A regional diabetic foot risk stratification early warning method and system

Abstract

This invention proposes a regional diabetic foot risk stratification and early warning method and system. The method includes: generating a diabetic foot risk score for the target patient at a unified regional scale; generating an early warning driving value for the target patient and mapping the early warning driving value to an intervention priority by combining it with a preset high-risk saturation value; receiving and processing the result data generated during the intervention task set process for the target patient, constructing a comprehensive health update index, and generating a resource allocation priority by combining the early warning status with the execution consistency index; generating a closed-loop management decision value based on the updated risk score and the execution consistency index; and determining the management intensity and management queue for the patient in the next cycle by comparing the closed-loop management decision value with a preset safety threshold. This invention is suitable for serving the needs of high-risk patients requiring priority intervention under conditions of multi-institutional collaborative management, multi-source data access, and limited medical resources within a region.

Description

Technical Field

[0001] This invention relates to the field of regional diabetic foot, and particularly to a regional diabetic foot risk stratification and early warning method and system. Background Technology

[0002] Diabetic foot is a common and serious complication of diabetes. Its development is usually associated with a combination of factors, including poor long-term blood sugar control, decreased nerve sensation in the foot, abnormal blood supply to the lower limbs, and inadequate daily foot care. If patients fail to recognize abnormalities such as redness, swelling, ulceration, and infection in the foot in the early stages, the condition often progresses rapidly, potentially leading to deep infection, tissue necrosis, and even amputation in severe cases.

[0003] Current management of diabetic foot in clinical and regional chronic disease management practices mainly relies on outpatient screening, regular follow-ups, and manual experience-based judgment. However, significant shortcomings remain in practical application. Firstly, different medical institutions often assess patient risk based on only a portion of data from blood glucose records, foot screening results, or follow-up information, lacking a unified quantitative integration mechanism. This leads to inconsistent risk assessments for the same patient across different institutions and time points, making it difficult to generate comparable and rankable risk results within the region. Secondly, existing management methods largely remain in a passive "problem-finding and then addressing" phase. There is a lack of continuity between risk assessment results and subsequent early warning triggers, intervention task assignments, patient follow-up tracking, and resource allocation adjustments. This results in a disconnect between front-end assessment and back-end execution. Even if a patient is identified as high-risk, intervention may be delayed due to untimely task implementation, inadequate follow-up examinations, or unreasonable allocation of regional outpatient and follow-up resources. Furthermore, in regional management scenarios, patients are scattered, data sources are diverse, and execution cycles are intertwined. If only static stratification or one-time judgment methods are used, it is difficult to reflect the real changes of patients after intervention, and it is also impossible to continuously adjust the management intensity and resource priority according to the patients' performance and changes in risk.

[0004] Therefore, existing technologies still lack a unified risk assessment, automatic early warning triggering, periodic status updates, and closed-loop execution guarantee for diabetic foot patients on a regional platform. It is difficult to simultaneously ensure the accuracy of risk identification, the timeliness of intervention push, and the continuity and targeting of regional resource allocation. Summary of the Invention

[0005] The purpose of this invention is to propose a regional diabetic foot risk stratification early warning method and system to solve the above-mentioned problems.

[0006] To achieve the above objectives, a regional diabetic foot risk stratification and early warning method is provided in a first aspect of the present invention, the method comprising the following steps: The blood glucose control indicators, foot examination indicators, and foot self-examination behavior indicators of the target patient are obtained, normalized, and then weighted according to preset weights to generate a diabetic foot risk score of the target patient at a unified regional scale. By comparing the diabetic foot risk score with a preset regional warning start threshold, a warning drive value for the target patient is generated, and the warning drive value is mapped to an intervention priority in combination with a preset high-risk saturation value; a corresponding warning state is generated according to the warning drive value, and a set of specific intervention tasks for the patient and doctor is generated according to the intervention priority; wherein, the preset high-risk saturation value is greater than the preset regional warning start threshold; Within a preset period, the system receives and processes the result data generated during the intervention task set for the target patient to construct a comprehensive health update index, and simultaneously calculates the patient's execution consistency index for the intervention task set; based on the diabetic foot risk score, the comprehensive health update index, and the execution consistency index, the system calculates the updated risk score of the target patient at the end of the current management period; and combines the warning status with the execution consistency index to generate resource allocation priorities. Based on the updated risk score and the consistency index of execution, a closed-loop risk index is constructed; the closed-loop risk index is weighted and merged with the resource allocation priority to generate a closed-loop management decision value; and based on the comparison between the closed-loop management decision value and the preset safety threshold, the management intensity and management queue for the patient in the next cycle are determined.

[0007] Preferably, the blood glucose control index is obtained by normalizing the average of the patient's most recent fasting blood glucose records. The foot examination indicators are derived from the cumulative results of abnormalities in foot nerve and vascular examinations, which are then normalized. The foot self-examination behavior index is derived from the number of times the patient effectively completed the self-examination task within a set number of days, after normalization.

[0008] Preferably, the preset weights are obtained by fitting historical case data of the region. The fitting sample consists of patient records of patients who have previously developed diabetic foot ulcers and patient records of patients who have not developed ulcers. The platform selects a set of parameters that can stably distinguish high-risk patients by comparing the deviation between the scoring results and the actual outcomes under different weight combinations, and the sum of the preset weights is 1.

[0009] Preferably, the step of generating a warning drive value for the target patient by comparing the diabetic foot risk score with a preset regional warning start threshold specifically involves: When the diabetic foot risk score is lower than the preset area warning start threshold, the warning drive value is equal to the diabetic foot risk score; when the diabetic foot risk score reaches or exceeds the preset area warning start threshold, the warning drive value is equal to the diabetic foot risk score plus an enhancement term proportional to the threshold exceedance.

[0010] Preferably, mapping the early warning drive value to an intervention priority specifically involves: When the warning drive value reaches or exceeds the preset high-risk saturation value, the intervention priority takes the maximum value, indicating that it enters the highest priority processing state; When the warning drive value is between 0 and a preset high-risk saturation value, the intervention priority increases linearly.

[0011] Preferably, generating the corresponding warning status based on the warning driving value specifically involves: When the warning drive value is greater than or equal to the preset area warning start threshold, the patient enters the warning state; when the warning drive value is less than the preset area warning start threshold, the patient remains in the routine follow-up state. The intervention task set includes at least the patient's task of uploading foot images and completing symptom questionnaires, as well as the doctor's task of generating high-risk markers and arranging foot-specific follow-up examinations.

[0012] Preferably, the result data includes foot images uploaded by the patient, symptom questionnaires filled out by the patient, and examination records entered by the doctor during follow-up or re-examination. The consistency index is obtained by accumulating the patient's completion time and completion status of the intervention task set according to preset task weights, and the value is between 0 and 1.

[0013] Preferably, the updated risk score is obtained by weighting and summing the complementary values ​​of the diabetic foot risk score, the comprehensive health update index, and the performance consistency index according to preset weights; The resource allocation priority is calculated only for patients who have entered the warning state, and is obtained by weighted summation of the complementary values ​​of the updated risk score and the execution consistency index.

[0014] Preferably, the closed-loop risk index is calculated based on the updated risk score and the feedback correction term based on the execution consistency indicator; the closed-loop management decision value is obtained by weighting the closed-loop risk index and the resource allocation priority according to preset weights. The feedback correction term is calculated based on the execution gap adjustment coefficient and the execution consistency index; the execution gap adjustment coefficient is used to control the impact of insufficient task execution on the risk index.

[0015] In a second aspect, the present invention provides a regional diabetic foot risk stratification early warning system, the system comprising: The data acquisition module is used to acquire the target patient's blood glucose control indicators, foot examination indicators, and foot self-examination behavior indicators, and after normalization, perform weighted calculations according to preset weights to generate the target patient's diabetic foot risk score at a unified regional scale. The early warning intervention module is used to generate an early warning drive value for the target patient by comparing the diabetic foot risk score with a preset regional early warning start threshold, and to map the early warning drive value to an intervention priority by combining it with a preset high-risk saturation value; and to generate a corresponding early warning state based on the early warning drive value, and to generate a set of specific intervention tasks for the patient and doctor based on the intervention priority; wherein, the preset high-risk saturation value is greater than the preset regional early warning start threshold; The periodic update module is used to receive and process the result data generated during the intervention task set of the target patient within a preset period, so as to construct a comprehensive health update index and at the same time count the patient's execution consistency index of the intervention task set; based on the diabetic foot risk score, the comprehensive health update index and the execution consistency index, calculate the update risk score of the target patient at the end of the current management period; and combine the warning status and the execution consistency index to generate resource allocation priority. The closed-loop management module is used to construct a closed-loop risk index based on the updated risk score and the consistency index of execution; to weight and merge the closed-loop risk index with the resource allocation priority to generate a closed-loop management decision value; and to determine the management intensity and management queue for the patient in the next cycle based on the comparison between the closed-loop management decision value and the preset safety threshold.

[0016] The beneficial technical effects of the present invention are at least as follows: This invention, centered on the core objective of regional diabetic foot risk stratification and early warning, establishes a continuous management scheme that integrates risk formation, risk triggering, risk updating, and implementation assurance. The scheme first incorporates blood glucose control, foot examination results, and foot self-examination behavior into a unified risk calculation framework, generating patient risk scores that can be directly compared across the region. Building upon this, it further introduces early warning driving values ​​and intervention priorities, enabling continuous risk values ​​to be transformed into specific early warning states and intervention tasks, thus connecting traditionally fragmented risk assessment and management actions. Subsequently, the system continuously receives new health data and task execution records generated by patients during the execution of triggered intervention tasks, using a periodic update mechanism to continuously correct patient risks. The updated risk status and execution gaps are jointly incorporated into resource allocation priority calculations, ensuring that medical resources are no longer statically allocated based on fixed stratification results, but rather dynamically adjusted according to changes in patient risk and execution status. Furthermore, in the closed-loop stage, the present invention integrates update risk, resource priority, and task execution consistency into a closed-loop risk index and a closed-loop management decision value, enabling the system to automatically determine whether a patient should be transferred to routine follow-up, re-enter the intervention process, or be upgraded to an enhanced management target, thereby forming a complete closed loop from initial assessment to task assignment, from execution tracking to resource rearrangement, and then to the next round of management decision-making.

[0017] Through the above technical solutions, diabetic foot management is no longer an isolated process of examination, reminders, and referrals, but a dynamic, tiered early warning process that runs continuously on a regional platform. This makes it more suitable for serving the needs of multi-institutional collaborative management, multi-source data access, and priority intervention for high-risk patients with limited medical resources within the region. Attached Figure Description

[0018] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0019] Figure 1 This is a flowchart of a regional diabetic foot risk stratification and early warning method according to the present invention. Detailed Implementation

[0020] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0021] like Figure 1As shown in the figure, an embodiment of the present invention provides a regional diabetic foot risk stratification and early warning method, the method comprising: S1 acquires the target patient's blood glucose control indicators, foot examination indicators, and foot self-examination behavior indicators, and after normalization, performs weighted calculations according to preset weights to generate the target patient's diabetic foot risk score at a unified regional scale.

[0022] Specifically, in a regional diabetic foot management scenario, the system first needs to generate a comparable and ranked unified risk score for each diabetic patient. This score is not a direct judgment of a single clinical outcome, but rather compresses three types of information most directly related to the formation of diabetic foot into a single calculation framework: one is blood glucose control, which reflects the degree of ongoing metabolic damage; another is foot examination, which reflects the current state of nerve and vascular abnormalities in the foot; and the third is the patient's self-examination behavior, which reflects their daily foot management level. The core challenge of regional management lies in the inconsistent data formats reported by different hospitals, clinics, and patients. Directly comparing raw values ​​would lead to inconsistent risk assessments for the same patient across different institutions. Therefore, the platform needs to extract data from each system and convert it to a unified scale before calculating the patient's risk score.

[0023] Furthermore, the platform extracts patient data from three data sources. Blood glucose control indicators The data comes from the hospital's testing system. The platform reads the patient's three most recent fasting blood glucose test records using their unique patient identification. When a patient visits different hospitals within the region, the system filters the three most recent valid records in reverse chronological order of test time. After removing duplicate uploads or records with missing sampling times, the average of the three results is calculated as the raw blood glucose control value. Foot examination indicators. The platform reads information from outpatient screening records or inpatient admission assessment records, which doctors fill out in their electronic medical records. Results of tactile examination using nylon filaments and palpation of the dorsalis pedis artery; if the nylon filament examination shows diminished or absent sensation, it is recorded as an abnormality; if the dorsalis pedis artery is weakened or not palpable, it is also recorded as an abnormality. The system accumulates these abnormalities to form the raw foot examination values. Foot self-examination behavioral indicators. From the patient follow-up mini-program, the platform reads the patient's recent... The self-inspection task record for feet submitted within the day includes uploading photos of the soles of the feet, indicating whether there is redness, swelling, or damage, and confirming whether shoes and socks have been changed as required. A single submission that simultaneously meets the requirements of taking photos and completing the form is counted as a valid self-inspection. The system calculates the total number of valid self-inspections to form the raw behavior value. Before entering the unified calculation, the above three indicators must be normalized according to the regional parameter table to map them to the same scale, thus ensuring that subsequent summation calculations are performed in the same evaluation space.

[0024] Furthermore, the risk scoring employs a linear comprehensive evaluation model from applied statistics. The fundamental idea behind this model is that when multiple factors contribute to the same risk outcome, and all factors have been mapped to the same scale, the overall risk can be expressed as a linear combination of the contributions of each factor. In the context of diabetic foot, glycemic control indicators... Foot examination indicators These are all positive risk factors; the higher the value, the higher the risk. Foot self-examination behavioral indicators It is a protective factor; the higher the value, the stronger the patient's ability to detect abnormalities. Therefore, it should be converted into a risk contribution item before being included in the calculation. Based on this derivation, the formula for calculating patient risk scores on the platform is as follows: ; in, Indicates the patient Diabetic foot risk score; Indicates the patient The normalized glycemic control index is obtained by averaging the three most recent fasting blood glucose records and then mapping them according to the regional parameter table. Indicates the patient The normalized foot examination indicators were obtained by mapping the abnormal items from nylon filament examination and dorsalis pedis artery palpation according to the regional parameter table. Indicates the patient The normalized foot self-examination behavior indicators, from recent The number of valid self-inspections per day is obtained by mapping according to the regional parameter table; , , The coefficients represent the weighting coefficients of the three indicators. These coefficients are obtained by fitting historical case data from the region. The fitted sample consists of records of patients who have previously developed diabetic foot ulcers and records of patients who have not developed ulcers. The platform selects a set of parameters that can stably distinguish high-risk patients by comparing the deviations between the scoring results and the actual outcomes under different weighting combinations, and meets the following requirements. Both sides of the formula are normalized proportional values, and the calculation structure is consistent. Taking a regional implementation scenario as an example, if a patient's three most recent fasting blood glucose records are... , , After averaging and mapping, we obtain The foot screening revealed abnormalities in the nylon filament examination and normal dorsalis pedis artery, which were obtained after mapping. Its proximity Completed in a day The first valid self-check, after mapping, yielded... If the weighting coefficients obtained from region fitting are , , Then we can obtain This result indicates that the patients were already at a high risk level under a uniform scale and could directly proceed to the subsequent early warning and intervention process.

[0025] Through the above process, the platform generates a unified risk score for every patient in the region. .

[0026] S2. By comparing the diabetic foot risk score with a preset regional warning start threshold, a warning drive value for the target patient is generated, and the warning drive value is mapped to an intervention priority in combination with a preset high-risk saturation value; and a corresponding warning state is generated according to the warning drive value, and a set of specific intervention tasks for the patient and doctor is generated according to the intervention priority; wherein, the preset high-risk saturation value is greater than the preset regional warning start threshold.

[0027] Specifically, S2 is based on the diabetic foot risk score. This continuous risk value is then converted into early warning status and intervention tasks within the regional management system. S1 has already compressed blood glucose control, foot screening, and foot self-examination behavior into a risk score on a unified scale. S2's task is to further transform this risk score into actionable early warning actions. A key characteristic of the diabetic foot scenario is that once the risk crosses a certain threshold, foot damage, infection, and ulcer formation accelerate significantly. Therefore, the system not only needs to identify who has entered the high-risk zone but also who has crossed the threshold and requires a faster response. Based on this scenario characteristic, the platform first... Convert to early warning driving value , and then Further mapping to intervention priorities Finally, an early warning status is generated. and intervention task set .

[0028] Furthermore, the construction of the early warning driving value originates from the classic idea of ​​piecewise linear penalty in statistical learning. Its original expression is typically used to highlight the additional cost of a variable exceeding a boundary. This step introduces this idea into the diabetic foot scenario, transforming the output of step one... As a basic risk, the portion exceeding the regional early warning threshold is considered as a risk exceeding the threshold requiring accelerated response, and then a new early warning driving value is formed through a linear enhancement method. The corresponding calculation formula is: ; in, Indicates the patient Early warning driving value; This represents the diabetic foot risk score output from step one; This indicates the regional early warning threshold, and its value is determined through the statistics of historical cases in the region. During the statistics, the risk scores of previous patients are analyzed in correspondence with the subsequent records of foot injuries, ulcers or infections, and a set of cutoff values ​​that can identify patients with rising risk earlier are selected. This represents the over-threshold amplification coefficient, used to control the amplification intensity after the risk score exceeds the threshold. The derivation logic of this formula is: when At that time, the patient was still in the general observation area, and the system remained... ;when When a patient enters the critical management zone for diabetic foot, the system adds an enhancement factor proportional to the extent of exceeding the threshold, further widening the numerical gap between high-risk patients. In a regional implementation scenario, if the regional platform sets [the risk level] based on historical data... , Then, when patient A obtains [the following information] in step one... When, it can be calculated When patient B's When the enhancement term is zero, therefore .

[0029] Furthermore, after obtaining the early warning driving value Afterwards, the platform continued to use the classic linear interpolation concept to... Convert to normalized intervention priority The original purpose of linear interpolation is to smoothly map a continuous variable to a target interval when the upper and lower boundaries are known. This step transforms this idea into a mapping process from "risk-driven values ​​to intervention priorities," allowing the system to continuously adjust the task intensity based on the patient's risk level. The platform pre-sets a high-risk saturation value. ,in When the warning drive value Reaching or exceeding When the intervention priority reaches its maximum value, it indicates that the highest priority processing state has been entered; when the warning drive value... lie in arrive During this period, the intervention priority increases linearly, ensuring a smooth and continuous change in priority before the risk value reaches saturation. The corresponding calculation formula is: ; in, Indicates the patient Normalized intervention priority; This represents the warning drive value calculated using the previous formula; Indicates a high-risk saturation value, when Once this value is reached, the system assigns the patient to the highest priority queue. The first formula characterizes the "accelerated risk after exceeding the threshold," while the second formula smoothly maps the accelerated risk outcome to "task execution intensity." For example, if the platform sets... Then patient A's After substituting, we can get Patient B's Then we get If another patient, C, receives [treatment / treatment] in step one... Then we first obtain from the first equation. Substituting into the second equation, we get... This indicates that patient C has reached the highest priority treatment status.

[0030] The platform is driven by early warning values. and intervention priority The two outputs of the cost generation step are the warning status. and intervention task set .

[0031] in, According to the platform and The comparison results are automatically generated: when At that time, the patient enters an early warning state; when During this period, patients maintained routine follow-up. Intervention task set Then it will be determined by the platform. The interval it is in is automatically assembled.

[0032] Patient-side tasks include uploading a foot photo, completing an abnormal symptom questionnaire, and confirming care adherence within a specified timeframe. Doctor-side tasks include generating a high-risk marker in the chronic disease management workbench, scheduling a specific foot re-examination, and, if necessary, referring the patient to a higher-level hospital's foot clinic. The platform caters to different... Different push rhythms are set for different intervals: when When the level is high, the patient's app sends a notification via instant message and SMS, while the doctor's app displays it at the top of the chronic disease management list; when If the level is low but still within the warning range, the system will use the standard follow-up reminder method within weekdays. Finally, this step will output... Indicates whether the patient has entered the early warning process, and outputs... These two outputs represent the specific set of intervention tasks for each patient, and will serve as direct inputs for subsequent periodic updates and resource optimization.

[0033] S3. Within a preset period, receive and process the result data generated during the intervention task set of the target patient to construct a comprehensive health update index, and simultaneously calculate the patient's execution consistency index for the intervention task set; based on the diabetic foot risk score, the comprehensive health update index, and the execution consistency index, calculate the updated risk score of the target patient at the end of the current management period; and combine the warning status and the execution consistency index to generate resource allocation priorities.

[0034] Specifically, in the regional diabetic foot management system, two key tasks have been completed in the previous phase: First, a patient risk score was calculated based on indicators such as blood glucose control, foot screening results, and foot self-examination behavior. The scoring is then used to generate early warning driving values. Intervention priority Warning status and intervention task set This completes the process from risk identification to intervention triggering. After the intervention is pushed out, patients will continuously generate new health information while performing these tasks, such as new plantar images, questionnaire symptom information, and doctor's follow-up records. If this information is no longer absorbed and updated by the system, the risk assessment will remain in a historical state and will not reflect the real changes. Therefore, in this stage, the system focuses on the information generated in step two. and It receives new data generated by patients performing tasks within a fixed time period, and realizes dynamic risk correction and optimized allocation of medical resources through risk update algorithm and resource allocation algorithm, so that the regional management system forms a closed loop of continuous operation.

[0035] Furthermore, when the system enters the periodic update phase, the platform first filters all... Patients were selected as the key focus of this cycle, and data on their performance in the intervention task set was collected. The execution records are stored within the system. Each task record includes the task type, issuance time, latest completion time, actual completion time, and result data path. Result data primarily comes from three sources: foot images uploaded by patients, symptom questionnaires completed by patients, and examination records entered by doctors during follow-up or re-examinations. For foot image data, the platform uses an image analysis module for recognition. This module employs a convolutional neural network structure. Its input layer receives the pixel matrix of the foot image uploaded by the patient, and then performs feature extraction through a combination of four sets of convolutional layers and pooling layers. Each set of convolutional layers uses... Convolutional kernels extract local texture information, pooling layers reduce feature dimensionality, and finally, a fully connected layer outputs the probability value of foot abnormalities. Simultaneously, an image segmentation algorithm is used to calculate the proportion of the suspected damaged area to the entire foot area. The system combines the abnormality probability with the area proportion to obtain an image abnormality score. Symptom questionnaire data comes from structured forms filled out by patients in a mini-program. The system reads fields such as pain, numbness, redness, swelling, effusion, and abnormal temperature, and generates a questionnaire abnormality score based on the number of abnormal items. Doctor's follow-up records come from the electronic medical record system, such as nylon tactile test results, dorsalis pedis artery palpation results, and skin condition descriptions. The system generates a follow-up abnormality score based on the number of abnormal items. To integrate this new health information into a unified computational structure, the platform combines image abnormality scores, questionnaire abnormality scores, and follow-up abnormality scores... , , The comprehensive health update index is formed by proportional weighting. .

[0036] At the same time, the platform also bases its decisions on task sets. Calculate the execution consistency index based on the execution status. The platform configures separate tasks for photo assignments, questionnaires, and review tasks. , , The task weighting is as follows: if completed on time, the full weight is counted; if completed after the latest completion time but within the delay tolerance window, half the weight is counted; if not completed, no weight is counted. Directly by The execution records are accumulated and the value is located in arrive Between. In this way, Represents changes in health status. This indicates a change in the quality of execution.

[0037] Furthermore, after obtaining and Afterwards, the platform assesses the basic risk score generated in step one. Periodic updates are performed. The underlying idea here comes from exponential smoothing and state correction methods in time series analysis. The original logic is to obtain an updated state that is not overly sensitive to short-term fluctuations but can absorb the latest changes by using a weighted combination of historical states and new observations. In the diabetic foot scenario, the platform adds an execution gap term to this classic approach, that is, it adds... This is added as a specific additional penalty to the update formula because failure to upload plantar photos, delayed completion of questionnaires, or incomplete follow-up examinations inherently indicate a break in the risk identification chain. Based on this modification, the patient's updated risk score at the end of the current cycle is calculated as follows: ; in, Indicates the patient Update the risk score at the end of the current management cycle; This represents the basic risk score formed in step one; This represents the comprehensive updated indicators formed by the patient within this cycle based on the results of the task performed in step two; This indicates that the patient's response to the set of intervention tasks in step two during this cycle is... The consistency indicators for execution. All three items use the same proportional scale, therefore they can be directly weighted and summed. Taking a regional implementation scenario as an example, a patient receives the following in step one: Step two assigned the patient three tasks: uploading photos, completing a questionnaire, and a follow-up examination. Within this period, the patient uploaded foot photos and completed the questionnaire on time, but the follow-up examination was completed later than required. The platform, according to the task weighting rules, awarded the patient a penalty. The image analysis module detected a small area of ​​ulceration and obvious local erythema in the forefoot region. An image anomaly score was then calculated. The questionnaire reported worsening of mild pain and numbness; abnormal scores were recorded. The doctor's follow-up record indicated decreased sensation in the nylon filaments and localized skin dryness and cracking; the abnormal score on the follow-up examination was recorded. Then, according to the aforementioned proportions, we can obtain .Will , , Substituting into the above equation, we can obtain .

[0038] Furthermore, after obtaining an updated risk score Subsequently, the platform further generates resource allocation priorities for regional resource pools. This calculation draws on the concept of a priority queue index from operations research, allowing the updated risk to be... This indicates the current medical necessity that should be addressed. This indicates the execution gap that should be caught up with, and then the warning status output in step two is used. This acts as a resource allocation gate, ensuring that resource optimization is only performed within the set of patients already under early warning management. The corresponding formula is as follows: ; in, Indicates the patient Priority of resource allocation; This indicates the warning status output in step two. When a patient enters the resource optimization sorting queue, During this period, the patient will maintain routine management and will not occupy high-priority resources; This represents the updated risk score obtained from the previous expression; This indicates the consistency index. Additional terms in the formula. This further explicitly propagates the execution gap to the resource prioritization layer, ensuring that among two patients in the same warning state with similar medical risks, the one with slower execution is more likely to be scheduled for a follow-up phone call, outpatient check-up, or remote consultation by the platform. Taking the aforementioned patient as an example, because they had already entered the warning state in step two, therefore... Substitute and achievable If another patient also has Around [amount], but its tasks for this cycle were almost entirely unfinished, for example... ,but This will be significantly increased, thus prioritizing access to foot care clinic follow-up appointments or remote specialist consultations when regional resources are limited. Finally, this step outputs an updated risk score. and resource allocation priority The former depicts the patient's true risk status at the end of the current cycle, while the latter is directly used for sorting in the regional doctor's workbench, allocating outpatient follow-up appointment slots, and scheduling remote consultations.

[0039] S4. Based on the updated risk score and the consistency index of execution, construct a closed-loop risk index; combine the closed-loop risk index with the resource allocation priority in a weighted manner to generate a closed-loop management decision value; and determine the management intensity and management queue for the patient in the next cycle based on the comparison between the closed-loop management decision value and the preset safety threshold.

[0040] Specifically, after completing the periodic health data updates and resource allocation optimization of the previous phase, the system has obtained the patient's updated risk score at the end of the current cycle. and resource allocation priority .in It reflects the risk level following changes in the patient's current foot health status, while This reflects the priority of patients in the allocation of regional medical resources. Meanwhile, the intervention task set generated in step two... It has been entered into the task log database, serving as the direct basis for statistical analysis of the execution status at this stage. Therefore, this stage focuses on... , and The input is used to determine whether the intervention measures have been effectively implemented, whether the existing resource allocation is still reasonable, and whether continued intensive management is necessary. Diabetic foot management has a significant long-term and continuous characteristic. The patient's risk may decrease after a single intervention, or it may increase again due to insufficient implementation or the appearance of new injuries. Therefore, the system needs to continuously monitor the implementation of the intervention and feed the results back to the risk assessment system, thus forming a complete closed-loop management structure. This structure is conceptually derived from feedback control theory in control engineering. Its basic principle is that after the system outputs control behavior, by detecting changes in the system state and correcting the control variables, the system gradually stabilizes. In this scenario, the risk score... This can be viewed as the system state, while medical intervention is the control behavior. The system continuously adjusts the management strategy for the next round by detecting the implementation of the intervention.

[0041] The system first uses the set of intervention tasks generated in step two. The system reads the patient's intervention execution records within the current management cycle from the task log database and combines them with the resource allocation priority given in step three. Determine the order of verification, prioritizing high-risk areas. The patient undergoes a closed-loop status check. In the task log database, each record includes a task number, task type, issuance time, latest completion time, actual completion time, and execution result. Patient-side tasks mainly include uploading plantar images, completing symptom questionnaires, and confirming nursing instructions; doctor-side tasks include follow-up records, plantar examinations, and outpatient follow-ups. The system further analyzes the task set... Calculate the execution consistency index based on the execution status. Specifically, the platform assigns weights to photo tasks, questionnaire tasks, and review tasks respectively. , , If completed on time, the full amount will be credited; if completed after the latest completion time but within the delay tolerance window, half the amount will be credited; if not completed, no amount will be credited. It is formed by summing the results of each task according to their weights, and the value is located in arrive Between. For example, if a patient completes the photo and questionnaire tasks on time within a cycle, but delays in completing the follow-up task, then... At the same time, the system will also obtain the patient's latest health status from the latest follow-up data and image recognition results, and use the updated risk score obtained in step three. This represents the current risk level. To integrate risk status with implementation progress to create feedback indicators, a closed-loop risk index is constructed. The mathematical structure of this index originates from the proportional feedback model in control theory, and its basic form is "system state multiplied by feedback correction term". In classical proportional feedback control, the control signal can be written as... ,in For system status, This is for feedback gain. Based on this idea, this patent will implement a gap. By introducing a correction term, the risk index is directly increased for tasks that are not performed adequately, thus ensuring that the system prioritizes patients with insufficient performance. This leads to the closed-loop risk index calculation formula: ; in, Indicates the patient The closed-loop risk index; This represents the updated risk score obtained in step three; Indicates the consistency index of task execution; This represents the performance gap adjustment coefficient, used to control the impact of insufficient task execution on the risk index. The derivation logic of this formula is as follows: when all tasks are completed as planned, near ,but near The risk index remains at When a task is significantly delayed or has incomplete parts, decline, A positive risk score increases the risk index proportionally, thereby raising the patient's priority in the management cohort. For example, a patient receives an updated risk score in step three. If the photo assignment, questionnaire assignment, and review assignment are all completed on time, then [the student] can [receive / benefit]. And thus obtain If another patient only completes the questionnaire task on time, but the photo task is delayed and the follow-up examination is not completed, then they can receive a weighted score. and set ,but This calculation shows that even if two patients have the same risk score, the patient who underperforms will have a higher closed-loop risk index.

[0042] After obtaining the closed-loop risk index Then, the system compares it with the resource allocation priority in step three. Used in conjunction with management decisions. Specifically, Used to maintain the original ordering basis for regional resource allocation. This reflects the latest risk changes after the implementation feedback in this cycle; when both are high, patients are prioritized for inclusion in the intensive management list. To compress these two types of information into a unified closed-loop management decision-making quantity, the platform further calculates the closed-loop management decision value. The calculation approach originates from a multi-indicator comprehensive ranking model. Its original logic is to combine two evaluation metrics already within the same scale into a single comprehensive indicator that can be directly ranked and used for threshold judgment. In this scenario, the platform will execute the feedback-based closed-loop risk index. As the primary factor, the resource allocation priority generated in step three will be prioritized. As a factor for maintaining resource continuity, it constitutes the following decision formula: ; in, Indicates the patient Closed-loop management decision values; This represents the closed-loop risk index calculated using the previous formula. This indicates the resource allocation priority given in step three. The reason for using this... and The combination of these is because the primary goal of the closed-loop phase is to reflect the actual changes in risk after the current cycle ends. It should have a dominant weight; at the same time, it needs to maintain the continuity of the resource allocation results from the previous step, and avoid frequent fluctuations in regional outpatient scheduling and follow-up queues within a single cycle. Secondary weights are retained. Taking a specific implementation scenario as an example, if patient A obtains... , If the photo and questionnaire tasks are completed on time this period, but the review task is delayed, then according to the task weight, the following can be obtained: and take Then we can obtain Further obtained If patient B also has and However, if only the questionnaire task is completed on time, and the other tasks are not completed, then... ,at this time Further obtained This shows that, when resources are roughly equal, insufficient execution significantly increases the decision value of closed-loop management, allowing patients to receive priority for follow-up examinations and intensive interventions in the next round of management.

[0043] Finally, the platform based on Proceed to the next round of closed-loop management decisions. The system sets two decision thresholds: if... If the patient's condition falls below a safe threshold, they will be transferred to a regular follow-up cohort and will only undergo periodic health monitoring; if... If the location is in the middle range, the system will automatically generate the next round of intervention tasks, including increasing the frequency of foot examinations, arranging remote follow-ups, or sending nursing guidance; if If the risk level exceeds the high-risk threshold, the system automatically generates a priority follow-up prompt in the doctor's workspace and arranges a follow-up examination at a specialist outpatient clinic or a remote expert consultation. Finally, this step outputs a closed-loop risk index. and closed-loop management decision value .in, Used to describe the overall risk level resulting from the combined effect of risk status and performance. This is used to characterize the intensity of the next round of management generated by the system. In this way, the platform forms a complete management closed loop: Step 1 calculates the basic risk, Step 2 generates the early warning status and task set, Step 3 updates the risk and optimizes resources based on the task execution results, and Step 4 merges the execution feedback and resource continuity into a new management decision, thereby realizing continuous monitoring, execution assurance and cyclical intervention in the management of diabetic foot areas.

[0044] This invention also provides a regional diabetic foot risk stratification early warning system, the system comprising: The data acquisition module is used to acquire the target patient's blood glucose control indicators, foot examination indicators, and foot self-examination behavior indicators, and after normalization, perform weighted calculations according to preset weights to generate the target patient's diabetic foot risk score at a unified regional scale. The early warning intervention module is used to generate an early warning drive value for the target patient by comparing the diabetic foot risk score with a preset regional early warning start threshold, and to map the early warning drive value to an intervention priority by combining it with a preset high-risk saturation value; and to generate a corresponding early warning state based on the early warning drive value, and to generate a set of specific intervention tasks for the patient and doctor based on the intervention priority; wherein, the preset high-risk saturation value is greater than the preset regional early warning start threshold; The periodic update module is used to receive and process the result data generated during the intervention task set of the target patient within a preset period, so as to construct a comprehensive health update index and at the same time count the patient's execution consistency index of the intervention task set; based on the diabetic foot risk score, the comprehensive health update index and the execution consistency index, calculate the update risk score of the target patient at the end of the current management period; and combine the warning status and the execution consistency index to generate resource allocation priority. The closed-loop management module is used to construct a closed-loop risk index based on the updated risk score and the consistency index of execution; to weight and merge the closed-loop risk index with the resource allocation priority to generate a closed-loop management decision value; and to determine the management intensity and management queue for the patient in the next cycle based on the comparison between the closed-loop management decision value and the preset safety threshold.

[0045] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0046] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or units may be electrical, mechanical, or other forms.

[0047] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0048] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A regional diabetic foot risk stratification and early warning method, characterized in that, The method includes: The blood glucose control indicators, foot examination indicators, and foot self-examination behavior indicators of the target patient are obtained, normalized, and then weighted according to preset weights to generate a diabetic foot risk score of the target patient at a unified regional scale. By comparing the diabetic foot risk score with a preset regional warning start threshold, a warning drive value for the target patient is generated, and the warning drive value is mapped to an intervention priority in combination with a preset high-risk saturation value; a corresponding warning state is generated according to the warning drive value, and a set of specific intervention tasks for the patient and doctor is generated according to the intervention priority; wherein, the preset high-risk saturation value is greater than the preset regional warning start threshold; Within a preset period, the system receives and processes the result data generated during the intervention task set for the target patient to construct a comprehensive health update index, and simultaneously calculates the patient's execution consistency index for the intervention task set; based on the diabetic foot risk score, the comprehensive health update index, and the execution consistency index, the system calculates the updated risk score of the target patient at the end of the current management period; and combines the warning status with the execution consistency index to generate resource allocation priorities. Based on the updated risk score and the consistency index of execution, a closed-loop risk index is constructed; the closed-loop risk index is weighted and merged with the resource allocation priority to generate a closed-loop management decision value; and based on the comparison between the closed-loop management decision value and the preset safety threshold, the management intensity and management queue for the patient in the next cycle are determined.

2. The regional diabetic foot risk stratification and early warning method according to claim 1, characterized in that, The blood glucose control index is derived from the average of the patient's most recent fasting blood glucose records after normalization. The foot examination indicators are derived from the cumulative results of abnormalities in foot nerve and vascular examinations, which are then normalized. The foot self-examination behavior index is derived from the number of times the patient effectively completed the self-examination task within a set number of days, after normalization.

3. The regional diabetic foot risk stratification and early warning method according to claim 1, characterized in that, The preset weights are obtained by fitting historical case data of the region. The fitting sample consists of patient records of patients who have previously developed diabetic foot ulcers and patient records of patients who have not developed ulcers. The platform selects a set of parameters that can stably distinguish high-risk patients by comparing the deviation between the scoring results and the actual outcomes under different weight combinations, and the sum of the preset weights is 1.

4. The regional diabetic foot risk stratification and early warning method according to claim 1, characterized in that, The method of generating a warning drive value for the target patient by comparing the diabetic foot risk score with a preset regional warning initiation threshold is as follows: When the diabetic foot risk score is lower than the preset area warning start threshold, the warning drive value is equal to the diabetic foot risk score; When the diabetic foot risk score reaches or exceeds the preset warning threshold, the warning driving value is equal to the diabetic foot risk score plus an enhancement term proportional to the magnitude of the exceedance.

5. The regional diabetic foot risk stratification and early warning method according to claim 4, characterized in that, The mapping of the early warning driving value to intervention priority specifically involves: When the warning drive value reaches or exceeds the preset high-risk saturation value, the intervention priority takes the maximum value, indicating that it enters the highest priority processing state; When the warning drive value is between 0 and a preset high-risk saturation value, the intervention priority increases linearly.

6. The regional diabetic foot risk stratification and early warning method according to claim 1, characterized in that, The step of generating the corresponding warning status based on the warning driving value specifically involves: When the warning drive value is greater than or equal to the preset area warning start threshold, the patient enters the warning state; when the warning drive value is less than the preset area warning start threshold, the patient remains in the routine follow-up state. The intervention task set includes at least the patient's task of uploading foot images and completing symptom questionnaires, as well as the doctor's task of generating high-risk markers and arranging foot-specific follow-up examinations.

7. The regional diabetic foot risk stratification and early warning method according to claim 1, characterized in that, The results data include foot images uploaded by the patient, symptom questionnaires filled out by the patient, and examination records entered by the doctor during follow-up or re-examination. The consistency index is obtained by accumulating the patient's completion time and completion status of the intervention task set according to preset task weights, and the value is between 0 and 1.

8. The regional diabetic foot risk stratification and early warning method according to claim 6, characterized in that, The updated risk score is obtained by weighting and summing the complementary values ​​of the diabetic foot risk score, the comprehensive health update index, and the performance consistency index according to preset weights. The resource allocation priority is calculated only for patients who have entered the warning state, and is obtained by weighted summation of the complementary values ​​of the updated risk score and the execution consistency index.

9. The regional diabetic foot risk stratification and early warning method according to claim 1, characterized in that, The closed-loop risk index is calculated based on the updated risk score and the feedback correction term based on the execution consistency indicator; The closed-loop management decision value is obtained by weighting the closed-loop risk index and resource allocation priority according to preset weights. The feedback correction term is calculated based on the execution gap adjustment coefficient and the execution consistency index; the execution gap adjustment coefficient is used to control the impact of insufficient task execution on the risk index.

10. A regional diabetic foot risk stratification and early warning system, characterized in that, The system includes: The data acquisition module is used to acquire the target patient's blood glucose control indicators, foot examination indicators, and foot self-examination behavior indicators, and after normalization, perform weighted calculations according to preset weights to generate the target patient's diabetic foot risk score at a unified regional scale. The early warning intervention module is used to generate an early warning drive value for the target patient by comparing the diabetic foot risk score with a preset regional early warning start threshold, and to map the early warning drive value to an intervention priority by combining it with a preset high-risk saturation value; and to generate a corresponding early warning state based on the early warning drive value, and to generate a set of specific intervention tasks for the patient and doctor based on the intervention priority; wherein, the preset high-risk saturation value is greater than the preset regional early warning start threshold; The periodic update module is used to receive and process the result data generated during the intervention task set of the target patient within a preset period, so as to construct a comprehensive health update index and at the same time count the patient's execution consistency index of the intervention task set; based on the diabetic foot risk score, the comprehensive health update index and the execution consistency index, calculate the update risk score of the target patient at the end of the current management period; and combine the warning status and the execution consistency index to generate resource allocation priority. The closed-loop management module is used to construct a closed-loop risk index based on the updated risk score and the consistency index of execution; to weight and merge the closed-loop risk index with the resource allocation priority to generate a closed-loop management decision value; and to determine the management intensity and management queue for the patient in the next cycle based on the comparison between the closed-loop management decision value and the preset safety threshold.

Images