Medical staff humanistic care perception degree identification method and system-level storage medium

Abstract

The invention discloses a medical personnel humanistic care perception degree identification method and a system-level storage medium, and relates to the technical field of medical personnel capability assessment. A low-annuity nurse exclusive narrative scene library is constructed, and a multi-dimensional data acquisition and preprocessing technology is combined, so that the medical personnel humanistic care perception degree is identified. According to the method, the perception degree of nurses in humanistic care can be comprehensively and accurately evaluated, the system can identify the advantages and deficiencies of the nurses in the aspects of estrus sharing ability, communication consciousness, respect consciousness, responsibility consciousness and the like through multi-dimensional feature extraction and fusion, and a personalized feedback report is generated; the reports not only provide clear self-cognition for nurses, but also make targeted training suggestions for nursing managers, thereby promoting personalized improvement and development of humanistic care ability of low-annuity nurses.

Description

Technical Field

[0001] This invention relates to the field of medical staff competency assessment technology, specifically to a method for identifying the degree of humanistic care perception among medical staff and a system-level storage medium. Background Technology

[0002] With the continuous development of medical services, patients' demand for humanistic care in nursing services is increasing. The humanistic care ability of medical staff, especially their empathy, communication awareness, respect, and sense of responsibility, directly affects patients' medical experience and satisfaction.

[0003] Traditional assessments of humanistic care provided by healthcare workers primarily rely on subjective evaluations or single-dimensional observations, which have the following limitations:

[0004] Assessment bias: Subjective evaluations are easily influenced by emotional factors and cannot quantify nurses' performance in specific dimensions such as empathy and sense of responsibility; observation of a single dimension ignores the correlation between behavior and situation, leading to distorted assessment results;

[0005] Poor dynamic adaptability: Clinical humanistic care scenarios are complex and ever-changing, and traditional methods are unable to dynamically adjust assessment criteria according to the difficulty of the scenario, resulting in a disconnect between assessment results and actual capabilities;

[0006] Insufficient training support: The assessment results are mostly summative scores, lacking specific analysis of nurses' skill gaps and recommendations for targeted training resources, thus failing to form a closed loop of assessment-feedback-improvement;

[0007] Low data reliability: Traditional data collection relies on manual recording, which carries the risk of omissions and errors, and outliers are not scientifically processed, affecting the accuracy of the assessment.

[0008] In response to the shortcomings of traditional technologies, this invention proposes a method for identifying the level of humanistic care perceived by medical staff, and the system-level storage medium is particularly important. Summary of the Invention

[0009] The purpose of this invention is to overcome the shortcomings of existing technologies by providing a method for identifying the level of humanistic care perception among medical staff and a system-level storage medium. This method, through the construction of a narrative task library covering high-frequency clinical humanistic care scenarios and combined with multi-channel data collection and preprocessing technologies, can quantify nurses' performance in narrative literacy and core humanistic care characteristics. Furthermore, through a custom feature fusion algorithm and dynamic decision-making model, it achieves graded identification of perception levels and generates feedback reports including strengths analysis, weaknesses identification, and personalized training suggestions. This method overcomes the subjectivity and one-sidedness of traditional assessments, providing a scientific tool for improving the capabilities of junior nurses, while also helping medical institutions optimize the quality of nursing services and build harmonious doctor-patient relationships.

[0010] To solve the above-mentioned technical problems, the present invention provides the following technical solution: Firstly, a method for identifying the degree of humanistic care perception among medical staff, the specific steps of which are as follows:

[0011] Steps for building a dedicated narrative scenario library: Build a dedicated narrative scenario library for junior nurses. Junior nurses are defined as nurses with ≤3 years of work experience. The narrative scenario library is designed based on high-frequency scenarios of clinical humanistic care. Each narrative scenario includes at least one core scenario element from patient communication, disease information disclosure, psychological counseling, and handling of doctor-patient conflicts. Each scenario has 3-5 narrative tasks with different difficulty levels. The difficulty levels are divided according to the patient's cooperation, the complexity of the disease, and the urgency of communication.

[0012] Multi-dimensional data collection steps: Through multi-channel data collection methods, obtain narrative expression data and behavioral feedback data of junior nurses in the process of completing narrative tasks in the narrative scenario library. The narrative expression data includes written narrative text and oral narrative audio transcription text. The behavioral feedback data includes facial expression feature data, voice tone feature data, and operation response time data.

[0013] Raw data preprocessing steps: The collected raw data is preprocessed, including data cleaning, data standardization and data format unification. Among them, data cleaning uses a custom outlier detection algorithm to handle outliers, multiple interpolation to handle missing values, and a custom standardization method to standardize numerical data and convert unstructured text data into structured vector data.

[0014] Multi-dimensional feature extraction and fusion steps: Based on the preprocessed data, the narrative literacy features and humanistic care perception core features of junior nurses are extracted. The narrative literacy features include feature parameters of three dimensions: narrative expression completeness, narrative comprehension accuracy, and narrative reflection depth. The humanistic care perception core features include feature parameters of four dimensions: empathy ability, communication awareness, respect awareness, and responsibility awareness. The feature parameters of each dimension are weighted and fused through a custom feature fusion algorithm to obtain a comprehensive feature vector.

[0015] Humanistic care perception level identification steps: Input the comprehensive feature vector into the preset fusion identification model, and use the fusion identification model to classify and identify the humanistic care perception level of junior nurses, and output the corresponding perception level. The perception level includes four levels: excellent, good, qualified, and needing improvement. The fusion identification model uses a custom decision fusion algorithm to determine the level.

[0016] Personalized feedback report generation steps: Generate a personalized feedback report based on the recognition results. The feedback report includes the scores of each dimension of features, analysis of strengths and weaknesses, and targeted training suggestions.

[0017] Furthermore, the specific implementation steps of the scenario dynamic optimization technology in the construction of the exclusive narrative scenario library are as follows: First, a scenario optimization team is formed to conduct a full-cycle clinical needs survey every 6 months. Through semi-structured interviews, on-the-job observation, and analysis of adverse nursing event cases, the team collects information on new humanistic care scenarios encountered by junior nurses in their actual work, as well as the adaptation issues of existing scenarios. Second, the collected scenario needs are categorized and organized, scored according to three dimensions: scenario frequency, degree of impact on nursing quality, and difficulty for junior nurses to cope. Duplicate or infrequently occurring scenarios with scores below a threshold are eliminated. Core elements of new scenarios are extracted to clarify the patient's physiological state, psychological needs, communication environment, and potential risk points within the scenario. Then, based on the original difficulty level... Based on the criteria for difficulty classification and the core elements of the new scenarios, 3-5 difficulty levels were set for each new scenario. The rationality of the scenarios was verified through pre-testing. Twenty junior nurses from different departments, including internal medicine, surgery, obstetrics and gynecology, and pediatrics, were selected to complete the new scenario tasks. A 5-point Likert scale was used to collect their evaluations of the scenario's authenticity and difficulty suitability. Finally, the optimization team collectively reviewed the new scenarios that passed the pre-test. After approval, they were included in the narrative scenario library. At the same time, scenarios with a usage rate of less than 10% in the past 6 months and those that were deemed out of touch with the current clinical nursing model by experts were deleted from the original scenario library. This ensured that the narrative scenario library always closely reflects the actual work of junior nurses and provides core support for the authenticity and validity of data in subsequent multi-dimensional data collection steps.

[0018] Furthermore, the quality control technology implementation steps in the multi-dimensional data acquisition process are as follows: First, all acquisition devices are uniformly calibrated within 72 hours before data acquisition. Second, a real-time monitoring mechanism is set up in the system background during the acquisition process to dynamically monitor the submission progress of written narrative text, the integrity of audio data, and the clarity of video data. If the submission of written narrative text times out, the interruption of audio data exceeds 5 seconds, or the blurriness of video data exceeds a preset threshold, a pop-up reminder and sound alarm are immediately sent to the acquisition terminal, and the acquisition time and device number of the abnormal data are recorded. Then, for the detected abnormal data, a graded supplementary acquisition mechanism is activated. If the written narrative text is incomplete, the nurse is notified to submit it within 24 hours. If the audio or video data is abnormal, the corresponding narrative task must be re-acquired within 48 hours. The supplementary acquisition data must be marked with a supplementary acquisition identifier and stored in association with the original data. Finally, within 24 hours after the data acquisition is completed, all original data is anonymized. The name, employee number, department, and personal identification information in the data are removed using the AES-256 encryption algorithm. At the same time, a unique data number is generated and encrypted and associated with the nurse's identity to ensure the standardization of the data acquisition process, the reliability of data quality, and the security of patient privacy.

[0019] Furthermore, the custom outlier detection algorithm in the raw data preprocessing step is specifically as follows: for numerical data such as communication duration, speech rate, and response time, an outlier detection formula based on scenario adaptability is used: ,in This is the outlier determination coefficient. The data value to be detected. This is the average of similar data in this scenario, calculated using historical data collected from junior nurses in the same department over the past three months. This represents the standard deviation of similar data in this scenario. For scene weighting coefficients, The data fluctuation correction coefficient, the The determination method is as follows: A two-level weighted assignment is performed based on the urgency of the scenario and the complexity of the illness. In the urgency dimension, extremely urgent scenarios are assigned a value of 1.2, urgent scenarios 1.0, and normal scenarios 0.8. In the illness complexity dimension, complex illnesses are assigned a value of 1.1, moderately complex illnesses 1.0, and simple illnesses 0.9. The result is calculated using a weighted summation formula. This algorithm incorporates scene characteristic weights, which solves the problem of misjudgment caused by traditional outlier detection algorithms ignoring scene differences, and improves the accuracy of data cleaning.

[0020] Furthermore, the custom feature fusion algorithm in the multi-dimensional feature extraction and fusion step specifically involves designing a weighted fusion formula for the multi-dimensional parameters of the core features of narrative literacy and humanistic care perception: ,in For the comprehensive feature vector values, The first for narrative literacy Each dimension of feature parameters, The first perception of humanistic care Each dimension of feature parameters, Weighting of each dimension of narrative literacy The weights of each dimension of humanistic care perception, the and The determination method is as follows: First, a two-layer feature weight judgment matrix is ​​constructed using the analytic hierarchy process (AHP). The first layer consists of the target layer and the criterion layer, while the second layer consists of the criterion layer and the indicator layer. Eight cross-disciplinary experts are invited to conduct pairwise comparisons and scores of the importance of each dimension, and the elements of the judgment matrix are determined using the 1-9 scaling method. Then, the largest eigenvalue and the corresponding eigenvector of the judgment matrix are calculated using the eigenvalue decomposition method. After normalizing the eigenvectors, the initial weights are obtained. Finally, the initial weights are corrected by combining the variance contribution rate of the historical identification data from the past six months, and the final weights are determined. , , , , , , All weights were verified to be valid through consistency testing; the algorithm achieves the organic integration of narrative literacy and humanistic care perception features by scientifically allocating the weights of each dimension, avoiding the one-sided influence of a single dimension feature on the recognition results.

[0021] Furthermore, the custom decision fusion algorithm in the humanistic care perception level identification step is specifically designed as follows: A dynamic weight fusion formula is designed based on the output results of the random forest and the improved BP neural network: ,in This is the quantified value corresponding to the final level of perception. The output quantization value of the improved BP neural network, The algorithm dynamically integrates weights; by dynamically adapting the recognition advantages of different models at each level, it overcomes the limitations of single-model recognition and improves the accuracy and stability of level recognition.

[0022] Furthermore, the specific implementation steps of the precise generation technology in the personalized feedback report generation process are as follows: First, based on the scores of each dimension feature in the multi-dimensional feature extraction and fusion step, a three-dimensional strength-weakness analysis matrix is ​​constructed, with 80 points as the strength threshold and 70 points as the weakness threshold. This clarifies the competency positioning of junior nurses in seven dimensions: narrative expression completeness, narrative comprehension accuracy, narrative reflection depth, empathy, communication awareness, respect awareness, and responsibility awareness, marking the strength, achievement, and weakness dimensions. Second, for the weakness dimensions, a pre-set dynamically updated training resource library is linked. This resource library stores training courses, typical case libraries, practical guides, and assessment question banks according to the seven dimensions. Each resource is marked with the appropriate competency improvement goals, training duration, applicable departments, and difficulty level. Then, combining the junior nurses' work departments, years of service, and specific performance of narrative tasks in the construction steps of the dedicated narrative scenario library, a collaborative filtering algorithm is used to select 3-5 most suitable training resources from the resource library to generate personalized training paths. The training paths are designed according to a scientific process of theoretical learning, case analysis, practical training, and assessment feedback, clearly defining the learning focus, time arrangement, and assessment methods for each stage. Finally, a training progress tracking module is added to the feedback report, setting core indicators for training effectiveness evaluation and providing a monthly review reminder function. At the same time, nursing managers can view nurses' training progress, assessment results, and ability improvement through the system backend, facilitating dynamic adjustments to the training plan based on actual results. This achieves a closed-loop management of identification, feedback, training, and improvement, significantly improving the practicality and operability of the feedback report.

[0023] Secondly, a system for identifying the level of humanistic care perceived by medical staff, which includes:

[0024] The narrative scenario library construction module: Constructs a narrative scenario library specifically for junior nurses. Junior nurses are defined as nurses with ≤3 years of work experience. The narrative scenario library is designed based on high-frequency scenarios of clinical humanistic care. Each narrative scenario includes at least one core scenario element from patient communication, disease information disclosure, psychological counseling, and handling of doctor-patient conflicts. Each scenario sets 3-5 narrative tasks with different difficulty levels. The difficulty levels are divided according to the patient's cooperation, the complexity of the disease, and the urgency of communication.

[0025] The multi-dimensional data acquisition module acquires narrative expression data and behavioral feedback data of junior nurses in the narrative scenario library during the process of completing narrative tasks through multiple data acquisition methods. The narrative expression data includes written narrative text and oral narrative audio transcription text. The behavioral feedback data includes facial expression feature data, voice tone feature data, and operation response time data.

[0026] The raw data preprocessing module preprocesses the collected raw data, including data cleaning, data standardization, and data format unification. The data cleaning uses a custom outlier detection algorithm to handle outliers, multiple interpolation to handle missing values, and a custom standardization method to standardize numerical data, converting unstructured text data into structured vector data.

[0027] The feature extraction and fusion module extracts narrative literacy features and humanistic care perception core features of junior nurses based on preprocessed data. The narrative literacy features include feature parameters in three dimensions: narrative expression completeness, narrative comprehension accuracy, and narrative reflection depth. The humanistic care perception core features include feature parameters in four dimensions: empathy ability, communication awareness, respect awareness, and responsibility awareness. The feature parameters in each dimension are weighted and fused through a custom feature fusion algorithm to obtain a comprehensive feature vector.

[0028] The humanistic care perception and recognition module inputs the comprehensive feature vector into a preset fusion recognition model, and uses the fusion recognition model to classify and identify the humanistic care perception level of junior nurses, and outputs the corresponding perception level level. The perception level level includes four levels: excellent, good, qualified, and needing improvement. The fusion recognition model uses a custom decision fusion algorithm to determine the level.

[0029] The feedback report generation module generates a personalized feedback report based on the recognition results. The feedback report includes the scores of each dimension of features, analysis of strengths and weaknesses, and targeted training suggestions.

[0030] The data storage module stores the exclusive narrative scene library, the collected raw data, the preprocessed data, the extracted feature parameters, the comprehensive feature vector, the recognition results, and the personalized feedback report.

[0031] Thirdly, a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the method for identifying the level of humanistic care perception of medical staff as described in any one of claims 1-6.

[0032] Compared with existing technologies, this method for identifying the level of humanistic care perceived by medical staff and its system-level storage medium have the following beneficial effects:

[0033] I. This invention constructs a narrative scenario library specifically for junior nurses and combines it with multi-dimensional data collection and preprocessing technology to comprehensively and accurately assess nurses' perception of humanistic care. Through multi-dimensional feature extraction and fusion, the system can identify nurses' strengths and weaknesses in empathy, communication awareness, respect, and responsibility, and generate personalized feedback reports. These reports not only provide nurses with a clear self-awareness but also provide nursing managers with targeted training suggestions, thereby promoting the personalized improvement and development of junior nurses' humanistic care capabilities.

[0034] Second, the method and system of this invention, by scientifically and objectively assessing the level of humanistic care perception of medical staff, helps medical institutions identify and improve weak links in nursing services. By improving nurses' humanistic care capabilities, the service quality of key aspects such as patient communication, disease information disclosure, psychological counseling, and handling of doctor-patient conflicts can be significantly improved, thereby enhancing patients' medical experience and satisfaction. This optimization of the patient-centered service model helps to build a more harmonious and trusting doctor-patient relationship and improve the overall image and social reputation of medical institutions.

[0035] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0037] Figure 1A flowchart for identifying the level of humanistic care perceived by medical staff;

[0038] Figure 2 A flowchart for a system to identify the level of humanistic care perceived by medical staff;

[0039] Figure 3 A flowchart of a computer-readable storage medium. Detailed Implementation

[0040] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0041] Example 1

[0042] like Figure 1 As shown, to address the high-frequency humanistic care needs in internal medicine clinical practice, a narrative scenario library specifically for junior nurses was constructed. This scenario uses patient information disclosure as its core element. Considering the characteristics of internal medicine patients—multiple underlying diseases and significant fluctuations in their conditions—three narrative tasks with different difficulty levels were set up. The difficulty levels were divided according to patient cooperation and the complexity of the patient's condition: Level 1 for high patient cooperation and simple condition; Level 2 for moderate patient cooperation and moderately complex condition; and Level 3 for low patient cooperation and complex condition. Scenario optimization involved a full-cycle clinical needs survey every six months, conducted through semi-structured interviews with junior internal medicine nurses, on-site observation of patient information disclosure practices, and analysis of adverse nursing events. Cases related to patient information disclosure were collected, including new scenarios and adaptation issues of existing scenarios. The collected needs were scored according to three dimensions: frequency of occurrence of the scenario, degree of impact on nursing quality, and difficulty for junior nurses to cope with. Duplicate and low-threshold scenarios were eliminated, and the core elements of the new scenarios were extracted. Twenty junior nurses from internal medicine, surgery, obstetrics and gynecology, and pediatrics were selected to complete the new scenario task. Their evaluation of the scenario's authenticity and difficulty suitability was collected using a 5-point Likert scale. After collective review and approval by the scenario optimization team, the patient information disclosure scenario was included in the narrative scenario library, while old scenarios with a usage rate of less than 10% in the past 6 months and that were out of touch with the current internal medicine nursing model were deleted from the original scenario library.

[0043] like Figure 2As shown, a multi-channel data collection method was used to obtain relevant data of the junior internal medicine nurse in the above three difficulty levels of narrative tasks: narrative expression data included written narrative text and oral narrative audio transcription text after completing the task; behavioral feedback data included facial expression feature data, voice tone feature data, and operation response time data during the communication process. Within 72 hours before data collection, all recording devices, video acquisition devices, and text input terminals were uniformly calibrated to ensure data collection accuracy. During the collection process, the system background real-time monitoring mechanism was activated to dynamically track the progress of written narrative text submission, the integrity of audio data, and the clarity of video data. When a timeout was detected in the submission of written narrative text or an interruption in audio data for more than 5 seconds, a pop-up reminder and sound alarm were immediately sent to the acquisition terminal, and the collection time and device number of the abnormal data were recorded. For abnormal data with interrupted audio, a graded re-collection mechanism was activated, notifying the nurse to re-complete the narrative task collection of the corresponding difficulty level within 48 hours. The re-collected data was marked with a re-collection identifier and stored in association with the original data. Within 24 hours of data collection, all raw data is anonymized using the AES-256 encryption algorithm to remove personal identification information such as name, employee number, and department, while generating a unique data number that is encrypted and associated with the nurse's identity.

[0044] like Figure 3 The collected raw data undergoes comprehensive preprocessing: a custom outlier detection algorithm is used to handle outliers in numerical data such as communication duration, speech rate, and response time. The formula is as follows: ,in This is the outlier determination coefficient. The data value to be detected. This represents the average of similar data in this scenario. This represents the standard deviation of similar data in this scenario. For scene weighting coefficients, A data fluctuation correction coefficient is applied to ensure that the data meets the requirements of scenario adaptability; multiple imputation methods are used to supplement and improve missing values ​​in the data; all numerical data are uniformly standardized using a custom standardization method to eliminate differences in units; unstructured data such as written narrative texts and oral narrative recordings are converted into structured vector data that can be used for feature extraction, laying the foundation for subsequent analysis.

[0045] Based on the preprocessed structured data, two core features of the nurse were extracted: first, narrative literacy features, including feature parameters in three dimensions: narrative expression completeness, narrative comprehension accuracy, and narrative reflection depth; second, core features of humanistic care perception, including feature parameters in four dimensions: empathy ability, communication awareness, respect awareness, and responsibility awareness. A custom feature fusion algorithm was used to weight and fuse the above seven dimensions of feature parameters. The weighted fusion formula is as follows: ,in For the comprehensive feature vector values, The first for narrative literacy Each dimension of feature parameters, The first perception of humanistic care Each dimension of feature parameters, Weighting of each dimension of narrative literacy By assigning weights to each dimension of humanistic care perception and comprehensively considering the importance of narrative literacy and each dimension of humanistic care perception, the nurse's comprehensive feature vector was finally obtained.

[0046] The above-mentioned comprehensive feature vectors are input into a preset fusion recognition model. This model uses a custom decision fusion algorithm, combining the outputs of random forest and improved BP neural network. The formula is as follows: ,in This is the quantified value corresponding to the final level of perception. The output quantization value of the improved BP neural network, To dynamically integrate weights, the nurse's level of humanistic care perception is classified and identified. The identification results are matched with the corresponding level from four levels: excellent, good, qualified, and need improvement. Finally, the nurse's level of humanistic care perception in the case of patient information is output as "good".

[0047] Based on the identification results, a personalized feedback report is generated: First, a three-dimensional strengths-weaknesses analysis matrix is ​​constructed based on the feature scores of each dimension, with 80 points as the strength threshold and 70 points as the weakness threshold, clarifying the nurse's competency positioning in seven dimensions—narrative expression completeness, narrative comprehension accuracy, narrative reflection depth, empathy, communication awareness, respect awareness, and responsibility awareness. For the weakness dimension of narrative reflection depth, a pre-set, dynamically updated training resource library is linked. Combining the nurse's internal medicine work background, two years of work experience, and specific performance in narrative tasks, a collaborative filtering algorithm is used to select three most suitable training resources from the resource library. A personalized training path is designed according to a scientific process of theoretical learning, case analysis, practical training, and assessment feedback. The theoretical learning stage focuses on mastering the narrative reflection framework; the case analysis stage breaks down three typical reflection cases; the practical training stage strengthens the application of reflection in simulated scenarios; and the assessment feedback stage evaluates effectiveness through simulated tasks. A training progress tracking module is added to the feedback report, setting core indicators for training effectiveness evaluation and providing a monthly review reminder function. Nursing management personnel can also view the nurse's training progress, assessment results, and competency improvement through the system backend.

[0048] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for identifying the degree of humanistic care perceived by medical staff, characterized in that, The specific steps of this method are as follows: Steps for building a dedicated narrative scenario library: Build a dedicated narrative scenario library for junior nurses. Each narrative scenario includes at least one core scenario element from patient communication, disease information disclosure, psychological counseling, and handling of doctor-patient conflicts. Each scenario also includes 3-5 narrative tasks with different difficulty levels. Multi-dimensional data collection steps: Through multi-channel data collection methods, obtain narrative expression data and behavioral feedback data of junior nurses in the process of completing narrative tasks in the narrative scenario library; Raw data preprocessing steps: The collected raw data is preprocessed, including data cleaning using a custom outlier detection algorithm to handle outliers, multiple interpolation to handle missing values, and a custom standardization method to standardize numerical data, converting unstructured text data into structured vector data; Multi-dimensional feature extraction and fusion steps: Based on the preprocessed data, the narrative literacy features and humanistic care perception core features of junior nurses are extracted. The narrative literacy features include feature parameters of three dimensions: narrative expression completeness, narrative comprehension accuracy, and narrative reflection depth. The humanistic care perception core features include feature parameters of four dimensions: empathy ability, communication awareness, respect awareness, and responsibility awareness. The feature parameters of each dimension are weighted and fused through a custom feature fusion algorithm to obtain a comprehensive feature vector. Humanistic care perception level identification steps: Input the comprehensive feature vector into the preset fusion identification model, and use the fusion identification model to classify and identify the humanistic care perception level of junior nurses, and output the corresponding perception level. The perception level includes four levels: excellent, good, qualified, and needing improvement. The fusion identification model uses a custom decision fusion algorithm to determine the level. Personalized feedback report generation steps: Generate a personalized feedback report based on the recognition results. The feedback report includes the scores of each dimension of features, analysis of strengths and weaknesses, and targeted training suggestions.

2. The method for identifying the degree of humanistic care perception among medical staff according to claim 1, characterized in that, The specific steps of the scenario dynamic optimization technology in the construction of the exclusive narrative scenario library are as follows: First, a scenario optimization team is formed to conduct a full-cycle clinical needs survey every 6 months. Through semi-structured interviews, on-the-job observation, and analysis of adverse nursing event cases, the team collects information on new humanistic care scenarios encountered by junior nurses in their actual work, as well as the adaptation issues of existing scenarios. Second, the collected scenario needs are categorized and organized, scored according to three dimensions: scenario frequency, degree of impact on nursing quality, and difficulty for junior nurses to cope. Duplicate or infrequent scenarios with scores below the threshold are eliminated. Core elements of new scenarios are extracted to clarify the patient's physiological state, psychological state, and other relevant factors within the scenario. The process involves understanding the needs, communication environment, and potential risks. Then, based on the existing difficulty gradient criteria and the core elements of the new scenarios, 3-5 difficulty gradients are set for each new scenario. A pre-test is conducted to verify the scenario's rationality. Twenty junior nurses from various departments, including internal medicine, surgery, obstetrics and gynecology, and pediatrics, complete the new scenario tasks. A 5-point Likert scale is used to collect their evaluations of the scenario's authenticity and difficulty suitability. Finally, the optimization team conducts a collective review of the new scenarios that pass the pre-test. Once approved, they are included in the narrative scenario library, while scenarios from the original scenario library with a usage rate of less than 10% in the past 6 months and deemed out of touch with current clinical nursing practices by experts are deleted.

3. The method for identifying the degree of humanistic care perception among medical staff according to claim 1, characterized in that, The quality control technology implementation steps in the multi-dimensional data acquisition process are as follows: First, all acquisition devices are uniformly calibrated within 72 hours before data acquisition. Second, a real-time monitoring mechanism is set up in the system background during the acquisition process to dynamically monitor the submission progress of written narrative text, the integrity of audio data, and the clarity of video data. If the submission of written narrative text times out, the interruption of audio data exceeds 5 seconds, or the blurriness of video data exceeds a preset threshold, a pop-up reminder and sound alarm are immediately sent to the acquisition terminal, and the acquisition time and device number of the abnormal data are recorded. Then, for the detected abnormal data, a tiered supplementary acquisition mechanism is activated. If the written narrative text is incomplete, the nurse is notified to submit it within 24 hours. If the audio or video data is abnormal, the corresponding narrative task needs to be re-acquired within 48 hours. The supplementary acquisition data needs to be marked with a supplementary acquisition identifier and stored in association with the original data. Finally, within 24 hours after the data acquisition is completed, all original data is anonymized. The name, employee number, department, and personal identification information in the data are removed by using the AES-256 encryption algorithm, and a unique data number is generated and encrypted and associated with the nurse's identity.

4. The method for identifying the degree of humanistic care perception among medical staff according to claim 1, characterized in that, The custom outlier detection algorithm in the raw data preprocessing step is specifically as follows: For numerical data such as communication duration, speech rate, and response time, an outlier detection formula based on scenario adaptability is used: ,in This is the outlier determination coefficient. The data value to be detected. This represents the average of similar data in this scenario. This represents the standard deviation of similar data in this scenario. For scene weighting coefficients, This is a correction factor for data fluctuations.

5. The method for identifying the degree of humanistic care perception among medical staff according to claim 1, characterized in that, The custom feature fusion algorithm in the multi-dimensional feature extraction and fusion step is specifically designed as follows: A weighted fusion formula is designed for the multi-dimensional parameters of the core features of narrative literacy and humanistic care perception. ,in For the comprehensive feature vector values, The first for narrative literacy Each dimension of feature parameters, The first perception of humanistic care Each dimension of feature parameters, Weighting of each dimension of narrative literacy Weighting of each dimension of humanistic care perception.

6. The method for identifying the degree of humanistic care perception among medical staff according to claim 1, characterized in that, The custom decision fusion algorithm in the humanistic care perception level identification step is specifically designed as follows: A dynamic weight fusion formula is designed based on the output results of the random forest and the improved BP neural network: ,in This is the quantified value corresponding to the final level of perception. The output quantization value of the improved BP neural network, For dynamic fusion weights.

7. The method for identifying the degree of humanistic care perception among medical staff according to claim 1, characterized in that, The specific implementation steps of the precise generation technology in the personalized feedback report generation process are as follows: First, based on the scores of each dimension feature in the multi-dimensional feature extraction and fusion step, a three-dimensional strength-weakness analysis matrix is ​​constructed, using 80 points as the strength threshold and 70 points as the weakness threshold. This clarifies the competency positioning of junior nurses in seven dimensions: narrative expression completeness, narrative comprehension accuracy, narrative reflection depth, empathy, communication awareness, respect awareness, and responsibility awareness, marking the strength, achievement, and weakness dimensions. Second, for the weakness dimensions, a pre-set dynamically updated training resource library is linked. Then, combined with the junior nurses'... The system details the narrative tasks within the steps of building a dedicated narrative scenario library, taking into account the work department, years of service, and the specific manifestations of the narrative tasks. A collaborative filtering algorithm is used to select 3-5 of the most suitable training resources from the library, generating personalized training paths. These paths are designed according to a scientific process of theoretical learning, case analysis, practical training, and assessment feedback, clearly defining the learning focus, time allocation, and assessment methods for each stage. Finally, a training progress tracking module is added to the feedback report, setting core indicators for evaluating training effectiveness and providing a monthly review reminder function. Simultaneously, nursing managers can view nurses' training progress, assessment results, and skill improvement status through the system backend.

8. A system for identifying the level of humanistic care perceived by medical staff, applicable to the method for identifying the level of humanistic care perceived by medical staff as described in any one of claims 1-7, characterized in that, The system includes: The narrative scenario library construction module: Constructs a narrative scenario library specifically for junior nurses. Junior nurses are defined as nurses with ≤3 years of work experience. The narrative scenario library is designed based on high-frequency scenarios of clinical humanistic care. Each narrative scenario includes at least one core scenario element from patient communication, disease information disclosure, psychological counseling, and handling of doctor-patient conflicts. Each scenario sets 3-5 narrative tasks with different difficulty levels. The difficulty levels are divided according to the patient's cooperation, the complexity of the disease, and the urgency of communication. The multi-dimensional data acquisition module acquires narrative expression data and behavioral feedback data of junior nurses in the narrative scenario library during the process of completing narrative tasks through multiple data acquisition methods. The narrative expression data includes written narrative text and oral narrative audio transcription text. The behavioral feedback data includes facial expression feature data, voice tone feature data, and operation response time data. The raw data preprocessing module preprocesses the collected raw data, including data cleaning, data standardization, and data format unification. The data cleaning uses a custom outlier detection algorithm to handle outliers, multiple interpolation to handle missing values, and a custom standardization method to standardize numerical data, converting unstructured text data into structured vector data. The feature extraction and fusion module extracts narrative literacy features and humanistic care perception core features of junior nurses based on preprocessed data. The narrative literacy features include feature parameters in three dimensions: narrative expression completeness, narrative comprehension accuracy, and narrative reflection depth. The humanistic care perception core features include feature parameters in four dimensions: empathy ability, communication awareness, respect awareness, and responsibility awareness. The feature parameters in each dimension are weighted and fused through a custom feature fusion algorithm to obtain a comprehensive feature vector. The humanistic care perception and recognition module inputs the comprehensive feature vector into a preset fusion recognition model, and uses the fusion recognition model to classify and identify the humanistic care perception level of junior nurses, and outputs the corresponding perception level level. The perception level level includes four levels: excellent, good, qualified, and needing improvement. The fusion recognition model uses a custom decision fusion algorithm to determine the level. The feedback report generation module generates a personalized feedback report based on the recognition results. The feedback report includes the scores of each dimension of features, analysis of strengths and weaknesses, and targeted training suggestions. The data storage module stores the exclusive narrative scene library, the collected raw data, the preprocessed data, the extracted feature parameters, the comprehensive feature vector, the recognition results, and the personalized feedback report.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for identifying the level of humanistic care perception of medical staff as described in any one of claims 1-7.

Images