Non-load exercise rehabilitation exercise decision support system for diabetic foot ulcer patient

Abstract

The invention belongs to the field of medical rehabilitation, and particularly discloses a diabetic foot ulcer patient non-load exercise rehabilitation exercise decision support system which comprises an evaluation grading module used for obtaining multi-mode ulcer data of a patient and scientifically grading diabetic foot ulcers of the patient in combination with a layering algorithm to obtain a grading result; the scheme generation module is used for matching a non-load exercise scheme suitable for the patient in a database according to the grading result and generating an initial exercise scheme; the exercise feedback module is used for sending the initial exercise scheme to a patient terminal, acquiring the exercise completion condition of the patient and performing personalized adjustment on the exercise intensity and frequency through a dynamic regulation and control mechanism; according to the method, static ulcer features and dynamic physiological features are combined through a multi-modal data fusion technology, automatic grading is realized based on an AI algorithm, and scientificity and consistency of grading evaluation are remarkably improved.

Description

Technical Field

[0001] The invention belongs to the field of medical rehabilitation, and in particular relates to a non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers. Background Art

[0002] Diabetic foot is one of the common serious complications of diabetes, mainly caused by neuropathy and vascular disease caused by diabetes. The incidence of diabetic foot ulcers is high and is accompanied by a high disability and mortality rate. Its treatment and rehabilitation management are clinical problems that need to be solved urgently;

[0003] However, traditional ulcer assessment mostly relies on manual experience and judgment, and subjective grading is based on ulcer location, size and infection status. It is easily affected by individual differences and insufficient data, making it difficult to achieve scientific and comprehensive stratified management. Summary of the invention

[0004] The purpose of the present invention is to provide a non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers, so as to solve the problem that traditional ulcer assessment proposed in the above background technology is based on subjective grading of ulcer location, size and infection status, which makes it difficult to achieve scientific and comprehensive hierarchical management.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers, comprising:

[0007] The evaluation and grading module is used to obtain the multimodal ulcer data of the patient, and scientifically grade the patient's diabetic foot ulcer in combination with the stratification algorithm to obtain the grading results;

[0008] A program generation module, used for matching a suitable non-weight-bearing exercise program for the patient in a database according to the classification result, and generating an initial exercise program;

[0009] The exercise feedback module is used to send the initial exercise plan to the patient terminal, obtain the patient's exercise completion status, and make personalized adjustments to the exercise intensity and frequency through a dynamic control mechanism.

[0010] Preferably, the evaluation and grading module performs hierarchical evaluation based on Wagner grading or Texas grading, and automatically generates grading results in combination with an AI algorithm, and the grading results include:

[0011] Mild ulcer: superficial ulcer, no infection or slight infection;

[0012] Moderate ulcer: The ulcer is deep and may involve muscle or tendon without serious infection;

[0013] Severe ulcers: extension of infection to bone, with risk of ischemia or vascular occlusion.

[0014] Preferably, the exercise rehabilitation recommendation module includes:

[0015] A non-weight-bearing exercise database for storing exercise programs suitable for patients with diabetic foot ulcers;

[0016] A program generating unit is used to match exercise items from the non-weight-bearing exercise database and generate a personalized rehabilitation plan.

[0017] Preferably, the multimodal ulcer data includes static ulcer data and dynamic physiological data;

[0018] The static ulcer data include ulcer site, ulcer diameter, ulcer depth, and infection status;

[0019] The dynamic physiological data include blood circulation conditions, plantar temperature distribution, and plantar pressure distribution.

[0020] Preferably, the dynamic control mechanism includes:

[0021] S1. Obtain and analyze new multimodal ulcer data after the patient completes exercise to obtain exercise feedback;

[0022] S2. Analyze the movement feedback through a reinforcement learning algorithm and adjust the movement plan in real time.

[0023] Preferably, the movement feedback module is also used to construct a patient's personal rehabilitation behavior model based on the patient's historical movement data and rehabilitation results;

[0024] The rehabilitation behavior model is used to dynamically optimize subsequent exercise plans by continuously learning the patient's exercise habits, pain tolerance, and rehabilitation performance.

[0025] Preferably, the system further comprises a health education module for generating personalized education content through the grading results and the exercise program, and sending the personalized education content to the patient terminal.

[0026] Preferably, the system further comprises an intelligent question-answering module for obtaining the patient's questions and retrieving corresponding answers from a database, and then sending the answers to the patient's terminal.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] The present invention uses multimodal data fusion technology to combine static ulcer characteristics with dynamic physiological characteristics, and realizes automatic grading based on AI algorithm, which significantly improves the scientificity and consistency of grading evaluation. It also generates exclusive exercise plans for different patients through dynamic control mechanism and personalized recommendation algorithm, and continuously optimizes the plans according to rehabilitation progress and feedback, effectively preventing secondary damage to ulcer wounds during exercise, reducing the risk of infection and ulcer deterioration, and significantly improving the patient's rehabilitation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0030] Figure 1 It is a system module block diagram of the present invention. DETAILED DESCRIPTION

[0031] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0032] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0033] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0034] As attached Figure 1 As shown:

[0035] Embodiment 1: This embodiment provides a non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers, including:

[0036] The evaluation and grading module is used to obtain the multimodal ulcer data of the patient, and scientifically grade the patient's diabetic foot ulcer in combination with the stratification algorithm to obtain the grading results;

[0037] The assessment and grading module performs hierarchical assessment based on Wagner grading or Texas grading, and automatically generates grading results in combination with AI algorithms. The grading results include:

[0038] Mild ulcer: superficial ulcer, no infection or slight infection;

[0039] Moderate ulcer: The ulcer is deep and may involve muscle or tendon without serious infection;

[0040] Severe ulcers: Extension of infection to bones, with risk of ischemia or vascular occlusion;

[0041] Specifically:

[0042] P(y|X)=softmax(Wf(X)+b)

[0043] Where: X represents the patient's multimodal ulcer data, y represents the grading result, f(X) represents the feature extraction function, the static image features are extracted by the AI ​​algorithm (CNN), and the temporal features are extracted in combination with the dynamic data of the foot, W represents the weight matrix, b represents the bias vector, and softmax represents the normalization function, which is used to calculate the probability distribution of different grading categories.

[0044] A program generation module, used for matching a suitable non-weight-bearing exercise program for the patient in a database according to the grading results, and generating an initial exercise program;

[0045] The exercise feedback module is used to send the initial exercise plan to the patient terminal, obtain the patient's exercise completion status, and make personalized adjustments to the exercise intensity and frequency through a dynamic control mechanism;

[0046] Dynamic control mechanisms include:

[0047] S1. Obtain and analyze new multimodal ulcer data after the patient completes exercise to obtain exercise feedback;

[0048] S2, analyze movement feedback through reinforcement learning algorithms and adjust movement plans in real time;

[0049] The deep Q network is used for implementation. The loss function of the deep Q network is defined as:

[0050] L(θ)=E[(Q(s,a;θ)-yt) 2 ]

[0051] Where: Q(s,a;θ) represents the Q-value function of the current policy network, which represents the expected reward for performing action a in state s;

[0052] yt=r+γmax a′ Q′(s′,a′;θ - )

[0053] Where: r represents the immediate feedback after the patient completes the exercise (such as pain score, completion rate), γ represents the discount factor used to balance short-term and long-term rewards, Q′(s′,a′; θ - ) represents the Q value function of the target network, represents the maximum expected reward of the next state, θ represents the parameters of the policy network, and θ - Represents the parameters of the target network, which are updated regularly.

[0054] Specifically, the recommended modules for sports rehabilitation include:

[0055] A non-weight-bearing exercise database for storing exercise programs suitable for patients with diabetic foot ulcers;

[0056] The program generation unit is used to match exercise items from the non-weight-bearing exercise database and generate a personalized rehabilitation plan.

[0057] By acquiring multimodal ulcer data, the grading is made more accurate and comprehensive. Then, according to the grading results, the appropriate exercise method is matched from the non-weight-bearing exercise database to generate a personalized rehabilitation plan, avoiding the "one-size-fits-all" problem of traditional rehabilitation plans. At the same time, the dynamic control mechanism can adjust the exercise intensity and frequency in real time according to patient feedback to ensure exercise safety and effectiveness, effectively prevent secondary damage to ulcer wounds during exercise, reduce the risk of infection and ulcer deterioration, and significantly improve the patient's rehabilitation efficiency.

[0058] Specifically, the multimodal ulcer data includes static ulcer data and dynamic physiological data;

[0059] Static ulcer data included ulcer site, ulcer diameter, ulcer depth, and infection status;

[0060] Dynamic physiological data include blood circulation, plantar temperature distribution, and plantar pressure distribution.

[0061] Specifically, the movement feedback module is also used to construct a patient's personal rehabilitation behavior model based on the patient's historical movement data and rehabilitation results;

[0062] The rehabilitation behavior model is used to dynamically optimize subsequent exercise programs by continuously learning the patient's exercise habits, pain tolerance, and rehabilitation performance;

[0063] The regression model is used, and the formula is:

[0064] R t =W h ·H t +W x ·X t +b

[0065] Where: R t represents the predicted value of the patient's rehabilitation effect at time t, H tIndicates the patient's historical movement completion status, X t Indicates the current motion feedback data, W h , W x represents the regression weight, and b represents the bias term.

[0066] Specifically, the system also includes a health education module, which is used to generate personalized education content through grading results and exercise plans, and send the personalized education content to the patient terminal; this module will develop detailed health education materials for patients of different grades in combination with their corresponding exercise plans; for example, for patients with mild ulcers, the education content may focus on the correct posture of exercise and daily foot care knowledge; for patients with severe ulcers, in addition to exercise-related guidance, it will also emphasize how to prevent the spread of infection and how to deal with emergencies, etc.; the education content is diverse in form, including graphic materials, video explanations, etc., which can help patients better understand and cooperate with rehabilitation exercises.

[0067] Specifically, the system also includes an intelligent question-and-answer module, which is used to obtain patients' questions and retrieve corresponding answers from the database, and then send the answers to the patient's terminal; this module has established a comprehensive knowledge base, covering the causes, symptoms, treatment methods, rehabilitation exercise knowledge, dietary precautions and other aspects of diabetic foot ulcers; when patients ask questions, the intelligent question-and-answer module can use natural language processing technology to accurately understand the intention of the question, and quickly retrieve the most matching answer from the knowledge base, and feedback to the patient in a clear and easy-to-understand manner; for example, the patient asks "What should I do if the pain in the ulcer area worsens during exercise?" The module will quickly call out relevant response measures from the knowledge base and reply to the patient.

[0068] From the above, we can see that: 1. Input data

[0069] The evaluation module collects multimodal ulcer data of patients through intelligent terminals or auxiliary hardware, including:

[0070] Static ulcer data:

[0071] Ulcer location: such as sole, dorsum, toes, etc.;

[0072] Ulcer diameter: the maximum transverse length of the ulcer;

[0073] Ulcer depth: How deep the ulcer penetrates into the skin or soft tissue;

[0074] Infection status: whether there are any signs of infection such as redness, swelling, exudation or pus.

[0075] Dynamic physiological data (obtained through sensors):

[0076] Temperature distribution of the sole of the foot: collected by infrared thermal imaging equipment;

[0077] Plantar pressure distribution: abnormal pressure areas are collected through plantar pressure sensors;

[0078] Blood circulation conditions: such as ankle-brachial index (ABI) or skin capillary response.

[0079] 2. Data processing flow

[0080] 2.1 Data Preprocessing

[0081] Standardize the collected static data of ulcers (such as ulcer diameter and infection grade) to ensure that different types of data (such as length and category data) have consistent numerical ranges;

[0082] Dynamic data (such as temperature distribution and pressure distribution) are used to extract features through convolutional neural networks (CNN).

[0083] 2.2 Grading Algorithm

[0084] Combining the standard logic of Wagner grading and Texas grading, ulcer depth, infection status and ischemia degree are included in the grading conditions, and automatic grading is achieved through an AI model.

[0085] (1) Wagner grading algorithm rules

[0086] Grade 0: no ulcers, but with foot deformity or risk factors;

[0087] Grade 1: Superficial ulcer, limited to the skin surface;

[0088] Grade 2: Deeper ulcer involving muscle or tendon;

[0089] Grade 3: ulcer with infection or abscess, possibly with bone infection;

[0090] Grade 4: local gangrene (such as toe gangrene);

[0091] Level 5: Gangrene of the entire foot, requiring amputation.

[0092] (2)Texas grading algorithm rules

[0093] Depth Level:

[0094] 0: no open ulcer;

[0095] 1: Superficial ulcer;

[0096] 2: The ulcer is deep into the muscle or tendon;

[0097] 3: The ulcer involves bone tissue or joints.

[0098] Stage classification:

[0099] A: No infection, no ischemia;

[0100] B: infection present;

[0101] C: ischemia present;

[0102] D: Infection and ischemia coexist.

[0103] 2.3 AI algorithm assistance

[0104] A deep learning model combining convolutional neural network (CNN) and long short-term memory network (LSTM) is introduced to fuse static and dynamic data and improve classification efficiency and accuracy.

[0105] Model Training

[0106] Use a multi-center diabetic foot patient dataset to train a classification model:

[0107] Static data is used to extract features such as boundaries and ulcer area through CNN;

[0108] Dynamic data uses LSTM to capture time series changes, such as dynamic changes in pressure distribution;

[0109] Finally, the classification results are generated through the fully connected layer and the softmax function.

[0110] 3. Grading results

[0111] The classification results are divided into the following three situations, and personalized follow-up processing suggestions are generated:

[0112] Mild ulcers:

[0113] Superficial ulcers, no infection or minimal infection;

[0114] Mild non-weight-bearing exercise is recommended, such as water exercise and upper limb rehabilitation exercises.

[0115] Moderate ulcers:

[0116] The ulcer is deep and may involve muscle or tendon without serious infection;

[0117] Moderate-intensity non-weight-bearing exercise, such as seated yoga and wheelchair exercises, is recommended, supplemented by nursing advice.

[0118] Severe ulcers:

[0119] Extension of infection to bone, with attendant risk of ischemia or vascular occlusion;

[0120] Exercise is not recommended for the time being. It is recommended to focus on nursing and debridement, and remind patients to seek medical attention as soon as possible;

[0121] For example: Input data:

[0122] Static ulcer data: location (sole), diameter (1.5 cm), depth (0.3 cm), infection status (slight redness and swelling, no pus);

[0123] Dynamic data: temperature distribution (no abnormal increase), pressure distribution (pressure is relatively uniform, without obvious concentration); ABI is normal (>0.9).

[0124] Grading results:

[0125] The system determined it to be a mild ulcer (Wagner grade 1) through the AI ​​grading model.

[0126] Personalized recommendation:

[0127] Recommended exercise: air bike exercise (15 minutes / time, 3 times a week), combined with upper limb strength training (10 minutes / time, 2 times a week);

[0128] Recommended care: Keep the wound clean and avoid direct weight bearing.

[0129] Example 2: (a) Basic information of the patient:

[0130] The patient, Mr. Li, a 65-year-old male, had been suffering from diabetes for 10 years and recently discovered foot ulcers.

[0131] (II) Assessment and grading:

[0132] Data Collection:

[0133] Static ulcer data: The ulcer site is the sole of the foot, the ulcer diameter is 15mm, the ulcer depth is 5mm, and the infection is slightly red and swollen.

[0134] Dynamic physiological data: Blood circulation showed that the blood flow rate was slightly lower than the normal range, the plantar temperature distribution around the ulcer was slightly abnormal, and the plantar pressure distribution at the corresponding position of the ulcer was high.

[0135] Grading evaluation: The above data are organized into a feature matrix and input into the CNN-based evaluation model. After calculation by the trained model, the output grading result is moderate ulcer.

[0136] (III) Plan generation:

[0137] Non-weight-bearing exercise database retrieval: The program generation unit screened out suitable exercise items from the non-weight-bearing exercise database based on the moderate ulcer grading results, and selected arm stretching exercises on a wheelchair (exercise type), with the exercise intensity set at 50% of the heart rate reserve, the exercise frequency set at 3 times a week, and each exercise duration of 20 minutes; and simple sitting balance training (exercise type), with the exercise intensity set at moderate difficulty, the exercise frequency set at 4 times a week, and each exercise duration of 15 minutes.

[0138] Personalized rehabilitation plan generation: Combine the above exercise programs to generate a personalized rehabilitation plan, including detailed arrangements such as exercise sequence and rest time.

[0139] (IV) Movement feedback and adjustment:

[0140] Exercise execution and data collection: After patient Li performed exercise for a week according to the initial exercise plan, the exercise feedback module collected new data. The new ulcer had a diameter of 14mm and a depth of 4.5mm. The infection was slightly red and swollen, and the blood circulation was improved. The average pain level during exercise was 3.

[0141] Dynamic control mechanism:

[0142] S1: The calculated changes in ulcer diameter were -1mm, ulcer depth -0.5mm, infection severity 0, and blood circulation was positive, indicating that exercise had a certain improvement effect on ulcers, but patients still had some pain.

[0143] S2: These changes are used as state inputs into the DQN-based reinforcement learning model. The model outputs an adjusted exercise plan based on the reward mechanism (positive rewards for ulcer improvement and negative rewards for pain severity), appropriately reducing the intensity of arm stretching exercises, reducing fatigue and pain during exercise, and increasing the frequency of sitting balance training to further improve the patient's balance ability and promote overall recovery.

[0144] (V) Health Education:

[0145] According to the moderate ulcer grading results and the adjusted exercise plan, the health education module generates personalized education content, including graphic materials, introducing how to reduce the pressure on the sole of the foot by adjusting the sitting posture during exercise, and how to observe the changes in ulcers and promptly feedback to doctors, and sends them to the patient's terminal.

[0146] (VI) Intelligent Question and Answer:

[0147] Patient Li had questions about dietary precautions during exercise and asked questions to the intelligent question-answering module through the terminal. The intelligent question-answering module analyzed the questions and retrieved relevant answers from the database about whether patients with diabetic foot ulcers should control sugar intake and increase protein intake, and fed back to the patient.

[0148] Importantly, it should be noted that the construction and arrangement of the present application shown in a plurality of different exemplary embodiments are only exemplary. Although only a few embodiments are described in detail in this disclosure, it should be readily understood by those who refer to this disclosure that many modifications are possible (e.g., the size, scale, structure, shape and ratio of various elements, and parameter values ​​(e.g., temperature, pressure, etc.), installation arrangement, use of materials, color, directional changes, etc.) without substantially departing from the novel teachings and advantages of the subject matter described in the application. For example, the element shown as integrally formed can be composed of multiple parts or elements, the position of the element can be inverted or otherwise changed, and the nature or number or position of the discrete element can be changed or changed. Therefore, all such modifications are intended to be included in the scope of the present invention. The order or sequence of any process or method steps can be changed or reordered according to alternative embodiments. In the claims, any "device plus function" clause is intended to cover the structure of the execution function described herein, and is not only structurally equivalent but also equivalent structure. Without departing from the scope of the present invention, other replacements, modifications, changes and omissions can be made in the design, operating conditions and arrangement of the exemplary embodiments. Therefore, the invention is not limited to a specific embodiment, but extends to several modifications still falling within the scope of the appended claims.

[0149] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention) may not be described.

[0150] It will be appreciated that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will be a routine task of design, fabrication, and production for those of ordinary skill having the benefit of this disclosure without undue experimentation.

[0151] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers, characterized in that: include: The evaluation and grading module is used to obtain the multimodal ulcer data of the patient, and scientifically grade the patient's diabetic foot ulcer in combination with the stratification algorithm to obtain the grading results; A program generation module, used for matching a suitable non-weight-bearing exercise program for the patient in a database according to the classification result, and generating an initial exercise program; The exercise feedback module is used to send the initial exercise plan to the patient terminal, obtain the patient's exercise completion status, and make personalized adjustments to the exercise intensity and frequency through a dynamic control mechanism.

2. A non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers according to claim 1, characterized in that: The evaluation and grading module performs hierarchical evaluation based on Wagner grading or Texas grading, and automatically generates grading results in combination with AI algorithms. The grading results include: Mild ulcer: superficial ulcer, no infection or slight infection; Moderate ulcer: The ulcer is deep and may involve muscle or tendon without serious infection; Severe ulcers: extension of infection to bone, with risk of ischemia or vascular occlusion.

3. The non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers according to claim 1, characterized in that: The sports rehabilitation recommendation module includes: A non-weight-bearing exercise database for storing exercise programs suitable for patients with diabetic foot ulcers; A program generating unit is used to match exercise items from the non-weight-bearing exercise database and generate a personalized rehabilitation plan.

4. The non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers according to claim 1, characterized in that: The multimodal ulcer data includes static ulcer data and dynamic physiological data; The static ulcer data include ulcer site, ulcer diameter, ulcer depth, and infection status; The dynamic physiological data include blood circulation conditions, plantar temperature distribution, and plantar pressure distribution.

5. The non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers according to claim 1, characterized in that: The dynamic control mechanism includes: S1. Obtain and analyze new multimodal ulcer data after the patient completes exercise to obtain exercise feedback; S2. Analyze the movement feedback through a reinforcement learning algorithm and adjust the movement plan in real time.

6. The non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers according to claim 1, characterized in that: The movement feedback module is also used to construct a patient's personal rehabilitation behavior model based on the patient's historical movement data and rehabilitation results; The rehabilitation behavior model is used to dynamically optimize subsequent exercise plans by continuously learning the patient's exercise habits, pain tolerance, and rehabilitation performance.

7. The non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers according to claim 1, characterized in that: The system also includes a health education module, which is used to generate personalized education content through the classification results and the exercise program, and send the personalized education content to the patient terminal.

8. The non-weight-bearing exercise rehabilitation training decision support system for patients with diabetic foot ulcers according to claim 1, characterized in that: The system also includes an intelligent question-answering module for obtaining the patient's questions and retrieving corresponding answers from a database, and then sending the answers to the patient's terminal.

Images