Release of technical guidelines for clinical evaluation of in vitro diagnostic reagents exempt from clinical trials

On September 24, the State Food and Drug Administration issued a notice on the technical guidelines for clinical evaluation of in vitro diagnostic reagents exempt from clinical trials (No.
74 of 2021)
.
The original text is as follows: Announcement of the State Food and Drug Administration on the issuance of technical guidelines for clinical evaluation of in vitro diagnostic reagents exempt from clinical trials (No.
74 of 2021) Measures” (State Administration for Market Regulation Order No.
48), the State Food and Drug Administration organized and formulated the “Guidelines for Clinical Evaluation of In Vitro Diagnostic Reagents Exempt from Clinical Trials” (see attachment), which is hereby issued
.
Hereby inform
.
Attachment: Technical guidelines for clinical evaluation of in-vitro diagnostic reagents exempt from clinical trials 1.
Purpose of preparation A methodological comparison study of the same variety of evaluation reagents and domestic marketed products proves that the reagents to be evaluated are substantially equivalent to the marketed products; or the method of comparing the reagents to be evaluated with reference measurement procedures or diagnostic accuracy standards is used to investigate the evaluation.
The rate of compliance/consistency between the reagent and the reference measurement procedure or diagnostic accuracy standard
.
This guideline aims to provide applicants with technical guidance for the clinical evaluation of in vitro diagnostic reagents that are exempt from clinical trials, and at the same time provide a reference for the technical review of this part of the data by the drug regulatory authority
.
2.
Scope of application This guideline applies to the clinical evaluation of products listed in the catalog of in vitro diagnostic reagents exempt from clinical trials (hereinafter referred to as the "catalog")
.
For in-vitro diagnostic reagents that are exempt from clinical trials, the applicant should submit the comparison materials between the declared product and the corresponding items in the "Catalog", which should be able to prove that the declared product is equivalent to the products described in the "Catalog"
.
For in vitro diagnostic reagents that are exempt from clinical trials, the applicant may conduct clinical evaluations in accordance with the requirements of this guideline, or conduct clinical trials in accordance with the requirements of the Technical Guidelines for In Vitro Diagnostic Reagents Clinical Trials
.
3.
Basic Principles For in vitro diagnostic reagents that are exempt from clinical trials, the applicant can compare the reagents to be evaluated with the domestic marketed products, and prove that the reagents to be evaluated are substantially equivalent to the marketed products, or with reference measurement procedures/diagnostic accuracy The standard test results have good consistency
.
The substantive equivalent mentioned here means that the intended use is the same and has the same safety and effectiveness
.
Applicants should use the final finalized reagents for clinical evaluation
.
Before clinical evaluation, the basic performance of the product should be determined, usually including the applicable sample type, specificity, precision, detection limit and/or quantification limit, measurement interval, positive judgment value, reference interval, etc.
, so as to be the reagent to be evaluated Provide basis for clinical evaluation
.
If the clinical evaluation fails to prove that the reagent to be evaluated is substantially equivalent to the domestic marketed product, or has good consistency with the reference measurement procedure/diagnostic accuracy standard test result, the application reagent should be evaluated through clinical trials
.
4.
Specific requirements (1) Methodological comparison study of the same species 1.
Selection of comparison reagents For in vitro diagnostic reagents that are free from clinical trials, if the method described in this guideline is used for methodological comparison, the comparison and analysis should be used to determine and Domestically marketed products with suitable evaluation reagents are used as comparison reagents
.
The applicant should first compare and analyze the intended use of the reagents to be evaluated and the products already on the market in China
.
Intended use refers to the general use or function of in vitro diagnostic reagents, including sample type, analyte, and indications
.
Indications refer to diseases or conditions that are diagnosed, prevented, predicted, treated by in vitro diagnostic reagents, or observed for prognosis, including applicable populations
.
If the intended use of the reagent to be evaluated is within the scope of the intended use of the product that has been marketed in China, the two are considered to have the same intended use
.
If there are functional differences between the reagents to be evaluated and the products already marketed in China in their intended uses (e.
g.
, used for auxiliary diagnosis and treatment monitoring respectively), and applicable population differences (e.
g.
, used for adults and children respectively), etc.
, the two are considered to have different functions.
Intended use
.
When the intended use is the same, the applicant shall continue to compare and analyze the basic principles, performance indicators, positive judgment value, reference interval, etc.
of the two
.
If there is a difference, the applicant needs to further evaluate whether the difference will cause a significant difference in the detection performance of the human sample
.
If the difference between the two does not have a significant impact on the detection performance of human samples, it is considered comparable
.
If the basic principles (methodology) of the two differ greatly or the performance indicators are quite different, the impact of the difference on the detection performance of human samples should be detailed
.
In summary, the comparison reagent must meet the following conditions: (1) The domestic market has been approved; (2) It has the same expected use as the reagent to be evaluated; (3) For quantitative reagents, the test results of the comparison reagent and the reagent to be evaluated should have The same measurement unit or the measurement units between the two can be mutually converted; (4) The reagent with less deviation from the test result of the reagent to be evaluated is preferred, and it is not recommended to select the reagent with inferior performance as the reagent to be evaluated
.
2.
Test site The test process is managed by the applicant and responsible for the authenticity, compliance and completeness of the test data
.
If overseas applicants need to conduct methodological comparison research conducted in China, they should conduct experiments in China through their agents in China
.
Applicants can also submit test materials completed overseas, and the content of the materials should meet the technical requirements of this guideline
.
Applicants should also fully consider the possible effects of differences in epidemiology, differences in the test population, differences in the environment and conditions of use, etc.
, on the research results
.
When necessary, the applicant should conduct supplementary research in my country based on the different factors
.
3.
Test personnel The test operators should be professional and technical personnel with corresponding test capabilities, and are familiar with the test procedures of the reagents to be evaluated and the comparison reagents
.
4.
Test samples should be tested with human samples consistent with the sample type claimed for the intended use, the background information of the sample should be clear, and the source of the sample should be traceable
.
Sample background information includes but is not limited to: sample source, unique and traceable number, sample type, and other background information related to the test (such as age, gender, interfering factors, etc.
); products with clear disease indications for reagent test results , The cases from which the included samples are derived should have clear clinical diagnosis information
.
The sample collection method and stability should meet the relevant requirements of the instructions of the reagents to be evaluated and the comparison reagents
.
In principle, clinical real samples should be used for research
.
When the true sample concentration cannot cover the detection range, the unavailable basis should be fully explained, and mixed samples obtained from patients with similar medical history should be used as appropriate.
Generally speaking, the mixed sample should not exceed 20% of the total sample size
.
If specific clinical samples cannot be obtained, manual preparation of samples such as dilution or processing to remove analytes to obtain low-concentration samples, and addition of analytes to obtain high-concentration samples can be used
.
Manual preparation of samples should fully consider the background information of the samples, matrix effects and other influencing factors
.
5.
Test time The test should take into account the actual clinical use and the impact of daytime precision.
Under the condition of ensuring the stability of the sample, set a reasonable duration, such as 3 to 20 days.
All samples should be blinded and adopted.
The reagents to be evaluated and the comparison reagents are tested separately, and the entire test should be quality controlled
.
6.
Test method The test method can refer to the technical guidance documents related to methodological comparison at home and abroad, and focus on the following content
.
6.
1 Quantitative products 6.
1.
1 Sample requirements The expected applicable population samples and interference samples whose concentration covers the linear/measurement interval should be selected for research
.
The research should include a certain number of samples, generally no less than 100 cases, and pay attention to the medical decision level and each concentration level in the measurement interval should contain a certain number of samples
.
Before the test, the limit of acceptable bias should be set in advance.
If the test results of the comparison study cannot meet the preset standard, the sample size can be appropriately expanded for evaluation
.
If the reagents to be evaluated and the comparison method show different performance for different population subgroups, different population subgroups should be stratified statistics, and each population should include at least 100 samples
.
If the reagent to be evaluated has different reference intervals for different populations, different populations should be stratified statistics, and at least 100 samples should be included in each population
.
Note: This does not include known physiological changes (such as female menstrual cycle, gender, age, etc.
), resulting in different reference intervals
.
The applicant should also fully consider various influencing factors based on the expected use of the product, the applicable population, clinical indications, genotype allocation of different subgroups, and acceptance criteria in clinical use, and adopt a reasonable method to determine the number of samples
.
6.
1.
2 Data Removal During the data collection process, all data should be recorded and checked in time
.
If it is determined that some abnormal results are caused by interpretable and acceptable reasons, the reasons should be recorded and removed from the data analysis
.
If the cause cannot be determined, the original result must be kept in the data set
.
6.
1.
3 Statistical analysis The applicant shall comprehensively consider the data characteristics and other factors, determine the specific statistical analysis method, and provide the basis for its selection
.
The statistical analysis of quantitative products is usually carried out in the following order
.
6.
1.
3.
1 Data mapping and review After data collection is completed, scatter plots and difference plots should be drawn first, and the data should be analyzed and reviewed.
Observe whether the data covers the linearity/measurement interval and whether there are outliers, and have a preliminary understanding Evaluate the potential variation characteristics between the measured values ​​of the reagent and the comparison reagent to determine how to better characterize these differences
.
Scatter diagrams and difference diagrams can show the comparison results of the reagent to be evaluated and the comparison reagent
.
Among them, the scatter chart should display all data, the x-axis represents the measurement result of the contrast reagent, the y-axis represents the measurement result of the reagent to be evaluated, and the x-axis and y-axis should use the same numerical range and spacing
.
The x-axis of the difference graph represents the measured concentration (such as the average value of the reagent to be evaluated and the comparison reagent), and the y-axis represents the difference between the measured value of the reagent to be evaluated and the contrast reagent (such as the absolute difference or percentage difference between the two) )
.
Applicants can choose a specific type of difference map according to the specific situation, and refer to relevant domestic and foreign guidance documents for details
.
6.
1.
3.
2 Calculate the correlation coefficient or the coefficient of determination 6.
1.
3.
3 Regression analysis The scatter diagram and the difference diagram should be used to determine whether the data meets the corresponding assumptions, and the best regression analysis method should be determined accordingly
.
Common regression analysis methods include Deming regression, Passing-Bablok regression analysis and least square regression
.
Regression analysis generally includes: obtaining a linear regression equation, b is the regression coefficient, a is the intercept, x represents the test result of the contrast reagent, y represents the test result of the reagent to be evaluated, and calculates the 95% confidence interval of the regression coefficient and the intercept
.
Hypothesis testing can also be used to evaluate the bias between the two
.
Generally, the intercept a and regression coefficient b in the regression equation are tested for hypothesis
.
6.
1.
3.
4 Evaluation of the bias of the reagent to be evaluated and the comparison reagent The bias of the reagent to be evaluated and the comparison reagent shall be evaluated through interval estimation
.
Generally, the expected bias at the medical decision level and its 95% confidence interval are compared with the limit of acceptable bias claimed by the applicant
.
The limit of acceptable bias is set by the applicant according to clinical needs after consulting the clinical institution, or by referring to relevant domestic and foreign standards
.
If the 95% confidence interval of the expected bias does not exceed the limit of acceptable bias claimed by the applicant, it means that the test results of the reagent to be evaluated and the comparison reagent meet the expected bias standard
.
If the 95% confidence interval of the expected bias exceeds the limit of acceptable bias claimed by the applicant, but contains the limit of acceptable bias, the applicant should continue to expand the sample size and/or analyze additional samples.
, To evaluate whether the bias of the test results of the reagent to be evaluated and the comparison reagent is acceptable
.
6.
1.
3.
5 Issues that should be paid attention to in the statistical analysis of quantitative products (1) If outliers appear, the reasons for the outliers should be analyzed, and two statistical analyses (including/excluding outliers) should be performed, such as two statistics The results of the analysis are inconsistent, and a reasonable analysis should be carried out
.
(2) The correlation coefficient can only indicate the closeness of the linear correlation between the reagent to be evaluated and the contrast reagent, not the consistency.
Therefore, the correlation coefficient cannot only be used to evaluate the consistency of the two
.
(3) Two straight lines should be displayed on the scatter plot, one is a straight line with y=x, and the other is a fitted regression line
.
6.
2 Qualitative products 6.
2.
1 Sample requirements The case from which the sample is derived should be representative of the expected applicable population
.
The research should include a certain number of positive samples and negative samples, generally no less than 50 cases each, and pay attention to include a certain number of samples near the positive judgment value and interference samples
.
The size of the research sample will affect the width of the credible interval, and the expansion of the sample size will result in a narrower credible interval
.
Applicants should fully consider various influencing factors based on the expected use of the product, applicable population, clinical indications, genotype allocation of different subgroups, and clinical acceptance criteria, and adopt reasonable methods to determine the number of samples
.
6.
2.
2 Check during data collection During the data collection process, all data should be recorded and checked in time
.
If it is determined that some abnormal results are caused by interpretable reasons, the reasons should be recorded and removed from the data analysis
.
If the cause cannot be determined, the original result must be kept in the data set
.
6.
2.
3 Statistical analysis The applicant shall comprehensively consider the data characteristics and other factors, determine the specific statistical analysis method, and provide the basis for its selection
.
Evaluate the consistency of the reagent to be evaluated and the comparison reagent
.
Summarize the test results of the reagent to be evaluated and the comparison reagent in the form of R×C table, and calculate the positive coincidence rate, negative coincidence rate and total coincidence rate of the two based on this
.
At the same time, the two-sided 95% confidence interval of the positive coincidence rate, negative coincidence rate and total coincidence rate was calculated
.
The following table 1 gives a sample of the recommended 2×2 table: Table 1 The positive coincidence rate of the 2×2 table of the reagent to be evaluated and the comparison reagent=100%×a/(a+c) The negative coincidence rate=100%×d /(D+b) Total coincidence rate=100% ×(a+d)/(a+b+c+d) Hypothesis test can also be performed at the same time to evaluate the consistency of the two, such as performing Kappa test and calculating the Kappa value two-sided 95% confidence interval and so on
.
6.
2.
4 Treatment of inconsistent results The inconsistency between the reagent to be evaluated and the comparison reagent may be due to the error of the reagent to be evaluated or the comparison reagent.
The inconsistent results can be confirmed by the "diagnostic accuracy standard" or other reasonable methods, and the reasons should be analyzed.
Should be included in statistical analysis
.
6.
3 Semi-quantitative products The semi-quantitative products mentioned in this guideline generally refer to reagents whose results are reported as multiple grades (such as negative, +, ++, or +++)
.
For the sample requirements, test time and sequence of such products, please refer to the relevant requirements of quantitative or qualitative products, depending on the characteristics of the product.

.
The statistical analysis method used for semi-quantitative products should also depend on the situation.
Generally, the following methods can be used to evaluate the consistency of the reagent to be evaluated and the comparison reagent
.
First, R×C is used to represent the test results of the reagent to be evaluated and the comparison reagent
.
In the R×C table, the grade numbers of the reagent to be evaluated and the comparison reagent are multiple grades, and the grade numbers of the two are equal, such as "negative, +, ++" and so on
.
Then, calculate the individual coincidence rate, total coincidence rate and 95% confidence interval of the two at each level, and also perform hypothesis testing, such as Kappa test, and calculate the two-sided 95% confidence interval of the Kappa value
.
6.
4 Different sample types If the reagent to be evaluated includes several sample types, a sample suitability study should be carried out
.
For samples that are comparable in research (such as serum and plasma), only one sample type can be selected for methodological comparison research
.
For samples that are not comparable in the research (such as blood and urine samples), no less than 100 samples should be selected in accordance with this guideline, and methodological comparison studies should be conducted for each sample type
.
The comparable samples mentioned here generally refer to the same performance indicators, the same positive judgment value, the same expected population, and the same clinical significance
.
Otherwise, it should be regarded as an incomparable sample
.
6.
5 Issues that should be paid attention to in statistical analysis 6.
5.
1 There should be reasonable reasons for sample removal, and it should not be removed at will
.
If it is necessary to exclude, the number and reason for the elimination should be explained
.
6.
5.
2 Reference should be made to authoritative technical guidelines, etc.
, according to the applicable statistical analysis model of the product, calculate the slope, intercept, expected bias of the medical decision level, positive coincidence rate, negative coincidence rate, total coincidence rate and its 95% confidence interval, etc.

.
6.
5.
3 If the comparison study adopts dilution or processing to remove analytes to obtain low-concentration samples, adding analytes to obtain high-concentration samples and other manual preparation samples, two statistical analyses (including/excluding manual preparation samples) shall be performed, and Reasonable analysis of statistical results
.
(2) Comparative studies with reference measurement procedures or diagnostic accuracy standards.
If there are reference measurement procedures or diagnostic accuracy standards, you can also refer to the relevant requirements in the chapter "(1) Methodological comparison studies of the same species" of this guideline.
Compare the evaluation reagent with the reference measurement program or diagnostic accuracy standard
.
Reference measurement procedures refer to measurement procedures that are accepted as providing measurement results suitable for their intended use
.
The diagnostic accuracy standard refers to the use of one method or a combination of multiple methods, including clinical information including laboratory tests, imaging tests, pathology and follow-up information, to define the criteria for the presence or absence of conditions, events, and features of interest
.
(3) Methodological comparison research related to change items For methodological comparison research related to change items, generally the product after the change is compared with the product before the change
.
If the product after the change is not comparable to the product before the change, the product after the change should be used to conduct a methodological comparison study with other products on the market.
For specific requirements, please refer to this guideline
.
V.
Clinical evaluation report requirements The report should at least include the following content: descriptive comparison analysis and comparison performance data between the reagent to be evaluated and the comparison reagent/reference measurement program or diagnostic accuracy standard
.
The report should be signed by the test personnel and signed and sealed by the applicant and/or agent
.
(1) Descriptive comparative analysis If the methodological comparative study of the same species is adopted, the applicant should submit detailed comparative analysis data of the reagent to be evaluated and the comparative reagent, and should detail the similarities and differences between the two, and submit supporting materials from the similarities and differences.

.
If there is a difference between the two, the applicant should also submit assessment data on whether the difference will have a significant impact on the detection performance of human samples
.
The comparative analysis items include but are not limited to the items listed in the appendix.
If there is an inapplicable item, the reason for the inapplicability should be explained.

.
If a reference measurement procedure or diagnostic accuracy standard is used as the comparison method, the applicant should explain in detail the reasons for choosing the reference measurement procedure or diagnostic accuracy standard as the comparison method, and submit detailed information about the selected reference measurement procedure or diagnostic accuracy standard.
method of operation, and other criteria
.
(2) Comparing performance data Comparing performance data should clearly describe the test design, test implementation, and statistical analysis, and include at least the following content
.
1.
Basic information
.
Including but not limited to: the name of the applicant, test personnel, test time and location, name of the reagent to be evaluated, comparison reagent/reference measurement procedure or diagnostic accuracy standard, product name of other reagents/instruments used in conjunction with, manufacturer, specification/ model and lot number
.
2.
Experimental design
.
The detailed description of the sample group, exclusion and exclusion criteria, sample size, statistical analysis and other methods to select the content
.
3.
The implementation of the experiment, including: 3.
1 Sample selection, including the number of cases and sample distribution
.
3.
2 Description of the test process
.
4.
Test management, including participation in staff, quality control, the data management, problems and treatment and other measures appear
.
5.
Statistical analysis and summary of evaluation results
.
Perform statistical analysis on the test results according to the determined statistical method, and reasonably evaluate the performance of the product
.
The applicant shall specify the basis for determining the acceptable bias limit or the acceptable consistency limit as claimed by the applicant, and explain in detail the calculation formulas and their sources in the statistical analysis, as well as the reasons for their selection
.
If statistical software is involved, the statistical software used and the software version number should be clarified
.
6.
The data summary table should summarize and describe the included samples in tabular form, and include at least the following content: traceable sample number, sample type, reagents to be evaluated and comparison reagents or reference measurement procedures/diagnostic accuracy standard test results, samples context (if applicable) and the like
.
The data summary table should be signed by the operator and traceable
.
The applicant should keep all the original test data for future reference
.
6.
Other materials In addition to the above evaluation reports, relevant documents on the evaluation of the test performance of the reagents to be evaluated on human samples can be submitted as supplementary materials
.
Attachment: Comparison and analysis items of reagents to be evaluated and comparison reagents Note: Comparison items include but are not limited to the above items, which can be added according to actual conditions
.