Clinical data statistical management analysis method and storage medium
Through standardized questionnaire design, automation tools and interactive charts, advanced missing value processing and multiple correction methods, the entry errors and data security problems in clinical data statistical analysis are solved, and the high accuracy and security of the data are achieved, and the high accuracy and security of the data are improved. The credibility and efficiency of the analysis results.
Patent Information
- Application Number
- CN202411185253.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-12-03
AI Technical Summary
The existing clinical data statistical analysis methods have entry errors, sample bias, inadequate data organization, poor interactiveness of visualization tools, low level of statistical summary details, and lack of encryption and access control of data storage and management, resulting in the reliability of data quality and analysis results and Security is affected.
Design standardized questionnaires, use automation tools and interactive charts, adopt advanced missing value processing methods and multiple comparison correction methods, perform data preprocessing and cross-validation, ensure data security and privacy protection, and establish feedback mechanisms to optimize data management processes .
By reducing entry errors and biases, improving data accuracy and representativeness, enhancing the credibility and transparency of analysis results, ensuring data security and ethical compliance, optimizing data management and analysis processes, and improving overall work quality and efficiency.
Smart Images

Figure CN119069056A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of clinical medical technology, and in particular to a clinical data statistical management and analysis method and a storage medium. Background Art
[0002] Clinical research uses scientific methods and standards to study and evaluate the causes of diseases, determine and evaluate disease diagnosis methods, and treat and prevent diseases, so as to develop and progress clinical medicine. In clinical research, statistical analysis of clinical data runs through the clinical research process. In modern clinical research and data analysis, the accuracy, completeness and security of data are crucial. However, existing technologies are deficient in certain aspects, which affect the quality of clinical data and analysis results:
[0003] Traditional data collection and entry methods rely on manual operations, which can lead to entry errors and data bias. In addition, if there is no strict standardization process during sample collection, it will lead to sample bias, affecting the representativeness of the data and research results.
[0004] Data sorting involves manual operations, which can easily lead to human errors and omissions. Traditional sorting methods cannot guarantee the systematicness and integrity of the data cleaning process. When dealing with missing values, simple filling strategies are used, which cannot fully utilize the information in the data and affect the reliability of the analysis results.
[0005] The interactivity of existing visualization tools makes it difficult for users to intuitively understand the complexity and trends of data. In addition, the low level of detail of statistical summaries limits the depth of data interpretation, which limits the credibility and practical application value of research results.
[0006] Traditional data storage and management methods lack encryption and access control measures, increasing the risk of data leakage and loss, resulting in the inability to promptly identify and resolve problems in data management and analysis, affecting overall work quality and efficiency.
[0007] Therefore, those skilled in the art provide a clinical data statistical management and analysis method and a storage medium to solve the problems raised in the above background technology. Summary of the invention
[0008] In view of the deficiencies of the prior art, the present invention provides a clinical data statistical management and analysis method and a storage medium to solve the problems raised in the above background technology.
[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: A clinical data statistical management and analysis method comprises the following steps:
[0010] S1. Design standardized questionnaires or data forms and conduct pilot tests before formal data collection to identify and correct potential problems;
[0011] S2. Use automated tools for data entry and organization to reduce manual errors, establish a systematic data cleaning process, and use advanced missing value processing methods to record processing strategies;
[0012] S3. Use interactive charts and visualization tools to help users understand the data;
[0013] S4. To ensure that the data analysis complies with the statistical assumptions, data preprocessing is performed to meet the assumptions, and multiple comparison correction methods are used to reduce the risk of false positives;
[0014] S5. Confirm the reliability of the results through cross-validation and external validation methods to combine the statistical results with the actual clinical situation, emphasize the clinical application value of the results, and explain in detail the steps and reasons in the data analysis process;
[0015] S6. Follow the standardized report format, ensure that each section is complete and logically clear, and fully analyze the significance and limitations of the results in the discussion, and propose specific future research directions;
[0016] S7. Use data encryption and access control measures to ensure data security, strictly comply with privacy protection regulations, ensure data management complies with ethical requirements, and regularly back up data to prevent data loss or damage;
[0017] S8. Establish a regular feedback mechanism to collect feedback from all aspects of data collection, analysis and reporting, implement continuous improvement based on the feedback, and optimize data management and analysis processes.
[0018] Preferably, in step S1, the internal consistency of the questionnaire is evaluated by a coefficient formula, and the coefficient algorithm formula is:
[0019]
[0020] Where k is the number of questions in the questionnaire, is the variance of the ith question, is the variance of the total score of the questionnaire;
[0021] In the S1 step, the independent sample t-test algorithm is used to compare whether there is a significant difference between the results of the pre-test and the results before formal data collection. The algorithm formula is:
[0022]
[0023] in, and is the sample mean, and is the variance of the sample, n1 and n2 are the sample sizes;
[0024] In step S1, potential patterns and errors in the data are identified by cluster analysis method. The algorithm formula is:
[0025]
[0026] Where k is the number of clusters, C i is the sample set of the i-th class, μ i is the center point of the i-th class, and x is the sample point.
[0027] Preferably, in the step S2, missing value interpolation and K nearest neighbor interpolation are used to improve the accuracy and effectiveness of data sorting and cleaning;
[0028] The operation steps of missing value interpolation are as follows:
[0029] Multiple imputation is an advanced method for dealing with missing values, which reduces uncertainty by generating multiple imputation datasets and then merging the results;
[0030] The formula for Bayesian interpolation is:
[0031] Generate an imputed dataset X with missing values i (i=1,2,…,M):
[0032] X i =X obs ∪X mis,i ,
[0033] Among them, X obs is the observed data, X mis,i is the missing value interpolation part of the i-th interpolation data set;
[0034] Analyze each imputed data set and calculate estimates and standard errors and
[0035] Merge results:
[0036]
[0037] The operation steps of the K nearest neighbor interpolation are:
[0038] Calculate the sample point x i To other sample points x j Distance:
[0039]
[0040] The k neighbors with the smallest distance calculate the interpolation value:
[0041]
[0042] Preferably, in step S3, high-dimensional data is converted into low-dimensional data by principal component analysis, and main features of the data are retained;
[0043] The principal component analysis algorithm steps and formula:
[0044] Compute the covariance matrix of the data:
[0045]
[0046] Where X is the data matrix, is the mean of each column of data, C is the covariance matrix;
[0047] Compute the eigenvalues and eigenvectors of the covariance matrix:
[0048] Cv i =λ i v i ,
[0049] Among them, v i is the eigenvector, λ i is the eigenvalue;
[0050] The eigenvectors with the largest eigenvalues constitute the new data space;
[0051] Convert the data:
[0052] X pca =XW,
[0053] Among them, W is a matrix consisting of the first k eigenvectors, X pca is the data matrix after dimensionality reduction.
[0054] Preferably, in step S3, high-dimensional data is converted into low-dimensional data by principal component analysis, and main features of the data are retained;
[0055] The principal component analysis algorithm steps and formula:
[0056] Compute the covariance matrix of the data:
[0057]
[0058] Where X is the data matrix, is the mean of each column of data, C is the covariance matrix;
[0059] Compute the eigenvalues and eigenvectors of the covariance matrix:
[0060] Cv i =λ i v i ,
[0061] Among them, vi is the eigenvector, λ i is the eigenvalue;
[0062] The eigenvectors with the largest eigenvalues constitute the new data space;
[0063] Convert the data:
[0064] X pca =XW,
[0065] Among them, W is a matrix consisting of the first k eigenvectors, X pca is the data matrix after dimensionality reduction.
[0066] Preferably, in step S5, a technique for evaluating model performance is used by cross-validation to perform training and testing by dividing the data set into multiple subsets;
[0067] The k-fold cross validation:
[0068] Dataset partitioning: Divide the dataset into k folds;
[0069] Training and testing: For each fold i, use the other folds as training sets and the i-th fold as the test set;
[0070] Model Evaluation:
[0071]
[0072] Among them, CV k is the average performance of k-fold cross validation, Performance i is the model performance index of the i-th fold;
[0073] The leave-one-out cross-validation is a special case of k-fold cross-validation, where k is equal to the total number of samples n;
[0074] Training and testing: For each sample i, use other samples as training sets and the i-th sample as test set;
[0075] Model Evaluation:
[0076]
[0077] Among them, LOOCV is the average performance index of leave-one-out cross validation, Performance i is the model performance indicator for the i-th sample.
[0078] Preferably, in step S6, the linear regression model diagnosis of the model diagnosis can be achieved by calculating the standardization of the residual and the influence index;
[0079] The standardized residuals are:
[0080]
[0081] Among them, r i is the residual of the ith observation, MSE is the mean square error, and h ii is the leverage value;
[0082] Impact indicators:
[0083]
[0084] Among them, e i is the standardized residual, p is the number of parameters in the model, and h ii is the leverage value.
[0085] Model fitness:
[0086] R 2 and the adjusted R 2 :
[0087]
[0088] Among them, SS res is the residual sum of squares, SS tot is the total sum of squares, n is the sample size, and p is the number of parameters in the model.
[0089] Preferably, in step S7, the role-based access control system is used to control access to system resources according to the user's role, and the role-based access control model formula is:
[0090] Role authorization:
[0091] RoleAssignments={(u,r)|u∈Uandr∈Rand(u,r)is assigned},
[0092] Among them, u is the user set, R is the role set, (u,r) represents the role r assigned to user u;
[0093] Permission allocation:
[0094] PermissionAssignments={(r,p)|r∈RandPand(r,p)is assigned},
[0095] Where P is a permission set, and (r,p) means that role r is granted permission p.
[0096] Preferably, in the step S8, when analyzing the effectiveness of the feedback mechanism, a statistical method may be used to evaluate the impact of the feedback;
[0097] Before-after comparative analysis: Calculate the changes in key indicators before and after feedback to evaluate the effectiveness of feedback;
[0098] ΔX=X post-feedback -X pre-feedback ,
[0099] Among them, X pre-feedback is the indicator value before feedback, X post-feedback is the indicator value after feedback, ΔX is the change in the indicator;
[0100] Paired sample t test: used to detect whether the change of indicators before and after feedback is significant;
[0101]
[0102] Where D is the mean of the paired differences, s D is the standard deviation of the paired differences, and n is the sample size.
[0103] A computer-readable storage medium storing at least one instruction, wherein when the at least one instruction is executed by a processor, the clinical data statistical management and analysis method according to any one of claims 1 to 9 is implemented.
[0104] The present invention provides a clinical data statistical management and analysis method and a storage medium. It has the following beneficial effects:
[0105] 1. The present invention reduces data entry errors and deviations through standardized tools and personnel training to ensure data accuracy and consistency. The pre-test and standardization process reduces sample selection bias and improves sample representativeness. Automated tools and systematic data cleaning processes reduce human errors and omissions and optimize data accuracy. At the same time, advanced missing value processing methods maintain data integrity.
[0106] 2. The present invention helps researchers and readers to intuitively understand the data and identify key trends and distribution characteristics through interactive data visualization tools and detailed statistical summaries, which helps to discover potential problems and opportunities. The result verification and focus on clinical practical significance improve the credibility of the analysis results. The detailed explanation of the analysis process and results improves the transparency and repeatability of the research. The detailed appendices support further review and verification and assist in decision-making.
[0107] 3. The present invention improves data security through data encryption and access control measures to prevent data leakage or loss, while privacy protection policies and compliance measures ensure that data management complies with ethical and legal requirements and protects the privacy of subjects. At the same time, regular feedback and continuous improvement measures optimize data management and analysis processes to improve overall work quality and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0108] Figure 1It is a flow chart of the present invention. DETAILED DESCRIPTION
[0109] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0110] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0111] The present invention is described in detail below in conjunction with the accompanying drawings:
[0112] Example:
[0113] Please see attached Figure 1 , an embodiment of the present invention provides a clinical data statistical management and analysis method, comprising the following steps: S1, designing a standardized questionnaire or data form, and conducting a preliminary test before formal data collection to identify and correct potential problems;
[0114] S2. Use automated tools for data entry and organization to reduce manual errors, establish a systematic data cleaning process, and use advanced missing value processing methods to record processing strategies;
[0115] S3. Use interactive charts and visualization tools to help users understand the data;
[0116] S4. To ensure that the data analysis complies with the statistical assumptions, data preprocessing is performed to meet the assumptions, and multiple comparison correction methods are used to reduce the risk of false positives;
[0117] S5. Confirm the reliability of the results through cross-validation and external validation methods to combine the statistical results with the actual clinical situation, emphasize the clinical application value of the results, and explain in detail the steps and reasons in the data analysis process;
[0118] S6. Follow the standardized report format, ensure that each section is complete and logically clear, and fully analyze the significance and limitations of the results in the discussion, and propose specific future research directions;
[0119] S7. Use data encryption and access control measures to ensure data security, strictly comply with privacy protection regulations, ensure data management complies with ethical requirements, and regularly back up data to prevent data loss or damage;
[0120] S8. Establish a regular feedback mechanism to collect feedback from all aspects of data collection, analysis and reporting, implement continuous improvement based on the feedback, and optimize data management and analysis processes.
[0121] S1 Standardized questionnaires or data forms help to collect data consistently, reduce personal biases and errors in records, ensure the smooth progress of the data collection process, and enhance the reliability of research results; S2: Automated tools reduce errors in manual input and improve data accuracy. Automated and systematic data cleaning processes speed up data processing and reduce manual intervention; S3 Interactive charts and visualization tools make data intuitive and easy to understand, helping researchers and readers quickly identify key trends and patterns, revealing the distribution and trends of data, and helping to identify potential problems and opportunities; S4 Data preprocessing ensures that the analysis meets statistical assumptions, improves the validity of the analysis results, and reduces the risk of false positive results. Ensure the accuracy of research results; S5 cross-validation and external validation methods confirm the reliability of the results, improve the clinical application value of the results, and make the results easier to understand and verify; S6 follow the standardized report format to ensure the integrity of the report content, clear logic, and easy for readers to understand, helping readers to fully understand the impact of the study and further research needs; S7 data encryption and access control measures improve data security, prevent data leakage or loss, ensure that data management complies with ethical and legal standards, and protect the privacy of subjects; S8: Regularly collect feedback and make improvements, optimize data management and analysis processes, improve overall work quality and efficiency, avoid repeated errors, and improve work efficiency and quality.
[0122] In step S1, the internal consistency of the questionnaire is evaluated by the coefficient formula. The coefficient algorithm formula is:
[0123]
[0124] Where k is the number of questions in the questionnaire, is the variance of the ith question, is the variance of the total score of the questionnaire;
[0125] In step S1, the independent sample t-test algorithm is used to compare whether there is a significant difference between the results of the pre-test and the results before formal data collection. The algorithm formula is:
[0126]
[0127] in, and is the sample mean, and is the variance of the sample, n1 and n2 are the sample sizes;
[0128] In step S1, cluster analysis is used to identify potential patterns and errors in the data. The algorithm formula is:
[0129]
[0130] Where k is the number of clusters, C i is the sample set of the i-th class, μ i is the center point of the i-th class, and x is the sample point.
[0131] The coefficient formula is used to evaluate the consistency between the questions in the questionnaire. A high internal consistency coefficient indicates that the correlation between the questions in the questionnaire is stronger and the internal consistency of the questionnaire is better, which helps to ensure that the measurement tool of the questionnaire can consistently reflect the constructs or traits that the researcher is concerned about, and improve the reliability of the data;
[0132] The independent sample t-test is used to compare whether the means of independent samples are significantly different. It helps to confirm whether there are significant differences between the data results of the pre-test and before the formal data collection, and to verify whether the questionnaire or data table has been optimized before the formal data collection.
[0133] Cluster analysis is used to group sample data into several subsets with similar characteristics, identify potential patterns and structures in the data, help correct problems in data input, improve the data sorting process, and improve the accuracy and efficiency of data analysis.
[0134] In step S2, missing value interpolation and K nearest neighbor interpolation are used to improve the accuracy and effectiveness of data sorting and cleaning;
[0135] The operation steps of missing value interpolation are:
[0136] Multiple imputation is an advanced method for dealing with missing values, which reduces uncertainty by generating multiple imputation datasets and then merging the results;
[0137] The formula for Bayesian interpolation is:
[0138] Generate an imputed dataset X with missing values i (i=1,2,…,M):
[0139] X i =X obs ∪X mis,i ,
[0140] Among them, X obs is the observed data, X mis,i is the missing value interpolation part of the i-th interpolation data set;
[0141] Analyze each imputed data set and calculate estimates and standard errors and
[0142] Merge results:
[0143]
[0144] The operation steps of K nearest neighbor interpolation are:
[0145] Calculate the sample point x i To other sample points x j Distance:
[0146]
[0147] The k neighbors with the smallest distance calculate the interpolation value:
[0148]
[0149] Bayesian imputation for multiple imputation handles missing data by generating multiple imputation data sets, each of which generates an estimated result, reduces uncertainty by merging the results, improves the reliability of the final estimate, and provides more precise estimates and standard errors;
[0150] K-nearest neighbor interpolation fills in missing values by using the mean of the closest sample points and using the local structure in the data to provide reasonable interpolation values. It can handle missing values while maintaining the data structure and reduce the interference of interpolation on the overall pattern of the data.
[0151] In step S3, high-dimensional data is converted into low-dimensional data through principal component analysis, and the main features of the data are retained;
[0152] Principal component analysis algorithm steps and formula:
[0153] Compute the covariance matrix of the data:
[0154]
[0155] Where X is the data matrix, is the mean of each column of data, C is the covariance matrix;
[0156] Compute the eigenvalues and eigenvectors of the covariance matrix:
[0157] Cv i =λ i vi ,
[0158] Among them, v i is the eigenvector, λ i is the eigenvalue;
[0159] The eigenvectors with the largest eigenvalues constitute the new data space;
[0160] Convert the data:
[0161] X pca =XW,
[0162] Among them, W is a matrix consisting of the first k eigenvectors, X pca is the data matrix after dimensionality reduction.
[0163] Calculate the covariance matrix of the data. The covariance matrix describes the linear relationship between the features in the data, reflecting the variance of the features and the degree of their common variation. The covariance matrix provides a measure of the distribution of the data and helps to determine the main directions in the data;
[0164] The eigenvectors represent the main directions of variation in the data, and the eigenvalues measure the degree of variation in these directions, helping to identify the most important dimensions in the data;
[0165] The function of the data transformation formula is to project the original data matrix X into a new space composed of the first k principal components to obtain the reduced-dimensional data Z. The reduced-dimensional data Z can reduce the computational complexity and data storage requirements while retaining the main features, which is helpful for subsequent data analysis and visualization.
[0166] In step S4, Z-score standardization is used to transform the data into a standard normal distribution with a mean of 0 and a standard deviation of 1:
[0167]
[0168] Among them, x i is the original data point, μ is the mean of the data, and σ is the standard deviation of the data;
[0169] In step S4, the non-normal distribution of the data is handled by logarithmic transformation:
[0170] x i,log =log(x i +c),
[0171] Among them, x i are the original data points and c is a constant used to handle zero values.
[0172] The Z-score standardization formula converts the data into a standard normal distribution with a mean of 0 and a standard deviation of 1. The data can be compared and analyzed on a uniform scale. Z-score standardization makes the data conform to these statistical assumptions and improves the accuracy and reliability of the analysis.
[0173] The function of the logarithmic transformation formula is to transform the right-skewed data into a form that is closer to a normal distribution. The data is more consistent with the normality assumption commonly used in analysis, making the data more symmetrical and meeting the requirements of many statistical analysis methods. This can avoid this situation and enable the data points to be logarithmically transformed.
[0174] In step S5, the cross-validation technique used to evaluate the performance of the model can be used to train and test by dividing the dataset into multiple subsets;
[0175] K-fold cross validation:
[0176] Dataset partitioning: Divide the dataset into k folds;
[0177] Training and testing: For each fold i, use the other folds as training sets and the i-th fold as the test set;
[0178] Model Evaluation:
[0179]
[0180] Among them, CV k is the average performance of k-fold cross validation, Performance i is the model performance index of the i-th fold;
[0181] Leave-one-out cross-validation is a special case of k-fold cross-validation, where k is equal to the total number of samples n;
[0182] Training and testing: For each sample i, use other samples as training sets and the i-th sample as test set;
[0183] Model Evaluation:
[0184]
[0185] Among them, LOOCV is the average performance index of leave-one-out cross validation, Performance i is the model performance indicator for the i-th sample.
[0186] The k-fold cross-validation formula trains and tests the model on different data subsets to evaluate the performance of the model on unseen data, reducing evaluation bias caused by improper data segmentation, making performance evaluation more stable and reliable, effectively utilizing data, and improving the effectiveness of model evaluation.
[0187] The leave-one-out cross-validation formula provides a model evaluation that is closest to the actual situation, especially when data is limited, it can obtain stable evaluation results, ensure that the model evaluation covers data points, and reduce randomness in the training and testing process.
[0188] In step S6, the linear regression model diagnosis of the model diagnosis can be achieved by calculating the standardization of the residuals and the influence index;
[0189] Standardized residuals:
[0190]
[0191] Among them, r i is the residual of the ith observation, MSE is the mean square error, and h ii is the leverage value;
[0192] Impact indicators:
[0193]
[0194] Among them, e i is the standardized residual, p is the number of parameters in the model, and h ii is the leverage value.
[0195] Model fitness:
[0196] R 2 and the adjusted R 2 :
[0197]
[0198] Among them, SS res is the residual sum of squares, SS tot is the total sum of squares, n is the sample size, and p is the number of parameters in the model.
[0199] The standardized residual formula helps identify outliers, adjusts the impact of leverage values, and makes residuals at different observation points directly comparable, helping to discover potential problems in the data;
[0200] The influence index formula combines the square of the residual and the leverage value to help identify observations that have a significant impact on the regression model fit. A high influence index indicates that the observation has a greater influence on the prediction and fit of the model.
[0201] RSS measures the difference between the model prediction and the actual observed value;
[0202] The total sum of squares formula measures the total variability of the observations and calculates the explanatory power of the model;
[0203] The coefficient of determination formula reflects the degree to which the model explains the variation in the data;
[0204] The adjusted coefficient of determination formula indicates that the model fits the data well and achieves a good balance between complexity and fit.
[0205] In step S7, the role-based access control system is used to control access to system resources based on the user's role. The role-based access control model formula is:
[0206] Role authorization:
[0207] RoleAssignments={(u,r)|u∈Uandr∈Rand(u,r)is assigned},
[0208] Among them, u is the user set, R is the role set, (u,r) represents the role r assigned to user u;
[0209] Permission allocation:
[0210] PermissionAssignments={(r,p)|r∈RandPand(r,p)is assigned},
[0211] Where P is a permission set, and (r,p) means that role r is granted permission p.
[0212] The role authorization formula determines the role of each user. Through the mapping relationship between users and roles (u, r), users obtain access rights through the assigned roles. Roles aggregate permissions, associate permissions with roles, and directly associate permissions with users, ensuring that users can only access resources allowed by their roles, thereby reducing the complexity and maintenance costs of permission management.
[0213] The authority allocation formula determines the authority of each role. Through the mapping relationship between roles and permissions (r, p), roles aggregate permissions, reducing management complexity and the probability of errors, and can effectively control and manage access rights to system resources, thereby enhancing system security.
[0214] In step S8, when analyzing the effectiveness of the feedback mechanism, statistical methods can be used to evaluate the impact of feedback;
[0215] Before-after comparative analysis: Calculate the changes in key indicators before and after feedback to evaluate the effectiveness of feedback;
[0216] ΔX=X post-feedback -X pre-feedback ,
[0217] Among them, X pre-feedback is the indicator value before feedback, X post-feedbackis the indicator value after feedback, ΔX is the change in the indicator;
[0218] Paired sample t-test: used to detect whether the change of indicators before and after feedback is significant;
[0219]
[0220] Where D is the mean of the paired differences, s D is the standard deviation of the paired differences, and n is the sample size.
[0221] The before-after comparison analysis formula calculates the change in key indicators before and after feedback, which can provide intuitive feedback on whether there is improvement after implementation. The change in indicators reflects the impact of feedback on related indicators, helping to determine whether the improvement measures have achieved the expected results.
[0222] The paired sample t-test formula is used to evaluate whether the changes in indicators before and after feedback are statistically significant. By testing whether the mean of the paired differences is significantly different from zero, it can be determined whether the feedback measures produce significant improvement effects.
[0223] A computer-readable storage medium storing at least one instruction, wherein when the at least one instruction is executed by a processor, the clinical data statistical management and analysis method according to any one of claims 1 to 9 is implemented.
[0224] By storing specific algorithm formulas and data processing steps, such as data standardization, missing value interpolation, principal component analysis, cross-validation, etc., it ensures that each step is performed according to the predetermined method to achieve efficient data management and analysis;
[0225] The computer-readable storage medium compiles the instructions into computer-executable code for execution by the processor, which performs data processing and analysis, including data cleaning, statistical analysis, result verification, and report generation, according to the instructions;
[0226] The storage medium may be a software program, a script file, a storage object in a database management system, etc., to ensure that the steps can be automatically executed on the computer to improve work efficiency.
[0227] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A clinical data statistical management and analysis method, characterized in that: The steps include: S1. Design standardized questionnaires or data forms and conduct pilot tests before formal data collection to identify and correct potential problems; S2. Use automated tools for data entry and organization to reduce manual errors, establish a systematic data cleaning process, and use advanced missing value processing methods to record processing strategies; S3. Use interactive charts and visualization tools to help users understand the data; S4. To ensure that the data analysis complies with the statistical assumptions, data preprocessing is performed to meet the assumptions, and multiple comparison correction methods are used to reduce the risk of false positives; S5. Confirm the reliability of the results through cross-validation and external validation methods to combine the statistical results with the actual clinical situation, emphasize the clinical application value of the results, and explain in detail the steps and reasons in the data analysis process; S6. Follow the standardized report format, ensure that each section is complete and logically clear, and fully analyze the significance and limitations of the results in the discussion, and propose specific future research directions; S7. Use data encryption and access control measures to ensure data security, strictly comply with privacy protection regulations, ensure data management complies with ethical requirements, and regularly back up data to prevent data loss or damage; S8. Establish a regular feedback mechanism to collect feedback from all aspects of data collection, analysis and reporting, implement continuous improvement based on the feedback, and optimize data management and analysis processes.
2. The clinical data statistical management and analysis method according to claim 1, characterized in that: In step S1, the internal consistency of the questionnaire is evaluated by the coefficient formula. The coefficient algorithm formula is: Where k is the number of questions in the questionnaire, is the variance of the ith question, is the variance of the total score of the questionnaire; In the S1 step, the independent sample t-test algorithm is used to compare whether there is a significant difference between the results of the pre-test and the results before formal data collection. The algorithm formula is: in, and is the sample mean, and is the variance of the sample, n1 and n2 are the sample sizes; In step S1, potential patterns and errors in the data are identified by cluster analysis method. The algorithm formula is: Where k is the number of clusters, C i is the sample set of the i-th class, μ i is the center point of the i-th class, and x is the sample point.
3. The clinical data statistical management and analysis method according to claim 1, characterized in that: In the step S2, missing value interpolation and K nearest neighbor interpolation are used to improve the accuracy and effectiveness of data sorting and cleaning; The operation steps of missing value interpolation are as follows: Multiple imputation is an advanced method for dealing with missing values, which reduces uncertainty by generating multiple imputation datasets and then merging the results; The formula for Bayesian interpolation is: Generate an imputed dataset X with missing values i (i=1,2,…,M): X i =X obs ∪X mis,i , Among them, X obs is the observed data, X mis,i is the missing value interpolation part of the i-th interpolation data set; Analyze each imputed data set and calculate estimates and standard errors and Merge results: The operation steps of the K nearest neighbor interpolation are: Calculate the sample point x i To other sample points x j Distance: The k neighbors with the smallest distance calculate the interpolation value:
4. The clinical data statistical management and analysis method according to claim 3, characterized in that: In the step S3, high-dimensional data is converted into low-dimensional data by principal component analysis, and the main features of the data are retained; The principal component analysis algorithm steps and formula: Compute the covariance matrix of the data: Where X is the data matrix, is the mean of each column of data, C is the covariance matrix; Compute the eigenvalues and eigenvectors of the covariance matrix: Cv i =λ i v i , Among them, v i is the eigenvector, λ i is the eigenvalue; The eigenvectors with the largest eigenvalues constitute the new data space; Convert the data: X pca =XW, Among them, W is a matrix consisting of the first k eigenvectors, X pca is the data matrix after dimensionality reduction.
5. The clinical data statistical management and analysis method according to claim 1, characterized in that: In the S4 step, Z-score standardization is used to convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1: Among them, x i is the original data point, μ is the mean of the data, and σ is the standard deviation of the data; In step S4, the non-normal distribution of the data is processed by logarithmic transformation: x i,log =log(x i +c), Among them, x i are the original data points and c is a constant used to handle zero values.
6. The clinical data statistical management and analysis method according to claim 1, characterized in that: In the step S5, a technique for evaluating model performance is used through cross-validation to perform training and testing by dividing the data set into multiple subsets; The k-fold cross validation: Dataset partitioning: Divide the dataset into k folds; Training and testing: For each fold i, use the other folds as training sets and the i-th fold as the test set; Model Evaluation: Among them, CV k is the average performance of k-fold cross validation, Performance i is the model performance index of the i-th fold; The leave-one-out cross-validation is a special case of k-fold cross-validation, where k is equal to the total number of samples n; Training and testing: For each sample i, use other samples as training sets and the i-th sample as test set; Model Evaluation: Among them, LOOCV is the average performance index of leave-one-out cross validation, Performance i is the model performance indicator for the i-th sample.
7. The clinical data statistical management and analysis method according to claim 1, characterized in that: In the step S6, the linear regression model diagnosis of the model diagnosis can be achieved by calculating the standardization of the residual and the influence index; The standardized residuals are: Among them, r i is the residual of the ith observation, MSE is the mean square error, and h ii is the leverage value; Impact indicators: Among them, e i is the standardized residual, p is the number of parameters in the model, and h ii is the leverage value. Model fitness: R 2 and the adjusted R 2 : Among them, SS res is the residual sum of squares, SS tot is the total sum of squares, n is the sample size, and p is the number of parameters in the model.
8. The clinical data statistical management and analysis method according to claim 1, characterized in that: In step S7, the role-based access control system is used to control access to system resources according to the user's role. The role-based access control model formula is: Role authorization: RoleAssignments={(u,r)|u∈Uandr∈Rand(u,r)is assigned}, Among them, u is the user set, R is the role set, (u,r) represents the role r assigned to user u; Permission allocation: PermissionAssignments={(r,p)|r∈RandPand(r,p)is assigned}, Where P is a permission set, and (r,p) means that role r is granted permission p.
9. The clinical data statistical management and analysis method according to claim 1, characterized in that: In step S8, when analyzing the effectiveness of the feedback mechanism, statistical methods can be used to evaluate the impact of the feedback; Before-after comparative analysis: Calculate the changes in key indicators before and after feedback to evaluate the effectiveness of feedback; ΔX=X post-feedback -X pre-feedback , Among them, X pre-feedback is the indicator value before feedback, X post-feedback is the indicator value after feedback, ΔX is the change in the indicator; Paired sample t test: used to detect whether the change of indicators before and after feedback is significant; Where D is the mean of the paired differences, s D is the standard deviation of the paired differences, and n is the sample size.
10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by a processor, the clinical data statistical management and analysis method according to any one of claims 1 to 9 is implemented.
