Regression method quantitative consistency sample size calculation software development method

Abstract

The invention discloses a regression method quantitative consistency sample size calculation software development method, which comprises the following steps of: optimizing quantitative consistency evaluation by introducing a clinical consistency boundary value, and designing a technical scheme which consists of two calculation models, namely a module I and a module II; the first module directly reads original data into a program, data parameters are obtained through calculation, and the sample size is obtained through formula calculation according to the parameters. The module II is used for directly inputting known data parameters and calculating the sample size by a formula according to the parameters; the designed quantitative consistency scheme is reviewed; after the scheme is reviewed and designed, a system manual is compiled; then system packaging is carried out; and system test cases and reviews are synchronously compiled and used for testing and reviewing. According to the quantitative consistency evaluation method, the scientificity and practicability of quantitative consistency evaluation are remarkably improved through the optimized regression method and by considering the clinical consistency boundary value, and the requirements of modern medical research and clinical diagnosis can be better met.

Description

Technical Field

[0001] The present invention relates to the technical field of software development, and in particular to a software development method for calculating quantitative consistency sample size using a regression method. Background Art

[0002] In medical and biostatistical research, accurate sample size calculation is crucial to ensure the statistical power and reliability of the research results. Insufficient or excessive sample size calculation may lead to waste of resources or inaccurate research results. There are many tools and software on the market that are specifically designed for sample size calculation and power analysis, such as PASS and MedCalc, each with its own unique advantages and features. For example, PASS software is known for its accurate estimation of sample size and test power under various statistical tests; while MedCalc is highly regarded for its wide application in biomedical research and its professional functions for quantitative consistency evaluation. In addition, there are some online platforms and websites, such as OpenEpi and Sample Size Calculators, which provide basic sample size calculation functions and are suitable for quick preliminary estimates, but they usually cannot provide advanced statistical functions such as PASS and MedCalc. It is worth noting that although these tools are powerful in sample size estimation, they do not support sample size calculation for quantitative consistency evaluation, which may be a limitation for researchers who need to conduct such evaluations.

[0003] PASS (Power Analysis and Sample Size) 15.0 software is a professional statistical software specially used for sample size calculation and test power estimation. It can estimate the sample size and confidence level under various statistical hypothesis tests. Figure 1The diagram of sample size calculation of PASS software is shown; A: "Test for Paired Means" module; B: "Pearson's Correlation Tests" module; C: "Linear Regression" module. Use keywords such as "Agreement", "Consistency Evaluation", "Correlation", "Linear Regression" to search for the built-in sample size calculation modules of PASS, and you can retrieve modules such as "Test for Paired Means", "Pearson's Correlation Tests", "Linear Regression" and "Confidence Interval for Linear Regression". The "Test for Paired Means" module calculates the sample size based on the paired t-test. The key parameters include the test level α, the test power 1-β, and the mean of the paired difference d. and the standard deviation s of d d .

[0004] MedCalc is a statistical analysis software designed for biomedical researchers, especially for biostatistical analysis and drawing high-quality charts. MedCalc is known for its user-friendly interface and powerful data processing capabilities. The software includes a function module for calculating the sample size required for quantitative consistency evaluation using the Bland-Altman plot method. Figure 2 The sample size calculation function module for quantitative consistency evaluation using the Bland-Altman method shown in the figure; and in the 2018 update, the Bland-Altman module introduced a new statistical function, which is to consider the impact of type I error (Type I error) and type II error (Type II error) on sample size when performing consistency evaluation. The addition of this function enables researchers to more accurately calculate the sample size required at a specific statistical power and significance level, thereby improving the reliability and efficiency of the research.

[0005] Both PASS and MedCalc provide advanced statistical analysis tools that enable researchers in the fields of medicine and biostatistics to perform accurate sample size calculations and power analysis. PASS's versatile modules support everything from simple paired mean tests to complex linear regression analyses, while MedCalc is particularly optimized for sample size calculations for quantitative consistency evaluations through its Bland-Altman plot module and new features that consider the effects of type errors. Although these tools differ in functionality, they are both designed to improve the statistical power of studies and the reliability of results.

[0006] Objective disadvantages of PASS: From a methodological perspective, this method has obvious limitations; from the perspective of sample size calculation, this method does not consider clinical boundaries with practical significance. The "Pearson's Correlation Tests" module calculates the sample size based on the statistical test of the Pearson's correlation coefficient, and the parameters include the test level α, the test power 1-β, the null hypothesis ρ0, and the alternative hypothesis ρ1. Similarly, this sample size calculation method may ignore the systematic errors between the results and lead to wrong conclusions. The "Linear Regression" and "Confidence Interval for Linear Regression" modules calculate the sample size based on whether the regression coefficient is statistically significant, not for quantitative consistency evaluation. In addition, from a methodological perspective, both the regression coefficient and the regression constant must meet the standard to be considered consistent; in addition, this method also does not consider the consistency boundaries with practical significance.

[0007] Objective shortcomings of MedCalc: There are many methodological problems in using Bland-Altman for quantitative consistency evaluation. In actual work, some problems were found. First, there is a lack of pre-determination of consistency evaluation criteria; second, there is a lack of sample size estimation; third, the application of statistical methods is insufficient or even inappropriate, and cannot fully meet the needs of consistency evaluation. For example, the methodological comparison does not give a 95% confidence interval for bias, the clinical trial of in vitro diagnostic reagents only evaluates the systematic error, and the application of the Bland-Altman method is not standardized. And MedCalc software still does not take clinical thresholds into consideration. Summary of the invention

[0008] To this end, the present invention provides a regression method for calculating the quantitative consistency sample size software development method to solve the above-mentioned problems in the prior art.

[0009] In order to achieve the above object, the present invention provides the following technical solutions:

[0010] According to a first aspect of the present invention, a method for developing software for calculating quantitative consistency sample size using regression method comprises the following steps:

[0011] S1. By introducing clinical consistency cutoffs, the quantitative consistency evaluation is optimized and a technical solution is designed, which consists of two calculation models, namely module 1 and module 2;

[0012] Module 1 directly reads the original data into the program, calculates the data parameters, and calculates the sample size based on the parameters using the formula;

[0013] Module 2 is to directly input the known data parameters and calculate the sample size based on the parameters by the formula;

[0014] S2. Review the design's quantitative consistency plan; if the design plan fails the review, continue to improve the design plan until it passes the review;

[0015] S3. After the design plan is reviewed, the system manual is written; then the system is packaged, the system source code, system manual, etc. are made into a system package, which is stored and submitted to the test team for system testing; and system test cases and reviews are written simultaneously for test review;

[0016] S4. System testing and modification: Testers conduct system testing and record test bugs in the BMS system. Developers modify the system based on the test records. After the system modification is completed, the system manual is updated and packaged into the warehouse again;

[0017] S5. Review the system that is packaged and put into storage again to confirm whether to conduct the next round of testing; if it is necessary, add another round of testing; otherwise, prepare a test report and end the testing process;

[0018] S6. After the testing process is completed, the tester will conduct a factory test on the system package in the project environment; after confirming that it can be released, the software system will be shipped out of the factory, the mini-program / software will be released to the outside world, and the system will be maintained.

[0019] Further: The specific calculation process of module one is as follows:

[0020] Step 1: Directly input the parameters required for the calculation formula into the program;

[0021] Step 2: Substitute the parameters into the formula of U boundary value;

[0022] Step 3: Use two formulas to calculate n0 and n1 respectively;

[0023] Step 4: Compare n0 and n1, and choose the larger one as n bmax1 ;

[0024] Step 5: According to n bmax1 Calculate the degrees of freedom v;

[0025] Step 6: Substitute the calculated degrees of freedom v into the formula for the t boundary value;

[0026] Step 7: Calculate n according to the two formulas t0 and n t1 ;

[0027] Step 8: Compare t0 and n t1, choose the larger one as n bmax2 ;

[0028] Step 9: To ensure the stability of the results, repeat Step 6-8;

[0029] Step 10: Until the obtained n bmax2 Same as last time, that is, get stable results and output the results.

[0030] Further: The specific calculation process of module 2 is as follows:

[0031] Step 1: Read the data obtained from the experiment directly into the program;

[0032] Step 2: Software based Calculate the required parameters and input the remaining manually set parameters into the program;

[0033] Step 3: Substitute the parameters into the formula of U boundary value;

[0034] Step 4: Use two formulas to calculate n0 and n1 respectively;

[0035] Step 5: Compare n0 and n1, and choose the larger one as n bmax1 ;

[0036] Step 6: According to n bmax1 Calculate the degrees of freedom v;

[0037] Step 7: Substitute the calculated degrees of freedom v into the formula for the t boundary value;

[0038] Step 8: Calculate n according to the two formulas t0 and n t1 ;

[0039] Step 9: Compare t0 and n t1 , choose the larger one as n bmax2 ;

[0040] Step 10: To ensure the stability of the results, repeat Step 6-8;

[0041] Step 11: Until the obtained n bmax2 Same as last time, that is, get stable results and output the results.

[0042] Further: During the calculation steps of module 1 and module 2, the calculation formula of n0 is:

[0043]

[0044] The calculation formula for n1 is:

[0045]

[0046] n t0 The calculation formula is:

[0047]

[0048] n t1 The calculation formula is:

[0049]

[0050] In the above four formulas: α is the test level; 1-β is the test efficiency; b0 is the regression constant of the expected regression model, which is usually 0; b1 is the regression coefficient of the expected regression model, which is usually 1; S yx is the residual standard deviation; x is the average value of the reference method sample measurement; S x is the standard deviation of the sample measurement value of the reference method; Δ0 is the cutoff value of the regression constant b0 in the clinically acceptable consistency limit (-Δ0, +Δ0); Δ1 is the cutoff value of the regression coefficient b1 in the clinically acceptable consistency limit (1-Δ1, 1+Δ1); t (n-2,1-α / 4) and t (n-2,1-β / 4) They represent the two-sided t distribution boundaries when the confidence levels are 1-α and 1-β, respectively.

[0051] Furthermore: The method of the present invention solves the sample size calculation formula for the quantitative consistency evaluation problem based on the regression coefficient and the regression constant.

[0052] Further: Monte-Carlo simulation and Bootstrap simulation methods were used to theoretically derive and verify the accuracy of the sample size calculation formula.

[0053] The present invention has the following advantages: the present invention significantly improves the scientificity and practicality of quantitative consistency evaluation, and can better meet the needs of modern medical research and clinical diagnosis;

[0054] Through the optimized regression method, the present invention overcomes the limitations of traditional linear regression and Pearson correlation test, especially when dealing with complex data structures and large data sets; it can more accurately identify and analyze data, reduce bias and errors, and the improved algorithm improves the accuracy of consistency evaluation, making the research results more reliable, thereby supporting more effective medical decision-making and policy making;

[0055] By considering the clinical consistency threshold, the present invention solves the problem of ignoring the actual clinical application needs in traditional methods; this makes the consistency evaluation more valuable for practical application and can directly reflect the clinical relevance of the method; it makes the research results more in line with the needs of clinical practice, enhances the practicality of the research and the support for clinical decision-making, and improves the safety and effectiveness of patient treatment;

[0056] Users can customize model parameters according to specific research needs or use raw data to automatically calculate parameters, which increases the flexibility and adaptability of research design; makes research design more personalized and precise, improves the applicability of the model and the accuracy of verification, thereby enhancing the reliability and validity of research results. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 Schematic diagram of sample size calculation using PASS software.

[0058] Figure 2 Schematic diagram of the sample size calculation function block for quantitative consistency evaluation using the Bland-Altman method.

[0059] Figure 3 The present invention is a flowchart of the method for developing software for calculating the quantitative consistency sample size using the regression method.

[0060] Figure 4 for Figure 3 Schematic diagram of the calculation process of module one.

[0061] Figure 5 for Figure 3 Schematic diagram of the calculation process of module 2. DETAILED DESCRIPTION

[0062] The following is an explanation of the implementation of the present invention by specific specific examples. People familiar with this technology can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. It should be understood that these embodiments are only to further illustrate the present invention and cannot be understood as limiting the scope of protection of the present invention. Technical engineers in this field can make some non-essential improvements and adjustments to the present invention based on the contents of the above invention; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] like Figure 3-5 As shown, the software development method for calculating the quantitative consistency sample size by regression method includes the following steps:

[0064] S1. By introducing clinical consistency cutoffs, the quantitative consistency evaluation is optimized and a technical solution is designed, which consists of two calculation models, namely module 1 and module 2;

[0065] Module 1 directly reads the original data into the program, calculates the data parameters, and calculates the sample size based on the parameters using the formula;

[0066] Module 2 is to directly input the known data parameters and calculate the sample size based on the parameters by the formula;

[0067] S2. Review the design's quantitative consistency plan; if the design plan fails the review, continue to improve the design plan until it passes the review;

[0068] S3. After the design plan is reviewed, the system manual is written; then the system is packaged, the system source code, system manual, etc. are made into a system package, which is stored and submitted to the test team for system testing; and system test cases and reviews are written simultaneously for test review;

[0069] S4. System testing and modification: Testers conduct system testing and record test bugs in the BMS system. Developers modify the system based on the test records. After the system modification is completed, the system manual is updated and packaged into the warehouse again;

[0070] S5. Review the system that is packaged and put into storage again to confirm whether to conduct the next round of testing; if it is necessary, add another round of testing; otherwise, prepare a test report and end the testing process;

[0071] S6. After the testing process is completed, the tester will conduct a factory test on the system package in the project environment; after confirming that it can be released, the software system will be shipped out of the factory, the mini-program / software will be released to the outside world, and the system will be maintained.

[0072] like Figure 4 The calculation process of module one is shown in FIG. 1 . The specific calculation process of module one is as follows:

[0073] Step 1: Directly input the parameters required for the calculation formula into the program;

[0074] Step 2: Substitute the parameters into the formula of U boundary value;

[0075] Step 3: Use two formulas to calculate n0 and n1 respectively;

[0076] Step 4: Compare n0 and n1, and choose the larger one as n bmax1 ;

[0077] Step 5: According to n bmax1 Calculate the degrees of freedom v;

[0078] Step 6: Substitute the calculated degrees of freedom v into the formula for the t boundary value;

[0079] Step 7: Calculate n according to the two formulas t0 and n t1 ;

[0080] Step 8: Compare t0 and n t1 , choose the larger one as n bmax2 ;

[0081] Step 9: To ensure the stability of the results, repeat Step 6-8;

[0082] Step 10: Until the obtained n bmax2 Same as last time, that is, get stable results and output the results.

[0083] like Figure 5 The calculation process of module 2 is shown in FIG. 1 . The specific calculation process of module 2 is as follows:

[0084] Step 1: Read the data obtained from the experiment directly into the program;

[0085] Step 2: Software based Calculate the required parameters and input the remaining manually set parameters into the program;

[0086] Step 3: Substitute the parameters into the formula of U boundary value;

[0087] Step 4: Use two formulas to calculate n0 and n1 respectively;

[0088] Step 5: Compare n0 and n1, and choose the larger one as n bmax1 ;

[0089] Step 6: According to n bmax1 Calculate the degrees of freedom v;

[0090] Step 7: Substitute the calculated degrees of freedom v into the formula for the t boundary value;

[0091] Step 8: Calculate n according to the two formulas t0 and n t1 ;

[0092] Step 9: Compare t0 and n t1 , choose the larger one as n bmax2 ;

[0093] Step 10: To ensure the stability of the results, repeat Step 6-8;

[0094] Step 11: Until the obtained n bmax2 Same as last time, that is, get stable results and output the results.

[0095] Use in Step 2 To fit a linear regression model to the measurements.

[0096] In the calculation steps of the above modules 1 and 2, the calculation formula of n0 is:

[0097]

[0098] The calculation formula for n1 is:

[0099]

[0100] n t0 The calculation formula is:

[0101]

[0102] n t1 The calculation formula is:

[0103]

[0104] In the above four formulas: α is the test level; 1-β is the test efficiency; b0 is the regression constant of the expected regression model, which is usually 0; b1 is the regression coefficient of the expected regression model, which is usually 1; S yx is the residual standard deviation; x is the average value of the reference method sample measurement; S x is the standard deviation of the sample measurement value of the reference method; Δ0 is the cutoff value of the regression constant b0 in the clinically acceptable consistency limit (-Δ0, +Δ0); Δ1 is the cutoff value of the regression coefficient b1 in the clinically acceptable consistency limit (1-Δ1, 1+Δ1); t (n-2,1-α / 4) and t (n-2,1-β / 4) They represent the two-sided t distribution boundaries when the confidence levels are 1-α and 1-β, respectively.

[0105] The regression method used in this invention is not just a simple application of the traditional linear regression model, but is optimized and customized for quantitative consistency evaluation. This method takes into account various variables and potential biases in the evaluation, making the consistency conclusions obtained more accurate and reliable. By improving the algorithm and calculation process, the new method can more effectively handle different data types and structures, especially when processing large data sets, the efficiency and accuracy are significantly improved.

[0106] In an embodiment of the present invention, the newly developed software / applet allows the user to directly input preset model parameters, or automatically calculate these parameters from the raw data; this flexibility enables the user to adjust and optimize the sample size calculation according to specific research needs, thereby adapting to different research designs and objectives; in addition, the software also includes a model validation function, and the user can test and verify the accuracy and applicability of the model through example data, which is very helpful for improving the confidence and accuracy of the research design.

[0107] Further: In this embodiment, in order to make the software more user-friendly, an intuitive interface design is also developed to ensure that all functions are easy to access and use. In addition, the software supports multi-platform operation, including online applications and mobile device applets, so that users can perform sample size calculations on different devices and environments, greatly improving its accessibility and convenience.

[0108] In this embodiment, software development and production is carried out by professional technicians according to the design instructions to build and test the system. The system that has passed the test can be released to scientific researchers for use. The specific software development and design process is as follows:

[0109] (1) Preparation and review of detailed design plan: Designers prepare the "Detailed Design Specification" based on the requirements analysis and other information, and relevant personnel and product development managers organize a technical review of the specification;

[0110] (2) Writing and reviewing system test cases: Testers write system test cases based on the baseline "Requirements Analysis" and "Detailed Design Specifications". Testers organize technical experts to review the document and make modifications based on expert opinions until it is passed.

[0111] (3) Write system manual: The development manager organizes personnel to write the "System User Manual" and "System Deployment Manual";

[0112] (4) System packaging: The development manager organizes personnel to prepare system source code, system manuals, etc. into a system package, which is then submitted to the test team for system testing;

[0113] (5) System testing and modification: Testers conduct system testing according to the "System Test Cases" and record test bugs in the BMS system. Developers modify the system according to the test records. After the system modification is completed, the system manual is updated and packaged into the warehouse again;

[0114] (6) Test confirmation: The test manager organizes system-related personnel to discuss and confirm whether to conduct the next round of testing; if it is necessary, an additional round of testing will be added; otherwise, a test report will be prepared and the test process will end;

[0115] (7) Factory testing and release: Testers conduct factory testing on system packages in the project environment; the test manager communicates with the development manager and project leader on system package release issues; after confirmation that the system package can be released, the test manager fills out the "System Release Confirmation Form" and the relevant responsible parties sign and confirm.

[0116] The quantitative consistency evaluation principle and statistical inference principle of the present invention derive a sample size formula for quantitative consistency evaluation based on a regression coefficient and a regression constant through formula transformation.

[0117] Quality control during the software development phase: To ensure the quality of software development, a "Software Quality Control Plan" is written to control the entire software development process; each quality control link should record accident reports, accident analysis and corrective measures in detail. Corrective measures are used to correct software products that do not meet requirements, avoid the recurrence of errors, and find out the root causes of errors. The specific measures are as follows:

[0118] ①Write the "Software Quality Control Measures Plan"

[0119] The plan should include the following: write evaluation criteria for software development activities and software products; determine the responsibilities and authorities of personnel in software development, integration, and testing activities; describe the standards, methods, languages, and tools used in each stage of the software development cycle; describe the key points of the review and describe the reference or compliance basis for carrying out activities to ensure software safety and quality.

[0120] ②Software quality control records

[0121] Software quality assurance record documents must be kept throughout the contract execution period; deviations and problems that occur in the software product or software task, together with their corrective measures, should be recorded in the defect registration file.

[0122] ③Independence of software quality assurance

[0123] The personnel responsible for software quality assurance assessment activities must be different from the personnel who carry out software development, software operation or responsible for software activities. Personnel involved in ensuring that the software product does not deviate from the contract must have sufficient resources, responsibilities, rights, and relative organizational independence to conduct quality assurance assessment activities objectively and impartially, and to take and verify corrective actions.

[0124] ④Tools, techniques and methods

[0125] Conduct software product deviation inspection and correction inspection activities according to the quality control plan. The software quality assurance plan must ensure that the development tools, technologies, methods and correction measures of its software products are within the scope of close monitoring and supervision of this quality assurance plan.

[0126] ⑤Source code and media management

[0127] The software quality control plan must take effective steps and measures to strengthen the maintenance, safekeeping, preservation and textual record of various versions of software products, including the management of software source code and media tools. The software quality control plan must ensure the implementation of these management steps and measures.

[0128] ⑥Quality Control Report

[0129] When the software product meets the requirements of testing, verification, and confirmation, a quality control report is written and the quality control process is completed only after the signatures of all relevant responsible parties.

[0130] The method of the present invention solves the sample size calculation formula for the quantitative consistency evaluation problem based on the regression coefficient and regression constant;

[0131] Combining clinical thresholds with statistical inference principles, taking clinical practicality and relevance into account in sample size calculations improves the practicality and accuracy of sample size calculations, ensures that research results have greater clinical application value, and directly supports the clinical decision-making process;

[0132] Monte-Carlo simulation and Bootstrap simulation methods were used to theoretically derive and verify the accuracy of the sample size calculation formula, which improved the reliability of the calculation formula and supported the scientificity and operability of the formula through empirical research;

[0133] In the embodiment of the present invention: Strategy for solving program errors:

[0134] Program errors usually include code bugs, calculation errors, or incompatibility of the software environment. When faced with these problems, first conduct a code review to identify and fix bugs through unit testing and integration testing. In addition, use a version control system to manage the software development process to ensure that every change can be tracked and backtracked. For complex calculation errors, use debugging tools and professional mathematical software to verify and adjust to ensure the accuracy of all calculation logic.

[0135] Parameter tuning involves optimization and refinement of the model, which is usually done based on analysis of simulation results. Parameter selection can be optimized by setting different parameter scenarios and running sensitivity analysis to determine the degree of influence of parameters on model output.

[0136] In addition, in this embodiment, data collected from actual applications can be used for simulation testing to verify the performance and accuracy of the model in real scenarios. Combined with case studies, the output of the model under different conditions is analyzed, and the model is adjusted to better adapt to the characteristics and needs of actual data.

[0137] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A software development method for calculating the quantitative consistency sample size using regression method, characterized in that: The following steps are involved: S1. By introducing the clinical consistency threshold U, the quantitative consistency evaluation is optimized and a technical solution is designed. The technical solution consists of two calculation models, namely module 1 and module 2; Module 1 directly reads the original data into the program, calculates the data parameters, and calculates the sample size based on the parameters using the formula; The specific calculation process of module 1 is as follows: Step 1: Directly input the parameters required for the calculation formula into the program; Step 2: Substitute the parameters into the formula of U boundary value; Step 3: Use two formulas to calculate n0 and n1 respectively; Step 4: Compare n0 and n1, and choose the larger one as n bmax1 ; Step 5: According to n bmax1 Calculate the degrees of freedom v; Step 6: Substitute the calculated degrees of freedom v into the formula for the t boundary value; Step 7: Calculate n according to the two formulas t0 and n t1 ; Step 8: Compare n t0 and n t1 , choose the larger one as n bmax2 ; Step 9: To ensure the stability of the results, repeat Step 6-8; Step 10: Until the obtained n bmax2 Same as last time, that is, get stable results and output the results; Module 2 is to directly input the known data parameters and calculate the sample size based on the parameters by the formula; The specific calculation process of module 2 is as follows: Step 1: Read the data obtained from the experiment directly into the program; Step 2: Software based Calculate the required parameters and input the remaining manually set parameters into the program; Step 3: Substitute the parameters into the formula of U boundary value; Step 4: Use two formulas to calculate n0 and n1 respectively; Step 5: Compare n0 and n1, and choose the larger one as n bmax1 ; Step 6: According to n bmax1 Calculate the degrees of freedom v; Step 7: Substitute the calculated degrees of freedom v into the formula for the t boundary value; Step 8: Calculate n according to the two formulas t0 and n t1 ; Step 9: Compare t0 and n t1 , choose the larger one as n bmax2 ; Step 10: To ensure the stability of the results, repeat Step 6-8; Step 11: Until the obtained n bmax2 Same as last time, that is, get stable results and output the results; S2. Review the design's quantitative consistency plan; if the design plan fails the review, continue to improve the design plan until it passes the review; S3. After the design plan is reviewed, the system manual is written; then the system is packaged, the system source code, system manual, etc. are made into a system package, which is stored and submitted to the test team for system testing; and system test cases and reviews are written simultaneously for test review; S4. System testing and modification: Testers conduct system testing and record test bugs in the BMS system. Developers modify the system based on the test records. After the system modification is completed, the system manual is updated and packaged into the warehouse again; S5. Review the system that is packaged and put into storage again to confirm whether to conduct the next round of testing; if it is necessary, add another round of testing; otherwise, prepare a test report and end the testing process; S6. After the testing process is completed, the tester will conduct a factory test on the system package in the project environment; after confirming that it can be released, the software system will be shipped out of the factory, the mini-program / software will be released to the outside world, and the system will be maintained.

2. The method for developing software for calculating the quantitative consistency sample size using regression method according to claim 1, characterized in that: During the calculation steps of Module 1 and Module 2, the calculation formula for n0 is: The calculation formula for n1 is: n t0 The calculation formula is: n t1 The calculation formula is: In the above four formulas: α is the test level; 1-β is the test efficiency; b0 is the regression constant of the expected regression model, which is usually 0; b1 is the regression coefficient of the expected regression model, which is usually 1; S yx is the residual standard deviation; x is the average value of the reference method sample measurement; S x is the standard deviation of the sample measurement value of the reference method; Δ0 is the cutoff value of the regression constant b0 in the clinically acceptable consistency limit (-Δ0, +Δ0); Δ1 is the cutoff value of the regression coefficient b1 in the clinically acceptable consistency limit (1-Δ1, 1+Δ1); t (n-2,1-α / 4) and t (n-2,1-β / 4) They represent the two-sided t distribution boundaries when the confidence levels are 1-α and 1-β, respectively.

3. The method for developing software for calculating the quantitative consistency sample size using regression method according to claim 1-2, characterized in that: The method of the present invention solves the sample size calculation formula for the quantitative consistency evaluation problem based on the regression coefficient and the regression constant.

4. The method for developing software for calculating the quantitative consistency sample size using regression method according to any one of claims 1 to 2, characterized in that: Monte-Carlo simulation and Bootstrap simulation methods were used to theoretically derive and verify the accuracy of the sample size calculation formula.

Images