A method, apparatus and equipment for processing data on the needs of terminally ill patients

Abstract

This invention provides a method, apparatus, and device for processing data on the needs of terminally ill patients, belonging to the field of medical data processing technology. It solves the problems of strong subjectivity and inability to quantify priorities in the assessment of the needs of terminally ill patients. The method includes: acquiring multi-source initial data from multiple terminally ill patients; extracting features from the multi-source initial data to obtain multi-dimensional raw feature values; performing standardized mapping processing on the multi-dimensional raw feature values ​​to obtain multi-dimensional standardized scores; determining a multi-dimensional weight vector based on the multi-dimensional standardized scores; determining initial comprehensive needs data for multiple terminally ill patients based on the multi-dimensional standardized scores and the multi-dimensional weight vector; correcting the initial comprehensive needs data for multiple terminally ill patients to obtain target comprehensive needs data for multiple terminally ill patients; and ranking the multiple patients according to the target comprehensive needs data to obtain a patient care priority ranking result. This solution shortens the assessment time and improves accuracy.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, and in particular to a method, apparatus, and equipment for processing data on the needs of terminally ill patients. Background Technology

[0002] Currently, patient needs assessment in the field of end-of-life care mainly relies on scales or qualitative interviews, with each dimension being relatively independent. Physiological dimensions commonly use the Numerical Rating Scale (NRS), Visual Analogue Scale (VAS), or Facial Expression Pain Scale; psychological dimensions employ the Self-Rating Anxiety Scale (SAS), Self-Rating Depression Scale (SDS), or Hospital Anxiety and Depression Scale (HADS); social dimensions use the Social Support Rating Scale (SSRS) or structured questionnaires to understand family care and economic status; spiritual dimensions often rely on open-ended interviews to identify feelings of meaning in life, unfulfilled wishes, or ritual needs. In recent years, wearable devices (such as smart bracelets) have begun to be used to collect parameters such as heart rate, respiration, and body movement, attempting to achieve objective pain monitoring, showing an overall trend from single-dimensional to multi-dimensional and from subjective to objective. However, existing technologies still have significant shortcomings.

[0003] First, there is a fragmentation of dimensions. Data from each dimension is stored and interpreted independently, making it impossible to integrate them into a unified quantitative indicator or generate a cross-dimensional comprehensive demand index or priority ranking. When patients simultaneously experience moderate pain and severe spiritual distress, clinical teams struggle to objectively compare the urgency of the two, easily leading to resource misallocation.

[0004] Secondly, the data sources are singular and subjective. Most assessments are based on patient self-reports or medical observations, which significantly reduces accuracy for patients who are unable to express themselves clearly, such as those with aphasia or confusion. Objective data from wearable devices are used only for inferences about single physiological indicators and are not linked to psychological, social, or spiritual dimensions.

[0005] Secondly, there is a lack of algorithmic support. The allocation of care resources (such as pain relief, psychological counseling, and spiritual care) relies on the personal experience of healthcare professionals and lacks dynamic ranking algorithms based on multi-dimensional weights. The pain, anxiety, social vulnerability, and spiritual urgency of different patients cannot be compared on the same scale, leading to decision-making difficulties.

[0006] Finally, the information from family interviews was not effectively utilized. Unstructured texts, such as behavioral changes, cuees about beliefs, and unexpressed wishes observed by family members, lacked the tools of natural language processing to extract quantitative features. Summary of the Invention

[0007] This invention provides a method, apparatus, and equipment for processing data on the needs of terminally ill patients, which solves the problems of strong subjectivity and inability to quantify priorities in the assessment of the needs of terminally ill patients.

[0008] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: This invention provides a method for processing data on the needs of terminally ill patients, including: Acquire multi-source initial data from multiple terminally ill patients, including patient self-assessment data, patient physiological signal data, and family interview text data; Feature extraction is performed on the multi-source initial data to obtain multi-dimensional original feature values, which include original feature values ​​of physiological dimension, psychological dimension, social dimension and spiritual dimension. The original multidimensional feature values ​​are standardized and mapped to obtain multidimensional standardized scores. Based on the multidimensional standardized scores, determine the multidimensional weight vector; Based on the multidimensional standardized scores and multidimensional weight vectors, the initial comprehensive needs data for multiple terminally ill patients are determined; The initial comprehensive needs data of the multiple terminally ill patients are corrected to obtain the target comprehensive needs data of the multiple terminally ill patients; Based on the comprehensive target needs data of multiple terminally ill patients, the patients are ranked to obtain the patient care priority ranking result.

[0009] Optionally, acquire multi-source initial data from multiple terminally ill patients, including: The patient's self-assessment data is obtained by collecting the patient's or family's answers to preset questions through an adaptive dialogue system. By using wearable devices, the patient's heart rate variability time-series signal is collected to obtain the patient's physiological signal data; The audio recordings of family interviews were processed to convert speech to text, resulting in the text data of the family interviews.

[0010] Optionally, feature extraction is performed on the multi-source initial data to obtain multi-dimensional original feature values, including: The patient's self-assessment data is processed to extract pain data, or the patient's physiological signal data is processed to transform the data to determine the original feature values ​​of the physiological dimensions. The patient self-assessment data was processed to extract psychological data and obtain the original feature values ​​of the psychological dimension. Entity recognition processing was performed on the family interview text data to obtain the original feature values ​​of the social dimension; The patient self-assessment data and family interview text data were processed to extract emotional data, and the original feature values ​​of the spiritual dimension were obtained.

[0011] Optionally, the multidimensional original feature values ​​are subjected to standardized mapping processing to obtain multidimensional standardized scores, including: According to the formula: , obtained the The physiological standardized score of each patient; in, For the first The physiological standardized score of each patient; For the first The original feature values ​​of the physiological dimensions of each patient; The preset physiological sensitivity coefficient has a value range of 2 to 5. The preset physiological half-activation point has a value of 0.3 to 0.7. According to the formula: , Get the first The standardized psychological scores of each patient; among them. For the first The standardized psychological scores of each patient; The preset weighting coefficients for the psychological scale range from 0.5 to 0.8. The preset voice emotion sensitivity coefficient has a value range of 1 to 4; The preset semi-activation point for voice emotion has a value range of 0.4 to 0.6; For the first The original total score of the psychological scale for each patient is obtained by summing the scale scores in the original feature values ​​of the psychological dimension, with a value range of 0 to 100. For the first The voice emotion feature score of each patient is obtained by extracting emotion features from the voice signal in the patient's self-assessment data, and the value ranges from 0 to 1. According to the formula: , The social standardization score is obtained; among which, For the first The social standardized score of each patient; The preset social low demand threshold has a value range of 10 to 20. The preset threshold for high social demand ranges from 70 to 85. For the first The original feature values ​​of the social dimension of each patient , For the first Preset weights for each social characteristic component. For the first The first patient's A social characteristic component; The total number of social characteristic components; According to the formula: , obtained the The standardized spiritual scores of each patient; among them. For the first The standardized spiritual score for each patient, and when the calculated result exceeds 100, Take 100; For the first The original feature values ​​of the spiritual dimension of each patient; This is the maximum possible value of the preset spiritual characteristics, ranging from 50 to 10.

[0012] Optionally, based on the multidimensional standardized scores, a multidimensional weight vector is determined, including: Based on the aforementioned multidimensional standardized scores, the information entropy data is determined; Based on the information entropy data, determine the objective weight data; Obtain the preset subjective weight vector data; The objective weight data and the subjective weight vector data are linearly weighted to obtain a multidimensional weight vector.

[0013] Optionally, based on the multidimensional standardized scores and multidimensional weight vectors, initial comprehensive needs data for multiple terminally ill patients are determined, including: According to the formula: Determine the first Initial comprehensive needs data for each patient; in, For the first Initial comprehensive needs data for each patient; Weights for physiological dimensions; Weights for psychological dimensions; Weighting for the social dimension; Weighting for the spiritual dimension; For the first The physiological standardized score of each patient; For the first The standardized psychological scores of each patient; For the first The social standardized score of each patient; For the first The patient's standardized spiritual score.

[0014] Optionally, the initial comprehensive needs data of the multiple terminally ill patients are revised to obtain target comprehensive needs data for the multiple terminally ill patients, including: According to the formula: Determine the first Comprehensive target needs data for each patient; in, For the first Comprehensive target needs data for each patient; No. Initial comprehensive needs data for each patient; For correction factor, The formula is: ; in, The preset correction amplitude coefficient ranges from 0.05 to 0.15. For the first The physiological standardized score of each patient; For the first The patient's standardized spiritual score; The preset social threshold; For the first The social standardized score of each patient; This is the function for finding the maximum value.

[0015] Optionally, based on the comprehensive target needs data of the multiple terminally ill patients, the multiple patients are ranked to obtain a patient care priority ranking result, including: Based on the numerical values ​​of the target comprehensive demand data, the multiple patients are sorted in descending order to obtain a patient care priority queue; among them, patients ranked higher indicate that their care needs are more urgent and they are given priority in receiving care resource allocation.

[0016] This invention also provides a data processing device for end-of-life patient needs, comprising: The acquisition module is used to acquire multi-source initial data from multiple terminally ill patients, including patient self-assessment data, patient physiological signal data, and family interview text data. The processing module is used to extract features from the multi-source initial data to obtain multi-dimensional original feature values, including original feature values ​​of physiological, psychological, social, and spiritual dimensions; perform standardized mapping processing on the multi-dimensional original feature values ​​to obtain multi-dimensional standardized scores; determine multi-dimensional weight vectors based on the multi-dimensional standardized scores; determine initial comprehensive needs data for multiple terminally ill patients based on the multi-dimensional standardized scores and multi-dimensional weight vectors; correct the initial comprehensive needs data for multiple terminally ill patients to obtain target comprehensive needs data for multiple terminally ill patients; and rank the multiple patients according to the target comprehensive needs data for multiple terminally ill patients to obtain a patient care priority ranking result.

[0017] This invention also provides a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when run by the processor, executes the above-described method.

[0018] The technical solution of the present invention has at least the following effects: The above-described solution of the present invention acquires multi-source initial data from multiple terminally ill patients, including patient self-assessment data, patient physiological signal data, and family interview text data; extracts features from the multi-source initial data to obtain multi-dimensional raw feature values, including raw feature values ​​for physiological, psychological, social, and spiritual dimensions; performs standardized mapping processing on the multi-dimensional raw feature values ​​to obtain multi-dimensional standardized scores; determines multi-dimensional weight vectors based on the multi-dimensional standardized scores; determines initial comprehensive needs data for multiple terminally ill patients based on the multi-dimensional standardized scores and multi-dimensional weight vectors; corrects the initial comprehensive needs data for multiple terminally ill patients to obtain target comprehensive needs data for multiple terminally ill patients; and ranks the multiple patients according to the target comprehensive needs data to obtain a patient care priority ranking result, thus shortening the assessment time and improving the accuracy. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method for processing data on the needs of terminally ill patients provided in an embodiment of the present invention; Figure 2 This is a structural diagram of the terminal patient needs data processing device provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the computing device provided in an embodiment of the present invention. Detailed Implementation

[0020] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0021] like Figure 1 As shown, an embodiment of the present invention proposes a method for processing data on the needs of terminally ill patients, including: Step 11: Obtain multi-source initial data from multiple terminally ill patients, including patient self-assessment data, patient physiological signal data, and family interview text data. Step 12: Extract features from the multi-source initial data to obtain multi-dimensional original feature values, which include original feature values ​​of physiological dimension, psychological dimension, social dimension and spiritual dimension. Step 13: Perform standardized mapping processing on the multidimensional original feature values ​​to obtain multidimensional standardized scores; Step 14: Determine the multidimensional weight vector based on the multidimensional standardized score; Step 15: Determine the initial comprehensive needs data for multiple terminally ill patients based on the multidimensional standardized scores and multidimensional weight vectors; Step 16: Correct the initial comprehensive needs data of the multiple terminally ill patients to obtain the target comprehensive needs data of the multiple terminally ill patients; Step 17: Based on the comprehensive target needs data of the multiple terminally ill patients, the multiple patients are ranked to obtain the patient care priority ranking result.

[0022] This invention proposes the above-mentioned technical solution, which constructs a four-dimensional assessment index system encompassing physiological, psychological, social, and spiritual aspects. It integrates multi-source data such as interactive questionnaires, physiological data from wearable devices, and family interview texts to calculate a comprehensive demand index and introduces a dimensional coupling correction factor for priority ranking. This reduces the time for a single assessment from approximately 45 minutes to less than 15 minutes, thereby improving assessment efficiency, enhancing assessment objectivity, and improving the feasibility of assessment for special patient groups.

[0023] In an optional embodiment of the present invention, step 11, obtaining multi-source initial data from multiple terminally ill patients, may include: Step 111: Collect the patient's or family's answers to preset questions through an adaptive dialogue system to obtain the patient's self-assessment data; Step 112: Collect the patient's heart rate variability time-series signal through a wearable device to obtain the patient's physiological signal data; Step 113: Perform speech-to-text processing on the family interview recordings to obtain the family interview text data.

[0024] In step 111 of this embodiment, the adaptive dialogue system has a pre-set question bank including items such as pain level, sleep quality, emotional state, and sense of meaning in life. For patients who can self-report, the system presents questions in the form of multiple-choice questions or short texts, such as "Please describe your current pain level using numbers from 0 to 10," directly obtaining a numerical score. For patients with difficulty expressing themselves, the system uses simplified options or has family members answer on their behalf. The text of answers to open-ended questions, such as "What are you most worried about right now?", is subsequently used for natural language processing. In the patient self-report data, scores for scale-type questions are directly recorded as raw scores, ranging from 0 to 100; open-ended answer texts are saved as strings for use in extracting psychological and spiritual dimension features.

[0025] In step 112, the wearable device, including a smart bracelet or mattress-type sensor, continuously acquires the patient's heart rate variability time-series signal at a sampling rate of at least 50 Hz. Heart rate variability refers to the small fluctuations in the intervals between successive heartbeats; its low-frequency component (0.04 to 0.15 Hz) is associated with sympathetic nerve activity, and its high-frequency component (0.15 to 0.40 Hz) is associated with parasympathetic nerve activity. Temporal and frequency domain features are extracted from the acquired signals, including the ratio of low-frequency power to high-frequency power. An increase in this ratio typically reflects enhanced sympathetic nerve excitability and is positively correlated with pain stress.

[0026] In step 113, the family interview recordings, after being recorded using specialized equipment, are transcribed into text using a speech recognition engine. The transcribed text includes descriptions from family members regarding the patient's daily behavior, emotional changes, diet, sleep, unexpressed wishes, and cultural beliefs. This text is saved as a structured document, with each line associated with a patient identifier and a recording timestamp. The speech-to-text process preserves the original spoken expression without semantic compression or summarization, allowing for subsequent use of a pre-trained language model for entity recognition and sentiment classification. The transcribed family interview text data, along with patient self-assessment data and physiological signal data, constitute a multi-source initial dataset.

[0027] In an optional embodiment of the present invention, step 12, which involves extracting features from the multi-source initial data to obtain multi-dimensional original feature values, may include: Step 121: Extract pain data from the patient's self-assessment data or convert the patient's physiological signal data to determine the original feature values ​​of the physiological dimensions; Step 122: Perform psychological data extraction processing on the patient self-assessment data to obtain the original feature values ​​of the psychological dimension; Step 123: Perform entity recognition processing on the family interview text data to obtain the original feature values ​​of the social dimension; Step 124: Perform emotional data extraction processing on the patient self-assessment data and family interview text data to obtain the original feature values ​​of the spiritual dimension.

[0028] In step 121 of this embodiment, the extraction of raw feature values ​​for the physiological dimension is divided into two cases. When the patient can self-assess, a numerical pain score is directly read from the patient's self-assessment data. This score uses a scale of 0 to 10, where 0 represents no pain and 10 represents the most severe pain. This value is directly used as the raw feature value for the physiological dimension. When the patient cannot self-assess, the obtained heart rate variability feature values ​​are used for transformation processing. Specifically, the heart rate variability feature values ​​are substituted into a pre-constructed pain prediction model: ; in, and These are regression coefficients obtained in advance through training with clinical samples; This represents the ratio of low-frequency power to high-frequency power; regardless of whether self-assessment or prediction methods are used, the final physiological primitive characteristic values ​​are uniformly denoted as... The value ranges from 0 to 1.

[0029] In step 122, the raw feature values ​​of the psychological dimension are extracted from the patient self-report data. The patient self-report data contains two types of information: one is the scores of structured scales, such as the Self-Rating Anxiety Scale (SAS) score and the Self-Rating Depression Scale (SDS) score. Each scale item is linearly transformed to obtain a score from 0 to 100, and the arithmetic mean of the two scores is taken as the raw total score of the psychological scale. Another type is unstructured speech emotion data, which consists of speech signals recorded by patients through an adaptive dialogue system while answering questions. Acoustic features such as Mel-frequency cepstral coefficients are extracted from the speech signals and input into a support vector machine classifier. This classifier is trained on clinically labeled anxiety levels and outputs speech emotion feature scores. The value ranges from 0 to 1, with higher scores indicating a more pronounced anxiety tendency. The original characteristic value of the psychological dimension is composed of the two components mentioned above, denoted as... .

[0030] In step 123, the original feature values ​​of the social dimension are obtained by performing entity recognition processing on the family interview text data. A pre-trained language model such as BERT is used to perform named entity recognition on the family interview text. This model can identify entity types related to caregiving, including caregiver identity entities such as children, spouse, and caregiver; care frequency entities such as daily, three times a week, and occasional; and economic status entities such as financial hardship, having a pension, and health insurance coverage. The identified entities are then converted into social feature components after rule matching. Each component takes a value of either 0 or 1. For example, if the entity "live alone" appears in the text, then... Otherwise, it is 0; entities experiencing "economic hardship" are... Otherwise, it is 0; if "children visit every day" appears, then... Otherwise, it is 0. All social feature components constitute the original feature vector of the social dimension. .

[0031] In step 124, the raw feature values ​​of the spiritual dimension are jointly extracted from open-ended response texts in the patient self-report data and family interview texts. Sentiment polarity analysis and topic modeling are performed on these two types of texts. Sentiment polarity analysis uses a dictionary-based method to calculate the frequency difference between positive and negative sentiment words in the text, obtaining a sentiment tendency score. Topic modeling uses the Latent Dirichlet Allocation method, with preset topics including: existential distress, lack of meaning in life, unfulfilled desires, and ritual needs. For each text, the model outputs the probability distribution of each topic. The sentiment polarity score is weighted and summed with the probability of each topic to obtain the raw feature values ​​of the spiritual dimension. The value ranges from 0 to ,in This is the preset maximum possible value, typically 100. This characteristic value reflects the overall intensity of the patient's spiritual needs.

[0032] In an optional embodiment of the present invention, step 13, which involves standardizing the original multidimensional feature values ​​to obtain a multidimensional standardized score, may include: Step 131, according to the formula: , obtained the The physiological standardized score of each patient; in, For the first The physiological standardized score of each patient; For the first The original feature values ​​of the physiological dimensions of each patient; The preset physiological sensitivity coefficient has a value range of 2 to 5. The preset physiological half-activation point has a value of 0.3 to 0.7. Step 132, according to the formula: , Get the first The standardized psychological scores of each patient; among them. For the first The standardized psychological scores of each patient; The preset weighting coefficients for the psychological scale range from 0.5 to 0.8. The preset voice emotion sensitivity coefficient has a value range of 1 to 4; The preset semi-activation point for voice emotion has a value range of 0.4 to 0.6; For the first The original total score of the psychological scale for each patient is obtained by summing the scale scores in the original feature values ​​of the psychological dimension, with a value range of 0 to 100. For the first The voice emotion feature score of each patient is obtained by extracting emotion features from the voice signal in the patient's self-assessment data, and the value ranges from 0 to 1. Step 133, according to the formula: , The social standardization score is obtained; among which, For the first The social standardized score of each patient; The preset social low demand threshold has a value range of 10 to 20. The preset threshold for high social demand ranges from 70 to 85. For the first The original feature values ​​of the social dimension of each patient , For the first Preset weights for each social characteristic component. For the first The first patient's A social characteristic component; The total number of social characteristic components; Step 134, according to the formula: , obtained the The standardized spiritual scores of each patient; among them. For the first The standardized spiritual score for each patient, and when the calculated result exceeds 100, Take 100; For the first The original feature values ​​of the spiritual dimension of each patient; This is the maximum possible value of the preset spiritual characteristics, ranging from 50 to 10.

[0033] In step 131 of this embodiment, the original physiological characteristic values ​​are... The mapping is to a standardized physiological score ranging from 0 to 100. This mapping process exhibits an S-shaped curve characteristic, when... When approaching 0, Approaching 0; when When it approaches 1, Approaching 100; when hour, .parameter The steepness of the control curve is adjusted; a higher value results in a higher score. The more sensitive the changes, the better. In clinical practice, A standardized score of 50 points represents the probability of moderate pain. When the probability of pain changes from 0.3 to 0.7, the standardized score increases from about 12 points to 88 points, which is consistent with the nonlinear characteristics of pain perception.

[0034] In step 132, the standardized psychological score integrates two sources: scale scores and vocal emotional features. (Scale portion) The contribution is direct and linear, ranging from 0 to 100, reflecting the patient's self-reported psychological state. The emotional component of speech is transformed by a logistic function, where... Controlling the sensitivity of speech emotion features to scores, As a semi-activation point for voice emotion, when This section contributes 50 points. Weighting coefficient. The weighting of the scale and voice is adjusted between 0.5 and 0.8, with a typical value of 0.7, meaning that the scale score is dominant and the voice emotion serves as a secondary correction, avoiding bias from a single data source. For example, a patient may have a high scale score but a calm voice emotion, yet the overall score is still high, reflecting the importance placed on subjective self-assessment.

[0035] In step 133, a piecewise quadratic function is used to process the original feature values ​​of the social dimension. The function is designed to output 0 points in the low-demand region, 100 points in the high-demand region, and use a square-law growth in the intermediate region. The square-law function ensures slow growth in the low-demand region, as... near The accelerating rate of increase reflects the cumulative effect of social vulnerability. (Preset threshold) Mild social needs are not scored. This indicates that a high level of demand equates to a perfect score.

[0036] In step 134, the original spiritual eigenvalues ​​are compressed using the square root function. The dynamic range. With Increase, The growth rate gradually slows down, meaning that initial spiritual clues contribute more to the score, while the marginal contribution of subsequent new clues decreases. hour ;when hour If the calculated result exceeds 100, it will be truncated to 100 to ensure a reasonable score range. For example... ,patient hour ; hour This mapping aligns with clinical experience that spiritual needs are easily identifiable in the early stages but require comprehensive judgment in later stages.

[0037] In an optional embodiment of the present invention, step 14, determining the multidimensional weight vector based on the multidimensional standardized score, may include: Step 141: Determine the information entropy data based on the multidimensional standardized score; Step 142: Determine objective weight data based on the information entropy data; Step 143: Obtain the preset subjective weight vector data; Step 144: Linearly weight the objective weight data and the subjective weight vector data to obtain a multidimensional weight vector.

[0038] In step 141 of this embodiment, the information entropy data is calculated based on the standardized scores of the current patient group across various dimensions. Let there be a total of... Each patient has standardized scores across four dimensions, including: For physiological standardized scores, For psychological standardized scores, As a social standardization score, This is a standardized score for spirituality. For consistent expression, it is recorded as follows: , , , ,Right now Indicates the first The patient in The standardized scores for each dimension, among which These correspond to physiological, psychological, social, and spiritual aspects, respectively. First, calculate the... The patient in The percentage of scores for each dimension: ; When a To avoid undefined logarithmic operations, we will... Replace with the smallest positive number ,at this time The approximation does not affect the entropy calculation result. Then calculate the first... Information entropy in one dimension: ; Information entropy The value ranges from 0 to 1. The larger the entropy value, the smaller the difference in the score of this dimension among patients, that is, the weaker the distinguishing ability of this dimension.

[0039] In step 142, the objective weight data is determined based on the degree of variation of information entropy. The degree of variation is defined as follows: The smaller the entropy value, the greater the degree of variation, and this dimension should be assigned a higher objective weight. Objective weight The calculation formula is: ; This formula makes the sum of the objective weights of each dimension equal to 1. The objective weights reflect the discrete nature of the data itself. For example, if the physiological standardized scores of all patients vary greatly, the objective weight of the physiological dimension will be higher.

[0040] In step 143, the pre-defined subjective weight vector data is determined by clinical experts using the analytic hierarchy process (AHP). Experts perform pairwise comparisons of the relative importance of the four dimensions, construct a judgment matrix, calculate the eigenvectors, and perform a consistency check. The subjective weight vector is denoted as... The sum of all components is 1. Typically, experts consider physical pain to be the most urgent, followed by spiritual needs, and then psychological and social needs. For example... , , , Subjective weighting reflects clinical consensus and ethical priorities, and does not change with variations in the patient population.

[0041] In step 144, the objective weights and subjective weights are combined using a linear weighted combination to obtain the final multidimensional weight vector. The combination formula is: ; in, This is a preset objective weighting coefficient, ranging from 0 to 1, with a typical value of 0.5. When... When entirely subjective weighting is used, when The weights are calculated using entirely objective weights. The weights for the four dimensions are... The sum of these values ​​equals 1, and they serve as the weight vector for subsequent calculations of the comprehensive demand index. This combined weighting method utilizes the statistical characteristics of the current patient population while preserving the value judgments of clinical experts.

[0042] In an optional embodiment of the present invention, step 15, determining the initial comprehensive needs data of multiple terminally ill patients based on the multidimensional standardized scores and multidimensional weight vectors, may include: Step 151, according to the formula: Determine the first Initial comprehensive needs data for each patient; in, For the first Initial comprehensive needs data for each patient; Weights for physiological dimensions; Weights for psychological dimensions; Weighting for the social dimension; Weighting for the spiritual dimension; For the first The physiological standardized score of each patient; For the first The standardized psychological scores of each patient; For the first The social standardized score of each patient; For the first The patient's standardized spiritual score.

[0043] In step 151 of this embodiment, the initial comprehensive demand data is fused from the standardized scores of the four dimensions using a weighted summation method. For each component of the final multidimensional weight vector, the following conditions must be met: ; The first Standardized scores for each patient's physiological, psychological, social, and spiritual aspects, each ranging from 0 to 100. Weighted summation formula: The significance is that it linearly superimposes the intensity of demand from the four dimensions according to their respective importance, resulting in a single quantitative indicator reflecting the urgency of the patient's overall needs. Since all standardized scores belong to the 0-100 scale, their weights sum to 1. The value range is also from 0 to 100, with higher scores indicating a more urgent need for comprehensive care. The linear weighting method has the advantages of simple calculation and strong interpretability; the contribution of each dimension can be analyzed independently, facilitating subsequent prioritization and resource allocation. For example, if a patient's physiological standardized score... Psychological standardized score Social standardization score Standardized spirituality score With a weight vector of [0.32, 0.23, 0.18, 0.27], the initial comprehensive demand data is calculated as follows: This score, falling between medium and high, suggests that the patient's spiritual needs and physical pain were the primary contributing factors. Initial comprehensive needs data. As a transitional result, it will be fed into subsequent correction steps. It should be noted that the weighted summation implicitly assumes the compensability between dimensions, that is, a low score in one dimension can be compensated by a high score in another dimension.

[0044] In an optional embodiment of the present invention, step 16, revising the initial comprehensive needs data of the plurality of terminally ill patients to obtain target comprehensive needs data of the plurality of terminally ill patients, may include: Step 161, according to the formula: Determine the first Comprehensive target needs data for each patient; in, For the first Comprehensive target needs data for each patient; No. Initial comprehensive needs data for each patient; For correction factor, The formula is: ; in, The preset correction amplitude coefficient ranges from 0.05 to 0.15. For the first The physiological standardized score of each patient; For the first The patient's standardized spiritual score; The preset social threshold; For the first The social standardized score of each patient; This is the function for finding the maximum value.

[0045] In step 161 of this embodiment, the target integrated demand data is obtained by applying a correction coefficient to the initial integrated demand data. In the correction formula... middle, As a correction factor, when a patient simultaneously experiences weak social support and prominent physiological or spiritual needs, Greater than 0, otherwise 0. Specifically, the correction condition is the social standardization score. Below the preset social threshold And physiological standardized score Higher than the preset physiological threshold Or spiritual standardized score Above the preset spiritual threshold Here The threshold for distinguishing between normal and weak social support is typically set at 30. and These values ​​are used to determine whether pain or spiritual distress has reached a moderate or higher level, with a typical value of 50 for both.

[0046] When the correction condition is met, the correction coefficient is calculated using the following formula: ; in, To correct the amplitude coefficient, its value ranges from 0.05 to 0.15, with a typical value of 0.10, and is used to control the overall upward fluctuation. Factor The maximum values ​​of physiological or spiritual scores were normalized to a range of 0 to 1. Higher scores indicate a larger factor, reflecting that vulnerable patients with high needs should receive a greater boost. Factor Social support score The lower the value, the larger the value. When the factor is 1, When the factor is 0, a positive correlation is achieved between the correction coefficient and the degree of weakness in social support.

[0047] Multiply the product of the two factors by The final correction coefficient is obtained. For example, let... , , , A patient , , Then it satisfies and ,calculate , , , This means the initial comprehensive demand index will increase by 4%. If the same patient... ,but , The target composite demand index rose by 6.7%. This will be used for subsequent ranking, ensuring that patients with weak social support but also high physical or spiritual needs receive an appropriate boost in the priority queue, thus avoiding underestimation of needs due to insufficient social support. If the social support score is above the threshold, or if it is below the threshold but neither the physical nor spiritual scores exceed their respective thresholds, then... , No corrections will be made.

[0048] In an optional embodiment of the present invention, step 17, which involves ranking the multiple terminally ill patients according to their comprehensive target needs data to obtain a patient care priority ranking result, may include: Step 171: Based on the numerical value of the target comprehensive demand data, the multiple patients are sorted in descending order to obtain a patient care priority queue; among them, patients ranked higher indicate that their care needs are more urgent and they are given priority in receiving care resource allocation.

[0049] In step 171 of this embodiment, the comprehensive requirement data of each objective is... Assuming there are currently a total of [number] [items / items], this will be used as the sorting criterion. Each patient, whose target comprehensive needs index constitutes a set. Sort the set in descending order, that is, rearrange the patients according to their numerical values ​​from largest to smallest, to obtain the patient care priority queue. ,in The patient with the highest overall target demand index. Secondly, and so on. For cases where the values ​​are equal, a secondary priority indicator can be used as an auxiliary comparison, such as the initial comprehensive demand index or the physiological standardized score. However, in general, the difference between the target comprehensive demand indices is sufficient to distinguish the priorities.

[0050] The priority queue signifies that patients ranked higher in the queue have a more urgent need for comprehensive care. Clinical palliative care teams can allocate limited care resources sequentially according to this queue order, for example, prioritizing pain management, psychological counseling, or spiritual support services for the patient at the top of the queue. Patients at the back of the queue indicate relatively milder current needs, allowing for delayed intervention or routine monitoring. This ranking method is significantly superior to traditional experience-based judgment because it integrates objective data from four dimensions: physiological, psychological, social, and spiritual, and eliminates information gaps between dimensions through dynamic weights and correction factors. Taking three patients as an example, if the calculation yields… , , The priority queue is then Patient 1, Patient 2, and Patient 3. When resources are limited, the care team should prioritize addressing Patient 1's spiritual distress and combined pain intervention, followed by Patient 2's social support and pain management. Patient 3 can be addressed later. Furthermore, the queue results can be presented using visualization tools, such as bar charts displaying each patient's need index with radar charts showing scores for each dimension, facilitating the team's understanding of the prioritization criteria. The final output of the priority ranking results is structured data, which can be imported into electronic medical record systems or palliative care information platforms to achieve closed-loop management from assessment to intervention.

[0051] A specific embodiment of the method for processing data on the needs of terminally ill patients provided in this invention is as follows: Step 1: Initial data acquisition from multiple sources.

[0052] Take, for example, three terminally ill patients admitted to a palliative care ward.

[0053] Patient A was able to self-assess. An adaptive dialogue system presented a pain rating scale, to which the patient answered "7 points." An anxiety self-assessment scale was also presented, resulting in a score of 65 points. The open-ended question, "What are you most worried about right now?" was answered with "Life is meaningless." Simultaneously, recordings of interviews with family members were collected and transcribed, revealing "Daily care from children, financially secure."

[0054] Patient B is unable to speak. Heart rate variability signals were collected using a smart bracelet at a sampling rate of 50Hz, and the ratio of low-frequency power to high-frequency power was calculated. The transcript of the family interview revealed that "the patient frequently groans at night, has disrupted sleep, lives alone, and is only cared for by a caregiver."

[0055] Patient C was able to self-assess, scoring 4 points for pain and 30 points for anxiety. In an open-ended response, they mentioned "hoping to write a will to leave to their children." These three types of data were stored as patient self-assessment data, physiological signal data, and family interview transcripts, respectively.

[0056] Step 2: Extraction of original feature values ​​for physiological dimensions.

[0057] For patients A and C, the numerical pain scores were directly read from the self-report data, divided by 10, and converted to the raw physiological characteristic values ​​in the range of 0 to 1. , .

[0058] For patient B, the heart rate variability characteristic value Substitute into the equation: ,in , Calculated The original physiological characteristic values ​​are denoted as follows: , , .

[0059] Step 3: Extraction of original feature values ​​for the psychological dimension.

[0060] The raw total score of the psychological scale was extracted from the patient's self-report data.

[0061] Patient A's anxiety self-rating score was 65, and their depression self-rating score was within the acceptable range (50 points). The average score was... ; Patient C scored 30 on both anxiety and depression self-ratings, with an average score of 30. ; Patient B's answers were submitted by a family member, resulting in missing scores on the scale; therefore, this is tentatively set as... That is, the level is moderate.

[0062] Simultaneously, Mel-frequency cepstral coefficients were extracted from the patient's speech and input into an SVM classifier to obtain speech emotion feature scores: Patient A's speech was calm. Patient B is unable to speak. Patient C's speech was steady. The original psychological characteristic values ​​consist of scale scores and voice scores.

[0063] Step 4: Extraction of original feature values ​​for the social dimension.

[0064] BERT entity recognition was performed on the family interview texts.

[0065] Patient A was identified as having "daily caregivers from their children" in the text, indicating no financial difficulties, and their social feature vector was taken. (Living alone) (Economic difficulties) (High-frequency care).

[0066] Patient B's text was identified as "living alone", "cared for by only a caregiver", and "no children". , (Assumption of economic hardship) .

[0067] Patient C's family did not provide any specific information, assuming good social support. , , The original feature vectors of society are denoted as follows: , , .

[0068] Step 5: Extraction of original feature values ​​for the spiritual dimension.

[0069] Emotional polarity analysis and topic modeling were performed on open-ended responses to patient self-reports and family interview texts.

[0070] Patient A's answer "Life is meaningless" categorized under "existential distress" with a probability of 0.9, indicating extremely negative emotions and primal spiritual characteristics. .

[0071] Patient B's family described him as "constantly wanting to see his grandson," with a probability of 0.7 for the theme "unfulfilled wishes." .

[0072] Patient C mentioned "hoping to finish writing their will," with a probability of 0.6 for the theme "unfulfilled wishes." Preset .

[0073] Step 6: Calculate the multidimensional standardized score.

[0074] A differential mapping formula is used.

[0075] Physiological dimension: ,Pick , ,have to , , .

[0076] Psychological dimension: ,Pick , , . Patient A: Patient B: Patient C: .

[0077] Social dimension: Using piecewise functions, setting... , The original social characteristic values ​​need to be converted into a weighted sum first. Let the weights be... (Living alone) (Economic difficulties) (High-frequency companion care reversed number). Calculation , , .but (because ), (because ), .

[0078] Spiritual dimension: ,Pick ,have to , , After truncation , , .

[0079] Step 7: Determine the dynamic weights.

[0080] Based on the standardized scores of the three patients, calculate the information entropy of each dimension. First, construct a matrix. :physiological ,psychology ,society spirituality Calculate the proportion of each dimension. Entropy Then obtain objective weights Calculations showed that physiological entropy values ​​were relatively low, with an objective weight of approximately 0.30; psychological entropy values ​​were moderate, with an objective weight of 0.25; social entropy values ​​were relatively high, with an objective weight of 0.20; and spiritual entropy values ​​had an objective weight of 0.25. The expert's subjective weight was preset to... .Pick Combined weights , , , The sum is 1.

[0081] Step 8: Calculate the initial comprehensive demand index.

[0082] Weighted summation: ; ; .

[0083] Step 9, Dimensional Coupling Correction.

[0084] Setting social thresholds Physiological threshold spiritual threshold Correction amplitude coefficient .

[0085] Patient A: ,and The correction conditions are met. , .

[0086] Patient B: The conditions are not met. .

[0087] Patient C: ,but and ,satisfy Therefore, it has been revised. , .

[0088] Step 10, Priority sorting.

[0089] The three patients were ranked in descending order of their target comprehensive needs index: , , The priority queue is patient B, patient A, and patient C. Output explanation: Patient B has extremely poor social support but the highest physiological needs, so priority should be given to strengthening social support and pain management. Patient A experiences significant spiritual distress and requires spiritual care combined with pain intervention. Patient C's needs are relatively mild, and arrangements can be made to complete the will.

[0090] The sorting results were consistent with the clinical judgment of the experts, and the overall calculation time was completed within 15 minutes.

[0091] This invention proposes the above-mentioned technical solution, which extracts four-dimensional original features (physiological, psychological, social, and spiritual) by collecting multi-source data such as patient self-assessment data, wearable device physiological signals, and family interview texts. Standardization is performed using differentiated formulas such as mapping, weighted fusion, piecewise quadratic functions, and square root compression. Dynamic weights are determined using an entropy-weighted analytic hierarchy process (AHP) to calculate a weighted comprehensive demand index and introduce a dimensional coupling correction factor. Finally, a priority queue is generated based on the descending order of the corrected index. This solution reduces the single assessment time from approximately 45 minutes manually to less than 15 minutes, and the consistency between the priority ranking results and expert consensus reaches over 92%. For patients unable to express themselves verbally, the accuracy of the assessment is improved by approximately 25% by predicting pain using heart rate variability signals. This solution provides palliative care teams with objective and quantitative criteria for demand prioritization, helping to rationally allocate care resources under limited conditions.

[0092] like Figure 2 As shown, this embodiment of the invention also provides a data processing device 20 for end-of-life patient needs, comprising: The acquisition module 21 is used to acquire multi-source initial data of multiple terminally ill patients, including patient self-assessment data, patient physiological signal data, and family interview text data. Processing module 22 is used to extract features from the multi-source initial data to obtain multi-dimensional original feature values, including original feature values ​​of physiological dimension, psychological dimension, social dimension, and spiritual dimension; perform standardized mapping processing on the multi-dimensional original feature values ​​to obtain multi-dimensional standardized scores; determine multi-dimensional weight vectors based on the multi-dimensional standardized scores; determine initial comprehensive needs data for multiple terminally ill patients based on the multi-dimensional standardized scores and multi-dimensional weight vectors; correct the initial comprehensive needs data for multiple terminally ill patients to obtain target comprehensive needs data for multiple terminally ill patients; and rank the multiple patients according to the target comprehensive needs data for multiple terminally ill patients to obtain a patient care priority ranking result.

[0093] Optionally, module 21 is specifically used for: The patient's self-assessment data is obtained by collecting the patient's or family's answers to preset questions through an adaptive dialogue system. By using wearable devices, the patient's heart rate variability time-series signal is collected to obtain the patient's physiological signal data; The audio recordings of family interviews were processed to convert speech to text, resulting in the text data of the family interviews.

[0094] Optionally, processing module 22 is specifically used for: The patient's self-assessment data is processed to extract pain data, or the patient's physiological signal data is processed to transform the data to determine the original feature values ​​of the physiological dimensions. The patient self-assessment data was processed to extract psychological data and obtain the original feature values ​​of the psychological dimension. Entity recognition processing was performed on the family interview text data to obtain the original feature values ​​of the social dimension; The patient self-assessment data and family interview text data were processed to extract emotional data, and the original feature values ​​of the spiritual dimension were obtained.

[0095] Optionally, the processing module 22 is also specifically used for: According to the formula: , obtained the The physiological standardized score of each patient; in, For the first The physiological standardized score of each patient; For the first The original feature values ​​of the physiological dimensions of each patient; The preset physiological sensitivity coefficient has a value range of 2 to 5. The preset physiological half-activation point has a value of 0.3 to 0.7. According to the formula: , Get the first The standardized psychological scores of each patient; among them. For the first The standardized psychological scores of each patient; The preset weighting coefficients for the psychological scale range from 0.5 to 0.8. The preset voice emotion sensitivity coefficient has a value range of 1 to 4; The preset semi-activation point for voice emotion has a value range of 0.4 to 0.6; For the first The original total score of the psychological scale for each patient is obtained by summing the scale scores in the original feature values ​​of the psychological dimension, with a value range of 0 to 100. For the first The voice emotion feature score of each patient is obtained by extracting emotion features from the voice signal in the patient's self-assessment data, and the value ranges from 0 to 1. According to the formula: , The social standardization score is obtained; among which, For the first The social standardized score of each patient; The preset social low demand threshold has a value range of 10 to 20. The preset threshold for high social demand ranges from 70 to 85. For the first The original feature values ​​of the social dimension of each patient , For the first Preset weights for each social characteristic component. For the first The first patient's A social characteristic component; The total number of social characteristic components; According to the formula: , obtained the The standardized spiritual scores of each patient; among them. For the first The standardized spiritual score for each patient, and when the calculated result exceeds 100, Take 100; For the first The original feature values ​​of the spiritual dimension of each patient; This is the maximum possible value of the preset spiritual characteristics, ranging from 50 to 10.

[0096] Optionally, the processing module 22 is also specifically used for: Based on the aforementioned multidimensional standardized scores, the information entropy data is determined; Based on the information entropy data, determine the objective weight data; Obtain the preset subjective weight vector data; The objective weight data and the subjective weight vector data are linearly weighted to obtain a multidimensional weight vector.

[0097] Optionally, the processing module 22 is also specifically used for: According to the formula: Determine the first Initial comprehensive needs data for each patient; in, For the first Initial comprehensive needs data for each patient; Weights for physiological dimensions; Weights for psychological dimensions; Weighting for the social dimension; Weighting for the spiritual dimension; For the first The physiological standardized score of each patient; For the first The standardized psychological scores of each patient; For the first The social standardized score of each patient; For the first The patient's standardized spiritual score.

[0098] Optionally, the processing module 22 is also specifically used for: According to the formula: Determine the first Comprehensive target needs data for each patient; in, For the first Comprehensive target needs data for each patient; No. Initial comprehensive needs data for each patient; For correction factor, The formula is: ; in, The preset correction amplitude coefficient ranges from 0.05 to 0.15. For the first The physiological standardized score of each patient; For the first The patient's standardized spiritual score; The preset social threshold; For the first The social standardized score of each patient; This is the function for finding the maximum value.

[0099] Optionally, the processing module 22 is also specifically used for: Based on the numerical values ​​of the target comprehensive demand data, the multiple patients are sorted in descending order to obtain a patient care priority queue; among them, patients ranked higher indicate that their care needs are more urgent and they are given priority in receiving care resource allocation.

[0100] It should be noted that this device is a device corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0101] like Figure 3 As shown, this embodiment of the invention also provides a computing device 30, including a processor 31, a memory 32, and a program or instructions stored in the memory 32 and executable on the processor 31. When the program or instructions are executed by the processor 31, they implement the various processes of the above-described embodiment of the method for processing data on the needs of terminally ill patients, and achieve the same technical effects. To avoid repetition, they will not be described again here. It should be noted that the computing device in this embodiment of the invention includes the aforementioned mobile electronic devices and non-mobile electronic devices.

[0102] The above are preferred embodiments of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for processing data on the needs of terminally ill patients, characterized in that, include: Acquire multi-source initial data from multiple terminally ill patients, including patient self-assessment data, patient physiological signal data, and family interview text data; Feature extraction is performed on the multi-source initial data to obtain multi-dimensional original feature values, which include original feature values ​​of physiological dimension, psychological dimension, social dimension and spiritual dimension. The original multidimensional feature values ​​are standardized and mapped to obtain multidimensional standardized scores. Based on the multidimensional standardized scores, determine the multidimensional weight vector; Based on the multidimensional standardized scores and multidimensional weight vectors, the initial comprehensive needs data for multiple terminally ill patients are determined; The initial comprehensive needs data of the multiple terminally ill patients are corrected to obtain the target comprehensive needs data of the multiple terminally ill patients; Based on the comprehensive target needs data of multiple terminally ill patients, the patients are ranked to obtain the patient care priority ranking result.

2. The method for processing terminally ill patient needs data according to claim 1, characterized in that, Obtain initial data from multiple sources from several terminally ill patients, including: The patient's self-assessment data is obtained by collecting the patient's or family's answers to preset questions through an adaptive dialogue system. By using wearable devices, the patient's heart rate variability time-series signal is collected to obtain the patient's physiological signal data; The audio recordings of family interviews were processed to convert speech to text, resulting in the text data of the family interviews.

3. The method for processing data on the needs of terminally ill patients according to claim 1, characterized in that, Feature extraction is performed on the multi-source initial data to obtain multi-dimensional original feature values, including: The patient's self-assessment data is processed to extract pain data, or the patient's physiological signal data is processed to transform the data to determine the original feature values ​​of the physiological dimensions. The patient self-assessment data was processed to extract psychological data and obtain the original feature values ​​of the psychological dimension. Entity recognition processing was performed on the family interview text data to obtain the original feature values ​​of the social dimension; The patient self-assessment data and family interview text data were processed to extract emotional data, and the original feature values ​​of the spiritual dimension were obtained.

4. The method for processing data on the needs of terminally ill patients according to claim 1, characterized in that, The multidimensional original feature values ​​are standardized and mapped to obtain multidimensional standardized scores, including: According to the formula: , obtained the The physiological standardized score of each patient; in, For the first The physiological standardized score of each patient; For the first The original feature values ​​of the physiological dimensions of each patient; The preset physiological sensitivity coefficient has a value range of 2 to 5. The preset physiological half-activation point has a value of 0.3 to 0.

7. According to the formula: , Get the first The standardized psychological scores of each patient; among them. For the first The standardized psychological scores of each patient; The preset weighting coefficients for the psychological scale range from 0.5 to 0.

8. The preset voice emotion sensitivity coefficient has a value range of 1 to 4; The preset semi-activation point for voice emotion has a value range of 0.4 to 0.6; For the first The original total score of the psychological scale for each patient is obtained by summing the scale scores in the original feature values ​​of the psychological dimension, with a value range of 0 to 100. For the first The voice emotion feature score of each patient is obtained by extracting emotion features from the voice signal in the patient's self-assessment data, and the value ranges from 0 to 1. According to the formula: , The social standardization score is obtained; among which, For the first The social standardized score of each patient; The preset social low demand threshold has a value range of 10 to 20. The preset threshold for high social demand ranges from 70 to 85. For the first The original feature values ​​of the social dimension of each patient. , For the first The pre-defined weights of each social characteristic component. For the first The first patient's A social characteristic component; The total number of social characteristic components; According to the formula: , obtained the The standardized spiritual scores of each patient; among them. For the first The standardized spiritual score for each patient, and when the calculated result exceeds 100, Take 100; For the first The original feature values ​​of the spiritual dimension of each patient; This is the maximum possible value of the preset spiritual characteristics, ranging from 50 to 10.

5. The method for processing data on the needs of terminally ill patients according to claim 1, characterized in that, Based on the aforementioned multidimensional standardized scores, a multidimensional weight vector is determined, including: Based on the aforementioned multidimensional standardized scores, the information entropy data is determined; Based on the information entropy data, determine the objective weight data; Obtain the preset subjective weight vector data; The objective weight data and the subjective weight vector data are linearly weighted to obtain a multidimensional weight vector.

6. The method for processing data on the needs of terminally ill patients according to claim 1, characterized in that, Based on the multidimensional standardized scores and multidimensional weight vectors, initial comprehensive needs data for multiple terminally ill patients are determined, including: According to the formula: Determine the first Initial comprehensive needs data for each patient; in, For the first Initial comprehensive needs data for each patient; Weights for physiological dimensions; Weights for psychological dimensions; Weighting for the social dimension; Weighting for the spiritual dimension; For the first The physiological standardized score of each patient; For the first The standardized psychological scores of each patient; For the first The social standardized score of each patient; For the first The patient's standardized spiritual score.

7. The method for processing data on the needs of terminally ill patients according to claim 1, characterized in that, The initial comprehensive needs data of the multiple terminally ill patients were revised to obtain the target comprehensive needs data of the multiple terminally ill patients, including: According to the formula: Determine the first Comprehensive target needs data for each patient; in, For the first Comprehensive target needs data for each patient; No. Initial comprehensive needs data for each patient; For correction factor, The formula is: ； in, The preset correction amplitude coefficient ranges from 0.05 to 0.

15. For the first The physiological standardized score of each patient; For the first The patient's standardized spiritual score; The preset social threshold; For the first The social standardized score of each patient; This is the function for finding the maximum value.

8. The method for processing terminally ill patient needs data according to claim 1, characterized in that, Based on the comprehensive target needs data of the multiple terminally ill patients, the patients are ranked to obtain a patient care priority ranking result, including: Based on the numerical values ​​of the target comprehensive demand data, the multiple patients are sorted in descending order to obtain a patient care priority queue; among them, patients ranked higher indicate that their care needs are more urgent and they are given priority in receiving care resource allocation.

9. A device for processing data on the needs of terminally ill patients, characterized in that, include: The acquisition module is used to acquire multi-source initial data from multiple terminally ill patients, including patient self-assessment data, patient physiological signal data, and family interview text data. The processing module is used to extract features from the multi-source initial data to obtain multi-dimensional original feature values, which include original feature values ​​of physiological dimension, original feature values ​​of psychological dimension, original feature values ​​of social dimension and original feature values ​​of spiritual dimension. The original multidimensional feature values ​​are standardized and mapped to obtain multidimensional standardized scores; multidimensional weight vectors are determined based on the multidimensional standardized scores; initial comprehensive needs data for multiple terminally ill patients are determined based on the multidimensional standardized scores and multidimensional weight vectors; the initial comprehensive needs data for multiple terminally ill patients are corrected to obtain target comprehensive needs data for multiple terminally ill patients; and multiple patients are ranked according to the target comprehensive needs data for multiple terminally ill patients to obtain a patient care priority ranking result.

10. A computing device, characterized in that, include: A processor, a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any one of claims 1 to 8.

Images