the development of accurate clinical nursing interventions for diabetic foot patients, and makes it difficult to achieve personalized and precise management. Ultimately, this impacts the quality of life of diabetic foot patients and exacerbates their condition. Summary of the Invention

[0004] To address the technical problem of inaccurate risk assessment of diabetic foot, the present invention aims to provide a multimodal data fusion system for high-risk assessment and care of diabetic foot. The specific technical solution adopted is as follows: This invention provides a multimodal data fusion-based high-risk assessment and care system for diabetic foot, the system comprising: The data acquisition module is used to acquire daily blood glucose and plantar pressure data for diabetic foot patients; The current blood glucose risk level acquisition module is used to obtain the current blood glucose risk level of diabetic foot patients based on the fluctuation of daily blood glucose data of diabetic foot patients in the current time period, as well as the changes of abnormal blood glucose data of diabetic foot patients that exceed the preset normal blood glucose range each day. The current overall abnormality acquisition module is used to acquire the current overall abnormality of diabetic foot patients based on the synchronization of abnormal blood glucose data and plantar pressure data change trends, the severity level assessed by the current hospital doctor, and the current blood glucose risk level. The data processing module is used to assess and manage high-risk diabetic foot based on the current overall level of abnormality.

[0005] Furthermore, the method for obtaining the current blood glucose risk level is as follows: For any diabetic foot patient, obtain the current reference abnormality analysis value of the diabetic foot patient based on the fluctuation of the patient's daily blood glucose data in the current time period; Based on the daily changes in abnormal blood glucose data that exceed the preset normal blood glucose range of the diabetic foot patient, obtain the abnormal blood glucose change analysis value of the diabetic foot patient. The product of the current reference anomaly analysis value and the blood glucose anomaly change analysis value is taken as the current blood glucose risk level of the diabetic foot patient.

[0006] Furthermore, the method for obtaining the current reference anomaly analysis value is as follows: For any day in the current time period, the difference between the maximum and minimum blood glucose data of the diabetic foot patient on that day is taken as the first difference; The result of normalizing the product of the standard deviation and the first difference of the blood glucose data of the diabetic foot patient on that day is taken as the first degree of blood glucose abnormality of the diabetic foot patient on that day. The average of the first abnormal blood glucose level for each day during the current time period for this diabetic foot patient is used as the current reference abnormality analysis value for this diabetic foot patient.

[0007] Furthermore, the method for obtaining the abnormal blood glucose change analysis values ​​is as follows: Based on the daily blood glucose data of the diabetic foot patient exceeding the preset normal blood glucose range, the second degree of blood glucose abnormality of the diabetic foot patient is obtained daily. The degree of the second blood glucose abnormality was fitted into a curve according to the time sequence, which was used as the first curve; The maximum points on the first curve are fitted into a straight line according to the time sequence, and the slope of the straight line is normalized to obtain the result as the blood glucose abnormality analysis value of the diabetic foot patient.

[0008] Furthermore, the method for obtaining the second degree of blood glucose abnormality is as follows: For any given day of the diabetic foot patient, the blood glucose data for that day is mapped onto a two-dimensional coordinate system in chronological order and fitted as a curve, which serves as the second curve; where the horizontal axis of the two-dimensional coordinate system represents time and the vertical axis represents blood glucose data. The area within the upper and lower baselines of the second curve, which is not within the preset normal blood glucose range, is obtained by integral method. This area is used as the second degree of blood glucose abnormality for the diabetic foot patient on that day.

[0009] Furthermore, the method for obtaining the current overall anomaly level is as follows: For any diabetic foot patient, the current overall impact of the diabetic foot disease can be obtained based on the synchronization of the abnormal blood glucose data and plantar pressure data trends of the diabetic foot patient. The first characteristic value is the product of the current blood glucose risk level and the current overall impact level. The product of the negative correlation between the current blood glucose risk level and the normalized result of the severity level assessed by the current hospital physician is used as the second characteristic value. The sum of the first and second eigenvalues ​​is taken as the current overall degree of abnormality for the diabetic foot patient.

[0010] Furthermore, the method for obtaining the current overall level of impact is as follows: The mean daily plantar pressure data of this diabetic foot patient was fitted into a curve according to time sequence, which is used as the pressure curve; All the maximum points on the pressure curve are taken as the first maximum points; all the maximum points on the first curve are taken as the second maximum points. For any first maximum point, obtain the number of days between the first maximum point and the corresponding days of each second maximum point, and use them as the interval reference duration; The minimum interval reference duration is used as the consistent analysis value for the first maximum point; The mean of all first maximum points, after negative correlation and normalization, is used as the current overall impact on the diabetic foot patient.

[0011] Furthermore, the method for assessing and caring for high-risk diabetic foot based on the current overall degree of abnormality is as follows: The risk level of diabetic foot patients is classified into different levels based on the current overall degree of abnormality. A predetermined number of diabetic foot patients with known high-risk levels of diabetic foot were used as the training and validation sets for the neural network. The trained neural network is obtained by using the training set and the validation set; A trained neural network is used to assess the high risk of diabetic foot in patients, thereby enabling them to receive proper care.

[0012] Furthermore, the method for classifying the high-risk level of diabetic foot patients based on the current overall degree of abnormality is as follows: When the current overall abnormality level is less than the preset first abnormality level threshold, the high risk of diabetic foot in the corresponding diabetic foot patient will be classified as low risk. When the current overall abnormality level is greater than or equal to the preset first abnormality level threshold and less than the preset second abnormality level threshold, the high risk of diabetic foot in the corresponding diabetic foot patient will be classified as medium risk. When the current overall abnormality level is greater than or equal to the preset second abnormality level threshold, the high risk of diabetic foot in the corresponding diabetic foot patient is classified as high risk.

[0013] Furthermore, the severity level is set by the physician based on the Wagner classification system.

[0014] The present invention has the following beneficial effects: This invention first obtains the current blood glucose risk level of diabetic foot patients based on the daily fluctuations of their blood glucose data within the current time period, as well as the changes in abnormal blood glucose data that exceed the preset normal range. This accurately reflects the current abnormal blood glucose status of diabetic foot patients, indirectly determining their current underlying disease condition. To accurately analyze the impact on diabetic foot patients, the invention further analyzes the influence of internal disease and external stress on diabetic foot patients based on the synchronization of abnormal blood glucose data and plantar pressure data trends, preparing for subsequent accurate analysis of the severity of diabetic foot. Then, combining the severity level assessed by the hospital physician with the current blood glucose risk level, the invention obtains the current overall abnormality level of diabetic foot patients, accurately reflecting the current severity of their diabetic foot condition. Based on this overall abnormality level, an accurate high-risk assessment of diabetic foot is conducted, which helps to promptly and accurately determine the severity of diabetic foot, enabling the timely development of nursing plans and targeted care, effectively reducing the risk of diabetic foot deterioration. Attached Figure Description

[0015] Figure 1 The diagram below shows a structural block diagram of a multimodal data fusion-based high-risk assessment and care system for diabetic foot, provided in one embodiment of the present invention. Figure 2 This is a flowchart illustrating a method for obtaining the current blood glucose risk level according to an embodiment of the present invention. Figure 3 This is a schematic diagram of a computer device provided according to an embodiment of the present invention. Detailed Implementation

[0016] The following description, in conjunction with the accompanying drawings, details the specific scheme of a multimodal data fusion-based high-risk assessment and nursing system for diabetic foot provided by this invention.

[0017] Example 1: This invention proposes a multimodal data fusion-based system for assessing and caring for high-risk diabetic foot. Please refer to [link / reference]. Figure 1 The diagram illustrates a structural block diagram of a multimodal data fusion-based high-risk assessment and care system for diabetic foot provided by an embodiment of the present invention. The system includes: a data acquisition module 10, a current blood glucose risk level acquisition module 20, a current overall abnormality level acquisition module 30, and a data processing module 40.

[0018] The data acquisition module 10 is used to acquire daily blood glucose and plantar pressure data of diabetic foot patients.

[0019] Specifically, diabetes is known to damage the autonomic and peripheral nerves, causing decreased or absent sensation in the feet. This makes it difficult for patients to detect minor injuries, delaying treatment and allowing small wounds to develop into serious problems. Simultaneously, diabetes is often accompanied by damage to both large and small blood vessels, leading to poor blood circulation in the feet. Insufficient blood supply significantly reduces wound healing ability and greatly increases the risk of infection and gangrene. Currently, the Wagner classification system is widely used in medicine to quantify the severity of diabetic foot. The Wagner classification system primarily classifies the condition from mild to severe into 5 grades based on ulcer depth and the extent of foot tissue necrosis, providing doctors with a unified assessment framework. The Wagner classification system is well-known and will not be elaborated upon further.

[0020] Considering that diabetic foot is a continuously evolving process in reality, directly quantifying its severity using the Wagner classification system would overlook the dynamic changes in the disease course and the inadequacy of integrating multi-source information, leading to inaccurate analysis of diabetic foot severity and affecting timely and accurate treatment. To conduct a comprehensive and accurate risk assessment for diabetic foot patients, this embodiment has them wear blood glucose monitoring devices and plantar pressure monitoring devices to acquire blood glucose and plantar pressure data at every moment of the day. This embodiment sets the blood glucose and plantar pressure data to be collected simultaneously, with a 15-minute time interval between two adjacent data collection moments. The implementer can adjust the time interval between adjacent data collection moments according to actual conditions; it is not limited here. It should be noted that this embodiment uses a single diabetic foot patient as an example for better description; all subsequent references to diabetic foot patients will refer to this patient.

[0021] Furthermore, by examining the cases of diabetic foot patients, we obtained the severity level of diabetic foot assessed by physicians using the Wagner classification system at each of the patients' hospital visits. This provides a comprehensive and structured data foundation for subsequent analysis, ensuring that assessments are based on objective clinical indicators and historical records, enabling more accurate and dynamic risk assessments and the generation of personalized care plans.

[0022] The current blood glucose risk level acquisition module 20 is used to acquire the current blood glucose risk level of diabetic foot patients based on the fluctuation of daily blood glucose data of diabetic foot patients in the current time period, as well as the changes in abnormal blood glucose data of diabetic foot patients that exceed the preset normal blood glucose range every day.

[0023] Specifically, it is known that worsening blood sugar levels can cause multiple harms. On the one hand, it directly weakens the healing ability of foot wounds in diabetic foot patients; on the other hand, it may also lead to a series of other complications such as infection, nerve damage, and poor blood circulation. When blood sugar levels fluctuate drastically in diabetic foot patients, the blood supply and oxygen transport function of their body tissues may be impaired, further aggravating the condition of the foot. Therefore, maintaining stable blood sugar levels is crucial for diabetic foot patients. Furthermore, it is necessary to monitor the direction of abnormal blood sugar data that exceeds the normal range; when abnormal blood sugar data shows an upward trend, it indicates a more severe blood sugar abnormality. Based on the above description, when conducting risk assessments for diabetic foot patients, a comprehensive analysis of blood sugar performance, including recent and long-term blood sugar data fluctuations, is necessary to accurately assess blood sugar-related risks.

[0024] Therefore, this embodiment obtains the current blood glucose risk level of diabetic foot patients based on the daily fluctuations of their blood glucose data within the current time period, as well as the changes in abnormal blood glucose data exceeding the preset normal range each day. The higher the current blood glucose risk level, the more severe the diabetic foot symptoms manifested by the patient's current blood glucose. It should be noted that this embodiment sets the current time period to 5 days. The implementer can set the duration of the current time period according to the actual situation, which is not limited here, but the end day of the current time period must be the current day. The preset normal blood glucose range is set by the staff according to the actual situation and is not limited here. In addition, in this embodiment, the diabetic foot patients have had their blood glucose and plantar pressure data monitored for at least 7 days.

[0025] Preferably, in one possible implementation of this embodiment, the method for obtaining the current blood glucose risk level is described in [reference needed]. Figure 2 The flowchart illustrates a method for obtaining the current blood glucose risk level provided in this embodiment, which includes the following steps: Step S201: For any diabetic foot patient, obtain the current reference abnormality analysis value of the diabetic foot patient based on the fluctuation of the daily blood glucose data of the diabetic foot patient in the current time period.

[0026] When the fluctuation of a diabetic foot patient's daily blood glucose data is greater during the current time period, it indicates that the patient's blood glucose is more abnormal. Therefore, in this embodiment, for any diabetic foot patient, the current reference anomaly analysis value is obtained based on the fluctuation of the patient's daily blood glucose data during the current time period; the larger the current reference anomaly analysis value, the more abnormal the patient's blood glucose is.

[0027] In one possible implementation of this embodiment, the current reference anomaly analysis value is obtained as follows: For any day within the current time period, the difference between the maximum and minimum blood glucose data of the diabetic foot patient on that day is taken as the first difference; the larger the first difference, the greater the blood glucose fluctuation of the diabetic foot patient on that day; to accurately characterize the blood glucose fluctuation of the diabetic foot patient on that day, the product of the standard deviation of the blood glucose data of the diabetic foot patient on that day and the first difference is further subjected to max-min normalization, and the result is taken as the first blood glucose abnormality degree of the diabetic foot patient on that day; here, the dataset corresponding to the max-min normalization method is the product of the standard deviation of the blood glucose data of a large-scale patient sample and the first difference every day; the larger the first blood glucose abnormality degree, the more abnormal the blood glucose of the diabetic foot patient on that day. The max-min normalization method is a well-known method and will not be elaborated further. To accurately characterize the current blood glucose abnormality of the diabetic foot patient, the mean of the first blood glucose abnormality degree of the diabetic foot patient every day within the current time period is taken as the current reference anomaly analysis value of the diabetic foot patient.

[0028] Step S202: Based on the daily changes in abnormal blood glucose data that exceed the preset normal blood glucose range for the diabetic foot patient, obtain the abnormal blood glucose change analysis value for the diabetic foot patient.

[0029] To more accurately analyze the current abnormal blood glucose status of the diabetic foot patient, this embodiment analyzes the changes in abnormal blood glucose data that exceed the preset normal blood glucose range each day. That is, it analyzes the trend of abnormal blood glucose data of the diabetic foot patient from a long-term perspective and obtains the abnormal blood glucose change analysis value of the diabetic foot patient. The larger the abnormal blood glucose change analysis value, the more the degree of abnormal blood glucose of the diabetic foot patient increases with time, which indirectly indicates that the current blood glucose of the diabetic foot patient is more abnormal.

[0030] In one possible implementation of this embodiment, the method for obtaining the abnormal blood glucose change analysis value is as follows: based on the daily blood glucose data of the diabetic foot patient exceeding the preset normal blood glucose range, the second degree of blood glucose abnormality for each day is obtained. The method for obtaining the second degree of blood glucose abnormality is as follows: for any given day of the diabetic foot patient, the blood glucose data for that day is mapped onto a two-dimensional coordinate system according to time sequence and fitted as a curve, which serves as the second curve. The horizontal axis of the two-dimensional coordinate system represents time, and the vertical axis represents blood glucose data. The area within the upper and lower baselines of the second curve, where the maximum and minimum blood glucose data are not within the preset normal blood glucose range, is obtained using an integration method. This area is taken as the second degree of blood glucose abnormality for that day of the diabetic foot patient. The fitting curve and the integration method are well-known techniques and will not be elaborated further. The greater the second degree of blood glucose abnormality, the greater the degree of abnormality corresponding to the abnormal blood glucose data for that day of the diabetic foot patient. Considering that the progression of diabetic foot disease is usually phased, meaning that abnormal blood glucose data may not show a continuous upward or downward trend, but rather exhibit phased fluctuations and changes, to analyze the changes in abnormal blood glucose data in this diabetic foot patient, a second abnormal blood glucose level is fitted as a curve according to time sequence, serving as the first curve. Then, the maximum points on the first curve are fitted as a straight line according to time sequence. A larger slope of the straight line indicates a greater abnormal blood glucose change in the diabetic foot patient. The slope of the straight line is then normalized to its minimum and maximum values, serving as the analytical value for the abnormal blood glucose change in this diabetic foot patient. The dataset corresponding to this minimum and maximum normalization method is the slope value fitted to all patients across all assessment time periods in a large-scale diabetic foot patient sample database. It should be noted that when the number of maximum points extracted from the first curve within the current time period is less than two, the first and last nodes of the corresponding time window of the first curve are directly selected and connected to obtain the slope. The methods for obtaining the fitted straight line and the slope are well-known techniques and will not be elaborated further.

[0031] Step S203: The product of the current reference abnormality analysis value and the blood glucose abnormality change analysis value is used as the current blood glucose risk level of the diabetic foot patient.

[0032] It is known that the larger the current reference anomaly analysis value and the larger the blood glucose abnormality analysis value, the more abnormal the blood glucose of the diabetic foot patient is. Therefore, in this embodiment, the product of the current reference anomaly analysis value and the blood glucose abnormality analysis value is used as the current blood glucose risk level of the diabetic foot patient.

[0033] This allows us to obtain the current glycemic risk level for each diabetic foot patient.

[0034] The current overall abnormality acquisition module 30 is used to acquire the current overall abnormality of diabetic foot patients based on the synchronization of abnormal blood glucose data and plantar pressure data change trends, the severity level assessed by the current hospital examiner, and the current blood glucose risk level.

[0035] Specifically, it is known that long-term poor blood sugar control can lead to peripheral neuropathy in diabetic foot patients, weakening or even damaging the sensory nerves in the foot. This makes diabetic foot patients less sensitive to minor wounds, abrasions, or pressure on the foot. Even minor injuries may go unnoticed, missing the opportunity for early treatment and ultimately leading to further deterioration and ulcer formation. Furthermore, considering that daily activities such as walking and standing exert continuous mechanical pressure on the foot, the greater the pressure, the more it gradually damages foot tissue, eventually leading to tissue damage and ulcer formation. Therefore, when assessing the risk of diabetic foot, it is necessary to fully consider both the effects of the underlying disease and external stress, avoiding misjudgments caused by a single-dimensional analysis.

[0036] As shown in step S2, the maximum point on the first curve represents a key change in blood glucose control, essentially a critical period of worsening or increased blood glucose fluctuations. The more similar the trend of pressure changes on the feet of diabetic foot patients is to the trend of the first curve—that is, the closer the peak of the plantar pressure data is to the day corresponding to the maximum point on the first curve in time—the more likely the internal disease (abnormal blood glucose) and external stress (plantar pressure) will create a double risk superposition, causing synergistic damage to foot health. Specifically, high plantar pressure directly aggravates foot tissue damage, while concurrent abnormal blood glucose weakens the repair and healing capacity of foot tissue. Under the combined effect of these two factors, the probability and rate of deterioration of foot ulcers will significantly increase, ultimately exacerbating the overall risk of diabetic foot. Therefore, this embodiment first analyzes the overall impact on diabetic foot patients based on the synchronization of abnormal blood glucose data and plantar pressure data trends.

[0037] Essentially, the root cause of the risk in diabetic foot patients lies in long-term abnormal blood glucose levels. Poor long-term blood glucose control can trigger a series of chain symptoms such as neuropathy and vascular damage, which together form the basis for the occurrence and deterioration of diabetic foot. Therefore, this embodiment, on the one hand, breaks down the manifestations of abnormal blood glucose to clarify the risks inherent in abnormal blood glucose itself; on the other hand, it performs a linkage analysis between abnormal blood glucose and plantar pressure (external stress) to explore the synergistic risks of internal disease and external factors, jointly covering different dimensions of the risk of diabetic foot and providing a foundation for comprehensive assessment.

[0038] This embodiment aims to accurately assess the severity of diabetic foot in real time. When a diabetic foot patient exhibits more severe abnormal blood glucose levels, it indicates a stronger destructive effect of the underlying disease on foot health. In this case, the assessment should focus more on the synergistic influence of internal and external factors, i.e., the overall impact on the diabetic foot patient, to avoid underestimating the risk due to neglecting synergistic effects. Conversely, when a diabetic foot patient exhibits relatively mild abnormal blood glucose levels, it indicates a weaker risk driven by the patient's core underlying disease. In this case, the assessment should shift to the patient's latest Wagner classification of diabetic foot, as the Wagner classification directly reflects the actual lesion status of the foot (e.g., ulcer depth, gangrene extent), providing a direct representation of immediate risk. Focusing on this indicator allows for a more precise identification of the key intervention points, avoiding the neglect of existing substantial foot damage due to excessive focus on blood glucose. Therefore, this embodiment obtains the current overall abnormality level of the diabetic foot patient based on the overall impact on the patient, the severity level assessed by the attending physician, and the current blood glucose risk level. The greater the current overall abnormality level, the more severe the diabetic foot condition.

[0039] Preferably, in one feasible way of this embodiment, the method for obtaining the current overall abnormality level is as follows: for any diabetic foot patient, the current overall impact level of the diabetic foot patient is obtained based on the synchronization of the abnormal blood glucose data and plantar pressure data change trends of the diabetic foot patient; the greater the current overall impact level, the greater the impact of the internal disease (abnormal blood glucose) and external stress (plantar pressure) on the diabetic foot patient, which indirectly reflects the greater risk of the current diabetic foot. The method for obtaining the overall impact level is as follows: The average daily plantar pressure data of the diabetic foot patient is fitted into a curve according to time sequence, serving as the pressure curve; all maximum points on the pressure curve are taken as the first maximum point; it should be noted that if no first maximum point is extracted due to the pressure curve being monotonous or without fluctuation, the point corresponding to the end of the current time period is selected as the first maximum point; all maximum points on the first curve are taken as the second maximum points; for any first maximum point, the number of days between the first maximum point and the corresponding days of each second maximum point is obtained, serving as the interval reference duration; the minimum interval reference duration is taken as the consistency analysis value of the first maximum point; it should be noted that if no second maximum point is extracted within the current time period, the interval reference duration is assigned a preset maximum matching penalty parameter (such as the fixed total number of days in the current time period, which is not limited here) to ensure the closed-loop operation of subsequent calculation processes. The smaller the consensus analysis value, the greater the impact of the intrinsic disease (abnormal blood glucose) and external stress (plantar pressure) on the diabetic foot patient within the day corresponding to the first maximum value. To comprehensively characterize the combined impact of the intrinsic disease and external stress currently experienced by the diabetic foot patient, the mean of all consensus analysis values ​​at the first maximum value is negatively correlated and normalized, which is then used as the overall degree of impact on the diabetic foot patient. In this embodiment, the negative of the mean of the consensus analysis values ​​is used as the power of an exponential function with the natural constant as the base. The output of this exponential function is the result of negatively correlated and normalized mean values. To determine the current overall abnormality of the diabetic foot patient and accurately analyze the severity of the diabetic foot condition, the product of the current glycemic risk level and the current overall impact level is used as the first characteristic value; the product of the negative correlation result of the current glycemic risk level and the normalized result of the severity level assessed by the current hospital physician is used as the second characteristic value. In this embodiment, the difference between a constant 1 and the current glycemic risk level is used as the negative correlation result of the current glycemic risk level; the ratio of the severity level to a preset maximum severity level is used as the normalized result of the severity level; the preset maximum severity level is set by professionals based on actual conditions and is not limited here. Then, the sum of the first and second characteristic values ​​is used as the current overall abnormality level of the diabetic foot patient.

[0040] This allows us to obtain the current overall degree of abnormality in patients with diabetic foot.

[0041] Data processing module 40 is used for high-risk assessment and care of diabetic foot based on the current overall degree of abnormality.

[0042] Specifically, it is known that the greater the overall abnormality level, the more severe the diabetic foot in the corresponding diabetic foot patient. Therefore, this embodiment classifies the high-risk level of diabetic foot in diabetic foot patients according to the current overall abnormality level. Specifically: when the current overall abnormality level is less than a preset first abnormality level threshold, the high-risk level of diabetic foot in the corresponding diabetic foot patient is classified as low risk; when the current overall abnormality level is greater than or equal to the preset first abnormality level threshold but less than a preset second abnormality level threshold, the high-risk level of diabetic foot in the corresponding diabetic foot patient is classified as medium risk; when the current overall abnormality level is greater than or equal to the preset second abnormality level threshold, the high-risk level of diabetic foot in the corresponding diabetic foot patient is classified as high risk. In this embodiment, the first abnormality level threshold is set to 0.3 and the second abnormality level threshold is set to 0.7. Implementers can set the size of the first and second abnormality level thresholds according to the actual situation, which is not limited here.

[0043] To accurately and efficiently predict the high-risk assessment of diabetic foot in patients, this embodiment uses a preset number of diabetic foot patients with known high-risk levels as the training and validation sets for the neural network. It should be noted that this embodiment sets the preset number to 30,000, with a training-to-validation ratio of 7:3. Implementers can adjust the preset number and the ratio of the training and validation sets according to actual circumstances; no limitation is imposed here. The neural network's inputs include basic information about the diabetic foot patient (such as age and gender), clinical symptom information (such as numbness, pain, ulceration, and signs of infection in the foot, directly reflecting the current health status of the foot), laboratory test results (key biochemical indicators related to diabetic foot, such as glycated hemoglobin, blood lipids, and renal function indicators, used to quantify the patient's basic physiological and pathological state), treatment and hospitalization information (such as the patient's previous diabetes treatment plan, diabetic foot intervention measures, length of hospital stay, and history of past complications, reflecting the patient's disease management process and historical risk), the severity level assessed by the physician, and daily blood glucose and plantar pressure data; the output is the high-risk assessment level of the diabetic foot patient; and the loss function is the cross-entropy function. It should be noted that before inputting the aforementioned multi-dimensional source data into the neural network, it is necessary to perform dimensionality reduction and tensor merging in a unified format: First, the clinical symptom information text and treatment history text are reduced to a fixed-width standard text vector through a predefined word embedding mapping; second, the preset discrete level labels (such as the severity level of doctor assessment and basic gender information) are converted into categorical scalars through one-hot fully connected encoding; based on the age in the basic information and the continuous numerical indicators in the laboratory test results, normalization is performed to generate numerical feature channels; the mean and variance within a fixed-length sliding window are extracted from the high-frequency time-series blood glucose data and plantar pressure data covering each day to construct representative time-series channels; finally, the generated text vectors, categorical scalars, numerical feature channels, and time-series channels are flattened and combined in a fixed channel order, thereby merging to establish a one-dimensional basic bottom-level input tensor with a unified dimensional size, thus ensuring that heterogeneous non-uniform dimensional modalities are accurately received by the underlying perceptual interaction structure; in addition, the neural network can extract the current overall abnormality level of diabetic foot patients through the input data. This embodiment uses gradient descent to train the neural network until the loss function converges. Simultaneously, a validation set is used to verify the robustness of the trained neural network, ensuring its reliability. Thus, the trained neural network is obtained. Both the neural network and the gradient descent method are well-known and will not be described further.

[0044] Finally, the trained neural network is used to assess the high risk of diabetic foot in patients, outputting a high-risk label for diabetic foot. This helps to determine the severity of diabetic foot in patients in a timely and accurate manner. Then, by referring to the nursing care plans of patients with the same high-risk label in the hospital's historical medical records, a reference nursing plan is provided for diabetic foot patients. This helps to accurately and efficiently develop nursing plans for diabetic foot patients, thereby providing timely and targeted care and effectively reducing the risk of diabetic foot deterioration.

[0045] In summary, this embodiment acquires blood glucose and plantar pressure data of diabetic foot patients; based on the fluctuation of blood glucose data within the current time period and the daily changes in abnormal blood glucose data exceeding the preset normal range, it obtains the current blood glucose risk level of diabetic foot patients; based on the synchronization of abnormal blood glucose and plantar pressure data changes, the severity level assessed by the current hospital physician, and the current blood glucose risk level, it obtains the current overall abnormality level of diabetic foot patients, enabling high-risk assessment and nursing care for diabetic foot. This invention, by acquiring the current overall abnormality level, facilitates timely and accurate assessment of diabetic foot risk in diabetic foot patients, thereby enabling timely and accurate development of clinical nursing measures and effectively preventing the aggravation of diabetic foot.

[0046] Example 2: The present invention also proposes a computer device, please refer to [link to relevant documentation]. Figure 3 The computer device includes a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402. When the processor 402 executes the computer program 403, the computer device enables the computer device to implement any of the aforementioned multimodal data fusion-based high-risk assessment and care systems for diabetic foot.

Claims

1. A multimodal data fusion-based high-risk assessment and care system for diabetic foot, characterized in that, The system includes: The data acquisition module is used to acquire daily blood glucose and plantar pressure data for diabetic foot patients; The current blood glucose risk level acquisition module is used to obtain the current blood glucose risk level of diabetic foot patients based on the fluctuation of daily blood glucose data of diabetic foot patients in the current time period, as well as the changes of abnormal blood glucose data of diabetic foot patients that exceed the preset normal blood glucose range each day. The current overall abnormality acquisition module is used to acquire the current overall abnormality of diabetic foot patients based on the synchronization of abnormal blood glucose data and plantar pressure data change trends, the severity level assessed by the current hospital doctor, and the current blood glucose risk level. The data processing module is used to assess and manage high-risk diabetic foot based on the current overall level of abnormality.

2. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 1, characterized in that, The method for obtaining the current blood glucose risk level is as follows: For any diabetic foot patient, obtain the current reference abnormality analysis value of the diabetic foot patient based on the fluctuation of the patient's daily blood glucose data in the current time period; Based on the daily changes in abnormal blood glucose data that exceed the preset normal blood glucose range of the diabetic foot patient, obtain the abnormal blood glucose change analysis value of the diabetic foot patient. The product of the current reference anomaly analysis value and the blood glucose anomaly change analysis value is taken as the current blood glucose risk level of the diabetic foot patient.

3. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 2, characterized in that, The method for obtaining the current reference anomaly analysis value is as follows: For any day in the current time period, the difference between the maximum and minimum blood glucose data of the diabetic foot patient on that day is taken as the first difference; The result of normalizing the product of the standard deviation and the first difference of the blood glucose data of the diabetic foot patient on that day is taken as the first degree of blood glucose abnormality of the diabetic foot patient on that day. The average of the first abnormal blood glucose level for each day during the current time period for this diabetic foot patient is used as the current reference abnormality analysis value for this diabetic foot patient.

4. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 2, characterized in that, The method for obtaining the abnormal blood glucose change analysis values ​​is as follows: Based on the daily blood glucose data of the diabetic foot patient exceeding the preset normal blood glucose range, the second degree of blood glucose abnormality of the diabetic foot patient is obtained daily. The degree of the second blood glucose abnormality was fitted into a curve according to the time sequence, which was used as the first curve; The maximum points on the first curve are fitted into a straight line according to the time sequence, and the slope of the straight line is normalized to obtain the result as the blood glucose abnormality analysis value of the diabetic foot patient.

5. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 4, characterized in that, The method for obtaining the second degree of blood glucose abnormality is as follows: For any given day of the diabetic foot patient, the blood glucose data for that day is mapped onto a two-dimensional coordinate system in chronological order and fitted as a curve, which serves as the second curve; where the horizontal axis of the two-dimensional coordinate system represents time and the vertical axis represents blood glucose data. The area within the upper and lower baselines of the second curve, which is not within the preset normal blood glucose range, is obtained by integral method. This area is used as the second degree of blood glucose abnormality for the diabetic foot patient on that day.

6. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 4, characterized in that, The method for obtaining the current overall anomaly level is as follows: For any diabetic foot patient, the current overall impact of the diabetic foot disease can be obtained based on the synchronization of the abnormal blood glucose data and plantar pressure data trends of the diabetic foot patient. The first characteristic value is the product of the current blood glucose risk level and the current overall impact level. The product of the negative correlation between the current blood glucose risk level and the normalized result of the severity level assessed by the current hospital physician is used as the second characteristic value. The sum of the first and second eigenvalues ​​is taken as the current overall degree of abnormality for the diabetic foot patient.

7. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 6, characterized in that, The method for obtaining the current overall impact level is as follows: The mean daily plantar pressure data of this diabetic foot patient was fitted into a curve according to time sequence, which is used as the pressure curve; All the maximum points on the pressure curve are taken as the first maximum points; all the maximum points on the first curve are taken as the second maximum points. For any first maximum point, obtain the number of days between the first maximum point and the corresponding days of each second maximum point, and use them as the interval reference duration; The minimum interval reference duration is used as the consistent analysis value for the first maximum point; The mean of all first maximum points, after negative correlation and normalization, is used as the current overall impact on the diabetic foot patient.

8. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 1, characterized in that, The method for assessing and caring for high-risk diabetic foot based on the current overall degree of abnormality is as follows: The risk level of diabetic foot patients is classified into different levels based on the current overall degree of abnormality. A predetermined number of diabetic foot patients with known high-risk levels of diabetic foot were used as the training and validation sets for the neural network. The trained neural network is obtained by using the training set and the validation set; A trained neural network is used to assess the high risk of diabetic foot in patients, thereby enabling them to receive proper care.

9. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 8, characterized in that, The method for classifying the high-risk level of diabetic foot patients based on the current overall degree of abnormality is as follows: When the current overall abnormality level is less than the preset first abnormality level threshold, the high risk of diabetic foot in the corresponding diabetic foot patient will be classified as low risk. When the current overall abnormality level is greater than or equal to the preset first abnormality level threshold and less than the preset second abnormality level threshold, the high risk of diabetic foot in the corresponding diabetic foot patient will be classified as medium risk. When the current overall abnormality level is greater than or equal to the preset second abnormality level threshold, the high risk of diabetic foot in the corresponding diabetic foot patient is classified as high risk.

10. The multimodal data fusion-based high-risk assessment and care system for diabetic foot as described in claim 1, characterized in that, The severity level is set by the physician based on the Wagner classification system.

Images