End-stage health care management method and system for elderly patients

Abstract

The invention relates to the technical field of end care, in particular to an end-stage health care management method and system for elderly patients. The method comprises the steps of determining a plurality of intervention time periods according to a pre-constructed ACP intervention model, and performing ACP plan investigation on ACP related personnel through preset investigation questions at the beginning of each intervention time period; in the ACP plan investigation process, acquiring emotion data of an investigated person, extracting a first emotion vector from the emotion data, and calculating investigation confidence based on the first emotion vector; grading the investigation confidence according to a preset grading standard, and adjusting the investigation question according to a grading result; determining an intervention measure in the next intervention time period according to a result of the ACP plan survey; and managing the end-stage health of the elderly patient based on the intervention measures. In the ACP investigation process, the investigation question is dynamically adjusted based on the emotion of the investigated person, and the investigation accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of hospice care, and in particular to a terminal health care management method and system for elderly patients. Background Art

[0002] Advance care planning (ACP) is one of the core contents of hospice care services. It refers to the process in which patients, in a fully conscious state and before losing their decision-making ability, fully obtain relevant information such as the disease, disease course, and end-of-life medical care measures, and then communicate with their family members and / or medical staff to choose their own medical care wishes at the end of their lives based on their life experience and values. At present, the specific operation method of advance care planning is mainly questionnaire surveys, which are completed through communication with patients, family members and doctors. However, questionnaire surveys mainly rely on the subjective consciousness of the respondents, and are easily affected by multiple factors such as the respondents' emotions, acceptance of advance medical care, and environmental interference, resulting in reduced accuracy of the questionnaire survey results and difficulty in reflecting the respondents' true wishes. Therefore, there is an urgent need for a terminal health care management method for elderly patients that can improve the accuracy of questionnaire surveys. Summary of the invention

[0003] The purpose of the present invention is to provide a terminal health care management method for elderly patients with an emotion recognition step, thereby realizing real-time emotion recognition of the respondents, adjusting the survey plan in time according to the recognition results, increasing the respondents' acceptance and helping to improve the accuracy of the survey.

[0004] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions: A method for managing end-of-life health care for elderly patients, comprising: Determine multiple intervention time periods according to the pre-constructed ACP intervention model, and at the beginning of each intervention time period, conduct an ACP plan survey on ACP-related personnel using preset survey questions; During the ACP survey, emotional data of the respondents are obtained, a first emotional vector is extracted from the emotional data, and survey confidence is calculated based on the first emotional vector; the survey confidence is graded according to a preset grading standard, and the survey topic is adjusted according to the grading result; Determine intervention measures in the next intervention period according to the results of the ACP program survey; and manage the terminal health of elderly patients based on the intervention measures.

[0005] As a preferred technical solution of this application, the method for constructing the ACP intervention model includes: Determine the intervention time period of the ACP program investigation based on the TTM theoretical model; Obtaining the ACP acceptance level of the respondents through a questionnaire survey; adjusting the intervention time period based on the ACP acceptance level; Obtaining the care knowledge of the respondents through semi-structured interviews, extracting a keyword set from the care knowledge, performing similarity comparison on the vectors in the keyword set, obtaining keywords whose similarity is less than or equal to a similarity threshold, and integrating these keywords into a subject category; forming a draft of the ACP intervention plan based on the intervention time period and the subject category; The Delphi expert inquiry method was used to modify and debug the initial draft of the ACP intervention program to obtain the ACP intervention model.

[0006] As a preferred technical solution of the present application, the method for extracting the first emotion vector from the emotion data includes: The emotional data of the person under investigation is obtained, including: mouth movement data, eye movement data, limb movement data and sitting posture data; the corresponding mouth movement vectors, eye movement vectors, limb movement vectors and sitting posture vectors are extracted using artificial intelligence technology, and the resultant vectors are spliced ​​into the first emotional vector.

[0007] As a preferred technical solution of the present application, the investigator records the emotional data of the respondent, applies natural language processing technology to extract keywords and encodes them, and obtains the corresponding first emotion vector.

[0008] As a preferred technical solution of the present application, the emotion data is continuously acquired multiple times within a preset time window through an image acquisition device; the emotion data is preprocessed to obtain the first emotion vector; the ratio of the modulus of each component in the first emotion vector and the size of the time window is spliced ​​into a first emotion change rate; the first emotion vector and the first emotion change rate are input into a pre-trained emotion recognition model, and the survey confidence and corresponding grade are output through the emotion recognition model.

[0009] As a preferred technical solution of the present application, the survey topics are preprocessed, and the topics of the survey topics are identified and classified through a topic modeling algorithm to determine whether the output topics contain negative topics; if negative topics exist, the emotions of the respondents are identified when the survey topic begins until the next survey topic begins.

[0010] As a preferred technical solution of the present application, adjusting the survey questions according to the grading results includes: the survey questions include direct questions, euphemistic questions and emotion relief questions; and setting a first number of emotion relief questions after the direct questions.

[0011] As a preferred technical solution of the present application, the survey questions are adjusted according to the grading results, including: determining the grading results as low, medium and high levels according to the grading standards, setting emotion relief questions after direct questions graded as low; setting euphemistic questions after direct questions graded as medium; and there is no need to set other forms of questions for direct questions graded as high.

[0012] As a preferred technical solution of the present application, the method of adjusting the survey questions according to the grading results includes: presetting a confidence threshold; if the survey confidence is less than or equal to the confidence threshold, adjusting the order of the subsequent direct questions to adjust the survey questions that do not contain negative topics to after the current direct questions; and simultaneously identifying the emotions of the respondents, and when the survey confidence is greater than the confidence threshold, continuing to investigate the direct questions.

[0013] The present application also provides a terminal health care management system for elderly patients, including: A setting module is used to determine multiple intervention time periods according to a pre-built ACP intervention model, and at the beginning of each intervention time period, display preset survey questions to ACP-related personnel to start the ACP plan survey; The emotion analysis module is used to obtain the emotion data of the respondents during the ACP survey, extract the first emotion vector from the emotion data, calculate and output the survey confidence based on the first emotion vector; and classify the survey confidence according to a preset classification standard, and adjust the survey topic according to the classification result; A program output module is used to determine the intervention measures in the next intervention time period according to the results of the ACP plan survey; and manage the terminal health of elderly patients based on the intervention measures.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. Elderly patients have different perceptions of their own health at different times, and their acceptance of the formulation and implementation of ACP plans is also different. To address this issue, we conducted ACP surveys on elderly patients, their families, and medical staff at different intervention time periods to timely understand their degree of awareness and acceptance of ACP, so as to formulate a step-by-step ACP plan, gradually improve the acceptance of elderly patients, better meet the actual needs of elderly patients, and enable elderly patients to receive more appropriate health care in the terminal stage.

[0015] 2. The ACP survey process often involves negative words such as death, which can easily cause emotional changes and communication difficulties in the respondents, resulting in reduced accuracy of the survey results. To address this issue, by judging the emotional state of the respondents when talking about negative questions, timely adjusting the survey questions and correcting the emotional state of the respondents, the interference of emotional changes on the survey results can be reduced, thereby improving the effectiveness and accuracy of the ACP plan.

[0016] 3. By designing a basic information questionnaire and an ACP perception questionnaire, the degree of ACP awareness among elderly terminal patients, their families, and medical staff was comprehensively collected; semi-structured interviews were used to obtain the care knowledge of the respondents, and the needs of ACP-related personnel for the ACP plan were further understood, providing data support for the formulation of a general ACP intervention plan. The interview content was converted into quantifiable data through natural language processing technology and other means, and scientific coding and subject classification were performed, which was conducive to a comprehensive analysis of the topics involved in the interview, which not only improved the efficiency of data processing, but also ensured the accuracy and objectivity of the analysis results. The first draft of the ACP intervention plan was modified and debugged through the Delphi expert letter inquiry method to ensure the reliability and effectiveness of the ACP intervention model in practical application.

[0017] 4. The emotions of the respondents change in many ways, which are difficult to be fully captured and recorded by the human eye. To address this problem, the image acquisition device is used to obtain the changes in the respondents' movements due to emotional changes, analyze the respondents' mouth movements, eye movements, limb movements, and sitting posture movements, and then use the emotion recognition model to identify the image data of these movement changes and derive the corresponding survey confidence. With the help of the artificial intelligence model, the changes in emotional data that are difficult for the human eye to observe can be captured, which is conducive to a comprehensive analysis of the respondents' emotional data and improves the accuracy of the survey. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of the end-of-life health care management approach for elderly patients; Figure 2 Schematic diagram of the process for constructing the ACP intervention model; Figure 3 Schematic diagram of the process of identifying the emotions of respondents for the emotion recognition model; Figure 4 This is a schematic diagram of the structure of the terminal health care management system for elderly patients. DETAILED DESCRIPTION

[0019] To make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be described clearly and completely in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them.

[0020] Therefore, the following detailed description of the embodiments of the present invention is not intended to limit the scope of the invention claimed for protection, but merely represents some embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features and technical solutions in the embodiments may be combined with each other.

[0022] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0023] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the invention product is usually placed when in use, or the orientation or positional relationship commonly understood by those skilled in the art. Such terms are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0024] Example 1: For example Figure 1-Figure 2 As shown, this embodiment provides a terminal health care management method for elderly patients, including: Multiple intervention time periods are determined based on the pre-built ACP intervention model. At the beginning of each intervention time period, the respondents are surveyed on the ACP plan using preset survey questions. It should be noted that although the ACP plan is implemented for elderly patients, it requires the participation and support of family members and the promotion and explanation of medical staff, so survey questions also need to be formulated for family members and medical staff.

[0025] In this embodiment, the respondents are relevant personnel participating in the ACP program, including elderly patients and their families, and medical staff involved in the end-of-life care of elderly patients. The ACP intervention model is established based on the TTM theoretical model, and the intervention time periods are determined to be five, including: In the early stage of intention, elderly patients believed that ACP plan was not necessary. The main purpose of the survey questions was to obtain the respondents' level of awareness and acceptance of ACP, and to provide a basis for judging whether to carry out the early knowledge dissemination and promotion of ACP plan.

[0026] In the intention period, elderly patients begin to think about completing the ACP plan, but there are factors that hinder actual action, such as lack of understanding of the ACP plan, subjective belief that it is not important, insufficient time to designate a plan, family opposition, etc. The main purpose of the survey questions is to obtain the respondents' ACP knowledge coverage, to determine the knowledge reinforcement points of the ACP plan, to minimize the interruption of the survey caused by insufficient ACP cognition during the ACP survey, and to improve the accuracy of the survey.

[0027] In the preparation stage, the elderly patients have already started or are preparing to start the development of the ACP plan. The main purpose of the survey questions is to obtain basic information of elderly patients, including but not limited to personal information, basic disease information, drug use information and living conditions.

[0028] During the action phase, elderly patients complete the survey and designation of the ACP plan based on their personal values ​​and wishes. The purpose of the survey questions is to develop a preliminary ACP intervention plan, mainly through a questionnaire compiled by experts to investigate the elderly patients' ACP willingness and end-of-life intervention measures, including but not limited to whether to receive palliative care in a hospice, whether to take excessive measures to prolong life, and whether to use drugs to relieve pain.

[0029] During the maintenance period, advance directives should be reviewed regularly and revised in a timely manner. Repeated evaluation and regular updating of advance directives is an ongoing process. After experiencing life events such as marriage, divorce, childbirth, and the death of a partner, the patient may need to re-appoint a medical decision-maker; or as medical standards advance, more effective treatments become available for the corresponding diseases, and the patient's expectations and choices for care will change accordingly. Therefore, the patient's values, beliefs, desired care goals, and choice of care location in the established ACP plan should be reviewed and revised regularly as the patient's condition changes.

[0030] ACP is a health management measure for the terminally ill elderly patients. When formulating an ACP plan, it is inevitable to talk about negative words such as death, which can easily cause strong psychological conflicts with the elderly patients' personal values, beliefs and cultural traditions. Therefore, there are certain communication difficulties, which can easily lead to a decrease in the accuracy of negative questions related to negative words such as death during the ACP plan survey.

[0031] Therefore, during the ACP program survey, it is necessary to pay attention to the emotional state of the respondents when talking about the negative questions. The emotional data of the respondents under investigation are obtained through manual judgment and recording by investigators, setting of emotional confirmation questions, and image recognition, and the survey confidence of the respondents is determined based on the emotional data. Mouth movement vectors, eye movement vectors, limb movement vectors, and sitting posture vectors are extracted from the emotional data and combined into a first emotional vector. The survey confidence can be obtained by calculating the module length of the first emotional vector, or by calculating the similarity between the first emotional vector and a preset conventional emotional vector. The conventional emotional vector is the emotional data collected from the respondents when they register for hospitalization.

[0032] The survey confidence is graded according to a preset grading standard, and the survey topic is adjusted according to the grading result. The grading standard is set by an expert in combination with expert experience and the statistical results of the survey confidence. Optionally, the median of the historical survey confidence is set as the first threshold, and the average value of the historical survey confidence greater than the median is recorded as the second threshold; the survey confidence less than or equal to the first threshold is graded as low, greater than the first threshold and less than or equal to the second threshold is graded as medium, and greater than the second threshold is graded as high. Optionally, the average value of the historical survey confidence is preset as the confidence threshold, and the survey confidence less than or equal to the confidence threshold is graded as low; the survey confidence greater than the confidence threshold is graded as high. When the survey confidence is graded as low or medium, it is considered that the respondent's emotions have changed negatively, which will affect the accuracy of the survey results. Therefore, it is necessary to take timely countermeasures to adjust the survey topic and correct the emotional state of the respondent to improve the authenticity and accuracy of the survey.

[0033] When a person has negative emotions, their mouth movements may include biting lips, drooping mouth corners, pursing lips, etc.; their eye movements may include pupil changes, increased blinking frequency, eyelid depression, increased eyeball rotation frequency, or continuous gaze at the ground, etc.; their limb movements may include continuous crossing of arms or increased frequency of crossing, crossing of legs and increased frequency of crossing, leg shaking, etc.; their sitting posture movements may include leaning forward, shrugging shoulders, fidgeting, frequent changes in sitting posture, etc. These movements may occur alone or in combination, forming frustrated or dissatisfied facial expressions and body movements, indicating that the person being investigated may have negative emotions such as anger, dissatisfaction, or disappointment.

[0034] As a preferred technical solution of the present application, the method for extracting the first emotion vector from the emotion data includes: obtaining the emotion data of the person being investigated, including: mouth movement data, eye movement data, limb movement data and sitting posture data; using artificial intelligence technology to extract the corresponding mouth movement vectors, eye movement vectors, limb movement vectors, sitting posture vectors, and splicing them into the first emotion vector.

[0035] Optionally, the investigators can record the actions of the respondents manually, and the recorded content includes: changes in mouth movements, such as the corners of the mouth sinking; changes in eye movements, such as blinking frequency, eye movement; changes in limb movements, such as crossing hands and arms; changes in sitting posture, such as upper body moving back and body leaning to one side. The emotional data includes descriptive words and / or sentences of the recorded actions of the respondents. Then, keywords are extracted through natural language processing technology, and the keywords are encoded using a bag-of-words model, etc. to obtain corresponding mouth movement vectors, eye movement vectors, limb movement vectors, and sitting posture vectors, which are spliced ​​into the first emotional vector.

[0036] Determine the intervention measures in the next intervention period based on the results of the ACP plan survey; manage the terminal health of elderly patients based on the intervention measures. The results of the ACP plan survey include changes in the respondents' acceptance of ACP, changes in their cognition, and terminal health management methods. The management methods include end-of-life medical decision entrustors, types of drugs, life-sustaining measures, etc.

[0037] Elderly patients have different perceptions of their own health at different times, and their acceptance of the formulation and implementation of ACP plans is also different. To address this issue, we conducted ACP surveys on elderly patients, their families, and medical staff at different intervention time periods to timely understand their degree of awareness and acceptance of ACP, so as to formulate a step-by-step ACP plan, gradually improve the acceptance of elderly patients, better meet the actual needs of elderly patients, and enable elderly patients to receive more appropriate health care in the terminal stage.

[0038] The ACP survey process often involves negative words such as death, which can easily cause emotional changes and communication difficulties in the respondents, resulting in reduced accuracy of the survey results. To address this issue, by judging the emotional state of the respondents when talking about negative questions, timely adjusting the survey questions and correcting the emotional state of the respondents, the interference of emotional changes on the survey results can be reduced, thus improving the effectiveness and accuracy of the ACP plan.

[0039] As the preferred technical solution of the present application, the method for constructing the ACP intervention model includes: determining the intervention time period of the ACP plan investigation based on the TTM theoretical model. Use platforms such as the China National Knowledge Infrastructure, Wanfang, VIP, and the China Biomedical Literature Database to search Chinese literature, and use databases such as Web of Science, Pubmed Medical Database, and the Cochrane Library to search foreign literature. Through literature research, sort out ACP-related concepts and the application of the TTM theoretical model in ACP, determine 5 intervention time periods based on the 5 stages of the TTM theoretical model, and provide a theoretical basis for the construction of the ACP intervention model for terminally ill elderly patients.

[0040] The respondents' ACP acceptance level is obtained through questionnaire survey; the intervention time period is adjusted based on the ACP acceptance level. For example, the ACP acceptance level is graded into low, medium, and high. Respondents graded as low have less ACP knowledge or are resistant to talking about ACP, so they need 5 intervention time periods; respondents graded as medium already have relevant ACP knowledge, so they can skip the pre-intention stage; respondents graded as high have rich ACP knowledge and are easy to accept the implementation of ACP, so they can skip the pre-intention stage and the intention stage. A statistical analysis is performed on the ACP acceptance level of all respondents, and the intervention time period corresponding to the grade with the largest number of people is selected for the construction of the ACP intervention model.

[0041] The questionnaire survey includes a basic information questionnaire and an ACP perception questionnaire. The basic information questionnaire includes but is not limited to three parts: basic demographic data survey, personal living conditions survey, and disease-related data survey. The basic information questionnaire is used to obtain the cultural background, educational background, and health perception of the respondents. Basic demographic data includes gender, age, religious beliefs, etc.; personal living conditions include monthly income, living conditions, etc.; disease-related data include disease types, payment methods, rescue experiences, etc. The ACP perception questionnaire is used to obtain the respondents' acceptance of ACP, including but not limited to the "Physical Health Perception Scale", "Chinese Values ​​Questionnaire", "Elderly Death Attitude Scale", and "Patient Advance Care Plan Acceptance Questionnaire". Each option in the questionnaire is set with a corresponding score, and the questionnaire score after the survey is completed is used as the ACP acceptance level.

[0042] Through semi-structured interviews, the care knowledge of the respondents was obtained, a keyword set was extracted from the care knowledge, the vectors in the keyword set were compared for similarity, the keywords whose similarity was less than or equal to the similarity threshold were obtained, and these keywords were integrated into a subject category; a preliminary draft of the ACP intervention plan was formed based on the intervention time period and the subject category.

[0043] The semi-structured interviews used a rough interview outline to limit the interview content to the ACP plan and to serve as a reminder for the interviewer. In the in-depth interviews, natural and open language was used to guide the respondents to share their cognition and experience of the care process, allowing the researcher and the interviewee to establish a natural communication in a natural and sincere situation, while recording the interview content.

[0044] The recordings or original text materials obtained from the interviews were transcribed into text within 24 hours and entered into the "Nvivo12" software for data management and data analysis. The analysis steps include: the investigators repeatedly listened to the recordings, removed irrelevant information from the survey questions and the respondents' answers, and transcribed them into text; during the transcription process, they also recorded their impressions and thoughts on the respondents to obtain a set of keywords. Generate initial coding, find keywords in the text, including important words in the original text, or self-summarized words, and then add them to the keyword set. Convert the keyword set into a vector through natural language processing techniques such as the bag-of-words model and save it. Find the topic category, compare the vectors corresponding to the keyword set for similarity, obtain keywords that meet the preset similarity threshold, and integrate these keywords into a topic category.

[0045] After integrating the subject categories and the intervention time periods, a preliminary draft of the ACP intervention program for terminally ill elderly patients was obtained, including: the timing of ACP intervention, treatment plan, treatment goals, the patient's outlook on life and death, the patient's wishes, the patient's preference for treatment measures, and the place of death.

[0046] Finally, the Delphi expert inquiry method was used to modify and debug the initial draft of the ACP intervention program, and ultimately an ACP intervention model for terminally ill elderly patients was formed.

[0047] An expert consultation questionnaire was customized based on the first draft of the ACP intervention plan. The expert consultation questionnaire included: a letter to experts, introducing the purpose and significance of the research project, the first draft of the ACP intervention plan and the basis for its formation; an expert opinion form for each level of indicators in the ACP intervention plan, with the Likert 5-level scoring method used to evaluate the importance of each level of indicators; an expert basic situation survey form; an expert authority self-assessment form, etc. Experts reviewed and provided feedback on the first draft of the ACP intervention plan. The first draft of the ACP intervention plan was adjusted based on the expert feedback to obtain the ACP intervention model.

[0048] By designing a basic information questionnaire and an ACP perception questionnaire, the degree of ACP awareness among elderly terminal patients, their families, and medical staff was comprehensively collected; the care knowledge of the respondents was obtained through semi-structured interviews, and the needs of ACP-related personnel for the ACP plan were further understood in depth, providing data support for the formulation of a universal ACP intervention program. The interview content was converted into quantifiable data through natural language processing technology and other means, and scientific coding and subject classification were carried out, which was conducive to a comprehensive analysis of the topics involved in the interview, which not only improved the efficiency of data processing, but also ensured the accuracy and objectivity of the analysis results. The first draft of the ACP intervention program was modified and debugged through the Delphi expert letter inquiry method to ensure the reliability and effectiveness of the ACP intervention model in practical application.

[0049] Example 2: During the survey, the emotional changes of the respondents can easily lead to inaccurate survey results. In particular, based on the traditional taboo on words such as "death", this type of negative question can easily cause respondents to resist. At the same time, the investigators need to be highly focused and carefully capture the changes in the respondents' emotions in real time, which still poses great difficulties for the investigators. Therefore, it is difficult to reflect the true situation of the ACP survey. Therefore, in order to improve the accuracy of the ACP survey, this example further adds the following technical features based on the technical solution of Example 1: For example Figure 3 As shown, as a preferred technical solution of the present application, the method for extracting the first emotion vector from the emotion data includes: After asking the respondent a negative question, the emotional data is obtained multiple times in a preset time window through an image acquisition device, and the emotional data includes the respondent's mouth image, eye image, limb image, and sitting image. The emotional data is preprocessed, including: Extract lip key frames from the first and last mouth images, obtain the downward movement distance of the mouth corners, the distance between the mouth corners and the lip thickness, and splice them into the mouth action vector. Calculate the modulus of the mouth action vector, and record the ratio of the modulus to the time window size as the mouth action change rate.

[0050] Extract pupil keyframes and eyelid keyframes from the first and last eye images to obtain pupil size, eyelid offset distance, and pupil downward distance; extract completely closed eyelid keyframes from each eye image and count them, and use the ratio of the count result to the time window as the blinking frequency; concatenate pupil size, eyelid offset distance, pupil downward distance, and blinking frequency into an eye motion vector. Calculate the modulus of the eye motion vector, and record its ratio to the time window size as the eye motion change rate.

[0051] Extract shoulder, elbow, and hand feature points from the first and last limb images and splice them into upper limb key frames; extract waist, knee, and ankle feature points and splice them into lower limb key frames; splice upper limb key frames and lower limb key frames into limb motion vectors. Determine whether there are crossover, jitter, and other motion features based on the image coordinates of the feature points. Calculate the modulus of the limb motion vectors, and record the ratio of the modulus to the time window size as the limb motion change rate.

[0052] Extract head feature points from the first and last sitting posture images, and splice them with upper limb key frames and lower limb key frames into a sitting posture vector. Calculate the modulus of the sitting posture vector, and record the ratio of the modulus to the time window size as the sitting posture change rate.

[0053] The mouth movement vector, eye movement vector, limb movement vector, and sitting posture vector are concatenated to be recorded as the first emotion vector. The mouth movement change rate, eye movement change rate, limb movement change rate, and sitting posture change rate are concatenated into a vector form to be recorded as the first emotion change rate.

[0054] The first emotion vector and the first emotion change rate are input into a pre-trained emotion recognition model, and the survey confidence is output through the emotion recognition model, and the classification is performed according to the classification standard. The emotion recognition model is established based on the ResNet18 residual network model, including 1 convolution layer, 1 maximum pooling layer, 1 residual network layer, 1 activation function layer, 1 regularization layer, and 1 fully connected layer as an output layer, and the cross entropy loss function is used as the judgment condition for parameter optimization of the emotion recognition model.

[0055] The emotions of the respondents change in many ways, which are difficult to be fully captured and recorded by the human eye. To address this problem, the image acquisition device is used to obtain the changes in the respondents' movements due to emotional changes, analyze the respondents' mouth movements, eye movements, limb movements, and sitting posture movements, and then use the emotion recognition model to identify the image data of these movement changes and derive the corresponding survey confidence. With the help of artificial intelligence models, emotional data changes that are difficult for the human eye to observe can be captured, which is conducive to a comprehensive analysis of the respondents' emotional data and improves the accuracy of the survey.

[0056] Conversations or surveys on negative topics such as death, dying, and dying vary among people of different cultural backgrounds, religious beliefs, and spiritual beliefs. At the same time, individuals in similar backgrounds also have differences in knowledge and cultural tolerance, which leads to differences in preferences between individuals. Therefore, when the survey topic involves negative topics, the respondents are more likely to have emotional changes.

[0057] As a preferred technical solution of the present application, it also includes: pre-processing the survey topic through natural language processing technology, including removing punctuation marks, numbers, special characters, etc. that are not related to the negative topic. The survey topic after pre-processing is segmented and represented as a bag-of-words model. Then, through a topic modeling algorithm, such as Top2Vec or latent Dirichlet distribution, the survey topic is identified and classified to determine whether the topic output by the algorithm contains negative topics related to keywords such as death and dying. If there is a negative topic, the emotion recognition of the respondent is performed at the beginning of the survey topic until the next survey topic begins. If there is no negative topic, emotion recognition is not required.

[0058] The survey confidence of all negative topics of different groups of people is statistically analyzed, and the educational background, religious background, etc. of different groups of people are statistically analyzed. The statistical results of survey confidence are used to analyze the differences between different groups in facing negative topics. For example, different groups of people are grouped according to their educational background, and the standard deviation of survey confidence in the groups reflects the degree of fluctuation, so as to analyze whether different individuals with the same educational background have similar cognition of negative topics.

[0059] By conducting thematic analysis on survey questions, obtaining negative questions and performing emotion recognition, while not recognizing questions on other topics, not only can we achieve targeted analysis of the emotional changes of different groups of people or individuals in response to negative questions, but we can also reduce the amount of data for emotion recognition and improve survey efficiency.

[0060] The survey questions include: direct questions, where negative themes appear directly in the description of the question, such as "Do you think it is necessary to communicate with your family and doctors in advance about your thoughts at the end of your life?"; euphemistic questions, which are multiple variations of the direct questions, require thinking and conversion of meaning to understand the negative theme, or include the negative theme in a large number of other descriptions, such as "In the long journey of life, as we slowly step into those more peaceful and profound days, have you ever thought about having a warm and detailed conversation with your closest family and professional medical caregivers to discuss those delicate ideas about the future and care?"; emotional relief questions, which are unrelated to the ACP survey and are used to relieve the emotions of the respondents, such as "If your pet walks towards you, do you want to touch its head or hold its face?". In order to directly reflect the respondents' acceptance of the ACP plan, the initial plan of the survey questions are all direct questions.

[0061] Preferably, a first number of emotion relief questions are set after the direct questions. The negative emotions of the respondents are relieved by the emotion relief questions or by setting entertainment measures, and the emotions of the respondents are recognized in real time by the emotion recognition model to obtain the survey confidence. When the survey confidence is greater than the confidence threshold, the direct questions are continued to be investigated. By setting emotion relief questions after each direct question, the emotional interference of the investigated direct questions on the subsequent questions is reduced, so that the investigation of each direct question is independent.

[0062] Preferably, the grading results are determined into three levels: low, medium and high according to the grading standard, and an emotion relief question is set after a direct question with a low grade; a euphemistic question is set after a direct question with a medium grade; and no other form of question is required for a direct question with a high grade.

[0063] Preferably, if the survey confidence is less than or equal to the confidence threshold, the order of the subsequent direct questions is adjusted, and the survey questions that do not contain negative topics are adjusted to after the current direct questions; at the same time, the emotions of the respondents are identified, and when the survey confidence is greater than the confidence threshold, the direct questions are continued to be investigated.

[0064] Preferably, the subsequent direct questions are adjusted to corresponding euphemistic questions. When the survey confidence of the survey question is less than the confidence threshold, it indicates that the respondents have a low acceptance of negative topics such as hospice care. By conducting a survey through euphemistic questions, the respondents' resistance to negative topics is reduced, which is conducive to improving the integrity of the survey.

[0065] In order to address the problem that respondents may be affected by negative topics in the questions during the ACP survey, resulting in emotional fluctuations, the emotional state of the respondents is obtained in real time, and the survey confidence grading results of the respondents are obtained through the emotion recognition model. During the survey, subsequent survey questions are dynamically adjusted according to the grading results. This is conducive to improving the comfort and acceptance of the respondents, helping the respondents to continue to participate in the survey when their emotions are relatively stable, and improving the effectiveness and accuracy of the survey.

[0066] Embodiment 3: This embodiment provides a terminal health care management system for elderly patients, a device and corresponding data and executable program storage medium.

[0067] For example Figure 4 As shown, a terminal health care management system for elderly patients includes: A setting module, for determining a plurality of intervention time periods according to a pre-built ACP intervention model, and displaying preset survey questions to ACP-related personnel at the beginning of each of the intervention time periods; The emotion analysis module is used to obtain the emotion data of the respondents during the ACP survey, extract the first emotion vector from the emotion data, calculate and output the survey confidence based on the first emotion vector; and classify the survey confidence according to a preset classification standard, and adjust the survey topic according to the classification result; A program output module is used to determine the intervention measures in the next intervention time period according to the results of the ACP plan survey; and manage the terminal health of elderly patients based on the intervention measures.

[0068] A terminal health care management device for elderly patients, the device comprising: a sensor for collecting various data related to the face and / or limbs of the respondent; at least one processor capable of processing various data related to the face and / or limbs of the respondent collected by the sensor and the questions and score data for the questionnaire; at least one memory for storing various data related to the face and / or limbs of the respondent collected by the sensor, the questionnaire questions and at least one program; when the at least one program is executed by the at least one processor, the at least one processor implements the terminal health care management method for the elderly patient, the processor and the memory can be connected by a bus or other means for storing at least one program; the memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory can include a high-speed random access memory, and can also include a non-transient memory, such as at least one disk storage device, a flash memory device or other non-transient solid-state storage device. In some embodiments, the memory optionally includes a memory remotely arranged relative to the processor, and these remote memories can be connected to the device via a network. Examples of the above-mentioned network include but are not limited to the Internet, corporate intranet, local area network, mobile communication network and combinations thereof. When the at least one program is executed by the at least one processor, the at least one processor implements the terminal health care management method for elderly patients.

[0069] A computer storage medium stores a program executable by a processor, wherein the program executable by the processor is used to implement the terminal health care management method for elderly patients when executed by the processor.

[0070] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0071] The above embodiments are only used to illustrate the present invention and are not intended to limit the technical solutions described in the present invention. Although the present invention has been described in detail with reference to the above embodiments, the present invention is not limited to the above specific implementation methods. Therefore, any modification or equivalent replacement of the present invention; and all technical solutions and improvements thereof that do not depart from the spirit and scope of the invention are included in the scope of the claims of the present invention.

Claims

1. A terminal health care management method for elderly patients, characterized in that: include: Determine multiple intervention time periods according to the pre-constructed ACP intervention model, and at the beginning of each intervention time period, conduct an ACP plan survey on ACP-related personnel using preset survey questions; During the ACP survey, emotional data of the respondents are obtained, a first emotional vector is extracted from the emotional data, and survey confidence is calculated based on the first emotional vector; the survey confidence is graded according to a preset grading standard, and the survey topic is adjusted according to the grading result; Determine intervention measures in the next intervention period according to the results of the ACP program survey; and manage the terminal health of elderly patients based on the intervention measures.

2. A terminal health care management method for elderly patients according to claim 1, characterized in that: The method for constructing the ACP intervention model includes: Determine the intervention time period of the ACP program investigation based on the TTM theoretical model; Obtaining the ACP acceptance level of the respondents through a questionnaire survey; adjusting the intervention time period based on the ACP acceptance level; Obtaining the care knowledge of the respondents through semi-structured interviews, extracting a keyword set from the care knowledge, performing similarity comparison on the vectors in the keyword set, obtaining keywords whose similarity is less than or equal to a similarity threshold, and integrating these keywords into a subject category; forming a draft of the ACP intervention plan based on the intervention time period and the subject category; The Delphi expert inquiry method was used to modify and debug the initial draft of the ACP intervention program to obtain the ACP intervention model.

3. A terminal health care management method for elderly patients according to claim 1, characterized in that: The method of extracting the first emotion vector from the emotion data comprises: The emotional data of the person under investigation is obtained, including: mouth movement data, eye movement data, limb movement data and sitting posture data; the corresponding mouth movement vectors, eye movement vectors, limb movement vectors and sitting posture vectors are extracted using artificial intelligence technology, and the resultant vectors are spliced ​​into the first emotional vector.

4. A terminal health care management method for elderly patients according to claim 3, characterized in that: The investigators record the emotional data of the respondents, extract keywords using natural language processing technology and encode them, so as to obtain the corresponding first emotion vector.

5. A terminal health care management method for elderly patients according to claim 3, characterized in that: Acquire the emotion data multiple times continuously through an image acquisition device within a preset time window; pre-process the emotion data to obtain the first emotion vector; splicing the ratio of the modulus of each component in the first emotion vector to the time window size into a first emotion change rate; The first emotion vector and the first emotion change rate are input into a pre-trained emotion recognition model, and the survey confidence and corresponding grade are output through the emotion recognition model.

6. A terminal health care management method for elderly patients according to claim 1, characterized in that: The survey topics are preprocessed, and the topics are identified and classified by a topic modeling algorithm to determine whether the output topics contain negative topics; if negative topics exist, the emotions of the respondents are identified when the survey topic begins, until the next survey topic begins.

7. A terminal health care management method for elderly patients according to claim 1, characterized in that: The step of adjusting the survey topics according to the grading results includes: the survey topics include direct topics, euphemistic topics, and emotion relief topics; and setting a first number of emotion relief topics after the direct topics.

8. A terminal health care management method for elderly patients according to claim 7, characterized in that: The adjusting of the survey questions according to the grading results includes: determining the grading results as low, medium and high according to the grading standards, setting emotion relief questions after direct questions with a low grade; setting euphemistic questions after direct questions with a medium grade; and there is no need to set other forms of questions for direct questions with a high grade.

9. A terminal health care management method for elderly patients according to claim 7, characterized in that: The method of adjusting the survey questions according to the grading results includes: presetting a confidence threshold; if the survey confidence is less than or equal to the confidence threshold, adjusting the order of subsequent direct questions to adjust the survey questions that do not contain negative topics to after the current direct questions; and simultaneously identifying the emotions of the respondents, and when the survey confidence is greater than the confidence threshold, continuing to investigate the direct questions.

10. A terminal health care management system for elderly patients, characterized in that: include: A setting module is used to determine multiple intervention time periods according to a pre-built ACP intervention model, and at the beginning of each intervention time period, display preset survey questions to ACP-related personnel to start the ACP plan survey; A sentiment analysis module, used for obtaining the sentiment data of the respondents during the ACP program survey, extracting a first sentiment vector from the sentiment data, and calculating and outputting the survey confidence based on the first sentiment vector; Used to classify the survey confidence according to a preset classification standard, and adjust the survey topic according to the classification result; A program output module, used to determine the intervention measures in the next intervention time period according to the results of the ACP plan survey; Manage the end-stage health of elderly patients based on the interventions described.

Images