Anning treatment and protection effect evaluation method

Abstract

The invention provides a peaceful treatment and care effect evaluation method, which belongs to the technical field of peaceful treatment and care effect evaluation, and comprises the following steps: establishing a physiological index anomaly detection model and removing abnormal data by collecting physiological index data and subjective score data of a patient in real time and applying technologies such as multivariate mixed decomposition and time sequence feature analysis. And establishing a function relationship between the physiological index stable component and the subjective score data, calculating a physiological fluctuation influence coefficient matrix, and performing weighted fusion calculation by using the corrected physiological stability coefficient matrix and subjective score coefficient matrix according to a personalized reference range in a patient baseline database to obtain a comprehensive score of the calming treatment and protection effect. And finally, according to a preset scoring interval, generating an evaluation result of the calming treatment and protection effect. According to the method, the physiological index data and the subjective score data are fully utilized, and the problem that the current calming treatment effect depends on subjective evaluation of medical staff and is difficult to reflect accurately is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of palliative care effect evaluation, and in particular, relates to a palliative care effect evaluation method. Background Art

[0002] There are some problems in the existing palliative care practice that need to be solved urgently. First, there is a lack of effective efficacy evaluation mechanism. At present, most palliative care projects rely on the subjective evaluation of medical staff, which is difficult to objectively reflect the treatment effect. Secondly, there are large individual differences, making it difficult to formulate targeted care plans. The physiological conditions, psychological needs and quality of life of different patients may vary greatly, but the existing evaluation methods are rarely targeted. Thirdly, there is a lack of dynamic monitoring and continuous optimization mechanisms. The demand for palliative care is constantly adjusted as the condition changes, but the existing methods make it difficult to achieve dynamic evaluation and real-time feedback.

[0003] In order to solve the above problems, it is urgent to develop a more scientific, comprehensive and personalized palliative care effect evaluation method. Summary of the invention

[0004] In view of this, the present invention provides a palliative care effect evaluation method, which can solve the problem that the existing palliative care effect depends on the subjective evaluation of medical staff and is difficult to accurately reflect the effect.

[0005] The present invention is achieved in that:

[0006] The present invention provides a method for evaluating the effect of palliative care, which comprises the following steps:

[0007] S10, real-time collection of patients' physiological index data and subjective scoring data of patients' palliative care effects;

[0008] S20, performing multivariate mixed decomposition calculation on the physiological index data to obtain stable components and variable components of each physiological index;

[0009] S30, calculating the time series characteristics of the physiological indicator data;

[0010] S40, based on the time series characteristics, establishing a physiological indicator abnormality detection model to remove abnormal data points;

[0011] S50, establishing a functional relationship between the stable components of each physiological index and each subjective scoring data in the subjective scoring data, recorded as a stable relationship, and establishing a physiological stability coefficient matrix and a subjective scoring coefficient matrix respectively according to the stable relationship;

[0012] S60, calculating a physiological fluctuation influence coefficient matrix based on the physiological index change components; and correcting the physiological stability coefficient matrix according to the physiological fluctuation influence coefficient matrix to obtain a corrected physiological stability coefficient matrix;

[0013] S70. Establish a patient baseline database and calculate personalized reference ranges based on the patient's age, gender, and disease type;

[0014] S80, performing weighted fusion calculation using the modified physiological stability coefficient matrix, the subjective scoring coefficient matrix, and the personalized reference range to obtain a comprehensive score for palliative care effects;

[0015] S90. Classify the comprehensive score of the palliative care effect according to a preset score range to generate a palliative care effect evaluation result.

[0016] On the basis of the above technical solution, a palliative care effect evaluation method of the present invention can also be improved as follows:

[0017] The time series characteristics include mean, standard deviation, kurtosis coefficient and skewness coefficient.

[0018] Among them, the preset scoring intervals include excellent interval, good interval, general interval, and interval for improvement.

[0019] Furthermore, the preset scoring intervals are specifically: excellent interval [90, 100], good interval [75, 90) , average interval [60, 75) , and interval for improvement [0, 60) .

[0020] Among them, the physiological indicator data include heart rate data, blood pressure data, respiratory rate data, body temperature data, and blood oxygen saturation data, which are objectively collected data.

[0021] The subjective scoring data include pain scoring data, quality of life scoring data and psychological state scoring data.

[0022] Furthermore, the subjective scoring data is collected through questionnaire surveys from patients and their families regarding the effectiveness of palliative care.

[0023] Wherein, the anomaly detection model is based on a local anomaly factor algorithm or a linear interpolation algorithm.

[0024] Wherein, the stable relationship adopts a polynomial regression model.

[0025] The multivariate mixed decomposition calculation is specifically implemented by using the method of empirical mode decomposition or principal component analysis or dynamic factor model.

[0026] The following is a detailed description of some of the calculation steps involved in the present invention:

[0027] 1. The multivariate mixed decomposition calculation model (S20) is specifically expressed as follows:

[0028] X(t)=S(t)+V(t)+n(t);

[0029] Where X(t) is the original physiological index time series; S(t) is the stable component; V(t) is the variable component; N(t) is the noise component; and t is the time variable.

[0030] The decomposition process uses the empirical mode decomposition (EMD) method:

[0031]

[0032] In the formula, c i (t) is the intrinsic mode function; r n (t) is the residual term; n is the number of decomposition levels, ranging from 3 to 8.

[0033] 2. The time series feature calculation (S30) is specifically expressed as follows:

[0034] Mean calculation:

[0035]

[0036] Standard deviation calculation:

[0037]

[0038] Kurtosis coefficient calculation:

[0039]

[0040] Skewness coefficient calculation:

[0041]

[0042] In the formula, x i is the time series data point; N is the total number of data points; μ is the mean; σ is the standard deviation.

[0043] 3. The anomaly detection model (S40) uses an improved local anomaly factor (LOF) algorithm:

[0044]

[0045] Where LOF k (p) is the local anomaly factor of point p; N k (p) is the set of k nearest neighbors of point p; lrd k(p) is the local reachability density of point p; k is the number of nearest neighbors, ranging from 5 to 20.

[0046] 4. Stable relationship modeling (S50) using polynomial regression model:

[0047]

[0048] Where Y j is the jth subjective rating; S i is the stable component of the ith physiological index; β0 is the intercept term; β i is the coefficient of the first-order term; il is the interaction term coefficient; ∈ is the random error term; m is the number of physiological indicators.

[0049] 5. Physiological stability coefficient matrix M s The structure is:

[0050]

[0051] In the formula, α ij represents the stable influence coefficient of the ith physiological index on the jth score; m is the number of physiological indexes; n is the number of subjective score types.

[0052] 6. Subjective rating coefficient matrix M p The structure is:

[0053]

[0054] In the formula, β ij It represents the weight coefficient of the i-th rating in the j-th subjective rating; k is the number of rating dimensions; n is the number of subjective rating types.

[0055] 7. Physiological fluctuation influence coefficient matrix M v Calculation (S60):

[0056]

[0057] where each element is calculated as:

[0058]

[0059] Where V i (t) is the change component of the ith physiological index; T is the observation time length; S i is the stable component of the i-th physiological index.

[0060] 8. Modified physiological stability coefficient matrix M s 'calculate:

[0061] M s ′=Ms ·(I-λM v );

[0062] Where I is the unit matrix; λ is the correction coefficient, and its value range is 0.1-0.5.

[0063] 9. Personalized reference range calculation (S70):

[0064] R i =[μ i -k i σ i , μ i +k i σ i ];

[0065] In the formula, R i is the reference range of the ith indicator; μ i is the baseline mean; σ i is the baseline standard deviation; k i is the adjustment coefficient, which is related to age and gender.

[0066] 10. Calculation of comprehensive score of palliative care effect (S80):

[0067]

[0068] In the formula, w ij is the comprehensive weight coefficient; f(x i , R i ) is the reference range normalization function; x i is the measured value; R i For reference range.

[0069] The preset scoring interval is specifically expressed as follows:

[0070] Score excellent = [90, 100];

[0071] Score good =[75,90);

[0072] Score normal =[60,75);

[0073] Score poor = [0, 60);

[0074] In the formula, Score excellent Score: Excellent range good Score is a good range. normal For the general interval; Score poor This is the area to be improved.

[0075] The steps for determining the preset scoring interval specifically include:

[0076] 1. Collect historical data to establish a benchmark scoring model:

[0077]

[0078] In the formula, S base Score for the benchmark; i is the score of the ith historical case; N is the total number of historical cases.

[0079] 2. Calculate the standard deviation of the ratings:

[0080]

[0081] In the formula, σ score is the scoring standard deviation.

[0082] 3. Establish a normal distribution model:

[0083]

[0084] Where f(x) is the probability density function of the score.

[0085] 4. Calculate the quantile:

[0086]

[0087] x q =F -1 (q);

[0088] Where P(x) is the cumulative distribution function; x q is the q quantile; F -1 is the inverse of the cumulative distribution function.

[0089] 5. Determine the interval boundary value:

[0090] B excellent =x 0.9 ;

[0091] B good =x 0.75 ;

[0092] B normal =x 0.6 ;

[0093] In the formula, B excellent is the lower bound of the excellent interval; B good is the lower bound of a good interval; B normal It is the lower bound of the general interval.

[0094] 6. Apply Delphi expert review method to make revisions:

[0095]

[0096] In the formula, B i ′ is the corrected i-th boundary value; B ij is the evaluation value of the jth expert on the i-th boundary; w j is the weight of the jth expert; M is the number of experts.

[0097] 7. Calculate the consistency coefficient of expert opinions:

[0098]

[0099] Where W is the Kendall harmony coefficient; S is the sum of squares of the expert ratings; and n is the number of boundaries to be evaluated.

[0100] 8. Determine the final interval:

[0101]

[0102] In the formula, is the final determined i-th boundary value; α is the compromise coefficient, and its value range is [0.3,0.7].

[0103] 9. Interval validity verification:

[0104]

[0105] Where V i is the distribution ratio of the i-th interval; N i is the number of samples falling into the i-th interval; N total is the total number of samples.

[0106] 10. Dynamic adjustment mechanism of interval:

[0107]

[0108] In the formula, is the i-th boundary value after the t-th adjustment; η is the learning rate, and its value range is [0.1, 0.3]; is the tth adjustment amount.

[0109] Compared with the prior art, the beneficial effects of the palliative care effect evaluation method provided by the present invention are:

[0110] 1. Achieve a comprehensive quantitative evaluation of physiological indicators and subjective feelings. This method not only collects the patient's physiological indicator data, such as heart rate, blood pressure, respiratory rate, etc., but also collects the patient's subjective rating of the effect of palliative care, such as pain, quality of life, and psychological state. By constructing a mathematical model between physiological indicator data and subjective rating data, a comprehensive quantitative analysis of the effect of palliative care is achieved.

[0111] 2. Identify individual differences and provide targeted evaluation. This method establishes a patient baseline database containing a large amount of medical record data, and calculates personalized reference ranges of physiological indicators based on the patient's age, gender, disease type and other characteristics. This personalized evaluation method can better reflect the actual situation of different patients and provide a basis for medical staff to formulate personalized care plans.

[0112] 3. Realize dynamic monitoring and continuous optimization. This method can not only evaluate a single palliative care, but also dynamically monitor the changing trend of treatment effects by real-time collection and analysis of physiological index data. At the same time, the evaluation results can also be fed back to medical staff to promote the continuous optimization of nursing measures and continuously improve the quality of palliative care services.

[0113] 4. Improve the objectivity and reliability of evaluation results. This method uses a series of advanced algorithms such as multivariate mixed decomposition, anomaly detection, and model building to conduct in-depth analysis of physiological indicator data, eliminate outliers, and establish a robust mathematical model. This ensures the objectivity and reliability of the evaluation results and provides a scientific basis for clinical decision-making.

[0114] In summary, the palliative care effect evaluation method proposed in the present invention makes full use of physiological index data and subjective scoring data, adopts multi-dimensional analysis technology, and realizes a comprehensive quantitative evaluation of the palliative care effect. This method can not only identify individual differences, but also has the characteristics of dynamic monitoring and continuous optimization, which solves the problem that the existing palliative care effect depends on the subjective evaluation of medical staff and is difficult to accurately reflect the effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] Figure 1 A flow chart of the method provided by the present invention; Figure 2 Schematic diagram of heart rate decomposition in the embodiment. DETAILED DESCRIPTION

[0116] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0117] like Figure 1FIG. 1 is a flow chart of a palliative care effect evaluation method provided by the present invention, and the method comprises the following steps:

[0118] S10, real-time collection of patients' physiological index data and subjective scoring data of patients' palliative care effects;

[0119] S20, performing multivariate mixed decomposition calculations on the physiological index data to obtain stable components and variable components of each physiological index;

[0120] S30, calculating time series characteristics of physiological indicator data;

[0121] S40, based on the time series characteristics, establish a physiological index abnormality detection model and remove abnormal data points;

[0122] S50, establishing a functional relationship between the stable components of various physiological indicators and each type of subjective scoring data in the subjective scoring data, recorded as a stable relationship, and establishing a physiological stability coefficient matrix and a subjective scoring coefficient matrix respectively according to the stable relationship;

[0123] S60, calculating a physiological fluctuation influence coefficient matrix based on the change components of various physiological indicators; and correcting a physiological stability coefficient matrix according to the physiological fluctuation influence coefficient matrix to obtain a corrected physiological stability coefficient matrix;

[0124] S70. Establish a patient baseline database and calculate personalized reference ranges based on the patient's age, gender, and disease type;

[0125] S80. Use the modified physiological stability coefficient matrix, subjective scoring coefficient matrix and personalized reference range to perform weighted fusion calculation to obtain a comprehensive score for palliative care effects;

[0126] S90. Grade the comprehensive score of the palliative care effect according to the preset score range to generate a palliative care effect evaluation result.

[0127] The specific implementation methods of the above steps are described in detail below:

[0128] Step S10: Real-time collection of patient physiological index data and subjective scoring data of the patient's palliative care effect. The purpose of this step is to obtain the original data required for evaluation. The specific implementation is as follows:

[0129] First, physiological signal acquisition equipment is used to record the patient's heart rate data, blood pressure data, respiratory rate data, body temperature data and blood oxygen saturation data in real time. These physiological index data can reflect the changes in the patient's physiological state and provide a basis for the subsequent evaluation of the effect of palliative care.

[0130] Secondly, we collected patients' subjective scores on the effect of their own palliative care through questionnaire surveys. Subjective scores include pain scores, quality of life scores, and psychological state scores. These subjective score data can reflect patients' subjective feelings about the treatment effect, and combined with physiological index data, they can more comprehensively evaluate the effect of palliative care.

[0131] By collecting physiological indicator data and subjective scoring data in real time, the original data required for evaluation can be obtained, laying the foundation for subsequent multi-dimensional analysis and comprehensive evaluation.

[0132] Step S20: Perform multivariate mixed decomposition calculation on the physiological index data to obtain the stable component and variable component of each physiological index. The purpose of this step is to decompose the physiological index data and extract the stable component and variable component of each index.

[0133] The specific implementation is as follows:

[0134] First, the original physiological index time series X(t) is decomposed using the multivariate mixed decomposition calculation model X(t)=S(t)+V(t)+N(t), where S(t) represents the stable component, V(t) represents the variable component, and N(t) represents the noise component.

[0135] The decomposition process uses the empirical mode decomposition (EMD) method, that is, Among them, c i (t) is the intrinsic mode function, r n (t) is the residual term, n is the number of decomposition levels, and its value range is 3-8.

[0136] Through EMD decomposition, the original physiological index data can be decomposed into stable components and variable components. The stable component S(t) represents the basic state of the physiological index, and the variable component V(t) represents the fluctuation of the physiological index. These two components provide an important basis for the subsequent evaluation of the effect of palliative care.

[0137] Step S30: Calculate the time series characteristics of the physiological index data, which include mean, standard deviation, kurtosis coefficient and skewness coefficient. The purpose of this step is to extract the statistical characteristics of the physiological index data to provide a basis for subsequent abnormality detection and model building.

[0138] The specific implementation is as follows:

[0139] 1. Calculate the mean Among them, x i is the time series data point, and N is the total number of data points. The mean reflects the average level of the physiological index.

[0140] 2. Calculate the standard deviation The standard deviation reflects the dispersion of physiological index data.

[0141] 3. Calculate the kurtosis coefficient The kurtosis coefficient reflects the steepness of the data distribution.

[0142] 4. Calculate the skewness coefficient The skewness coefficient reflects the degree of skewness of the data distribution.

[0143] By calculating the above time series statistical characteristics, the distribution characteristics of physiological index data can be fully described. These characteristics provide an important basis for subsequent anomaly detection and model building.

[0144] Step S40: Based on the time series characteristics, a physiological index abnormality detection model is established to remove abnormal data points. The purpose of this step is to identify and remove abnormal values ​​in the physiological index data to improve the reliability of the data.

[0145] The specific implementation is as follows:

[0146] The improved local outlier factor (LOF) algorithm is used for anomaly detection. The algorithm formula is: Among them, LOF k (p) is the local anomaly factor of point p, N k (p) is the set of k nearest neighbors of point p, lrd k (p) is the local reachability density of point p, k is the number of neighbors, and its value range is 5-20.

[0147] Through this algorithm, outliers that deviate significantly from the normal data distribution can be identified based on the time series characteristics of physiological indicator data. These outliers may be caused by measurement errors, instrument failures, or abnormal physiological states, and need to be removed from the original data to ensure the reliability of subsequent analysis.

[0148] Step S50: Establish a functional relationship between the stable components of each physiological index and each subjective scoring data in the subjective scoring data, which is recorded as a stable relationship, and establish a physiological stability coefficient matrix and a subjective scoring coefficient matrix according to the stable relationship. The purpose of this step is to establish a mathematical model between the stable components of physiological indicators and subjective scores, so as to provide a basis for subsequent comprehensive evaluation.

[0149] The specific implementation is as follows:

[0150] 1. Use the polynomial regression model to establish the functional relationship between the stable component of physiological indicators and various subjective scores, that is, Among them, Y j is the jth subjective rating, S i is the stable component of the ith physiological index, β0 is the intercept term, β iand γ il are the linear term coefficient and the interaction term coefficient respectively, ∈ is the random error term, and m is the number of physiological indicators.

[0151] 2. Based on the above stability relationship modeling results, construct the physiological stability coefficient matrix M s and subjective rating coefficient matrix M p Among them, M s The structure is α ij represents the stable influence coefficient of the ith physiological index on the jth score; M p The structure is β ij It represents the weight coefficient of the i-th rating in the j-th subjective rating, k is the number of rating dimensions, and n is the number of subjective rating types.

[0152] By establishing a stable relationship model between physiological stability components and subjective scores, the influence of each physiological index on different scoring dimensions can be quantified. These coefficient matrices provide an important basis for subsequent comprehensive evaluation.

[0153] Step S60: Based on the physiological index change components, the physiological fluctuation influence coefficient matrix is ​​calculated; and the physiological stability coefficient matrix is ​​corrected according to the physiological fluctuation influence coefficient matrix to obtain a corrected physiological stability coefficient matrix. The purpose of this step is to consider the impact of physiological index changes on the effect of palliative care and improve the accuracy of the evaluation model.

[0154] The specific implementation is as follows:

[0155] 1. Calculate the physiological fluctuation influence coefficient matrix Among them, each element V i (t) is the change component of the ith physiological index, T is the observation time length, S i is the stable component of the ith physiological index. These coefficients reflect the influence of the fluctuation of physiological indexes on each subjective scoring index.

[0156] 2. According to the physiological fluctuation influence coefficient matrix M v The original physiological stability coefficient matrix M s After correction, the corrected physiological stability coefficient matrix M is obtained. s ′=M s ·(I-λM v ). Where I is the unit matrix, λ is the correction coefficient, and its value range is 0.1-0.5.

[0157] By introducing the impact of fluctuations in physiological indicators on the effectiveness of palliative care, the relationship between each physiological indicator and the subjective score can be more accurately portrayed, thereby improving the reliability of the evaluation model.

[0158] Step S70: Establish a patient baseline database and calculate a personalized reference range based on the patient's age, gender, and disease type. The purpose of this step is to establish personalized evaluation reference standards for different patient characteristics to provide a basis for comprehensive evaluation.

[0159] The specific implementation is as follows:

[0160] 1. Establish a patient baseline database containing a large number of medical records, recording basic information such as the patient's age, gender, disease type, as well as corresponding physiological index data and subjective scoring data.

[0161] 2. Calculate the personalized reference range R for each physiological indicator based on the patient's age, gender, disease type and other characteristics i =[μ i -k i σ i ,μ i +k i σ i ]. Among them, μ i and σ i are the mean and standard deviation of the indicator in the baseline data, k i is the adjustment coefficient, which is related to factors such as age and gender.

[0162] By establishing a baseline database containing rich clinical data and calculating personalized reference ranges based on patient characteristics, a more reasonable and accurate reference standard can be provided for subsequent comprehensive evaluation, which helps to improve the pertinence and applicability of the evaluation results.

[0163] Step S80: Use the modified physiological stability coefficient matrix, the subjective scoring coefficient matrix and the personalized reference range to perform weighted fusion calculation to obtain a comprehensive score of the palliative care effect. The purpose of this step is to combine physiological indicators and subjective scoring data to obtain an overall palliative care effect evaluation result.

[0164] The specific implementation is as follows:

[0165] 1. Calculate the comprehensive score of palliative care effect Among them, w ij is the comprehensive weight coefficient, f(x i ,R i ) is the reference range normalization function, x i is the measured value, R i For reference range.

[0166] 2. Modified physiological stability coefficient matrix M s ′ and subjective rating coefficient matrix M pThe relationship between physiological indicators and subjective scores was quantified respectively. By weighting the two, we can comprehensively consider the two dimensions of physiological state and subjective feelings and obtain an overall palliative care effect evaluation result Score.

[0167] 3. In order to make the scoring results more interpretable, the reference range normalization function f(x i ,R i ). This function is based on the measured value x i With personalized reference range R i The relationship between them should be appropriately adjusted.

[0168] By weighted integration of physiological indicators and subjective scoring data, combined with personalized reference ranges, a scoring result that comprehensively reflects the effect of palliative care can be obtained, which provides a basis for subsequent graded evaluation.

[0169] Step S90: grading the palliative care effect comprehensive score according to the preset scoring range to generate a palliative care effect evaluation result. The purpose of this step is to grade and evaluate the palliative care effect according to the comprehensive scoring result to provide a clear judgment basis for clinical application.

[0170] The specific implementation is as follows:

[0171] 1. The following scoring ranges are preset:

[0172] Excellent interval Score excellent =[90,100];

[0173] Good interval Score good =[75,90);

[0174] General interval Score normal =[60,75);

[0175] Score of the area to be improved poor =[0,60);

[0176] 2. Compare the calculated comprehensive score Score with the above-mentioned preset range to determine the level of the patient's palliative care effect evaluation result.

[0177] 3. When determining the specific value of the preset scoring range, the following factors are taken into consideration:

[0178] (1) Collect historical data to establish a benchmark scoring model Calculate the standard deviation of the score And determine the quantile x based on the normal distribution model q =F -1 (q).

[0179] (2) Using the Delphi expert review method, the initially determined interval boundary values ​​were revised And calculate the consistency coefficient of expert opinions

[0180] (3) Compromise between statistical analysis results and expert opinions to determine the final scoring range boundaries Where α is the compromise coefficient, and its value range is [0.3, 0.7].

[0181] (4) Verify the validity of the determined scoring interval And establish a dynamic adjustment mechanism To ensure the rationality and applicability of the evaluation results.

[0182] Through the above steps, an objective and fair graded evaluation of the effect of palliative care can be carried out based on the comprehensive scoring results. This will provide clear effect feedback to clinicians and patients and help further optimize and improve palliative care measures.

[0183] In summary, the hospice care effect evaluation method proposed in this invention makes full use of physiological index data and subjective scoring data, adopts a series of algorithms and analysis techniques such as multivariate mixed decomposition, anomaly detection, model building, and finally obtains a comprehensive and objective comprehensive scoring result. This method can reflect individual differences, taking into account the two dimensions of physiological state and subjective feelings, and provides an innovative evaluation method for clinical medical practice.

[0184] Specifically, the principle of the present invention is: a multi-dimensional analysis technology combining physiological index data and subjective scoring data is adopted, and its core principle is as follows:

[0185] First, the original data required for evaluation is obtained by real-time collection of patients' physiological index data, such as heart rate, blood pressure, respiratory rate, etc., as well as subjective scoring data, such as pain, quality of life, and psychological state. Physiological index data can reflect changes in the patient's physiological state, while subjective scoring data reflects the patient's subjective feelings about the treatment effect. Combining the two types of data for analysis can more comprehensively evaluate the effect of palliative care.

[0186] Secondly, the physiological index data were decomposed into stable components and variable components. The stable components represent the basic state of the physiological indexes, and the variable components represent the fluctuations of the physiological indexes. This laid the foundation for the subsequent establishment of a relationship model between physiological indexes and subjective scores.

[0187] Thirdly, statistical analysis methods are used to extract the time series characteristics of physiological index data, such as mean, standard deviation, kurtosis coefficient and skewness coefficient, etc. These characteristics quantify the distribution characteristics of physiological index data and provide a basis for anomaly detection and model building.

[0188] Based on the above analysis results, the present invention established a mathematical model between the stable component of physiological indicators and subjective scores, and quantified the influence of each physiological indicator on different scoring dimensions. At the same time, the influence of physiological indicator fluctuations on the effect of palliative care was also considered, and the physiological stability coefficient matrix was corrected to improve the accuracy of the model.

[0189] In addition, the present invention establishes a patient baseline database containing a large amount of medical record data, and calculates personalized physiological index reference ranges based on the patient's age, gender, disease type and other characteristics. This personalized evaluation method can better reflect the actual situation of different patients.

[0190] Finally, the present invention weights and fuses the above analysis results to obtain a comprehensive score of the palliative care effect. At the same time, the results are graded according to the preset score range to generate a clear palliative care effect evaluation result. This quantitative evaluation method can not only objectively reflect the treatment effect, but also provide a basis for medical staff and promote the continuous optimization of nursing measures.

[0191] In summary, the palliative care effect evaluation method of the present invention makes full use of physiological indicator data and subjective scoring data, adopts a series of advanced algorithms such as multivariate mixed decomposition, anomaly detection, and model building, and realizes a comprehensive quantitative evaluation of the palliative care effect.

[0192] A specific embodiment 1 of the present invention is provided below. Each step in this embodiment 1 is described in detail as follows: Step S10: Real-time collection of patient physiological indicator data and subjective scoring data of the patient's palliative care effect.

[0193] The purpose of this step is to obtain the original data required for evaluation. First, the physiological signal acquisition equipment is used to record the patient's heart rate data HR(t), blood pressure data BP(t), respiratory rate data RR(t), body temperature data T(t) and blood oxygen saturation data Sp2(t) in real time. These physiological index data can reflect the changes in the patient's physiological state and provide a basis for the subsequent evaluation of the effect of palliative care.

[0194] Secondly, we collected the patients' subjective scores on the effect of their own palliative care through questionnaire surveys. Subjective scores include pain score Pain(t), quality of life score QoL(t) and psychological state score Psy(t). These subjective score data can reflect the patient's subjective feelings about the treatment effect, and combined with physiological index data, the effect of palliative care can be more comprehensively evaluated.

[0195] In summary, by collecting physiological indicator data and subjective scoring data in real time, the original data required for evaluation can be obtained, laying the foundation for subsequent multi-dimensional analysis and comprehensive evaluation.

[0196] Step S20: performing multivariate mixed decomposition calculations on the physiological index data to obtain stable components and variable components of each physiological index.

[0197] The purpose of this step is to decompose the physiological index data and extract the stable components and variable components of each index. First, the original physiological index time series X(t) is decomposed using the multivariate mixed decomposition calculation model X(t) = S(t) + V(t) + N(t). Among them, S(t) represents the stable component, V(t) represents the variable component, and N(t) represents the noise component.

[0198] The decomposition process uses the empirical mode decomposition (EMD) method, that is, Among them, c i (t) is the intrinsic mode function, r n (t) is the residual term, n is the number of decomposition levels, and its value range is 3-8.

[0199] The core idea of ​​EMD decomposition is to decompose the original signal layer by layer into intrinsic mode functions (IMFs) of different time scales and a residual term. Each IMF represents the fluctuation characteristics of the signal on a certain time scale. By continuously decomposing and filtering out high-frequency components, the low-frequency stable component S(t) and the fluctuating component V(t) can be finally obtained.

[0200] Through EMD decomposition, the original physiological index data can be decomposed into a stable component S(t) and a variable component V(t). The stable component S(t) represents the basic state of the physiological index, and the variable component V(t) represents the fluctuation of the physiological index. These two components provide an important basis for the subsequent evaluation of the effect of palliative care.

[0201] Step S30: Calculate the time series characteristics of the physiological indicator data, wherein the time series characteristics include mean, standard deviation, kurtosis coefficient and skewness coefficient.

[0202] The purpose of this step is to extract the statistical characteristics of physiological indicator data and provide a basis for subsequent abnormality detection and model building.

[0203] 1. Calculate the mean Among them, x i is the time series data point, N is the total number of data points. The mean μ reflects the average level of physiological indicators.

[0204] 2. Calculate the standard deviation The standard deviation σ reflects the degree of dispersion of physiological index data.

[0205] 3. Calculate the kurtosis coefficient The kurtosis coefficient K reflects the steepness of the data distribution. When K>0, the data distribution is sharper, and when K<0, the data distribution is flatter.

[0206] 4. Calculate the skewness coefficient The skewness coefficient S reflects the degree of skewness of the data distribution. When S>0, the data distribution is right-skewed, and when S<0, the data distribution is left-skewed.

[0207] By calculating the above time series statistical characteristics, the distribution characteristics of physiological index data can be fully described. These characteristics provide an important basis for subsequent anomaly detection and model building.

[0208] Step S40: Based on the time series characteristics, a physiological index abnormality detection model is established to remove abnormal data points.

[0209] The purpose of this step is to identify and remove outliers in the physiological index data to improve the reliability of the data. The improved local outlier factor (LOF) algorithm is used for anomaly detection. The core idea of ​​the LOF algorithm is to determine whether the sample point is an outlier based on the density relationship between the sample point and its k nearest neighbor points.

[0210] The formula of the LOF algorithm is:

[0211] in:

[0212] LOF k (p) is the local outlier factor of point p. When LOF k When (p)>1, point p is an abnormal point.

[0213] N k (p) is the set of k nearest neighbors of point p.

[0214] lrd k (p) is the local reachability density of point p. The local reachability density reflects the density of the neighborhood around the sample point.

[0215] k is the number of nearest neighbors, ranging from 5 to 20.

[0216] Through this algorithm, outliers that deviate significantly from the normal data distribution can be identified based on the time series characteristics of physiological indicator data. These outliers may be caused by measurement errors, instrument failures, or abnormal physiological states, and need to be removed from the original data to ensure the reliability of subsequent analysis.

[0217] Step S50: establishing a functional relationship between the stable components of each physiological index and each type of subjective scoring data in the subjective scoring data, recorded as a stable relationship, and establishing a physiological stability coefficient matrix and a subjective scoring coefficient matrix respectively according to the stable relationship.

[0218] The purpose of this step is to establish a mathematical model between the stable components of physiological indicators and subjective scores, providing a basis for subsequent comprehensive evaluation.

[0219] 1. Use the polynomial regression model to establish the functional relationship between the stable component of physiological indicators and various subjective scores, that is, in:

[0220] Y j is the jth subjective score, such as pain score Pain(t), quality of life score QoL(t) and psychological state score Psy(t).

[0221] S i is the stable component of the i-th physiological index, such as the stable component of heart rate HR S (t), blood pressure stability component BP S (t) etc.

[0222] β0 is the intercept term, β i and γ il are the linear term coefficient and the interaction term coefficient, respectively.

[0223] ∈ is a random error term.

[0224] m is the number of physiological indicators.

[0225] 2. Based on the above stability relationship modeling results, construct the physiological stability coefficient matrix M s and subjective rating coefficient matrix M p .in:

[0226] M s The structure is α ij It represents the stable influence coefficient of the i-th physiological index on the j-th score.

[0227] M p The structure is β ij It represents the weight coefficient of the i-th category score in the j-th subjective score.

[0228] k is the number of rating dimensions, and n is the number of subjective rating types.

[0229] By establishing a stable relationship model between physiological stability components and subjective scores, the influence of each physiological index on different scoring dimensions can be quantified. These coefficient matrices provide an important basis for subsequent comprehensive evaluation.

[0230] Step S60: Calculate the physiological fluctuation influence coefficient matrix based on the physiological index change components; and modify the physiological stability coefficient matrix according to the physiological fluctuation influence coefficient matrix to obtain a modified physiological stability coefficient matrix.

[0231] The purpose of this step is to consider the impact of changes in physiological indicators on the effectiveness of palliative care and improve the accuracy of the evaluation model.

[0232] 1. Calculate the physiological fluctuation influence coefficient matrix Among them, each element in:

[0233] V i (t) is the variation component of the ith physiological index.

[0234] T is the length of observation time.

[0235] S i is the stable component of the i-th physiological index.

[0236] These coefficients δ ij It reflects the influence of fluctuations in physiological indicators on each subjective scoring indicator.

[0237] 2. According to the physiological fluctuation influence coefficient matrix M v The original physiological stability coefficient matrix M s After correction, the corrected physiological stability coefficient matrix M is obtained. s ′=M s ·(I-λM v ).in:

[0238] I is the identity matrix.

[0239] λ is the correction coefficient, and its value range is 0.1-0.5.

[0240] By introducing the impact of fluctuations in physiological indicators on the effectiveness of palliative care, the relationship between each physiological indicator and the subjective score can be more accurately portrayed, thereby improving the reliability of the evaluation model.

[0241] Step S70: Establish a patient baseline database and calculate a personalized reference range based on the patient's age, gender, and disease type.

[0242] The purpose of this step is to establish personalized evaluation reference standards based on different patient characteristics and provide a basis for comprehensive evaluation.

[0243] 1. Establish a patient baseline database containing a large number of medical records, recording the patient's age, gender, disease type and other basic information, as well as the corresponding physiological index data and subjective scoring data.

[0244] 2. Calculate the personalized reference range R for each physiological indicator based on the patient's age, gender, disease type and other characteristics i =[μ i -k i σ i , μ i +k i σ i ].in:

[0245] μ i and σ i are the mean and standard deviation of the indicator in the baseline data, respectively.

[0246] k i It is the adjustment coefficient, which is related to factors such as age and gender.

[0247] By establishing a baseline database containing rich clinical data and calculating personalized reference ranges based on patient characteristics, a more reasonable and accurate reference standard can be provided for subsequent comprehensive evaluation, which helps to improve the pertinence and applicability of the evaluation results.

[0248] Step S80: Perform weighted fusion calculation using the corrected physiological stability coefficient matrix, the subjective scoring coefficient matrix, and the personalized reference range to obtain a comprehensive score for palliative care effects.

[0249] The purpose of this step is to integrate physiological indicators and subjective scoring data to obtain an overall evaluation result of palliative care effectiveness.

[0250] 1. Calculate the comprehensive score of palliative care effect

[0251] in:

[0252] w ij is the comprehensive weight coefficient.

[0253] is the element in the modified physiological stability coefficient matrix.

[0254] is the element in the subjective rating coefficient matrix.

[0255] f(x i , R i ) is the reference range normalization function, according to the measured value x iWith personalized reference range R i The relationship between them should be appropriately adjusted.

[0256] m is the number of physiological indicators, and n is the number of subjective scoring types.

[0257] 2. Modified physiological stability coefficient matrix M s ′ and subjective rating coefficient matrix M a The relationship between physiological indicators and subjective scores was quantified respectively. By weighting the two, we can comprehensively consider the two dimensions of physiological state and subjective feelings and obtain an overall palliative care effect evaluation result Score.

[0258] 3. Introduce the reference range normalization function f(x i , R i ), according to the measured value x i With personalized reference range R i The relationship between the two factors is analyzed, and the scores are appropriately adjusted to make the results more interpretable.

[0259] By weighted integration of physiological indicators and subjective scoring data, combined with personalized reference ranges, a scoring result that comprehensively reflects the effect of palliative care can be obtained, which provides a basis for subsequent graded evaluation.

[0260] Step S90: grading the comprehensive score of the palliative care effect according to a preset scoring range to generate a palliative care effect evaluation result.

[0261] The purpose of this step is to grade and evaluate the effectiveness of palliative care based on the comprehensive scoring results, providing a clear basis for judgment for clinical applications.

[0262] 1. The following scoring ranges are preset:

[0263] Excellent interval Score excellent =[90,100];

[0264] Good interval Score good =[75,90);

[0265] General interval Score normal =[60,75);

[0266] Score of the area to be improved poor =[0,60);

[0267] 2. Compare the calculated comprehensive score Score with the above-mentioned preset range to determine the level of the patient's palliative care effect evaluation result.

[0268] When determining the specific value of the preset scoring range, the following factors are taken into consideration:

[0269] (1) Collect historical data to establish a benchmark scoring model Score i is the score of the ith historical case, and N is the total number of historical cases.

[0270] (2) Calculate the standard deviation of the score

[0271] (3) Rating-based normal distribution model Calculate the quantile x q =F -1 (q).

[0272] (4) Determine the interval boundary value: B excellent =x 0.9 , B good =x 0.75 , B normal =x 0.6 .

[0273] (5) Use the Delphi expert review method to modify the initially determined boundary values Calculate the coefficient of agreement of expert opinions

[0274] (6) Compromise between statistical analysis results and expert opinions to obtain the final interval boundaries Where α is the compromise coefficient, and its value range is [0.3, 0.7].

[0275] (7) Verify the validity of the determined scoring interval And establish a dynamic adjustment mechanism Where η is the learning rate, and its value range is [0.1, 0.3].

[0276] Through the above steps, an objective and fair graded evaluation of the effect of palliative care can be carried out based on the comprehensive scoring results. This will provide clear effect feedback to clinicians and patients and help further optimize and improve palliative care measures.

[0277] In order to further better understand and implement the present invention, the following provides Example 2 of a specific application scenario of the present invention: A 65-year-old male patient with advanced lung cancer, Mr. Yao, was admitted to the hospice care center of a hospital in a certain city in a certain province. In order to comprehensively evaluate the effect of Mr. Yao's hospice care, the center adopted the evaluation method proposed in the present invention for evaluation.

[0278] First, the medical staff collected Yao's physiological index data and subjective scoring data in real time. The physiological index data included heart rate HR(t), systolic blood pressure SBP(t), respiratory rate RR(t), body temperature T(t) and blood oxygen saturation SpO2(t). The subjective scoring data included pain score Pain(t), quality of life score QoL(t) and psychological state score Psy(t). The specific collection results are shown in Table 1.

[0279] Table 1 Data collection table

[0280]

[0281]

[0282] Next, the collected physiological index data is subjected to multivariate mixed decomposition to obtain the stable component and variable component of each index. Taking heart rate HR(t) as an example, the decomposition results are as follows: Figure 2 As shown, the original heart rate data HR(t) is decomposed into the stable component HR S (t) and the variable component HR V (t). The stable component reflects the basic state of the heart rate, and the variable component represents the fluctuation of the heart rate. Similarly, other physiological indicators were also decomposed in the same way.

[0283] On this basis, the time series characteristics of each physiological indicator data were calculated, including mean, standard deviation, kurtosis coefficient and skewness coefficient, as shown in Table 2. These characteristics quantify the distribution characteristics of physiological indicator data and provide a basis for subsequent anomaly detection.

[0284] Table 2 Data characteristics table

[0285] index Mean Standard Deviation Kurtosis Skewness coefficient HR 91.0 2.28 -0.59 0.07 SBP 133.7 3.01 -0.76 0.16 RR 21.5 1.12 0.35 0.12 T 36.95 0.12 0.40 0.35 <![CDATA[SpO2]]> 92.5 1.05 -1.11 -0.43

[0286] Based on the time series characteristics of physiological index data, the improved local outlier factor (LOF) algorithm was used to detect anomalies in the data. The results showed that there were no outliers in Yao's physiological index data during the observation time, and subsequent model establishment and comprehensive evaluation could be carried out.

[0287] Next, the present invention establishes a mathematical model between the stable component of physiological indicators and subjective scores. Taking the pain score Pain(t) as an example, it is related to the stable component of heart rate HR S (t), stable systolic blood pressure component SBP S The relationship between (t) is as follows:

[0288] Pain(t)=8.72-0.11HR S (t)-0.09SBP S (t)+0.02HR S (t)SBPS (t)+∈;

[0289] Among them, β0 = 8.72, β1 = -0.11, β2 = -0.09, γ = 0.02, ∈ is a random error term. Through a similar modeling process, the functional relationship between each subjective scoring index and the stable component of the physiological index can be obtained, thereby constructing the physiological stability coefficient matrix M s and subjective rating coefficient matrix M p .

[0290] In addition, the present invention also considers the impact of changes in physiological indicators on the effect of palliative care. Taking the heart rate change component HRV(t) as an example, its influence coefficient δ on the pain score Pain(t) 11 The calculation is as follows:

[0291]

[0292] Similarly, the influence coefficients of other physiological index change components on each subjective score can be calculated, thereby constructing the physiological fluctuation influence coefficient matrix Mv.

[0293] Based on the above analysis results, the present invention modifies the original physiological stability coefficient matrix Ms to obtain the modified matrix Ms′:

[0294]

[0295] Next, the present invention established a patient baseline database, recording a large number of similar patients' physiological index data and subjective scoring data. According to Yao's age of 65, male gender and disease type lung cancer, his personalized physiological index reference range was calculated, as shown in Table 3.

[0296] Table 2 Reference range of personalized indicators

[0297] index Reference range HR [84,98] SBP [125,145] RR [19,24] T [36.8,37.2] SpO2 [91,96]

[0298] With the above analysis basis, the present invention can calculate the comprehensive score of Yao's palliative care effect:

[0299]

[0300] Among them, wij is the comprehensive weight coefficient, and f(xi,Ri) is the reference range normalization function. After calculation, Yao's comprehensive score Score = 83.2 points.

[0301] According to the preset scoring range, the score falls in the good range [75,90), indicating that Yao's palliative care is effective. Based on this result, medical staff can continue to optimize Yao's nursing measures to further improve his quality of life. At the same time, the center can also conduct similar dynamic assessments on other patients to provide a basis for the continuous improvement of palliative care services.

[0302] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for evaluating the effect of palliative care, characterized in that: The following steps are involved: S10, real-time collection of patients' physiological index data and subjective scoring data of patients' palliative care effects; S20, performing multivariate mixed decomposition calculation on the physiological index data to obtain stable components and variable components of each physiological index; S30, calculating the time series characteristics of the physiological indicator data; S40, based on the time series characteristics, establishing a physiological indicator abnormality detection model to remove abnormal data points; S50, establishing a functional relationship between the stable components of each physiological index and each subjective scoring data in the subjective scoring data, recorded as a stable relationship, and establishing a physiological stability coefficient matrix and a subjective scoring coefficient matrix respectively according to the stable relationship; S60, calculating a physiological fluctuation influence coefficient matrix based on the physiological index change components; and correcting the physiological stability coefficient matrix according to the physiological fluctuation influence coefficient matrix to obtain a corrected physiological stability coefficient matrix; S70. Establish a patient baseline database and calculate personalized reference ranges based on the patient's age, gender, and disease type; S80, performing weighted fusion calculation using the modified physiological stability coefficient matrix, the subjective scoring coefficient matrix, and the personalized reference range to obtain a comprehensive score for palliative care effects; S90. Classify the comprehensive score of the palliative care effect according to a preset score range to generate a palliative care effect evaluation result.

2. A palliative care effect evaluation method according to claim 1, characterized in that: The time series characteristics include mean, standard deviation, kurtosis coefficient, and skewness coefficient.

3. A palliative care effect evaluation method according to claim 1, characterized in that: The preset scoring intervals include an excellent interval, a good interval, a general interval, and an interval for improvement.

4. A palliative care effect evaluation method according to claim 3, characterized in that: The preset scoring intervals are specifically: excellent interval [90, 100], good interval [75, 90), average interval [60, 75), and interval for improvement [0, 60).

5. A palliative care effect evaluation method according to claim 1, characterized in that: The physiological indicator data include heart rate data, blood pressure data, respiratory rate data, body temperature data, and blood oxygen saturation data, which are objectively collected data.

6. A palliative care effect evaluation method according to claim 1, characterized in that: The subjective scoring data includes pain scoring data, quality of life scoring data and psychological state scoring data.

7. A palliative care effect evaluation method according to claim 6, characterized in that: The subjective scoring data are collected through questionnaires from patients and their families regarding the effectiveness of palliative care.

8. A palliative care effect evaluation method according to claim 1, characterized in that: The anomaly detection model is based on a local anomaly factor algorithm or a linear interpolation algorithm.

9. A palliative care effect evaluation method according to claim 1, characterized in that: The stable relationship adopts a polynomial regression model.

10. A palliative care effect evaluation method according to claim 1, characterized in that: The multivariate mixed decomposition calculation is specifically implemented by using the method of empirical mode decomposition or principal component analysis or dynamic factor model.

Images