A hospital nursing quality and efficiency evaluation method and system

Abstract

The application discloses a hospital nursing quality and efficiency evaluation method and system, and relates to the technical field of hospital quality management, and specifically relates to a hospital nursing quality and efficiency evaluation method and system. The application introduces information entropy dynamic weight, calculates the maximum Lyapunov index, combines the recovery index to construct a chaotic nursing resilience index, and fuses the three views of patients, safety and management to improve the quantification of the distinguishing degree and the recovery capacity. The application uses a clustering algorithm to evaluate the nursing quality and efficiency, combines fuzzy membership and firefly mixed optimization, and uses a target function containing time attenuation and resilience penalty to enhance global search and convergence. The application maps the membership degree to an individual quality index, weights and sums up the individual quality index to obtain a hospital score, calculates a confidence interval and structured output, and obtains efficient and accurate nursing quality and efficiency evaluation results, which facilitates system docking and decision-making.

Description

Technical Field

[0001] This invention relates to the field of hospital quality management technology, specifically to a method and system for evaluating the quality and effectiveness of hospital nursing care. Background Technology

[0002] With the rapid development of medical informatization and big data technologies, hospital nursing quality evaluation has gradually shifted from traditional paper records and manual scoring to a multi-source data-driven intelligent assessment system. Methods based on statistical analysis, machine learning, and weighted scoring have emerged, enabling the initial integration of nursing records, physiological indicators, and patient feedback. Simultaneously, the widespread adoption of real-time data collection via IoT devices and electronic medical record systems provides a solid foundation for time-series data analysis. Furthermore, with the deep integration of artificial intelligence, chaos theory, and optimization algorithms, nursing quality and effectiveness evaluation will evolve towards real-time, personalized, and precise approaches, providing strong support for refined management and assessment.

[0003] However, existing nursing feature extraction methods generally rely on fixed or empirical weights, are difficult to reflect the information content and time dynamics of indicators, and lack characterization of potential nonlinear recovery capabilities in time series; existing clustering scoring models are sensitive to initial values, are difficult to simultaneously take into account time correlation and individual chaotic differences, and conventional clustering cannot guarantee global optimality; existing nursing quality and efficiency evaluations often lack quantifiable and reliable hospital-level summary indicators; individual-level evaluations are difficult to take into account individual differences and overall uncertainty. Summary of the Invention

[0004] To address the above issues and overcome the shortcomings of existing technologies, this invention provides a method and system for evaluating the quality and effectiveness of hospital nursing care. Addressing the problems of existing nursing feature extraction methods generally relying on fixed or empirical weights, failing to reflect the information content and temporal dynamics of indicators, and lacking characterization of potential nonlinear recovery capabilities in time series, this solution introduces dynamic weights based on information entropy to adaptively quantify the information contribution of each indicator and employs… The algorithm calculates the maximum This approach constructs a chaotic nursing resilience index by combining an index penalty term with recovery-type indicators, and fuses multi-perspective features with fixed view weights based on patient, safety, and management views. This achieves simultaneous characterization of indicator importance, temporal nonlinearity, and multi-dimensional perspectives. Addressing the problems of existing clustering scoring models being susceptible to initial value sensitivity, difficulty in simultaneously considering temporal correlation and individual chaotic differences, and the inability of conventional clustering to guarantee global optimum, this solution utilizes... A stable starting point is obtained by initializing the cluster centers and generating a fuzzy membership matrix. A time decay regularization term and a chaos resilience penalty are introduced into the objective function to preserve temporal information and penalize resilience deviations, thereby characterizing the quality changes and resilience consistency over time. A combination of fuzzy membership update and the firefly algorithm is used to iteratively optimize the cluster centers and membership. The firefly mechanism provides global search capability for the population, and fuzzy membership guarantees soft assignment and uncertainty expression. The two work together to reduce the risk of getting trapped in local minima and accelerate convergence, ultimately obtaining a more robust, time-consistent clustering result and membership matrix that takes into account chaotic characteristics. To address the shortcomings of existing nursing quality and efficiency evaluations, which often lack quantifiable and reliable hospital-level summary indicators and output formats that can be integrated with the system, and the difficulty of individual-level evaluations in taking into account individual differences and overall uncertainty, this scheme maps the soft assignment results into intervalized and comparable individual scores by constructing a patient quality index based on the final membership and patient contribution values. Then, a weighted summary is used to obtain the total hospital nursing quality score and the confidence interval is calculated based on the sample standard deviation.

[0005] The technical solution adopted in this invention is as follows: A method for evaluating the quality and effectiveness of hospital nursing care, the method comprising the following steps:

[0006] Step S1: Construct a nursing quality dataset by collecting nursing records, physiological indicators and patient feedback, aligning and merging multi-source data into a time-series three-dimensional data structure, and standardizing and imputing missing values ​​for each indicator to obtain a standardized data tensor that is free of missing values ​​and comparable.

[0007] Step S2: Extract nursing features. First, calculate dynamic weights using the information entropy method. Then, calculate the chaotic nursing resilience index for recovery-type time-series indicators. Finally, concatenate sub-vectors according to the patient view, safety view, and management view. Perform fixed view weight fusion to generate the final feature vector for each patient.

[0008] Step S3: Construct a clustering scoring model. First, initialize the cluster centers and membership matrix, construct a joint objective function that includes fuzzy distance, time decay, and chaos penalty terms, and then use a combination of fuzzy membership update and firefly algorithm to iteratively optimize the cluster centers and membership until convergence, and obtain the final clustering results and membership matrix.

[0009] Step S4: Nursing quality and effectiveness evaluation. Calculate the quality index for each patient and summarize it into the hospital's total nursing quality score according to the weights. At the same time, calculate the confidence interval of the total nursing quality score. Finally, encapsulate the hospital's total score, membership matrix, and confidence interval and output them as the evaluation results.

[0010] Further, in step S1, constructing the nursing quality dataset specifically includes the following steps:

[0011] Step S11: Construct the basic data tensor. First, collect nursing records, physiological indicators, and patient feedback forms; then, categorize by timestamp and patient... ,Nurse Merge the indicator names; construct a three-dimensional array as the original tensor;

[0012] Step S12: Standardize, Read The output array; loops along the indicator dimensions, finding the minimum and maximum values ​​of each indicator across all time and patient dimensions; performs a standardized formula on each element; iterates through the missing positions in the data and fills them with the global mean of the corresponding indicator.

[0013] Further, in step S2, the extraction of nursing features specifically includes the following steps:

[0014] Step S21: Calculate dynamic weights for information entropy, read... The output is standardized data; for each indicator, its standardized observations across the entire time domain and all patients are first mapped to a probability distribution, and then the normalized entropy value, temporal fluctuation intensity, and cross-patient dispersion of the indicator are jointly characterized to form dynamic weights.

[0015] Step S22: Extraction of Chaotic Nursing Resilience Index, Read The output standardized data, for each patient The time series is segmented. Algorithm for extracting local maxima An index was constructed by combining the horizontal and trend stability of recovery-related indicators to create a chaotic nursing resilience index.

[0016] Step S23: Feature vector fusion, read all features; concatenate sub-vectors according to patient view, security view, and management view respectively; perform fixed view weight fusion to generate the final feature vector for each patient.

[0017] Further, in step S3, the construction of the clustering scoring model specifically includes the following steps:

[0018] Step S31: Initialize clustering parameters, read the generated fusion vector, and use... The algorithm selects initial cluster centers; it also generates a random initial membership matrix and normalizes it row by row so that the sum of each row is 1. Finally, it outputs the initial membership matrix and the initial cluster center set.

[0019] Step S32: Construct a joint objective function based on fuzzy distance, time decay, and chaos penalty term;

[0020] Step S33: Update the clustering. The Firefly algorithm is used to update the membership degree and cluster center in a hybrid manner. The specific operation process is as follows: Read the initial membership degree matrix and the initial cluster center set; enter the iteration loop and set the maximum number of iterations; first update the entire membership degree matrix according to the standard fuzzy formula; then treat each cluster center as a firefly and execute the firefly position update formula for each cluster center; calculate the new objective function value after each iteration and determine whether the convergence condition is met; save the final membership degree matrix and cluster center set, and output the clustering result.

[0021] Furthermore, in step S4, the nursing quality and effectiveness evaluation specifically includes the following steps:

[0022] Step S41: Quality level mapping, read the final membership matrix and cluster center set; calculate the contribution value of each patient to the objective function; calculate the nursing quality index based on the contribution value; and then map the nursing quality index to the four-level nursing quality level;

[0023] Step S42: Calculate the confidence interval, read the nursing quality index and weight vector; sum the weighted values ​​to obtain the total hospital nursing quality score; calculate the sample standard deviation; calculate the confidence interval of the total nursing quality score based on the normal distribution formula;

[0024] Step S43: Output the evaluation results. Integrate all the final evaluation results and encapsulate them for output.

[0025] The present invention provides a hospital nursing quality and effectiveness evaluation system, including a nursing quality dataset construction module, a nursing feature extraction module, a clustering scoring model construction module, and a nursing quality and effectiveness evaluation module;

[0026] The module for constructing the nursing quality dataset collects nursing records, physiological indicators, and patient feedback, aligns and merges multi-source data into a time-series three-dimensional data structure, standardizes and imputes missing values ​​for each indicator to obtain a standardized data tensor that is free of missing values ​​and comparable, and sends the data to the module for extracting nursing features.

[0027] The nursing feature extraction module receives data sent by the nursing quality dataset construction module, first calculates dynamic weights using the information entropy method, then calculates the chaotic nursing resilience index for recovery-type time-series indicators, and finally concatenates sub-vectors according to the patient view, safety view, and management view respectively; and performs fixed view weight fusion to generate the final feature vector for each patient, and sends the data to the clustering scoring model construction module.

[0028] The clustering scoring model construction module receives data sent by the nursing feature extraction module, first initializes the cluster center and membership matrix, constructs a joint objective function including fuzzy distance, time decay and chaos penalty term, and then uses a combination of fuzzy membership update and firefly algorithm to iteratively optimize the cluster center and membership until convergence, obtains the final clustering result and membership matrix, and sends the data to the nursing quality and efficiency evaluation module.

[0029] The nursing quality and efficiency evaluation module receives data sent by the clustering scoring model construction module, calculates the quality index for each patient and aggregates them into the hospital's total nursing quality score according to weights, calculates the confidence interval of the total nursing quality score, and finally encapsulates and outputs the hospital's total score, membership matrix and confidence interval as the evaluation result.

[0030] The beneficial effects achieved by the present invention using the above solution are as follows:

[0031] (1) To address the problems that existing nursing feature extraction methods generally rely on fixed or empirical weights, are difficult to reflect the information content and temporal dynamics of indicators, and lack characterization of potential nonlinear recovery capabilities in time series, this scheme introduces dynamic weights based on information entropy to adaptively quantify the information contribution of each indicator and adopts... The algorithm calculates the maximum The index is used to construct a chaotic nursing resilience index by combining an index penalty term with recovery-type indicators. Furthermore, multi-perspective feature fusion with fixed view weights is performed according to the three views of patient, safety, and management. This enables the simultaneous characterization of indicator importance, temporal nonlinearity, and multi-dimensional perspectives, thereby improving feature discrimination, enhancing the quantitative perception of patients' recovery capabilities, and making subsequent modeling more robust and interpretable.

[0032] (2) To address the problems of existing clustering scoring models being susceptible to initial value sensitivity, difficulty in simultaneously considering temporal correlation and individual chaotic differences, and the inability of conventional clustering to guarantee global optimality, this solution adopts... A stable starting point is obtained by initializing the cluster centers and generating a fuzzy membership matrix. A time decay regularization term and a chaotic resilience penalty are introduced into the objective function to preserve temporal information and penalize resilience deviations, thereby characterizing the quality changes and resilience consistency over time. A combination of fuzzy membership update and the firefly algorithm is adopted to iteratively optimize the cluster centers and membership. The firefly mechanism provides global search capability for the population, and fuzzy membership guarantees soft assignment and uncertainty expression. The two work together to reduce the risk of getting trapped in local minima and accelerate convergence, ultimately obtaining a more robust, time-consistent clustering result and membership matrix that takes into account chaotic characteristics.

[0033] (3) In view of the shortcomings of existing nursing quality and efficiency evaluation, which often lack quantifiable and reliable hospital-level summary indicators and output formats that can be integrated with the system, and the difficulty of individual-level evaluation in taking into account individual differences and overall uncertainty, this solution maps the soft allocation results into intervalized and comparable individual scores by constructing a patient quality index based on the final membership degree and patient contribution value; then, it uses weighted aggregation to obtain the total hospital nursing quality score and calculates the confidence interval based on the sample standard deviation, which not only provides interpretable grading conclusions, but also gives statistical confidence for risk assessment and decision support; finally, it outputs in a structured manner, which is convenient for seamless integration with information systems, reports or quality improvement processes, and enhances practicality and transparency. Attached Figure Description

[0034] Figure 1 A schematic diagram illustrating a hospital nursing quality and efficiency evaluation method provided by the present invention;

[0035] Figure 2 This is a schematic diagram of a hospital nursing quality and efficiency evaluation system provided by the present invention;

[0036] Figure 3 This is a schematic diagram of step S1;

[0037] Figure 4 This is a schematic diagram of step S2;

[0038] Figure 5 This is a schematic diagram of step S3;

[0039] Figure 6 This is a schematic diagram of step S4.

[0040] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0041] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0042] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0043] Example 1, see Figure 1 The present invention provides a method for evaluating the quality and effectiveness of hospital nursing care, which includes the following steps:

[0044] Step S1: Construct a nursing quality dataset by collecting nursing records, physiological indicators and patient feedback, aligning and merging multi-source data into a time-series three-dimensional data structure, and standardizing and imputing missing values ​​for each indicator to obtain a standardized data tensor that is free of missing values ​​and comparable.

[0045] Step S2: Extract nursing features. First, calculate dynamic weights using the information entropy method. Then, calculate the chaotic nursing resilience index for recovery-type time-series indicators. Finally, concatenate sub-vectors according to the patient view, safety view, and management view. Perform fixed view weight fusion to generate the final feature vector for each patient.

[0046] Step S3: Construct a clustering scoring model. First, initialize the cluster centers and membership matrix, construct a joint objective function that includes fuzzy distance, time decay, and chaos penalty terms, and then use a combination of fuzzy membership update and firefly algorithm to iteratively optimize the cluster centers and membership until convergence, and obtain the final clustering results and membership matrix.

[0047] Step S4: Nursing quality and effectiveness evaluation. Calculate the quality index for each patient and summarize it into the hospital's total nursing quality score according to the weights. At the same time, calculate the confidence interval of the total nursing quality score. Finally, encapsulate the hospital's total score, membership matrix, and confidence interval and output them as the evaluation results.

[0048] Example 2, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S1, the construction of the nursing quality dataset specifically includes the following steps:

[0049] Step S11: Construct the basic data tensor. First, collect nursing records, physiological indicators, and patient feedback forms; then use... By timestamp, patient ,Nurse Merge the four dimensions of indicator names; construct a three-dimensional structure. An array, as a raw tensor, is represented as follows:

[0050] ;

[0051] The nursing records include nursing duration, medication records, number of medication errors, and shift handover records; the physiological indicators include heart rate, blood pressure, and blood oxygen; and the patient feedback forms include pain scores and satisfaction levels. Represents the initial three-dimensional raw data tensor; Indicates a point in time ,patient ,index The original observations at the location; Indicates the length of the data collection time window, in hours; This indicates the total number of valid patient cases within the current period; The total number of indicators includes nursing duration, medication records, number of medication errors, shift handover records, heart rate, blood pressure, blood oxygen, pain score, and satisfaction. Represents the real number field;

[0052] Step S12: Standardize, Read Output Array; along the index Dimensional loop, for each Find this indicator separately in the entire Find the minimum and maximum values ​​on the plane; perform standardization on an element-by-element basis; iterate through the data and fill in missing positions using the global mean of the corresponding index; if all indices are 0 or constants, automatically skip the standardization process for that index and set all values ​​of that index to 0.5; the input for this sub-step is the original tensor. The output is a standardized tensor without missing values, represented as follows:

[0053] ;

[0054] Among them, if Missing, for indicators The mean of the non-missing observations is calculated, and the missing values ​​of the indicator are filled with the mean. This represents the standardized observation values; Indicators Minimum across all times and patients Indicators Maximum value over all time and patients; denominator is the index The full range of differences; the mean in the missing imputation formula is the sum of the entire tensor divided by the total number of elements. ;when When the standardization result is set to 0.5.

[0055] Example 3, see Figure 1 and Figure 4This embodiment is based on the above embodiment. In step S2, the extraction of nursing features specifically includes the following steps:

[0056] Step S21: Calculate dynamic weights for information entropy, read... Output For each indicator First, its standardized observations across the entire time domain and all patients are mapped to a probability distribution. Then, the normalized entropy value, temporal fluctuation intensity, and cross-patient dispersion of this indicator are jointly characterized to form a dynamic weight. The probability distribution is expressed as: ;in, Indicators At the point of time ,patient Probabilistic observations at; It is a very small constant; when At that time, it was agreed Based on this, calculate the indicators. Normalized entropy value: ; Indicators The normalized entropy value, with a range of values ​​of ; It is the natural logarithm; The smaller the value, the better the indicator. The greater the amount of information, the stronger the discrimination ability; further calculation of indicators Temporal fluctuation intensity and cross-patient dispersion intensity: ; ; The normalized entropy value, temporal fluctuation intensity, and cross-patient dispersion are integrated into a coherent information content. ;in, Indicators Time fluctuation intensity The larger the value, the more drastic the changes in the indicator over time, and the stronger the dynamic information. Indicators The cross-patient discrete intensity; Indicates at a point in time Above, indicators The mean of all patient observations; Indicates the same patient Indicators at adjacent time points The change in; This represents the total count of the time difference items; Indicators The amount of collaborative information; and The adjustment coefficients for the time-varying term and the cross-patient discrete term are used to balance the time-varying and discrete terms, respectively. Finally, the co-information of all indicators is normalized to obtain the final dynamic weights. ;in, Indicators The final dynamic weight; satisfies ;

[0057] Step S22: Extraction of Chaotic Nursing Resilience Index, Read The output standardized data, for each patient The time series is segmented. Algorithm for extracting local maxima An index was constructed by combining the level stability and trend stability of recovery-related indicators to create a chaotic nursing resilience index; firstly, patients were... The time series is sliced ​​locally according to a preset window length, and the maximum value is calculated for each local segment. The index yields the first... Local index: ;in, Indicates the patient In the The largest time slice index, Indicates the time slice index. This represents the limit operation as the time step approaches infinity. Indicates the patient In the Adjacent embedded trajectories within a time slice The Euclidean distance difference at time , The initial distance between adjacent trajectories is represented; then, each local index is fused according to its time decay weight to obtain the patient-level index. index: ;in, Indicates the patient Overall maximum index, Indicates the patient The total number of local time slices, This is the time decay coefficient; An exponential function with the natural constant as its base; then, for the set of recovery indicators: ={Heart rate, blood pressure, blood oxygen, pain score}; Construct a recovery consistency score: ;in, Indicates the patient In recovery indicators On the recovery consistency score, This is the balance coefficient between the horizontal and trend terms; finally, the patient-level... The chaotic nursing resilience index is obtained by coupling the index with the recovery consistency score: ;in, Indicates the patient Chaotic nursing resilience index, For exponential penalty terms; This indicates taking the maximum value; when When this occurs, it indicates that the patient's condition is highly volatile, and the punishment is intensified; when At this time, the penalty remains at 1, without further reducing the resilience value;

[0058] Step S23: Feature vector fusion: Read all features; concatenate sub-vectors according to patient view, security view, and management view; then multiply by fixed weights; finally generate... The fusion vectors are represented as follows:

[0059] ;

[0060] in, Indicates post-fusion patients The final feature vector; , , These are sub-vectors for the three views, The patient view is represented, including: nursing resilience index, pain score, and satisfaction. This displays a safety view, including: the number of dosing errors and dosing records; The management view includes: nursing duration, shift handover records, heart rate, blood pressure, and blood oxygen saturation. Indicates the patient view weight; Weights for security views; To manage view weights; Indicates the total dimension.

[0061] By performing the above operations, this solution addresses the problems of existing nursing feature extraction methods that generally rely on fixed or empirical weights, struggle to reflect the information content and temporal dynamics of indicators, and lack characterization of potential nonlinear recovery capabilities in time series. Instead, it introduces dynamic weights based on information entropy to adaptively quantify the information contribution of each indicator and employs... The algorithm calculates the maximum The index is used to construct a chaotic nursing resilience index by combining an index penalty term with recovery-type indicators. Furthermore, multi-perspective feature fusion with fixed view weights is performed according to the three views of patient, safety, and management. This enables the simultaneous characterization of indicator importance, temporal nonlinearity, and multi-dimensional perspectives, thereby improving feature discrimination, enhancing the quantitative perception of patients' recovery capabilities, and making subsequent modeling more robust and interpretable.

[0062] Example 4, see Figure 1 and Figure 5 This embodiment is based on the above embodiment. In step S3, the construction of the clustering scoring model specifically includes the following steps:

[0063] Step S31: Initialize clustering parameters and read the generated data. A fusion vector, using The algorithm selects initial cluster centers; it also generates a random initial membership matrix, normalizes it row-wise so that the sum of each row is 1, and finally outputs the initial membership matrix. and the initial set of cluster centers , means as follows:

[0064] ;

[0065] in, Let the initial membership matrix be denoted as . OK List; Indicates the patient Belongs to quality grade The initial membership degree; This indicates the preset quality level, including 4 levels: Excellent, Good, Average, and Needs Improvement. Represents the initial set of cluster centers; and Same dimensions;

[0066] Step S32: Construct a joint objective function. Based on fuzzy distance, time decay, and chaos penalty term, construct a joint objective function, as follows:

[0067] ;

[0068] in, Let represent the overall objective function to be minimized; the first term is the fuzzy distance loss; the second term is the time decay regularization term. The third term represents the time decay coefficient; the fourth term is the chaos resilience penalty term. Indicates the penalty coefficient; This indicates taking the Euclidean distance; For patients In time slices fused feature vectors; Indicates the first The class resilience mean is calculated by weighting the initial membership matrix; constraint: for each , and ;

[0069] Step S33: Update clustering. Use the Firefly algorithm to update membership and cluster centers in a hybrid manner. The specific operation process is as follows: Read the initial... and Enter the iterative loop, performing a maximum of 200 iterations; first, update the entire membership matrix according to the standard fuzzy formula; then, treat each cluster center as a firefly and execute the firefly position update formula for each cluster center; after each iteration, calculate the new objective function value and determine whether the convergence condition is met; save the final membership matrix and cluster center set, and output the clustering result, as shown below:

[0070] ;

[0071] in, Indicates the patient To the Distance from class center Indicates the patient To the Distance from class center; Representing time slices Lower patient To the Distance from class center Representing time slices Lower patient To the Distance from class center; This is a fuzzy index, with a value range of [value range missing]. ; To attract the base number, the initial value is set to 1; It is the light absorption coefficient; Two centers and The square of the Euclidean distance; The random perturbation step size, Indicates the value in Random numbers in the model; brightness is negatively correlated with the objective function value, the higher the brightness, the better. The lower the value, the stronger the attraction; 'l' represents the category index. Indicates the first The class's resilience mean; update order: first update the membership matrix. Then calculate the center brightness. , In the current iteration, with the th Class center calculation The obtained local cost, for each pair of centers , ,if Then execute the firefly position update formula to move the center. After each iteration, the step size is increased by 0.01. and Fine-tuning is performed; the convergence condition is two consecutive... The decrease is less than Or the number of iterations reaches 200.

[0072] By performing the above operations, this solution addresses the problems of existing clustering scoring models being susceptible to initial value sensitivity, difficulty in simultaneously considering temporal correlation and individual chaotic differences, and the inability of conventional clustering to guarantee global optimality. It further addresses these issues by using... A stable starting point is obtained by initializing the cluster centers and generating a fuzzy membership matrix. A time decay regularization term and a chaotic resilience penalty are introduced into the objective function to preserve temporal information and penalize resilience deviations, thereby characterizing the quality changes and resilience consistency over time. A combination of fuzzy membership update and the firefly algorithm is adopted to iteratively optimize the cluster centers and membership. The firefly mechanism provides global search capability for the population, and fuzzy membership guarantees soft assignment and uncertainty expression. The two work together to reduce the risk of getting trapped in local minima and accelerate convergence, ultimately obtaining a more robust, time-consistent clustering result and membership matrix that takes into account chaotic characteristics.

[0073] Example 5, see Figure 1 and Figure 6 This embodiment is based on the above embodiment. In step S4, the evaluation of nursing quality and effectiveness specifically includes the following steps:

[0074] Step S41: Quality grade mapping, read the final membership matrix and cluster center set; for each patient Calculate its contribution to the objective function. And based on contribution value Calculate the nursing quality index Mapped to the fourth level of nursing quality, it is represented as follows:

[0075] ;

[0076] in, Indicates the patient The nursing quality index; A predefined score is assigned to each level, with "excellent" being... Good Medium It needs to be improved to 0.3; For patients For the overall objective function The contribution value, The largest contribution value to patients globally;

[0077] Step S42: Calculate the confidence interval, read the nursing quality index and weight vector; calculate the weighted sum to obtain the total hospital nursing quality score. Calculate the sample standard deviation; apply the normal distribution formula to obtain... The output consists of the hospital's total score and the confidence interval, as shown below:

[0078] ;

[0079] in, This indicates the overall score for hospital nursing quality; Patient weights, defaulting to ; for Standard deviation; The total number of patients; 1.96 is the 95% confidence coefficient for a standard normal distribution; These are the upper and lower limits of the confidence interval;

[0080] Step S43: Output the evaluation results. Integrate all the final evaluation results, encapsulate the output, and represent the output vector as follows:

[0081] ;

[0082] in, Represents the evaluation result vector. This represents the hospital's total score. This is the final membership matrix; This represents the confidence interval.

[0083] By performing the above operations, this solution addresses the shortcomings of existing nursing quality and effectiveness evaluations, which often lack quantifiable and reliable hospital-level aggregate indicators and output formats that can be integrated with systems, and whose individual-level evaluations struggle to balance individual differences and overall uncertainty. It maps soft-assignment results into interval-based, comparable individual scores by constructing a patient quality index based on final membership degree and patient contribution value. Then, a weighted aggregation is used to obtain the total hospital nursing quality score, and confidence intervals are calculated based on the sample standard deviation. This provides interpretable grading conclusions and statistical confidence levels for risk assessment and decision support. Finally, the structured output facilitates seamless integration with information systems, reports, or quality improvement processes, enhancing practicality and transparency.

[0084] Example 6, see Figure 2 Based on the above embodiments, this embodiment provides a hospital nursing quality and efficiency evaluation system, including a nursing quality dataset construction module, a nursing feature extraction module, a clustering scoring model construction module, and a nursing quality and efficiency evaluation module.

[0085] The module for constructing the nursing quality dataset collects nursing records, physiological indicators, and patient feedback, aligns and merges multi-source data into a time-series three-dimensional data structure, standardizes and imputes missing values ​​for each indicator to obtain a standardized data tensor that is free of missing values ​​and comparable, and sends the data to the module for extracting nursing features.

[0086] The nursing feature extraction module receives data sent by the nursing quality dataset construction module, first calculates dynamic weights using the information entropy method, then calculates the chaotic nursing resilience index for recovery-type time-series indicators, and finally concatenates sub-vectors according to the patient view, safety view, and management view respectively; and performs fixed view weight fusion to generate the final feature vector for each patient, and sends the data to the clustering scoring model construction module.

[0087] The clustering scoring model construction module receives data sent by the nursing feature extraction module, first initializes the cluster center and membership matrix, constructs a joint objective function including fuzzy distance, time decay and chaos penalty term, and then uses a combination of fuzzy membership update and firefly algorithm to iteratively optimize the cluster center and membership until convergence, obtains the final clustering result and membership matrix, and sends the data to the nursing quality and efficiency evaluation module.

[0088] The nursing quality and efficiency evaluation module receives data sent by the clustering scoring model construction module, calculates the quality index for each patient and aggregates them into the hospital's total nursing quality score according to weights, calculates the confidence interval of the total nursing quality score, and finally encapsulates and outputs the hospital's total score, membership matrix and confidence interval as the evaluation result.

[0089] It should be noted that, in this document, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0090] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0091] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for evaluating the quality and effectiveness of hospital nursing care, characterized in that: The method includes the following steps: Step S1: Construct a nursing quality dataset by collecting nursing records, physiological indicators and patient feedback, aligning and merging multi-source data into a time-series three-dimensional data structure, and standardizing and imputing missing values ​​for each indicator to obtain a standardized data tensor that is free of missing values ​​and comparable. Step S2: Extract nursing features. First, calculate dynamic weights using the information entropy method. Then, calculate the chaotic nursing resilience index for recovery-type time-series indicators. Finally, concatenate sub-vectors according to the patient view, safety view, and management view. Perform fixed view weight fusion to generate the final feature vector for each patient. Step S3: Construct a clustering scoring model. First, initialize the cluster centers and membership matrix, construct a joint objective function that includes fuzzy distance, time decay, and chaos penalty terms, and then use a combination of fuzzy membership update and firefly algorithm to iteratively optimize the cluster centers and membership until convergence, and obtain the final clustering results and membership matrix. Step S4: Nursing quality and effectiveness evaluation. Calculate the quality index for each patient and summarize it into the hospital's total nursing quality score according to the weights. At the same time, calculate the confidence interval of the total nursing quality score. Finally, encapsulate the hospital's total score, membership matrix, and confidence interval and output them as the evaluation results.

2. The method for evaluating the quality and effectiveness of hospital nursing care according to claim 1, characterized in that: In step S1, constructing the nursing quality dataset specifically includes the following steps: Step S11: Construct the basic data tensor. First, collect nursing records, physiological indicators, and patient feedback forms; then, categorize by timestamp and patient... ,Nurse Merge the indicator names; construct a three-dimensional array as the original tensor; Step S12: Standardize, Read The output array; loops along the indicator dimensions, finding the minimum and maximum values ​​of each indicator across all time and patient dimensions; performs a standardized formula on each element; iterates through the missing positions in the data and fills them with the global mean of the corresponding indicator.

3. The method for evaluating the quality and effectiveness of hospital nursing care according to claim 1, characterized in that: In step S2, the extraction of nursing features specifically includes the following steps: Step S21: Calculate dynamic weights for information entropy, read... The output is standardized data; for each indicator, its standardized observations across the entire time domain and all patients are first mapped to a probability distribution, and then the normalized entropy value, temporal fluctuation intensity, and cross-patient dispersion of the indicator are jointly characterized to form dynamic weights. Step S22: Extraction of Chaotic Nursing Resilience Index, Read The output standardized data, for each patient The time series is segmented. Algorithm for extracting local maxima An index was constructed by combining the horizontal and trend stability of recovery-related indicators to create a chaotic nursing resilience index. Step S23: Feature vector fusion, read all features; concatenate sub-vectors according to patient view, security view, and management view respectively; perform fixed view weight fusion to generate the final feature vector for each patient.

4. The method for evaluating the quality and effectiveness of hospital nursing care according to claim 1, characterized in that: In step S3, constructing the clustering scoring model specifically includes the following steps: Step S31: Initialize clustering parameters, read the generated fusion vector, and use... The algorithm selects initial cluster centers; it also generates a random initial membership matrix and normalizes it row by row so that the sum of each row is 1. Finally, it outputs the initial membership matrix and the initial cluster center set. Step S32: Construct a joint objective function based on fuzzy distance, time decay, and chaos penalty term; Step S33: Update the clustering. The Firefly algorithm is used to update the membership degree and cluster center in a hybrid manner. The specific operation process is as follows: Read the initial membership degree matrix and the initial cluster center set; enter the iteration loop and set the maximum number of iterations; first update the entire membership degree matrix according to the standard fuzzy formula; then treat each cluster center as a firefly and execute the firefly position update formula for each cluster center; calculate the new objective function value after each iteration and determine whether the convergence condition is met; save the final membership degree matrix and cluster center set, and output the clustering result.

5. The method for evaluating the quality and effectiveness of hospital nursing care according to claim 1, characterized in that: In step S4, the nursing quality and effectiveness evaluation specifically includes the following steps: Step S41: Quality level mapping, read the final membership matrix and cluster center set; calculate the contribution value of each patient to the objective function; calculate the nursing quality index based on the contribution value; and then map the nursing quality index to the four-level nursing quality level; Step S42: Calculate the confidence interval, read the nursing quality index and weight vector; sum the weighted values ​​to obtain the total hospital nursing quality score; calculate the sample standard deviation; calculate the confidence interval of the total nursing quality score based on the normal distribution formula; Step S43: Output the evaluation results, integrate all the final evaluation results, and encapsulate the output.

6. A hospital nursing quality and efficiency evaluation system, used to implement the hospital nursing quality and efficiency evaluation method as described in any one of claims 1-5, characterized in that: It includes modules for constructing nursing quality datasets, extracting nursing features, constructing clustering scoring models, and evaluating nursing quality and effectiveness.

7. A hospital nursing quality and efficiency evaluation system according to claim 6, characterized in that: The module for constructing the nursing quality dataset collects nursing records, physiological indicators, and patient feedback, aligns and merges multi-source data into a time-series three-dimensional data structure, standardizes and imputes missing values ​​for each indicator to obtain a standardized data tensor that is free of missing values ​​and comparable, and sends the data to the module for extracting nursing features. The nursing feature extraction module receives data sent by the nursing quality dataset construction module, first calculates dynamic weights using the information entropy method, then calculates the chaotic nursing resilience index for recovery-type time-series indicators, and finally concatenates sub-vectors according to the patient view, safety view, and management view respectively. It then performs fixed view weight fusion to generate the final feature vector for each patient, and sends the data to the module for building the clustering scoring model; The clustering scoring model construction module receives data sent by the nursing feature extraction module, first initializes the cluster center and membership matrix, constructs a joint objective function including fuzzy distance, time decay and chaos penalty term, and then uses a combination of fuzzy membership update and firefly algorithm to iteratively optimize the cluster center and membership until convergence, obtains the final clustering result and membership matrix, and sends the data to the nursing quality and efficiency evaluation module. The nursing quality and efficiency evaluation module receives data sent by the clustering scoring model construction module, calculates the quality index for each patient and aggregates them into the hospital's total nursing quality score according to weights, calculates the confidence interval of the total nursing quality score, and finally encapsulates and outputs the hospital's total score, membership matrix and confidence interval as the evaluation result.

Images