A method and system for predicting the burden risk of a family caregiver of a minor stroke patient

Abstract

The application discloses a kind of small stroke patient family caregiver burden risk prediction method and system, comprising: constructing research cohort and collecting the baseline clinical data of patient acute phase and the baseline information of family caregiver;Through statistical analysis, the independent correlation of the prediction variable of caregiver medium-severe burden is screened out;Based on the variable and its regression coefficient screened out, a nomogram prediction model is constructed, the score of each variable value is converted, the individual risk probability is predicted by accumulating total score;And ROC curve and Bootstrap resampling method are used to verify the model.The system includes data input, prediction calculation and result output module, and can be integrated into clinical information system.The application can realize individualization, quantification early warning of future burden risk of family caregiver of small stroke patient in acute phase, provide key decision support for implementing precise nursing intervention, effectively improve nursing resource efficiency and family care quality.

Description

Technical Field

[0001] This invention relates to the field of medical and health management technology, specifically to a method and system for predicting the risk of burden on family caregivers of minor stroke patients. Background Technology

[0002] Stroke is the leading cause of death and disability among Chinese residents, with ischemic stroke accounting for approximately 85%. In recent years, with the significant improvement in the construction of stroke centers and the level of acute-phase treatment, the proportion of disabling strokes has been declining year by year. However, minor strokes (also known as non-disabling ischemic cerebrovascular events, referring to ischemic events with a National Institutes of Health Stroke Scale score ≤5 and mild symptoms) have become the most common type of stroke, accounting for more than 46.4% of ischemic stroke cases. Minor strokes are not "benign" diseases; they have a high risk of recurrence and often leave behind non-limb functional symptoms such as cognitive impairment, emotional disorders (such as depression and anxiety), and post-stroke fatigue. These symptoms are insidious and easily overlooked, but they seriously impair patients' quality of life and long-term prognosis, thus significantly increasing their long-term care needs.

[0003] Against this backdrop, approximately 90% of minor stroke patients return home after discharge, with their care primarily falling to family caregivers. Caregivers must adapt to this role change quickly, undertaking complex tasks such as daily care, rehabilitation supervision, emotional support, and medical coordination – a long, multi-dimensional, and stressful process. Domestic and international research indicates that family caregivers generally bear a significant burden during the caregiving process, including physical fatigue, psychological distress (such as anxiety and depression), social limitations, and economic pressure. This burden not only harms the caregivers' own physical and mental health and quality of life but may also indirectly affect the patient's rehabilitation adherence and ultimate outcome.

[0004] However, current research and practice regarding the burden on family caregivers of minor stroke patients have significant limitations: First, existing studies are mostly cross-sectional surveys, focusing on describing the burden at a specific point in time, lacking longitudinal follow-up data on the dynamic changes in burden over time, making it difficult to reveal its evolutionary patterns and the differences in influencing factors at different stages. Second, in clinical practice and nursing management, there is a lack of effective tools for early screening and individualized risk prediction. Healthcare professionals can usually only intervene retrospectively after burden symptoms appear, lacking means to identify high-risk caregivers and provide preventative support before patient discharge or in the early stages of the disease.

[0005] In the field of predictive modeling, nomograms, as a technique that transforms multivariate regression models into intuitive and visual scoring tools, have been widely used in clinical research such as cancer prognosis and chronic disease risk prediction. They can integrate multiple predictive factors and individually predict the probability of a specific event by simply adding the scores of each factor, demonstrating good clinical applicability and interpretability. However, this technique has not yet been applied to the risk prediction of the burden on family caregivers of minor stroke patients. Current nursing assessments mostly rely on general scales for current status scoring, which cannot achieve quantitative prediction of future burden risk. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method and system for predicting the burden risk of family caregivers in minor stroke patients. This method overcomes the deficiencies of existing technologies and enables individualized and quantitative early warning of the future burden risk of family caregivers in the acute phase of minor stroke patients. It provides key decision support for implementing precision nursing interventions and effectively improves the efficiency of nursing resources and the quality of family care.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for predicting the risk of burden on family caregivers of patients with minor stroke includes the following steps:

[0009] S1. Study Cohort Construction and Data Collection: Based on the pre-set inclusion and exclusion criteria, acute-phase minor stroke patients and their primary family caregivers were selected to form the study cohort; during the acute phase of the patients' illness, baseline clinical data of the patients and baseline information of the family caregivers were collected.

[0010] S2. Predictive variable screening: Statistical analysis was performed on the baseline clinical data of the patients and the baseline information of the family caregivers. Univariate analysis and multivariate logistic regression analysis were combined to screen out predictive variables that were independently associated with the occurrence of moderate to severe care burden in the family caregivers within 6-12 months after the onset of the disease.

[0011] S3. Construction of nomogram prediction model: Based on the independent predictors selected in step S2 and their corresponding regression coefficients in multivariate logistic regression analysis, a nomogram prediction model is constructed using statistical calculation software; the nomogram model maps different values ​​of each predictor to corresponding scores, and obtains a total score by accumulating the scores of each item, and maps the total score to the predicted probability of the family caregiver experiencing moderate to severe burden;

[0012] S4. Model Performance Validation and Optimization: The discrimination of the nomogram model was evaluated using receiver operating characteristic (ROC) curve analysis, and the internal validation of the model was performed using the Bootstrap resampling method. The corrected C-index was calculated to evaluate the predictive consistency and stability of the model.

[0013] Preferably, in step S1, the patient's baseline clinical data includes at least: demographic data, education level, cognitive function assessment results, neurological deficit severity score, activities of daily living score, and emotional state score;

[0014] The baseline information of the family caregiver includes at least: demographic data, relationship with the patient, daily care time, monthly family income level, emotional state score, coping style score, and social support score.

[0015] Preferably, the independently related predictor variables in step S2 include at least five of the following:

[0016] (a) The gender of the family caregiver is female;

[0017] (b) The family caregiver had a depressive disorder at baseline;

[0018] (c) The relationship between the family caregiver and the patient is that of spouse;

[0019] (d) The patient had post-stroke cognitive impairment at baseline;

[0020] (e) The patient's educational level is high school or above.

[0021] Preferably, the construction of the nomogram prediction model in step S3 specifically includes:

[0022] Using the rms package in R, the results of the multivariate logistic regression analysis were taken as input, and the nomogram function was used to generate a visual nomogram containing all independent predictor variables; the rating scale in the nomogram was set proportionally according to the regression coefficients of each variable.

[0023] Preferably, the method further includes step S5: model application.

[0024] The constructed and validated nomogram model was applied to new minor stroke patients and their family caregivers. By collecting predictive variable information of the new patients and caregivers in the acute phase, and summing the scores against the nomogram model, the individualized risk probability of the caregiver developing moderate to severe burden in the next 6-12 months was output to guide early clinical intervention.

[0025] Preferably, the application scenario of the model described in step S5 is as follows: before the patient is discharged from the hospital, medical staff perform an assessment to identify caregivers with high burden risk, and develop personalized nursing support plans, health education and psychological intervention programs accordingly.

[0026] This invention also discloses a risk prediction system for the burden on family caregivers of minor stroke patients, comprising:

[0027] Data input interface module: used to receive characteristic data of target minor stroke patients and their family caregivers during the acute phase. The characteristic data includes: patient's education level, cognitive function status, caregiver's gender, relationship with the patient, and emotional state.

[0028] Data processing and calculation module: Stores a pre-trained nodal plot prediction model, which is constructed and validated based on the above method; the data processing and calculation module is used to call the model to perform calculations based on the input feature data, and output various scores and a total score;

[0029] Risk prediction and output module: used to convert the total score into the corresponding moderate to severe care burden risk probability value, and present the prediction results to the user in at least one of the following forms: numbers, percentages, risk levels or visualization charts.

[0030] Preferably, the system is an electronic evaluation tool, which includes the following forms:

[0031] (a) Software modules integrated into hospital information systems or stroke center management platforms;

[0032] (b) Standalone mobile terminal application;

[0033] (c) Internet-based online risk assessment web platform.

[0034] This invention provides a method and system for predicting the risk of burden on family caregivers of minor stroke patients. It offers the following advantages: By integrating objective clinical indicators of the patient during the acute phase with baseline characteristics of the caregiver, a multi-factor prediction model is constructed. This model can calculate a quantified future risk probability for each specific caregiver, thus advancing the assessment of caregiver burden in minor stroke from a "current situation description" stage to a "risk prediction" stage. This allows for a significant shift in clinical intervention to before patient discharge, achieving a fundamental transformation from passive response to proactive prevention. By converting the regression equation into a visualized nomogram, healthcare professionals can complete a rapid assessment in just a few minutes without complex calculations, with results clearly presented in the form of risk probability and rank. This greatly lowers the barrier to entry, enabling seamless integration into busy clinical workflows and discharge preparation services, significantly improving the efficiency of nursing assessment and management. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the prior art will be briefly introduced below.

[0036] Figure 1 This is a flowchart illustrating the overall method of the present invention;

[0037] Figure 2 This is a flowchart illustrating the construction process of the line graph model in this invention.

[0038] Figure 3 The visualized column chart generated in Embodiment 1 of this invention;

[0039] Figure 4 ROC curve of the nomogram prediction model in Embodiment 1 of the present invention;

[0040] Figure 5 The prediction system architecture diagram of this invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0042] Example 1, as Figures 1 to 4 As shown in the figure, this embodiment provides a method for predicting the risk of burden on family caregivers of minor stroke patients. The core of this method is to transform the patient and caregiver characteristics that are readily available in clinical practice into a nomogram that can be used for individualized risk prediction through scientific statistical modeling.

[0043] Specifically, the following steps are included:

[0044] S1. Study Cohort Construction and Data Collection: Based on pre-defined inclusion and exclusion criteria, acute-phase minor stroke patients and their primary family caregivers were selected to form the study cohort. During the acute phase of the stroke, baseline clinical data of the patients and baseline information of the family caregivers were collected; specifically including:

[0045] S101: Cohort Establishment: Based on pre-defined criteria, acute minor stroke patients and their primary family caregivers who were hospitalized in the Department of Neurology at Shanghai Tenth People's Hospital from August 2019 to June 2020 were consecutively included, forming a total of 248 research cohorts.

[0046] S102: Baseline clinical data collection for patients (acute phase, within 2 weeks of onset): including:

[0047] (1) Population sociology data: records of gender and age.

[0048] (2) Educational level: recorded as "junior high school or below" or "senior high school or above".

[0049] (3) Cognitive function assessment: The Chinese version of the Montreal Cognitive Assessment Scale was used for assessment. A total score of <26 points was defined as having post-stroke cognitive impairment.

[0050] (4) Severity of neurological deficit: assessed using the National Institutes of Health Stroke Scale.

[0051] (5) Daily living activities: assessed using the Barthel Index.

[0052] (6) Emotional state: The Hamilton Depression Rating Scale (HAMD-17) and Hamilton Anxiety Rating Scale were used for assessment.

[0053] S103: Baseline Information Collection of Family Caregivers (Simultaneous): Includes:

[0054] (1) Population sociology data: records of gender and age.

[0055] (2) Relationship with the patient: recorded as spouse, child or other.

[0056] (3) Daily care time: Record the average number of daily care hours.

[0057] (4) Monthly household income level: recorded as <5,000 yuan or ≥5,000 yuan.

[0058] (5) Emotional state: The HAMD-17 was used for assessment, and a total score >7 was defined as having a depressive disorder.

[0059] (6) Coping methods: A simple coping method questionnaire was used to assess positive and negative coping tendencies.

[0060] (7) Social support: The Social Support Rating Scale was used for assessment.

[0061] S104: Outcome Definition and Follow-up: The Caregiver Burden Questionnaire (CBI) was used for follow-up assessment 6-12 months after the onset of illness. A total CBI score >32 was defined as "moderate to severe care burden" and served as a binary outcome event in this predictive model. A total of 225 pairs were followed up effectively.

[0062] S2. Predictive Variable Screening: Statistical analysis was performed on the baseline clinical data of the patients and the baseline information of the family caregivers. A combination of univariate analysis and multivariate logistic regression analysis was used to screen predictive variables independently associated with the occurrence of moderate to severe caregiving burden in the family caregivers within 6-12 months after the onset of illness. Specifically, these included:

[0063] S201: Univariate Analysis: Univariate analysis was performed on all baseline variables of patients and caregivers collected in S1 and the "moderate to severe burden" outcome observed during the follow-up period. For continuous variables, the independent samples t-test was used if they conformed to a normal distribution, and the Mann-Whitney U test was used if they did not. For categorical variables, the chi-square test was used. A p-value < 0.1 was set as statistically significant to initially screen potential predictors.

[0064] S202: Multivariate Logistic Regression Analysis: All variables with P < 0.1 in the univariate analysis were included in the binary logistic regression model. A stepwise forward method was used for variable selection, with P < 0.05 as the inclusion criterion and P > 0.10 as the exclusion criterion, to finally determine the independent predictors.

[0065] S203: Identification of Independent Predictors: Regression analysis ultimately identified five independent predictors of moderate to severe burden at 6-12 months (P < 0.05), namely caregiver gender (female) (OR: 4.362, 95% CI: 1.667-11.412), caregiver depressive disorder (OR: 3.222, 95% CI: 1.210-8.578), spousal relationship (OR: 2.763, 95% CI: 1.223-6.240), post-stroke cognitive impairment in patients (OR: 2.371, 95% CI: 1.228-4.577), and patient education level (OR: 0.406, 95% CI: 0.216-0.763). The OR value being less than 1 here indicates that patients with an educational level of "junior high school or below" are statistically a protective factor against burden risk; that is, compared to patients with an educational level of "high school or above," their caregivers have a lower risk of experiencing moderate to severe burden. In other words, a patient's educational level of "high school or above" is a risk factor. The specific definitions, values, and statistical results of each variable are shown in the table below.

[0066] Table 1. Multivariate analysis of baseline variables and moderate to severe caregiver burden (CBI > 32).

[0067]

[0068] S3. Nodal plot prediction model construction: Based on the independent predictor variables selected in step S2 and their corresponding regression coefficients in multivariate logistic regression analysis, a nodal plot prediction model is constructed using statistical calculation software; the nodal plot model maps different values ​​of each predictor variable to corresponding scores, and obtains a total score by accumulating the scores, mapping the total score to the predicted probability of the family caregiver experiencing moderate to severe burden; specifically including:

[0069] S301: Select software and tools: Use the R language and the rms package for model visualization and construction.

[0070] S302: Model Input: Use the results of the multivariate logistic regression model (containing the above 5 variables and their regression coefficients) obtained in S2 as input.

[0071] S303: Generate Nomogram: Call the nomogram() function in the rms package to generate a visual nomogram based on the input multivariate logistic regression model. The function automatically converts the influence of each variable into a 0-100 "score" axis based on the regression coefficients of each variable. For example, "Careerer Gender: Female" is assigned approximately 100 points. For the patient's education level, since the OR value is less than 1, "Junior High School and below" corresponds to a lower score (protective factor), and "High School and above" corresponds to a higher score (risk factor).

[0072] S304: Establishing a score-risk mapping: The generated nodal plot (see appendix) Figure 3 The graph, from top to bottom, contains five rating axes for the five predictor variables. Each variable receives a score (0-100 points) on the "Points" axis above it, based on its value (e.g., "Yes / No," "Junior High and Below / High School and Above"). The scores for all variables are summed to obtain the "Total Points" score. Finally, the total score is projected onto the "Risk of Moderate-to-Severe Burden" axis at the bottom. By summing the scores for each individual's five variables on the "Total Score" axis, the probability of developing a moderate to severe burden can be found on the "Predicted Risk" axis.

[0073] S4. Model Performance Validation and Optimization: Receiver Operating Characteristic (ROC) curve analysis was used to evaluate the discrimination of the nomogram model, and the Bootstrap resampling method was used for internal model validation. The corrected C-index was calculated to assess the model's predictive consistency and stability. Specifically, this includes:

[0074] S401: Discrimination Assessment: Based on follow-up data from 225 cases, the risk probability predicted by the nomogram model for each caregiver was calculated. Using this probability as the test variable and the actual burden status as the state variable, the receiver operating characteristic (ROC) curve of the nomogram prediction model was plotted, as shown below. Figure 4 As shown. The area under the curve was calculated. In this embodiment, the AUC was 0.843 (95% CI: 0.789-0.897), indicating that the model has good discriminative ability.

[0075] S402: Internal Validation and Calibration:

[0076] Internal validation was performed using the Bootstrap repeated sampling method. The number of resampling operations was set to 1000. After each sampling, the model was refitted on the new samples, and its C-index on the calibration samples was calculated.

[0077] The average of 1000 C-index calculations is used to obtain the corrected C-index. In this embodiment, the corrected C-index is 0.814, which is close to the original C-index (0.843), indicating that the model is less overly optimistic and the prediction performance is stable.

[0078] The calibration curves show a good agreement between the model's predicted probabilities and the actual observed frequencies.

[0079] Through the steps outlined above, a nomogram model incorporating five key predictor variables was successfully constructed and validated. This model effectively utilizes clinical and caregiver information from the acute phase of minor stroke patients to individually predict the risk of moderate to severe caregiving burden on their family caregivers within the next 6-12 months.

[0080] This invention integrates objective clinical indicators of patients in the acute phase (education level, cognitive state) with baseline characteristics of caregivers (gender, relationship, depressive state) to construct a multifactor predictive model. This model can calculate a quantified future risk probability (e.g., 70%) for each specific caregiver, advancing the assessment of caregiver burden in minor stroke from a "current situation description" to a "risk prediction" stage for the first time. This individualized prediction allows for a significant shift in clinical intervention to before patient discharge, achieving a fundamental transformation from passive response to proactive prevention.

[0081] And this is achieved by transforming the regression equation into a visual nomogram. For example... Figure 3 As shown, healthcare professionals do not need to memorize formulas or perform complex calculations. They can complete a risk assessment within 1-2 minutes using only five easily accessible indicators through a simple "lookup-addition-location" process. This significantly lowers the barrier to entry, allowing for seamless integration into busy clinical workflows and discharge preparation services, and substantially improving the efficiency of nursing assessment and management.

[0082] Furthermore, the model of this invention is not based on theoretical assumptions, but rather rooted in rigorous clinical research data. Constructed and validated using scientific statistical methods (Logistic regression, Bootstrap validation), the model demonstrates excellent discriminative ability (AUC = 0.843) and good stability (corrected C-index = 0.814). These quantitative indicators prove that the model has high accuracy and reliability, and its predictive results can provide a solid scientific basis for clinical decision-making, avoiding errors from subjective experience-based judgments. This invention expands its focus from the single dimension of the patient to the dual entity of "patient-caregiver," embodying the core ideas of the modern biopsychosocial medical model. By identifying high-risk caregivers early, nursing staff can implement precise and proactive intervention packages based on the predictive results, such as targeted health education, psychological counseling, skills training, or social resource linkages. This not only effectively prevents or reduces caregiver burnout, improving their caregiving ability and quality of life, but also ultimately, through a stable and healthy family support system, it feeds back into and optimizes the patient's recovery outcome, forming a virtuous cycle with significant social benefits.

[0083] Example 2, as Figure 5 As shown, the present invention also discloses a risk prediction system for the burden on family caregivers of minor stroke patients, comprising:

[0084] Data input interface module: used to receive characteristic data of the target minor stroke patient and their family caregiver during the acute phase. The characteristic data includes: patient's education level, cognitive function status, caregiver's gender, relationship with the patient, and emotional state. The specific implementation of the data input interface module is as follows:

[0085] Designed as an intuitive graphical user interface (GUI), it includes controls such as drop-down menus, radio buttons, or sliders for data collection.

[0086] (1) Patient's educational level: The options are "junior high school or below" or "high school or above".

[0087] (2) Patient cognitive impairment: The options are "Yes (MoCA < 26 points)" or "No (MoCA ≥ 26 points)". The system can provide an interface with the hospital's cognitive assessment system to automatically obtain and judge the MoCA score.

[0088] (3) Caregiver gender: The options are "male" or "female".

[0089] (4) Relationship between caregiver and patient: The options are “spouse”, “children” or other.

[0090] (5) Caregiver depression: The options are "Yes (HAMD>7)" or "No (HAMD≤7)". A simplified depression screening scale (such as PHQ-2) can be integrated for rapid initial screening.

[0091] In this embodiment, the data input interface module has built-in verification logic to ensure that all required fields are filled in and in the correct format before they can be submitted to the data processing and calculation module.

[0092] The data processing and calculation module contains the algorithm logic of the nomogram prediction model constructed and verified in Example 1. It is used to call the model to perform calculations based on the input feature data, and output various scores and a total score; the specific implementation includes:

[0093] (1) Model encapsulation: The scoring rules (i.e., the scores corresponding to each variable value) and the total score-risk probability mapping relationship corresponding to the nodal graph in Example 1 are encoded into a lookup table or mathematical function within the system. For example, a scoring dictionary is established: {"Patient's education level_junior high school and below": 48, "Patient's cognitive impairment_yes": 52, …}.

[0094] (2) Calculation process: After receiving the input data, the module automatically queries the dictionary to obtain the scores of each variable, calculates the total score, and then calculates the final risk probability value (between 0 and 1) based on the preset risk probability transformation curve (based on the Logistic regression equation).

[0095] (3) Technology selection: This module can be implemented based on backend technologies such as Python (e.g., Flask / Django framework), Java or .NET, or the computing logic can be encapsulated as a RESTful API for frontend to call.

[0096] Risk Prediction and Output Module: This module converts the total score generated by the data processing and calculation module into a corresponding probability value for moderate to severe care burden risk, and presents the prediction results to the user in at least one of the following formats: numbers, percentages, risk levels, or visual charts. Specific implementation includes:

[0097] (1) Multiple formats for result display:

[0098] a. Number / Percentage: Directly displays the predicted risk probability, such as "Moderate to severe burden risk: 70%".

[0099] b. Risk level: The risk level is classified according to the preset threshold (e.g., <30% is low risk, 30%-70% is medium risk, and >70% is high risk) and visually indicated by color (green / yellow / red).

[0100] c. Visual charts: Displays a simplified diagram of the line graph, highlighting the current individual's score position and the total score landing point.

[0101] (2) Structured report output: Automatically generate a brief assessment report, including: a summary of input data, the calculated total score, the risk probability / level, and systematic care recommendations based on the risk level (e.g., "High-risk recommendations: strengthen psychological support, provide care skills training, and contact community respite services").

[0102] This system is adaptable to various software and hardware environments, and its specific implementation forms include, but are not limited to, the following three:

[0103] (1) Software modules integrated into hospital information systems or stroke center management platforms

[0104] Application scenarios: As a functional plug-in for electronic medical record systems or stroke clinical pathway management systems.

[0105] Implementation method: By using the development interface provided by the hospital information system, the data input interface of this system is embedded into the "Discharge Assessment" or "Continuing Care" section of the nurse workstation or doctor workstation.

[0106] Data input can be partially automated, such as automatically extracting the patient's education level from the electronic medical record and obtaining the MoCA score from the examination results system. Nurses only need to supplement the caregiver's relevant information.

[0107] The assessment results are automatically saved to the patient database and can generate care plan entries or trigger consultation / referral processes.

[0108] (2) Standalone mobile terminal application

[0109] Application scenarios: For use by community nurses and case managers during home visits, or for caregivers' family members to conduct preliminary self-screening.

[0110] Implementation: Develop iOS and Android versions of the app. The interface is simple and guides users step-by-step through answering five questions. Clicking the "Calculate" button immediately displays the risk results and basic recommendations. A basic history function can be included to observe risk trends.

[0111] (3) Internet-based online risk assessment web platform

[0112] Application scenarios: Targeting a wider range of the public and primary healthcare institutions, as a health management or science popularization tool.

[0113] Implementation: Establish a dedicated website accessible to users via a browser. Display usage instructions and a privacy statement in a prominent location. After users complete an online questionnaire, the system backend calculates and returns the results page in real time. Expandable features include links to risk-based personalized educational resources and information on online support groups.

[0114] This embodiment details the specific structure, implementation method, and application process of the prediction system. This system transforms complex prediction models into stable, easy-to-use, and integrable software tools, effectively solving the "last mile" application problem of nomogram models in clinical practice, realizing the technical solution protected by the claims, and possessing significant practical value.

[0115] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the risk of burden on family caregivers of minor stroke patients, characterized in that, Includes the following steps: S1. Study Cohort Construction and Data Collection: Based on the pre-set inclusion and exclusion criteria, acute-phase minor stroke patients and their primary family caregivers were selected to form the study cohort; during the acute phase of the patients' illness, baseline clinical data of the patients and baseline information of the family caregivers were collected. S2. Predictive variable screening: Statistical analysis was performed on the baseline clinical data of the patients and the baseline information of the family caregivers. Univariate analysis and multivariate logistic regression analysis were combined to screen out predictive variables that were independently associated with the occurrence of moderate to severe care burden in the family caregivers within 6-12 months after the onset of the disease. S3. Construction of nomogram prediction model: Based on the independent predictors selected in step S2 and their corresponding regression coefficients in multivariate logistic regression analysis, a nomogram prediction model is constructed using statistical calculation software; the nomogram model maps different values ​​of each predictor to corresponding scores, and obtains a total score by accumulating the scores of each item, and maps the total score to the predicted probability of the family caregiver experiencing moderate to severe burden; S4. Model Performance Validation and Optimization: The discrimination of the nomogram model was evaluated using receiver operating characteristic (ROC) curve analysis, and the internal validation of the model was performed using the Bootstrap resampling method. The corrected C-index was calculated to evaluate the predictive consistency and stability of the model.

2. The method for predicting the risk of burden on family caregivers of minor stroke patients according to claim 1, characterized in that: In step S1, the patient's baseline clinical data includes at least: demographic information, education level, cognitive function assessment results, neurological deficit severity score, activities of daily living score, and emotional state score. The baseline information of the family caregiver includes at least: demographic data, relationship with the patient, daily care time, monthly family income level, emotional state score, coping style score, and social support score.

3. The method for predicting the risk of burden on family caregivers of minor stroke patients according to claim 1, characterized in that: The independently relevant predictors mentioned in step S2 include at least five of the following: (a) The gender of the family caregiver is female; (b) The family caregiver had a depressive disorder at baseline; (c) The relationship between the family caregiver and the patient is that of spouse; (d) The patient had post-stroke cognitive impairment at baseline; (e) The patient's educational level is high school or above.

4. The method for predicting the risk of burden on family caregivers of minor stroke patients according to claim 1, characterized in that: The construction of the nomogram prediction model in step S3 specifically includes: Using the rms package in R, the results of the multivariate logistic regression analysis were taken as input, and the nomogram function was used to generate a visual nomogram containing all independent predictor variables; the rating scale in the nomogram was set proportionally according to the regression coefficients of each variable.

5. The method for predicting the risk of burden on family caregivers of minor stroke patients according to claim 1, characterized in that: The method further includes step S5: Model application: The constructed and validated nomogram model was applied to new minor stroke patients and their family caregivers. By collecting predictive variable information of the new patients and caregivers in the acute phase, and summing the scores against the nomogram model, the individualized risk probability of the caregiver developing moderate to severe burden in the next 6-12 months was output to guide early clinical intervention.

6. The method for predicting the risk of burden on family caregivers of minor stroke patients according to claim 5, characterized in that: The specific application scenario of the model described in step S5 is as follows: before the patient is discharged from the hospital, medical staff conduct an assessment to identify caregivers with high burden risk, and develop personalized nursing support plans, health education and psychological intervention programs accordingly.

7. A risk prediction system for the burden on family caregivers of minor stroke patients, characterized in that, include: Data input interface module: used to receive characteristic data of target minor stroke patients and their family caregivers during the acute phase. The characteristic data includes: patient's education level, cognitive function status, caregiver's gender, relationship with the patient, and emotional state. Data processing and calculation module: stores a pre-trained nodal plot prediction model, which is constructed and verified based on the method of any one of claims 1 to 4; the data processing and calculation module is used to call the model to perform calculations based on the input feature data, and output various scores and a total score; Risk prediction and output module: used to convert the total score into the corresponding moderate to severe care burden risk probability value, and present the prediction results to the user in at least one of the following forms: numbers, percentages, risk levels or visualization charts.

8. The predictive system for the risk of burden on family caregivers of minor stroke patients according to claim 7, characterized in that: The system is an electronic assessment tool, and its forms include: (a) Software modules integrated into hospital information systems or stroke center management platforms; (b) Standalone mobile terminal application; (c) Internet-based online risk assessment web platform.

Images