Clinical depression and melanoma patient nursing intervention system and method

Abstract

The invention discloses a clinical depression and melanoma patient nursing intervention system and method, which utilizes big data and artificial intelligence technology to establish a digital twinborn model according to the information, illness state, depression state and the like of a patient, collects existing cases for training to continuously optimize the model, and improves the precision of the digital twinborn model. And then similar cases are recommended by utilizing the database, the patient information and the digital twinborn model, a nursing scheme is generated, and then the nursing scheme is input into the digital twinborn model to be tested and predicted and is continuously improved. According to the mode, the efficiency of making the nursing scheme can be greatly improved, personalized customization can be carried out according to the actual condition of the patient, and the nursing effect is greatly improved. Meanwhile, the physiological data of the patient is detected by adopting the wearable equipment, whether the patient is in a depression state or in a depression mood is speculated by utilizing the physiological data, and psychological nursing is carried out through nursing intervention measures set in advance, so that the depression state of the patient is effectively improved.

Description

Technical Field

[0001] The invention relates to software technology, and in particular to a nursing intervention system and method for clinically depressed patients with melanoma. Background Art

[0002] Melanoma is a serious malignant tumor of the skin that progresses rapidly and is difficult to treat, causing great physical pain and psychological pressure on patients. At the same time, clinical studies have shown that many melanoma patients experience depression during their illness, which not only affects the patient's quality of life and treatment compliance, but may also aggravate the condition and have an adverse effect on the patient's prognosis.

[0003] Existing studies have revealed that depression is correlated with malignant melanoma. The chemotherapy effect of malignant melanoma patients with depression is worse than that of malignant melanoma patients without depression. The connection between DUSP11, malignant melanoma and depression has been found, but its interaction and significance need further study.

[0004] At present, the care for melanoma patients mainly focuses on the physiological treatment of the disease, but not enough attention is paid to the psychological depression of patients. At the same time, the existing care methods lack systematicity and pertinence, cannot effectively alleviate the depressive symptoms of patients, and cannot meet the comprehensive physiological and psychological needs of patients. At the same time, they cannot reveal the relationship and significance between DUSP11 and depression-induced malignant melanoma, which has become a technical problem that needs to be solved at present. Summary of the invention

[0005] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is to provide a nursing intervention system and method for clinically depressed patients with melanoma, which can formulate personalized physiological care and psychological care plans according to the characteristics of the patients.

[0006] To achieve the above object, the present invention provides a nursing intervention system for clinically depressed patients with melanoma, comprising: Information collection module, used to collect patient information as well as existing case information, existing treatment and nursing information; The digital twin model generates a digital model based on the patient's physical data, historical cases, DUSP11 data, depression history, and treatment plan. The digital twin model is then combined with existing treatment and nursing plans to recommend treatment and nursing plans suitable for the patient. Different nursing plans are also evaluated and simulated to predict subsequent results. Nursing planning module, used to generate nursing plans and optimize them in combination with treatment plans; Nursing implementation and evaluation module, which is used to monitor the implementation of nursing plans and conduct evaluations, and continuously optimize and adjust nursing plans based on the evaluation and implementation status; Wearable devices are hardware worn on patients. They collect the patient's heart rate, electrocardiogram, respiratory rate, blood pressure, skin resistance, sleep conditions, and exercise conditions, and combine these parameters to analyze whether depression is likely to exist. If so, nursing access is performed according to the preset program to achieve psychological care.

[0007] As a further improvement of the present invention, the digital twin model includes: The patient's basic physiological information, including the patient's physical information and medical history obtained through patient examination, pathology and other information. The physical information includes height, weight, IBM disability status, and the medical history mainly includes other diseases that the patient has or has had, as well as the patient's personality, hobbies, and specialties; Melanoma conditions, including tumor characteristics, treatment process, and medical history. Tumor characteristics include tumor location, size, morphology, pathological type, and genetic characteristics. The treatment process includes treatment methods and treatment response. The genetic characteristics are an analysis of genes related to melanoma. Depression-related parameters, including psychological assessment scores, indicators related to emotional state, social support, and depression history; DUSP11 status, including DUSP11 historical status, DUSP11 current status, and the relationship between DUSP11 data and depression and melanoma status. The current and past DUSP11 expression and methylation status of the patient are collected, and these data are associated with depression and melanoma status according to the proposed relationship function. Nursing plan recommendation: compare the collected information with similar nursing plan cases in the database, and score the similar cases. The nursing plan with the highest score will be the recommended nursing plan. Nursing simulation: simulate the nursing plan based on the recommended nursing plan combined with similar cases, nursing results of similar cases, and the simulation model set by the system, and give corresponding nursing result predictions, optimization suggestions, and precautions; First, as much patient information as possible is collected, and then a digital twin model is established. Through the patient information and the digital twin model, similar cases are found in the database, and then evaluated and recommended nursing plans are output. The user confirms, modifies, and optimizes the recommended nursing plan into a proposed nursing plan, and then inputs the proposed nursing plan into the digital twin model for simulation. The digital twin model outputs prediction results, precautions, and optimization suggestions. The user makes adjustments based on the results of the digital twin model until a satisfactory nursing plan is obtained; then nursing begins, and patient data, nursing data, nursing effects, and feedback from medical staff and patients are collected regularly or irregularly to monitor the implementation of the nursing plan; throughout the process, patient information is collected through wearable devices, and depression is judged based on this information. If depression is likely, timely intervention is made according to the system settings.

[0008] The present invention also discloses a nursing intervention method for patients with clinical depression and melanoma, comprising the following steps: S100, collecting the patient's physical information, physiological data, exercise status, personality and emotional information, historical cases, DUSP11 expression and methylation status information; S200, establish a digital twin model, recommend similar plans based on patient information and the digital twin model, and output a nursing plan; S300, nursing plan planning, adjust the nursing plan output by the system, input it into the digital twin model for simulation and evaluation, and improve the nursing plan based on the evaluation; S400, implementation and evaluation of nursing plans. Implement nursing care according to the nursing plan, and evaluate the effectiveness of the nursing plan based on the feedback from medical staff and patients to continuously optimize the nursing plan.

[0009] As a further improvement of the present invention, S100 also includes: S110, collecting patient information, including collecting the patient's height, weight, IBM, and disability information, and using smart wearable devices to detect the patient's physiological data and exercise conditions, and inferring whether the patient is in a suspected depression or suspected of having depression through these data. If so, depression nursing intervention is performed according to a pre-set nursing plan, and then the improvement of depression is judged in combination with the parameters collected subsequently; the physiological data includes blood pressure, heart rate, electrocardiogram, respiratory rate, skin resistance, and sleep conditions; S120, collecting the patient's personality and emotional information, including depression, personality, hobbies, specialties, and emotional experiences; at the same time, collecting as much historical information as possible, including historical information on depression, personality, hobbies, specialties, and emotional experiences; S130. Historical cases are to collect some historical case information of patients before they visit the hospital, including melanoma tumor information, depression status, and other diseases. Melanoma tumor information includes historical and current tumor characteristics, treatment methods, treatment response, and treatment effect; tumor characteristics include tumor location, size, morphology, pathological type, and genetic characteristics. Genetic characteristics include BRAF mutation, NRAS mutation, c-KIT mutation, DUSP11 expression, and methylation status; other diseases include current and historical illnesses, treatment plans for the disease, and recovery status; The information on S140 and DUSP11 expression and methylation status is collected as current and historical data.

[0010] As a further improvement of the present invention, in S110, the method of inferring the suspected depression state and evaluating the effect after nursing is as follows: S111, index normalization processing Heart rate (HR): set normalized heart rate NHR = (HR-60) / (100-60); Sleep duration (SL): Normalized sleep duration NSL = (SL-7) / (9-7); Sleep interruption count (SI): Normalized sleep interruption count NSI = SI / 2 (the more the number, the greater the value); Movement steps (SS): Normalized movement steps NSS = (SS-5000) / (10000-5000); Exercise duration (ST): Normalized exercise duration NST = (ST-30) / (60-30); Skin conductivity (SC): Normalized skin conductivity NSC = (SC-10) / (20-10); Blood pressure: systolic blood pressure is SBP, diastolic blood pressure is DBP, normalized systolic blood pressure NSBP=(SBP-90) / (140-90), normalized diastolic blood pressure NDBP=(DBP-60) / (90-60); ECG: Assume that the frequency of abnormal ECG index is AECG, and the normalized frequency of abnormal ECG index is NAECG=AECG / 10%; Respiratory frequency (RF): Normalized respiratory frequency NRF = (RF-12) / (20-12); S112. Construct a formula for estimating depression status (D represents depression status index) D=ω1×NHR+ω2×NSL+ω3×NSI+ω4×NSS+ω5×NST+ω6×NSC+ω7×NSBP+ω8×NDBP+ω9×NAECG+ω10×NRF, where ω1, ω2, ω3, ω4, ω5, ω6, ω7, ω8, ω9, ω10 are the weight coefficients corresponding to each indicator; S113. Determine the calibration value of the depression state After a large number of clinical sample tests, data from known depressed patients and non-depressed patients were collected to calculate the average value of the depression state index D of the depressed patients. depressed and standard deviation S depressed , and the average value of the depression state index D of non-depressed patients non-depressed and standard deviation S non-depressed ; Set a threshold value D for depression threshold , determined by: D threshold =D non-depressed +k×S non-depressed Where k is a coefficient determined according to actual conditions, and is usually selected through statistical analysis and weighing the misjudgment rate.

[0011] As a further improvement of the present invention, S200 also includes: S210, Data Collection and Integration Collect comprehensive patient data from multiple sources including the hospital's electronic medical record system, laboratory test database, and smart wearable device monitoring platform; including but not limited to: basic patient information, disease-related information, depression assessment scale scores, blood test indicators, DUSP11 expression and methylation status, physiological data collected by smart wearable devices, patient social environment information, and patient personality, hobbies, emotional experiences, and emotions; Clean and pre-process the collected data, remove duplicate, erroneous or incomplete data, and integrate data from different sources to ensure data consistency and accuracy; S220, Feature Engineering Extract and construct features from preprocessed data, and standardize or normalize continuous features to make different features comparable in value, which is convenient for model learning; S230, Model Selection and Architecture Design The architecture combines the convolutional neural network (CNN) and the long short-term memory network (LSTM) in deep learning. CNN is used to automatically extract local features in the data, and LSTM is used to process time series data and capture long-term dependencies. The input layer receives the multi-dimensional feature vector after feature engineering. The middle layer consists of multiple convolutional layers and LSTM layers for feature learning and pattern recognition. The output layer is set according to the specific task. S240, model training Divide the dataset into training set, validation set and test set; the training set is used for model learning, the validation set is used to adjust hyperparameters and monitor the model training process, and the test set is used to finally evaluate the model performance; Select appropriate optimization algorithms and loss functions, and iteratively train the model based on training data; During the training process, the early stopping method is used to stop training when the loss on the validation set no longer decreases significantly to prevent overfitting; S250, Hyperparameter Tuning Tune the model’s hyperparameters; S260, Model Evaluation and Validation Use the test set to evaluate the trained model and use a variety of evaluation indicators to comprehensively evaluate the performance of the model; Perform visual analysis, such as plotting a comparison chart of predicted values ​​and true values, viewing the model's attention weight distribution, and gaining an in-depth understanding of the model's decision-making process and potential problems; S270, Model Fusion and Integration Try to integrate multiple trained models, such as random forest and Adaboost in ensemble learning, to combine the prediction results of multiple models and improve the stability and generalization ability of the model; Combined with a rule-based expert system, prior knowledge and experience in the clinical field are converted into rules and integrated with the model's prediction results to enhance the reliability and interpretability of the model; S280, Model Deployment and Update Deploy the trained digital twin model to the clinical application system to achieve real-time data input and prediction output; Establish a monitoring mechanism for the model and regularly collect new data to retrain and update the model to adapt to changes in patient conditions and treatment methods, ensuring the accuracy and effectiveness of the model.

[0012] As a further improvement of the present invention, S300 also includes: S310, recommended solutions S311. Compare the patient data categories and corresponding values ​​with existing cases, first find the cases with the most identical or similar data categories; then compare the parameters of similar or identical categories, and identify the data items with parameters within a certain range as identical data items; sort by the number of identical data items, select the first few cases as references, and use the nursing plans recorded in them as the basic plan for nursing plan planning, and subsequent users can improve on this plan; You can also weight the categories and parameters, and finally rank them by weighted scores, and just select the top few items; S312, screening corresponding cases according to the set conditions, scoring different data categories of each case in advance, sorting by the selected data category scores during screening, and selecting the top few as recommendations; S313, setting nursing plan planning conditions in the system, then matching according to the patient's condition, and finally combining and outputting the corresponding recommended nursing plan; S320, Digital Twin Model Simulation S321. Input the selected nursing plan into the digital twin model for simulation, and output the evaluation conclusion and precautions; S322. Based on the simulation results, output the deficiencies or defects of the solution or areas for improvement, and provide optimization recommendations; S323, predicting each stage after the implementation of the nursing plan, and outputting the judgment of abnormality of the prediction indicators of DUSP11 expression and methylation status; DUSP11 expression index: The expression of DUSP11 is obtained through blood testing, recorded as DUSP11_E. The upper and lower floating thresholds are set based on the average value of the normal population. When DUSP11_E exceeds this range, it is considered abnormal; DUSP11 methylation status indicator: The methylation degree of the DUSP11 gene is detected, recorded as DUSP11_M, and the threshold is set based on the normal mean. If it exceeds this range, it is considered abnormal.

[0013] As a further improvement of the present invention, S320 also includes: S324, specific simulation process S324.1 Digital coding of nursing plans Digitally encode the various care plans so that the model can understand and process them, and also encode the specific details of each care plan accordingly; S324.2 Establishing nursing program parameter library Create a database containing various care regimen parameters; S324.3 model input interface design Design a dedicated input interface in the digital twin model to ensure that the coding and parameters of the nursing plan can be accurately and efficiently input into the model. The input interface should have data verification and error correction functions to prevent erroneous or incomplete data from entering the model and affecting the accuracy of the simulation results. S324.4 Simulation scene setting Set the initial conditions of the simulation based on the patient’s current status and historical data; S324.5 Nursing plan input and model operation The digitally encoded nursing plan and its parameters are input into the digital twin model through the input interface. After receiving the nursing plan, the model combines the initial conditions and the built-in algorithm logic to start simulating the changes in the patient's physiological and psychological state after receiving the nursing plan. S324.6 Determination of evaluation indicators for simulation results Define a series of evaluation indicators to measure the effectiveness of the simulation results and the potential effects of the nursing plan, including the degree of improvement of depressive symptoms, the recovery of physiological indicators, and the changing trends of DUSP11 expression and methylation status; S324.7 Multi-round simulation and optimization Conduct multiple rounds of simulations on the same nursing plan, fine-tune the input parameters each time, and observe the changing trend of the results; at the same time, combine and input different nursing plans to explore the best nursing plan combination; S324.8 Result analysis and decision support Conduct detailed analysis on the large amount of data generated by the simulation, compare the simulation results of different nursing plans or combinations of plans, and evaluate their effectiveness in improving patients' depression and promoting recovery; based on the analysis results, provide decision support to medical staff, recommend the best nursing plan or make suggestions for improving existing plans; S330, expert evaluation, through experts to evaluate and improve the current nursing plan, or through experts to improve the nursing plan of previous cases, through these improvements to optimize the nursing plan, and continuously improve the nursing plan; S340, dynamic adjustment, dynamically optimize and adjust the nursing plan according to system data update, weight adjustment, and manual adjustment.

[0014] As a further improvement of the present invention, S400 also includes: S410. Provide nursing care to patients according to the nursing plan and evaluate the nursing effect regularly or irregularly; S411. After a period of nursing, re-evaluate the monitoring indicators of melanoma, DUSP11 expression and methylation status, and the comprehensive depression evaluation index Y; set the nursing effect improvement index P = ((Y_Pre-Y_Post) / Y_Pre) × 100%, where Y_Pre is the comprehensive depression evaluation index before nursing, and Y_Post is the index after nursing. When P ≥ 30%, the current nursing method is considered to be effective and should be continued; when P < 30%, the nursing method should be changed; S412, immediate intervention and recording When the monitoring data of the smart wearable device indicates that immediate intervention is needed, the smart wearable device intervenes according to the program; after each intervention, the patient's immediate response and the changes in the parameters of the smart wearable device over a period of time are recorded, and these changes are parameterized; S420, Nursing plan optimization and adjustment Various existing nursing plans are input into the digital twin model, and the model gives scores based on the predicted effects. Plans with scores higher than 80 points are listed as preferred plans. After the preferred plan is implemented, the patient's status and indicator changes are continuously monitored. If the actual effect does not meet expectations, the reasons are analyzed and the model parameters and nursing plans are adjusted. S430, intelligent terminal assisted treatment and nursing Smart wearable devices are connected to the patient’s mobile phones and other smart terminals, and timely nursing intervention is provided through smart terminals to help patients relieve depression.

[0015] As a further improvement of the present invention, it also includes: The association between S500, DUSP11 and depression, the relationship between DUSP11 and methylation and depression was established by regression statistics; S510. Data collection: Collect data on DUSP11 expression and methylation level of a large number of patients with depression and malignant melanoma, as well as the corresponding depression assessment scale scores; S520, variable definition: DUSP11 expression and methylation level were set as independent variables, and depression assessment scale score was set as dependent variable; S530, select regression model: first consider using a simple linear regression model. If the data presents a nonlinear relationship, try using polynomial regression, logistic regression or a more complex nonlinear regression model. S540, Model Fitting: Fit the selected regression model using the collected data; S550, Model evaluation: Use evaluation indicators to judge the goodness of fit of the model; you can also perform residual analysis to check whether the assumptions of the model are met; S560, Model Diagnosis: Check whether there are multicollinearity, outliers, and high leverage points; S570, model optimization: If there are problems with the model, try to optimize it by changing variables, adding or deleting variables, or selecting a different model; S580 Final model determination: Select the optimal model as the final relationship model.

[0016] The beneficial effects of the present invention are: The present invention uses big data and artificial intelligence technology to establish a digital twin model based on the patient's information, condition, depression state, etc., and collects existing cases for training to continuously optimize the model and improve the accuracy of the digital twin model. Then, the database, patient information, and digital twin model are used to recommend similar cases and generate nursing plans, and then the nursing plans are input into the digital twin model for testing, prediction, and continuous improvement. This method can greatly improve the efficiency of making nursing plans, and can be customized according to the actual situation of the patient to greatly improve the nursing effect. At the same time, the present invention uses wearable devices to detect the patient's physiological data, uses physiological data to infer whether the patient is in a depressed state or depressed mood, and performs psychological care through nursing intervention measures set in advance to effectively improve the patient's depressive state. At the same time, the present invention also records the data of DUSP11 expression and methylation status indicators and depression, and attempts to find the relationship between them to provide a basis for subsequent research, and also provide clearer indicators for nursing effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the nursing intervention system; Figure 2 It is a schematic diagram of the construction of the digital twin model; Figure 3 It is a flowchart of the nursing intervention method; Figure 4 It is a flowchart showing the process of information collection steps. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0019] See also Figure 1 , Figure 4 The nursing intervention system of this embodiment includes: The information collection module is software installed on a computer or smart wearable device, which is used to collect patient information, existing case information, existing treatment and nursing information. The collected patient information includes the patient's physical parameters, historical cases, DUSP11 data, physiological data, exercise conditions, suspected depression conditions, depression nursing intervention conditions, etc.; the collected existing cases include the basic conditions, conditions, treatment conditions, depression history, DUSP11 data, etc. of existing patients, which are used to continuously optimize the model. Existing treatment and nursing information includes treatment and nursing plans optimized for some existing cases, some new treatment and nursing plans, some new intervention plans for depression, etc., so that the entire system is always in a data update state to adapt to the development of medical care technology.

[0020] The digital twin model is software installed on a computer or smart wearable device. It generates a digital person based on the patient's physical data, historical cases, DUSP11 data, depression history, treatment plan, etc. This mainly relies on big data and artificial intelligence technology. Through a large amount of data, the model is domesticated and the corresponding digital twin model is generated using patient information. Subsequently, the digital twin model can be combined with existing treatments and nursing plans to recommend treatments and nursing plans suitable for patients. At the same time, different nursing plans can be evaluated, simulated and predicted for subsequent results to find a better nursing plan for the patient. At the same time, the nursing plan can be adjusted based on the feedback from medical staff, experts and patients, and the adjusted plan and results can be evaluated and predicted through the digital twin model.

[0021] The nursing planning module is software installed on computers or smart wearable devices to generate nursing plans, such as physiological nursing plans, psychological nursing plans, etc., and optimize the nursing plans in combination with treatment plans. For example, chemotherapy may cause hair loss and loss of appetite, and the nursing plans can provide some nutritional supplement plans, psychological counseling plans, etc. During use, the nursing plans can be adjusted through expert evaluation, dynamic adjustment, digital twin model simulation, etc., in an effort to provide a nursing plan that is more suitable for the patient.

[0022] The nursing implementation and evaluation module is a software installed on a computer or smart wearable device to monitor the implementation of the nursing plan. On the one hand, it can confirm whether the nursing plan is being implemented based on the feedback from medical staff and patients; on the other hand, it can evaluate the nursing plan through the collected data and compare it with the results of pre-simulation predictions, thereby continuously optimizing the digital twin model; at the same time, during use, the nursing plan can be continuously optimized and adjusted based on the evaluation and implementation status.

[0023] The communication and education platform is a software installed on computers or smart wearable devices for communication between medical staff, between patients, and between doctors and patients. It also provides educational materials and platforms for medical staff and patients. On the one hand, this method can strengthen communication and education between medical staff, and on the other hand, it can also increase the efficiency of doctor-patient communication and patient education. At the same time, medical staff and patients can use this platform to provide feedback on treatment, nursing effects and suggestions, so as to continuously optimize the nursing plan, provide patients with a better treatment environment and improve treatment effects.

[0024] The wearable device is a hardware device worn by the patient. In this embodiment, it is preferably a smart wearable device. The wearable device collects the patient's heart rate, electrocardiogram, respiratory rate, blood pressure, skin resistance, sleep status, and exercise status, and analyzes whether there is depression based on these parameters. If there is depression, nursing access is performed according to the preset program to achieve psychological care.

[0025] See also Figure 2, the digital twin model includes: The patient's basic physiological information, through the patient's examination, pathology and other information to obtain the patient's physical information, past medical history, the physical information includes height, weight, IBM disability, etc., the past medical history mainly includes other diseases that the patient has or has had. Of course, it can also include the patient's personality, interests, hobbies, specialties, etc. Personality can be evaluated through the Eysenck Personality Questionnaire (EPQ).

[0026] Melanoma conditions, including tumor characteristics, treatment process, and historical medical history. Tumor characteristics include tumor location, size, morphology, pathological type, and genetic characteristics; the treatment process includes treatment methods and treatment responses, such as whether chemotherapy is used, the patient's physical response and effect after chemotherapy, and what type of treatment is used for depression; the genetic characteristics can be used to analyze genes related to melanoma, such as BRAF, NRAS, c-KIT, and DUSP11 expression and methylation status.

[0027] Depression-related parameters include psychological assessment scale scores, emotional state-related indicators, social support, and depression history. The psychological assessment scale can use the Self-Rating Anxiety Scale (SAS) and the Self-Rating Depression Scale (SDS) to assess the patient's depression and anxiety status; emotional state-related indicators include heart rate, blood pressure, respiratory rate, galvanic skin response (GSR), emotional intensity, etc., which can be assessed using the Emotional-Social Loneliness Inventory (ESLI), Positive and Negative Affect Scale (PANAS), and Positive State Scale (POMS); social support can be further assessed using the Social Support Rating Scale (SSRS) and the Multidimensional Perceived Social Support Scale (MSPSS); depression history records mainly obtain the patient's depression history through self-reports by the patient and family members, historical pathology, depression diagnosis and treatment, etc., to facilitate subsequent psychological care.

[0028] DUSP11 status, including DUSP11 historical status, DUSP11 current status, and the correlation between DUSP11 data and depression and melanoma status. The patient's current and previous DUSP11 expression and methylation status are collected, and these data are associated with depression and melanoma status according to the proposed relationship function for subsequent verification and continuous improvement, in an effort to find the accurate correlation between DUSP11 expression and methylation status and the development of depression and melanoma.

[0029] Nursing plan recommendation: compare the collected information with similar nursing plan cases in the database, and score the similar cases. The top three nursing plans are recommended. Specifically, the collected patient parameters and information can be compared with the database to find similar patients, such as cases with a similarity of 80 or more, and then ranked according to the nursing effects recorded by the system. This nursing effect can be scored manually in advance, and the top three nursing effect scores are recommended.

[0030] Nursing simulation simulates the nursing plan based on the recommended nursing plan combined with similar cases, nursing results of similar cases, simulation models set by the system, etc., and gives corresponding nursing result predictions, optimization suggestions, and precautions. The precautions are mainly based on the nursing plan prediction that a certain plan has a low score and is not effective or has a low score (medical staff, patient score) in most (such as 70%) nursing cases. At this time, the user is prompted to choose carefully.

[0031] When using it, first collect as much patient information as possible, then build a digital twin model, find similar cases in the database through the patient information and the digital twin model, then evaluate and output the recommended nursing plan, the user confirms, modifies, and optimizes the recommended nursing plan into a proposed nursing plan, and then inputs the proposed nursing plan into the digital twin model for simulation. The digital twin model outputs prediction results, precautions, optimization suggestions, etc. The user can adjust according to the results of the digital twin model until a satisfactory nursing plan is obtained. Then start nursing, and collect patient data, nursing data, nursing effects, feedback from medical staff and patients regularly or irregularly to monitor the implementation of the nursing plan; if there are deficiencies in the nursing plan, it can be modified manually or by experts, and the nursing effect can be fed back regularly or irregularly, and the plan with poor nursing effect can be improved. During the whole process, patient information is collected through wearable devices, and depression is judged through this information. If depression is found to be possible, timely intervention is carried out according to the system settings, such as notifying family members and medical staff to contact and guide in time, playing music, and intelligent robots initiating voice chats, etc., in this way, depression is interfered with and the patient's depressive state is improved.

[0032] See also Figure 3-Figure 4 The nursing intervention method of this embodiment includes the following steps: S100, Information Collection S110, collecting patient information. The collection of patient information mainly includes collecting the patient's height, weight, IBM, and disability information on site or during physical examination. At the same time, the patient's physiological data and exercise conditions are detected by smart wearable devices. These data are used to infer whether the patient is in a suspected depression or suspected of having depression. If so, depression care intervention is performed according to the pre-set care plan, and then the improvement of depression is judged in combination with the parameters collected later. The physiological data includes blood pressure, heart rate, electrocardiogram, respiratory rate, skin resistance, sleep conditions, etc. Of course, it also includes academic qualifications, educational experience, etc.

[0033] S120. Collect the patient's personality and emotional information, including depression, personality, hobbies, specialties, and emotional experiences. Depression and personality can be quantitatively judged using existing scales. For example, depression can be assessed using the Self-Rating Anxiety Scale (SAS) and the Self-Rating Depression Scale (SDS); the Emotion-Social Loneliness Scale (ESLI), the Positive and Negative Affect Scale (PANAS), and the Positive and Negative Affect Scale (POMS) can be used to assess the patient's emotions; social support can be further assessed using the Social Support Rating Scale (SSRS) and the Multidimensional Perceived Social Support Scale (MSPSS). Emotional experiences can be quantitatively tested using the Relationship Assessment Scale (RAS), the Adult Attachment Scale (AAS), and the Love Attitude Scale (LAS). Hobbies and specialties can be tested using scales such as the Holland Vocational Interest Test (RIASEC) and the Gardner Multiple Intelligence Theory-related Assessment (MI).

[0034] At the same time, as much historical information as possible is collected, including historical information on depression, personality, hobbies, specialties, and emotional experiences. Historical records are mainly obtained through self-reports of patients and family members, historical pathology, depression diagnosis and treatment, etc., in order to facilitate subsequent psychological care.

[0035] S130, historical cases, mainly collect some historical cases of patients before the visit, including melanoma information, depression status, other diseases, melanoma information including historical and current tumor characteristics, treatment methods, treatment response, treatment effect, etc.; tumor characteristics include tumor location, size, morphology, pathological type, gene characteristics, gene characteristics include BRAF mutation, NRAS mutation, c-KIT mutation, DUSP11 expression and methylation status, etc.; treatment methods, treatment response, treatment effect are related, treatment methods include targeted therapy, immunotherapy, chemotherapy, radiotherapy, surgical treatment, etc. Depression status includes current depression status, historical depression cases, current and historical depression treatment and nursing measures, etc. Other diseases mainly include current and historical illness, treatment plan and recovery status of the disease, etc., mainly focusing on diseases that affect melanoma and depression.

[0036] S140, DUSP11 expression and methylation status, mainly collect current and historical data, which will be mainly used for subsequent research on the association between DUSP11 expression and methylation status and depression, melanoma, and strive to find academic breakthroughs.

[0037] S200, Building a Digital Twin Model S210, Data Collection and Integration Comprehensive patient data is collected from multiple sources, including the hospital's electronic medical record system, laboratory test database, and smart wearable device monitoring platform. Including but not limited to: basic patient information (age, gender, genetic background, etc.), disease-related information (melanoma stage, pathological type, treatment history, etc.), depression assessment scale scores (SDS, HAMD, etc.), blood test indicators (DUSP11 expression and methylation status), physiological data collected by smart wearable devices (heart rate, sleep, exercise, etc.), patient social environment information (family relationships, work pressure, etc.), as well as patient personality, hobbies, emotional experience, emotions, etc.

[0038] Clean and preprocess the collected data, remove duplicate, erroneous or incomplete data, and integrate data from different sources to ensure data consistency and accuracy.

[0039] S220, Feature Engineering Extract and construct features from the preprocessed data. For example, extract features such as average heart rate and heart rate variability from heart rate data; extract total sleep time and deep sleep ratio from sleep data; and encode the patient's treatment plan into a quantifiable feature vector.

[0040] Standardize or normalize continuous features to make different features comparable in value, which is convenient for model learning.

[0041] S230, Model Selection and Architecture Design Consider using a combination of convolutional neural networks (CNN) and long short-term memory networks (LSTM) in deep learning. CNN is used to automatically extract local features in the data, and LSTM is used to process time series data and capture long-term dependencies.

[0042] The input layer receives the multi-dimensional feature vector after feature engineering processing. The middle layer consists of multiple convolutional layers and LSTM layers for feature learning and pattern recognition. The output layer is set according to specific tasks, such as predicting changes in depression status and treatment effects.

[0043] S240, model training The dataset is divided into training set, validation set and test set. The training set is used for model learning, the validation set is used to adjust hyperparameters and monitor the model training process, and the test set is used to finally evaluate the model performance.

[0044] Select a suitable optimization algorithm (such as Adam optimizer) and loss function (such as mean square error, cross entropy, etc.), and iteratively train the model based on the training data.

[0045] During the training process, the early stopping method is used to stop training when the loss on the validation set no longer decreases significantly to prevent overfitting.

[0046] S250, Hyperparameter Tuning Tune the model's hyperparameters, such as convolution kernel size, number of layers, number of LSTM units, learning rate, etc. You can use grid search, random search, or model-based hyperparameter search methods (such as Hyperopt) to find the optimal hyperparameter combination.

[0047] S260, Model Evaluation and Validation The trained model is evaluated using the test set, and a variety of evaluation indicators, such as accuracy, recall, F1 value, root mean square error (RMSE), are used to comprehensively evaluate the performance of the model.

[0048] Perform visual analysis, such as plotting a comparison chart of predicted values ​​and true values, viewing the model's attention weight distribution, etc., to gain an in-depth understanding of the model's decision-making process and potential problems.

[0049] S270, Model Fusion and Integration Try to fuse multiple trained models, such as using random forest, Adaboost and other methods in ensemble learning, to combine the prediction results of multiple models to improve the stability and generalization ability of the model.

[0050] Combined with a rule-based expert system, prior knowledge and experience in the clinical field are converted into rules and integrated with the model's prediction results to enhance the reliability and interpretability of the model.

[0051] S280, Model Deployment and Update Deploy the trained digital twin model to the clinical application system to achieve real-time data input and predictive output.

[0052] Establish a monitoring mechanism for the model and regularly collect new data to retrain and update the model to adapt to changes in patient conditions and treatment methods, ensuring the accuracy and effectiveness of the model.

[0053] Of course, other methods can also be used to establish a digital twin model, as long as it can achieve accurate simulation of the patient.

[0054] S300, Nursing Program Planning S310, recommended solutions S311. Compare the patient data categories and corresponding values ​​with existing cases. First, find the cases with the most identical or similar data categories (parameter names, such as height, weight, etc.), such as 20. Then compare the parameters of similar or identical categories. Parameters within a certain range are identified as identical data items, such as parameter differences within 20%. Sort by the number of identical data items, select the first three, and use these three cases as references. The nursing plans recorded in them can be used as the basic plan for nursing plan planning, and subsequent users can improve on this plan.

[0055] Of course, you can weight the categories and parameters, and finally rank them by weighted scores, and just choose the top three items. The specific weighting method is determined based on actual needs, and can be obtained through a limited number of experiments or set manually.

[0056] S312, select the corresponding cases according to the set conditions, such as selecting by tumor characteristics first, treatment method first, etc. It is still the same to first select the corresponding candidate cases according to the procedure of S311, and then perform conditional screening in these cases. Of course, it is also possible to score and screen the different data categories of each case in advance, sort them by the selected data category scores, and select the first three as recommendations.

[0057] S313, nursing plan planning conditions can be set in the system, such as nursing plans for different levels of pain, nursing plans for different depression states, etc., and then matched according to the patient's condition, and finally combined to output the corresponding recommended nursing plan. This step and S311 can be set in parallel or in sequence or only one of them can be set.

[0058] S320, Digital Twin Model Simulation S321. Input the selected nursing plan into the digital twin model for simulation, and output evaluation conclusions, precautions, etc.

[0059] S322. Based on the simulation results, output the deficiencies or defects of the solution or areas for improvement, and provide optimization recommendations.

[0060] S323. Predict each stage after the implementation of the nursing plan and output the changes in the predictive indicators of DUSP11 expression and methylation status, mainly to determine whether it is abnormal.

[0061] DUSP11 expression index: The expression of DUSP11 is obtained through blood testing and recorded as DUSP11_E. Based on the average value of the normal population, the upper and lower floating thresholds are set, such as ±30%. When DUSP11_E exceeds this range, it is considered abnormal.

[0062] DUSP11 methylation status indicator: Detect the methylation level of DUSP11 gene, recorded as DUSP11_M. Also based on the normal mean, set a threshold, such as ±20%, and it is considered abnormal if it exceeds this range.

[0063] S324, specific simulation process S324.1 Digital coding of nursing plans The various care plans are digitally coded so that the model can understand and process them. For example, psychological care can be coded as "01", exercise care as "02", dietary care as "03", etc. The specific details of each care plan, such as the frequency of psychological care, the intensity and type of exercise care, and the specific recipes of dietary care, are also subdivided and coded accordingly.

[0064] S324.2 Establishing nursing program parameter library Create a database containing various parameters of the care program. These parameters can include the duration, intensity, frequency, stage of implementation, etc. For example, for psychological care, the parameters can be the number of consultations per week and the duration of each consultation; for exercise care, the parameters can be the type of exercise per day (such as walking, yoga, aerobics), exercise duration, and intensity level.

[0065] S324.3 model input interface design Design a dedicated input interface in the digital twin model to ensure that the coding and parameters of the nursing plan can be accurately and efficiently input into the model. The input interface should have data verification and error correction functions to prevent erroneous or incomplete data from entering the model and affecting the accuracy of the simulation results.

[0066] S324.4 Simulation scene setting The initial conditions of the simulation are set according to the patient's current status and historical data. This includes the patient's disease severity, depression assessment score, physiological indicators, treatment progress, etc. The simulation scenario should also consider possible external factors, such as seasonal changes, social environment, etc., which may have a potential impact on the patient.

[0067] S324.5 Nursing plan input and model operation The digitally encoded nursing plan and its parameters are input into the digital twin model through the input interface. After receiving the nursing plan, the model combines the initial conditions and the built-in algorithm logic to simulate the changes in the patient's physiological and psychological state after receiving the nursing plan.

[0068] S324.6 Determination of evaluation indicators for simulation results A series of evaluation indicators are defined to measure the effectiveness of the simulation results and the potential effect of the nursing plan. These indicators can include the degree of improvement of depressive symptoms (reflected by the change in the depression assessment scale score), the recovery of physiological indicators (such as improvement in heart rate, blood pressure, sleep quality, etc.), and the change trend of DUSP11 expression and methylation status.

[0069] S324.7 Multi-round simulation and optimization In order to obtain more reliable and comprehensive simulation results, multiple rounds of simulation can be performed on the same nursing plan, and the input parameters can be fine-tuned each time to observe the changing trend of the results. At the same time, different nursing plans can also be combined and input to explore the best nursing plan combination.

[0070] S324.8 Result analysis and decision support Conduct detailed analysis of the large amount of data generated by the simulation. Compare the simulation results of different nursing plans or combinations of plans to evaluate their effectiveness in improving patients' depression and promoting recovery. Based on the analysis results, provide decision support for medical staff, recommend the best nursing plan or make suggestions for improving existing plans.

[0071] S330, expert evaluation, through experts to evaluate and improve the current nursing plan, or through experts to improve the nursing plan of previous cases, through these improvements to optimize the nursing plan, and continuously improve the nursing plan.

[0072] S340, dynamic adjustment, including system adjustment, weight adjustment, and manual adjustment. System adjustment mainly involves the database constantly updating data. Once a better nursing plan is found, it will be recommended. After the user accepts it, the nursing plan can be optimized and adjusted. Weight adjustment mainly involves assigning weights to each data category and the corresponding parameters. In practice, the weights are adjusted according to the actual situation. After adjustment, the recommended plan or the recommended optimized plan is output. Users can adjust the nursing plan according to the recommended information. Manual adjustment mainly involves medical staff optimizing and correcting the nursing plan during its execution, such as manually adjusting the nursing plan based on patient feedback, nursing effects, and treatment effects.

[0073] S400, Implementation and evaluation of nursing plans S410. Provide care to patients according to the nursing plan and evaluate the effectiveness of the care regularly or irregularly.

[0074] S411. After a period of nursing (e.g., 2 weeks), re-evaluate the comprehensive depression evaluation index Y. Set the nursing effect improvement index P = ((Y_Pre-Y_Post) / Y_Pre) × 100%, where Y_Pre is the comprehensive depression evaluation index before nursing and Y_Post is the depression evaluation index after nursing. When P ≥ 30%, the current nursing method is considered to be effective and should be continued; when P < 30%, the nursing method should be changed. The evaluation methods for melanoma, DUSP11 expression, and methylation status monitoring indicators are the same as the above scheme.

[0075] S412, immediate intervention and recording When the monitoring data of the smart wearable device indicates that immediate intervention is needed, the smart wearable device automatically starts functions such as music playback, digital human interaction, and remote chat.

[0076] After each intervention, the patient's immediate response and the changes in smart wearable device parameters within the next 1-2 hours were recorded. For example, whether the heart rate returned to the normal range after the intervention, whether the number of sleep interruptions decreased, etc., and these changes were parameterized.

[0077] S420, Nursing plan optimization and adjustment Various existing nursing plans are input into the digital twin model, and the model gives scores based on the predicted results. Plans with scores higher than 80 points are listed as preferred plans. Weights are assigned to each data category and parameter in advance, and scores are calculated based on the weights.

[0078] After implementing the optimal plan, continuously monitor the patient's status and indicator changes. If the actual effect does not meet expectations, analyze the reasons and adjust the model parameters and nursing plan.

[0079] Intelligent terminal assisted treatment and nursing Smart wearable devices are connected to the patient's mobile phones and other smart terminals, and push videos, music, conversations and other content through smart terminals to help patients relieve depression.

[0080] Association between S500, DUSP11 and depression The relationship between DUSP11, methylation and depression can be established through regression statistical methods.

[0081] S510. Data collection: Collect data on DUSP11 expression and methylation levels, as well as corresponding depression assessment scale scores (such as SDS or HAMD scores) from a large number of patients with depression and malignant melanoma.

[0082] S520. Variable definition: DUSP11 expression and methylation level are set as independent variables, and depression assessment scale score is set as dependent variable.

[0083] S530. Select a regression model: You can first consider using a simple linear regression model. If the data presents a nonlinear relationship, you can try using polynomial regression, logistic regression, or a more complex nonlinear regression model, such as the generalized additive model (GAM).

[0084] S540, Model Fitting: Fit the selected regression model using the collected data.

[0085] S550, Model evaluation: The goodness of fit of the model is judged by evaluation indicators such as R-squared value, mean square error (MSE), mean absolute error (MAE), etc. Residual analysis can also be performed to check whether the assumptions of the model are met.

[0086] S560, Model Diagnosis: Check whether there are problems such as multicollinearity, outliers, high leverage points, etc.

[0087] S570, Model optimization: If there are problems with the model, try to optimize it by changing variables, adding or deleting variables, selecting a different model, etc.

[0088] S580 Final model determination: Select the optimal model as the final relationship model.

[0089] For example, when using a linear regression model, you might end up with an equation of the following form: Depression score = β0 + β1 × DUSP11 expression + β2 × DUSP11 methylation level + ε; Among them, β0 is the intercept, β1 and β2 are regression coefficients, and ε is the error term.

[0090] You can also use the following methods: Assuming that the depression level (Y) is negatively correlated with the expression of DUSP11 (X1) and positively correlated with the methylation level of DUSP11 (X2), we can construct the following function: Y=a×X1+b×X2+c; a and b are correlation coefficients, representing the influence weights of DUSP11 expression and methylation level on depression level. c is a constant term.

[0091] In order to determine the values ​​of a, b, and c, it is necessary to collect DUSP11 expression levels, methylation levels, and corresponding depression assessment data from a large number of patients, and then perform fitting calculations using statistical methods (such as linear regression analysis).

[0092] For example, by analyzing a set of sample data, the following function may be obtained: Y=-0.5×X1+0.3×X2+20; this means that when the DUSP11 expression level increases by one unit, the depression level may decrease by 0.5 units; when the DUSP11 methylation level increases by one unit, the depression level may increase by 0.3 units. The constant term 20 represents the baseline depression level when both the DUSP11 expression level and the methylation level are 0.

[0093] Preferably, in S110, the method of inferring the suspected depression state and evaluating the effect after nursing is as follows: S111, index normalization processing Heart rate (HR): The normal heart rate range is generally 60-100 beats / minute. Set the normalized heart rate NHR = (HR-60) / (100-60).

[0094] Sleep duration (SL): Assuming that the normal sleep duration for adults is 7-9 hours, normalized sleep duration NSL = (SL-7) / (9-7).

[0095] Sleep interruptions (SI): Assume that the normal number of sleep interruptions is less than 2. Normalized sleep interruptions NSI = SI / 2 (the more the number, the larger the value).

[0096] Sports steps (SS): Assuming that a normal healthy adult takes 5000-10000 steps per day, the normalized sports steps NSS = (SS-5000) / (10000-5000).

[0097] Exercise duration (ST): Assuming that the normal exercise duration is 30-60 minutes per day, normalized exercise duration NST = (ST-30) / (60-30).

[0098] Skin Conductivity (SC): Assuming normal skin conductivity ranges from 10-20 microSiemens. Normalized Skin Conductivity NSC = (SC-10) / (20-10).

[0099] Blood pressure: SBP is systolic pressure, DBP is diastolic pressure, the normal range of systolic pressure is 90-140 mmHg, the normal range of diastolic pressure is 60-90 mmHg. Normalized systolic pressure NSBP = (SBP-90) / (140-90), normalized diastolic pressure NDBP = (DBP-60) / (90-60).

[0100] ECG (assuming the frequency of abnormal ECG indicators is AECG): the normal range is assumed to be 0-10%. Normalized ECG abnormal indicator frequency NAECG = AECG / 10%.

[0101] Respiratory rate (RF): The normal respiratory rate range is 12-20 times / minute. Normalized respiratory rate NRF = (RF-12) / (20-12).

[0102] S112. Construct a formula for estimating depression status (D represents depression status index) D=ω1×NHR+ω2×NSL+ω3×NSI+ω4×NSS+ω5×NST+ω6×NSC+ω7×NSBP+ω8×NDBP+ω9×NAECG+ω10×NRF, where ω1, ω2, ω3, ω4, ω5, ω6, ω7, ω8, ω9, and ω10 are the weight coefficients corresponding to each indicator.

[0103] These weight coefficients need to be determined through a large amount of clinical data and statistical analysis. For example, based on some studies showing that sleep has a greater impact on depression, ω2 can be set to 0.2 and ω3 to 0.15; exercise conditions also have a certain correlation, ω4 is set to 0.1, and ω5 is set to 0.1; other indicators such as heart rate ω1=0.1, skin conductivity ω6=0.05, blood pressure related ω7=0.05, ω8=0.05, ECG ω9=0.1, and respiratory rate ω10=0.05.

[0104] S113. Explanation of the relationship between indicators and depression status Heart rate (HR): People with depression may have abnormal heart rates. For example, during an anxious depressive episode, the heart rate may increase, while during a severe depression that causes a decrease in physical function, the heart rate may decrease.

[0105] Sleep duration (SL) and sleep interruptions (SI): Sleep disorders are common symptoms of depression. Shortened sleep duration and increased sleep interruptions are often associated with depressive states. Patients may have difficulty falling asleep and waking up early, which can lead to insufficient sleep duration and increased sleep interruptions.

[0106] Number of exercise steps (SS) and exercise duration (ST): Depression may lead to reduced activity in patients. Reduced exercise may in turn aggravate depression. Normal exercise can promote the secretion of neurotransmitters and improve emotional state.

[0107] Skin conductivity (SC): Skin conductivity is related to the activity of the autonomic nervous system. During depression-related mood changes (such as anxiety and tension), the autonomic nervous system is excited, which may lead to an increase in skin conductivity.

[0108] Blood pressure and ECG: Long-term depression may affect the cardiovascular system. Blood pressure may fluctuate, and the frequency of abnormal ECG indicators may increase. For example, depression may lead to sympathetic nerve excitement, increase blood pressure, and may also cause changes in cardiac electrical activity, such as premature beats and T wave changes.

[0109] Respiratory rate (RF): People with depression may experience changes in their respiratory rate. For example, when they are excited or anxious, they may hyperventilate; when they are depressed, their breathing may become slow and shallow.

[0110] S114. Determine the calibration value of the depression state After a large number of clinical sample tests (assuming the sample size is n), data from known depressed patients and non-depressed patients were collected to calculate the average value D of the depression state index D of the depressed patients. depressed and standard deviation S depressed , and the average value of the depression state index D of non-depressed patients non-depressed and standard deviation S non-depressed .

[0111] Set a threshold value D for depression threshold , can be determined by: D threshold =D non-depressed +k×S non-depressed Where k is a coefficient determined according to the actual situation, usually selected through statistical analysis and weighing the false positive rate. For example, if you want to minimize false negatives (i.e. misclassifying depressed patients as non-depressed patients), k can take a smaller value, such as 1.5; if you want to balance false positives and false negatives, k can take a value of about 2.

[0112] Assume that after actual testing D non-depressed =0.3, S non-depressed =0.1, if k=2, then D threshold =0.3+2×0.1+0.5. That is, when the calculated depression state index D≥0.5, it is preliminarily judged that the person may be in a depressive state.

[0113] S115. Continuous verification and adjustment This calibration value needs to be continuously verified and adjusted in subsequent research and practice. With the increase of sample size, in-depth research and a deeper understanding of the physiological mechanism of depression, the weight coefficient and calibration value may need to be changed to improve the accuracy of the inference of depression state.

[0114] Nursing intervention can be playing a piece of music for music therapy, or playing a pre-recorded voice to encourage, or medical staff, family members, etc. actively calling to guide, or sending signals from smart wearable devices to smart wearable devices, and smart wearable devices push some stress-relieving games to smartphones to divert patients' attention, etc. After the nursing intervention is completed, the heart rate, sleep duration, number of sleep interruptions, number of exercise steps, exercise duration, skin conductivity and other data are recollected for re-evaluation. If the D value drops by 0.01 or more, it is judged that the nursing intervention is effective and registered in the system; otherwise, it is recorded as ineffective and included in the system to facilitate the subsequent improvement of the psychological care plan.

[0115] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by technicians in the field to which this application belongs.

[0116] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. The nursing intervention system for patients with clinical depression and melanoma is characterized by: include: Information collection module, used to collect patient information as well as existing case information, existing treatment and nursing information; The digital twin model generates a digital model based on the patient's physical data, historical cases, DUSP11 data, depression history, and treatment plan. The digital twin model is then combined with existing treatment and nursing plans to recommend treatment and nursing plans suitable for the patient. Different nursing plans are also evaluated and simulated to predict subsequent results. Nursing planning module, used to generate nursing plans and optimize them in combination with treatment plans; Nursing implementation and evaluation module, which is used to monitor the implementation of nursing plans and conduct evaluations, and continuously optimize and adjust nursing plans based on the evaluation and implementation status; Wearable devices are hardware worn on patients. They collect the patient's heart rate, electrocardiogram, respiratory rate, blood pressure, skin resistance, sleep conditions, and exercise conditions, and combine these parameters to analyze whether depression is likely to exist. If so, nursing access is performed according to the preset program to achieve psychological care.

2. The nursing intervention system for clinically depressed patients with melanoma according to claim 1, characterized in that: The digital twin model includes: The patient's basic physiological information, including the patient's physical information and medical history obtained through patient examination and pathological information. The physical information includes height, weight, and IBM disability status. The medical history mainly includes other diseases that the patient has or has had, as well as the patient's personality, interests, hobbies, and specialties; Melanoma conditions, including tumor characteristics, treatment process, and medical history. Tumor characteristics include tumor location, size, morphology, pathological type, and genetic characteristics. The treatment process includes treatment methods and treatment response. The genetic characteristics are an analysis of genes related to melanoma. Depression-related parameters, including psychological assessment scores, indicators related to emotional state, social support, and depression history; DUSP11 status, including DUSP11 historical status, DUSP11 current status, and the relationship between DUSP11 data and depression and melanoma status. The current and past DUSP11 expression and methylation status of the patient are collected, and these data are associated with depression and melanoma status according to the proposed relationship function. Nursing plan recommendation: compare the collected information with similar nursing plan cases in the database, and score the similar cases. The nursing plan with the highest score will be the recommended nursing plan. Nursing simulation: simulate the nursing plan based on the recommended nursing plan combined with similar cases, nursing results of similar cases, and the simulation model set by the system, and give corresponding nursing result predictions, optimization suggestions, and precautions; First, as much patient information as possible is collected, and then a digital twin model is established. Through the patient information and the digital twin model, similar cases are found in the database, and then evaluated and recommended nursing plans are output. The user confirms, modifies, and optimizes the recommended nursing plan into a proposed nursing plan, and then inputs the proposed nursing plan into the digital twin model for simulation. The digital twin model outputs prediction results, precautions, and optimization suggestions. The user makes adjustments based on the results of the digital twin model until a satisfactory nursing plan is obtained; then nursing begins, and patient data, nursing data, nursing effects, and feedback from medical staff and patients are collected regularly or irregularly to monitor the implementation of the nursing plan; throughout the process, patient information is collected through wearable devices, and depression is judged based on this information. If depression is likely, timely intervention is made according to the system settings.

3. Nursing intervention method for patients with clinical depression and melanoma, characterized by: The steps include: S100, collecting the patient's physical information, physiological data, exercise status, personality and emotional information, historical cases, DUSP11 expression and methylation status information; S200, establish a digital twin model, recommend similar plans based on patient information and the digital twin model, and output a nursing plan; S300, nursing plan planning, adjust the nursing plan output by the system, input it into the digital twin model for simulation and evaluation, and improve the nursing plan based on the evaluation; S400, implementation and evaluation of nursing plans. Implement nursing care according to the nursing plan, and evaluate the effectiveness of the nursing plan based on the feedback from medical staff and patients to continuously optimize the nursing plan.

4. The nursing intervention method for clinically depressed patients with melanoma according to claim 3, characterized in that: Also included in the S100: S110, collecting patient information, including collecting the patient's height, weight, IBM, and disability information, and using smart wearable devices to detect the patient's physiological data and exercise conditions, and inferring whether the patient is in a suspected depression or suspected of having depression through these data. If so, depression nursing intervention is performed according to a pre-set nursing plan, and then the improvement of depression is judged in combination with the parameters collected subsequently; the physiological data includes blood pressure, heart rate, electrocardiogram, respiratory rate, skin resistance, and sleep conditions; S120, collecting the patient's personality and emotional information, including depression, personality, hobbies, specialties, and emotional experiences; at the same time, collecting as much historical information as possible, including historical information on depression, personality, hobbies, specialties, and emotional experiences; S130. Historical cases are to collect some historical case information of patients before they visit the hospital, including melanoma tumor information, depression status, and other diseases. Melanoma tumor information includes historical and current tumor characteristics, treatment methods, treatment response, and treatment effect; tumor characteristics include tumor location, size, morphology, pathological type, and genetic characteristics. Genetic characteristics include BRAF mutation, NRAS mutation, c-KIT mutation, DUSP11 expression, and methylation status; other diseases include current and historical illnesses, treatment plans for the disease, and recovery status; The information on S140 and DUSP11 expression and methylation status is the collection of current and historical data.

5. The nursing intervention method for clinically depressed patients with melanoma according to claim 4, characterized in that: In S110, the suspected depressive state and the evaluation method after care are as follows: S111, index normalization processing Heart rate (HR): set normalized heart rate NHR = (HR-60) / (100-60); Sleep duration (SL): Normalized sleep duration NSL = (SL-7) / (9-7); Sleep interruption count (SI): Normalized sleep interruption count NSI = SI / 2 (the more the number, the greater the value); Movement steps (SS): Normalized movement steps NSS = (SS-5000) / (10000-5000); Exercise duration (ST): Normalized exercise duration NST = (ST-30) / (60-30); Skin conductivity (SC): Normalized skin conductivity NSC = (SC-10) / (20-10); Blood pressure: systolic blood pressure is SBP, diastolic blood pressure is DBP, normalized systolic blood pressure NSBP=(SBP-90) / (140-90), normalized diastolic blood pressure NDBP=(DBP-60) / (90-60); ECG: Assume that the frequency of abnormal ECG index is AECG, and the normalized frequency of abnormal ECG index is NAECG=AECG / 10%; Respiratory frequency (RF): Normalized respiratory frequency NRF = (RF-12) / (20-12); S112. Construct a formula for estimating depression status (D represents depression status index) D=ω1×NHR+ω2×NSL+ω3×NSI+ω4×NSS+ω5×NST+ω6×NSC+ω7×NSBP+ω8×NDBP+ω9×NAECG+ω10×NRF, where ω1, ω2, ω3, ω4, ω5, ω6, ω7, ω8, ω9, ω10 are the weight coefficients corresponding to each indicator; S113. Determine the calibration value of the depression state After a large number of clinical sample tests, data from known depressed patients and non-depressed patients were collected to calculate the average value of the depression state index D of the depressed patients. depressed and standard deviation S depressed , and the average value of the depression state index D of non-depressed patients non-depressed and standard deviation S non-depressed ; Set a threshold value D for depression threshold , determined by: D threshold =D non-depressed +k×S non-depressed Where k is a coefficient determined according to actual conditions, and is usually selected through statistical analysis and weighing the misjudgment rate.

6. The nursing intervention method for clinically depressed patients with melanoma according to claim 3, characterized in that: Also included in the S200: S210, Data Collection and Integration Collect comprehensive patient data from multiple sources including the hospital's electronic medical record system, laboratory test database, and smart wearable device monitoring platform; including but not limited to: basic patient information, disease-related information, depression assessment scale scores, blood test indicators, DUSP11 expression and methylation status, physiological data collected by smart wearable devices, patient social environment information, and patient personality, hobbies, emotional experiences, and emotions; Clean and pre-process the collected data, remove duplicate, erroneous or incomplete data, and integrate data from different sources to ensure data consistency and accuracy; S220, Feature Engineering Extract and construct features from preprocessed data, and standardize or normalize continuous features to make different features comparable in value, which is convenient for model learning; S230, Model Selection and Architecture Design The architecture combines the convolutional neural network (CNN) and the long short-term memory network (LSTM) in deep learning. CNN is used to automatically extract local features in the data, and LSTM is used to process time series data and capture long-term dependencies. The input layer receives the multi-dimensional feature vector after feature engineering. The middle layer consists of multiple convolutional layers and LSTM layers for feature learning and pattern recognition. The output layer is set according to the specific task. S240, model training Divide the dataset into training set, validation set and test set; the training set is used for model learning, the validation set is used to adjust hyperparameters and monitor the model training process, and the test set is used to finally evaluate the model performance; Select appropriate optimization algorithms and loss functions, and iteratively train the model based on the training data; During the training process, the early stopping method is used to stop training when the loss on the validation set no longer decreases significantly to prevent overfitting; S250, Hyperparameter Tuning Tune the model’s hyperparameters; S260, Model Evaluation and Validation Use the test set to evaluate the trained model and use a variety of evaluation indicators to comprehensively evaluate the performance of the model; Perform visual analysis, such as plotting a comparison chart of predicted values ​​and true values, viewing the model's attention weight distribution, and gaining an in-depth understanding of the model's decision-making process and potential problems; S270, Model Fusion and Integration Try to fuse multiple trained models and combine the prediction results of multiple models to improve the stability and generalization ability of the model; Combined with a rule-based expert system, prior knowledge and experience in the clinical field are converted into rules and integrated with the model's prediction results to enhance the reliability and interpretability of the model; S280, Model Deployment and Update Deploy the trained digital twin model to the clinical application system to achieve real-time data input and prediction output; Establish a monitoring mechanism for the model and regularly collect new data to retrain and update the model to adapt to changes in patient conditions and treatment methods, ensuring the accuracy and effectiveness of the model.

7. The nursing intervention method for clinically depressed patients with melanoma according to claim 3, characterized in that: The S300 also includes: S310, recommended solutions S311. Compare the patient data categories and corresponding values ​​with existing cases, first find the cases with the most identical or similar data categories; then compare the parameters of similar or identical categories, and identify the data items with parameters within a certain range as identical data items; sort by the number of identical data items, select the first few cases as references, and use the nursing plans recorded in them as the basic plan for nursing plan planning, and subsequent users can improve on this plan; You can also weight the categories and parameters, and finally rank them by weighted scores, and just select the top few items; S312. Filter corresponding cases according to the set conditions, score different data categories of each case in advance, sort them according to the selected data category scores during the screening, and select the top few as recommendations; S313, setting nursing plan planning conditions in the system, then matching according to the patient's condition, and finally combining and outputting the corresponding recommended nursing plan; S320, Digital Twin Model Simulation S321. Input the selected nursing plan into the digital twin model for simulation, and output the evaluation conclusion and precautions; S322. Based on the simulation results, output the deficiencies or defects of the solution or areas for improvement, and provide optimization recommendations; S323, predicting each stage after the implementation of the nursing plan, and outputting the judgment of abnormality of the prediction indicators of DUSP11 expression and methylation status; DUSP11 expression index: The expression of DUSP11 is obtained through blood testing, recorded as DUSP11_E. The upper and lower floating thresholds are set based on the average value of the normal population. When DUSP11_E exceeds this range, it is considered abnormal; DUSP11 methylation status indicator: The methylation degree of the DUSP11 gene is detected, recorded as DUSP11_M, and the threshold is set based on the normal mean. If it exceeds this range, it is considered abnormal.

8. The nursing intervention method for clinically depressed patients with melanoma according to claim 7, characterized in that: The S320 also includes: S324, specific simulation process S324.1 Digital coding of nursing plans Digitally encode various care plans so that the model can understand and process them, and also encode the specific details of each care plan accordingly; S324.2 Establishing nursing program parameter library Create a database containing various care regimen parameters; S324.3 model input interface design Design a dedicated input interface in the digital twin model to ensure that the coding and parameters of the nursing plan can be accurately and efficiently input into the model. The input interface should have data verification and error correction functions to prevent erroneous or incomplete data from entering the model and affecting the accuracy of the simulation results. S324.4 Simulation scene setting Set the initial conditions of the simulation based on the patient’s current status and historical data; S324.5 Nursing plan input and model operation The digitally encoded nursing plan and its parameters are input into the digital twin model through the input interface. After receiving the nursing plan, the model combines the initial conditions and the built-in algorithm logic to start simulating the changes in the patient's physiological and psychological state after receiving the nursing plan. S324.6 Determination of evaluation indicators for simulation results Define a series of evaluation indicators to measure the effectiveness of the simulation results and the potential effects of the nursing plan, including the degree of improvement of depressive symptoms, the recovery of physiological indicators, and the changing trends of DUSP11 expression and methylation status; S324.7 Multi-round simulation and optimization Conduct multiple rounds of simulations on the same nursing plan, fine-tune the input parameters each time, and observe the changing trend of the results; at the same time, combine and input different nursing plans to explore the best nursing plan combination; S324.8 Result analysis and decision support Conduct detailed analysis on the large amount of data generated by the simulation, compare the simulation results of different nursing plans or combinations of plans, and evaluate their effectiveness in improving patients' depression and promoting recovery; based on the analysis results, provide decision support to medical staff, recommend the best nursing plan or make suggestions for improving existing plans; S330, expert evaluation, through experts to evaluate and improve the current nursing plan, or through experts to improve the nursing plan of previous cases, through these improvements to optimize the nursing plan, and continuously improve the nursing plan; S340, dynamic adjustment, dynamically optimize and adjust the nursing plan according to system data update, weight adjustment, and manual adjustment.

9. The nursing intervention method for clinically depressed patients with melanoma according to claim 3, characterized in that: The S400 also includes: S410. Provide nursing care to patients according to the nursing plan and evaluate the nursing effect regularly or irregularly; S411. After a period of nursing, re-evaluate the monitoring indicators of melanoma, DUSP11 expression and methylation status, and the comprehensive depression evaluation index Y; set the nursing effect improvement index P = ((Y_Pre-Y_Post) / Y_Pre) × 100%, where Y_Pre is the comprehensive depression evaluation index before nursing, and Y_Post is the index after nursing. When P ≥ 30%, the current nursing method is considered to be effective and should be continued; when P < 30%, the nursing method should be changed; S412, immediate intervention and recording When the monitoring data of the smart wearable device indicates that immediate intervention is needed, the smart wearable device intervenes according to the program; after each intervention, the patient's immediate response and the changes in the parameters of the smart wearable device over a period of time are recorded, and these changes are parameterized; S420, Nursing plan optimization and adjustment Various existing nursing plans are input into the digital twin model, and the model gives scores based on the predicted effects. Plans with scores higher than 80 points are listed as preferred plans. After the preferred plan is implemented, the patient's status and indicator changes are continuously monitored. If the actual effect does not meet expectations, the reasons are analyzed and the model parameters and nursing plans are adjusted. S430, intelligent terminal assisted treatment and nursing Smart wearable devices are connected to the patient’s mobile phone, and timely nursing intervention is provided through smart terminals to help patients relieve depression.

10. The nursing intervention method for clinically depressed patients with melanoma according to claim 3, characterized in that: Also includes: The association between S500, DUSP11 and depression, the relationship between DUSP11 and methylation and depression was established by regression statistics; S510. Data collection: Collect data on DUSP11 expression and methylation level of a large number of patients with depression and malignant melanoma, as well as the corresponding depression assessment scale scores; S520, variable definition: DUSP11 expression and methylation level were set as independent variables, and depression assessment scale score was set as dependent variable; S530, select regression model: first consider using a simple linear regression model. If the data presents a nonlinear relationship, try using polynomial regression, logistic regression or a more complex nonlinear regression model. S540, Model Fitting: Fit the selected regression model using the collected data; S550, Model evaluation: Use evaluation indicators to judge the goodness of fit of the model; you can also perform residual analysis to check whether the assumptions of the model are met; S560, Model Diagnosis: Check whether there are multicollinearity, outliers, and high leverage points; S570, model optimization: If there are problems with the model, try to optimize it by changing variables, adding or deleting variables, or selecting a different model; S580 Final model determination: Select the optimal model as the final relationship model.

Images