RB/T 151-2016 Guidelines for the evaluation of measurement uncertainty for quantitative detection of food microorganisms RB/T151-2016 |tt||Standard compression package decompression password: www.bzxz.net
This standard specifies the evaluation and expression methods of measurement uncertainty for quantitative detection of food microorganisms. This standard applies to the measurement uncertainty evaluation activities of quantitative analysis of food and environmental samples used to monitor processed and stored food, and also applies to instrumental quantitative analysis methods that can replace conventional microbial colony counting methods. This standard does not apply to the most likely value counting method and the analysis method of low-content microorganisms. The measurement uncertainty evaluation of quantitative analysis of animal feed products also refers to this standard.

ICS03.120.00
Certification and Accreditation Industry Standard of the People's Republic of China RB/T151—2016
Guidelines for the estimation of measurement uncertainty of foodmicrobiologicalquantitativedetection(ISO/TS 19036:2006 Microbiology of food and animal feedstuffs-Guidelines for the estimation of measurementuncertainty forquantitativedeterminations,MOD)Published on September 22, 2016
National Certification and Accreditation Administration
Implementation on April 1, 2017
This standard was drafted in accordance with the rules of GB/T1.1—2009. RB/T151—2016
This standard is revised and adopted by the redrafting method based on ISO/TS19036:2006 “Guide for the assessment of uncertainty in quantitative microbiology of food and animal feed” (English version). The technical differences between this standard and ISO/TS19036:2006 "Guide to the assessment of uncertainty in quantitative microbiology of food and animal feed" (English version) are as follows:
All contents related to animal feed in ISO/TS19036:2006 are deleted, and only the contents related to food are retained; normative reference document chapters are added;
"ISO/IEC17025:2005" in ISO/TS19036:2006 is replaced by \GB/T270252008\; 2.1 "Note 3" in ISO/TS19036:2006 is modified to the main text; 4.2 and 5 \intralaboratory standard deviation of reproducibility in ISO/TS19036:2006 are replaced by \GB/T270252008\ The standard also makes the following editorial changes: a, b, c, d are used in sequence to represent the reproducibility of the sample. Replace the "\i\\ii\\ii\\iv" in the "category" column in Table A.1 to Table A.5 of the English version of ISO/TS19036:2006 to make it consistent with the symbols used for the classification of matrices in A.2.2; delete the preface and foreword of the international standard.
This standard is proposed and managed by the National Certification and Accreditation Administration. The drafting unit of this standard: Shandong Exit-Entry Inspection and Quarantine Bureau of the People's Republic of China. The main drafters of this standard: Ma Weixing, Lin Chao, Wang Manxia, ​​Lei Zhiwen. 1 Scope
Quantitative detection of food microorganisms
Measurement is not Guide to Certainty Assessment
This standard specifies the evaluation and expression methods of measurement uncertainty for quantitative detection of food microorganisms RB/T151—2016
This standard applies to the measurement uncertainty assessment activities for quantitative analysis of food and environmental samples used to monitor processed and stored food, and also applies to instrumental quantitative analysis methods that can replace conventional microbial colony counting methods. This standard does not apply to the most likely value counting method and the analysis method of low-level microorganisms. The measurement uncertainty assessment of quantitative analysis of animal feed products also refers to this standard. 2 Normative References
The following documents The following documents are indispensable for the application of this document. For any dated reference, only the version with the date is applicable to this document. For any undated reference, the latest version (including all amendments) is applicable to this document. GB/Z22553—2010 Guide for the evaluation of measurement uncertainty using estimates of repeatability, reproducibility and trueness GB/T27025-2008 General requirements for the competence of testing and calibration laboratories ISO16140:2003 Microbiology off and animal feeding stuffs Protocol for the validation of alternative mcthods ISO/IEC Guide 98-3:2008 Uncertainty in measurement Part 3: Guide to the expression of uncertainty in measurement (GUM: 1995, Guide to the expression of uncertainty in measurement) 3 Terms and definitions
The following terms and definitions apply to this document. 3.1
measurement uncertainty
uncertainty (of measurement) parameter associated with the result of a measurement that characterizes the dispersion of the values ​​that can reasonably be attributed to the measurand. NOTE 1 This parameter can be the standard deviation or a multiple thereof, or the half-width of a confidence interval for a given confidence probability. NOTE 2 Measurement uncertainty generally consists of a number of components, some of which can be calculated from the statistical distribution of the measurement result and characterized by the experimental standard deviation, while others can be calculated from an assumed probability distribution based on experience or other information and also characterized by the standard deviation. The measurement result is the best estimate of the value of the measurand. All components of uncertainty are related to the dispersion, including those arising from systematic effects (e.g. those associated with corrections and reference measurement standards). 3.2
standarduncertainty
standard uncertainty
measurement uncertainty expressed in terms of standard deviation. 3.3
combined standard uncertaintyCombined standard uncertainty
When the measurement result is obtained from the values ​​of several other quantities, the standard uncertainty is calculated according to the variance and covariance of the other quantities. 1
RB/T151—2016
Expanded uncertaintyexpandeduncertaintyU
The quantity that determines the interval of the measurement result, and most of the values ​​reasonably assigned to the measured value are in this interval. Note: This interval is also called the coverage probability or confidence interval probability (confidence probability). The expanded uncertainty must have a clear expansion interval. To combine a specific confidence probability with the interval specified by the expanded uncertainty, it is necessary to make external or implicit assumptions about the probability distribution represented by the measurement result and its standard uncertainty. The possible confidence probability of this interval is only valid to the extent that these assumptions are verified. The expanded uncertainty U is calculated from the combined standard uncertainty u ((3.3)) and the coverage factor k (3.5): U = ku(y). 3.5
Coverage factor
coverage faetnr
The value of the coverage factor is usually taken as a multiple of the combined standard uncertainty, and the product of the two is the expanded uncertainty. Note: The coverage factor is within the range of
.
The difference between the expected value of the result and the acceptable reference value Note: Compared with random error, bias is the systematic difference of the points. It is the difference between the acceptable reference value and the expected value. 4 Principles
4.1 Overall method to evaluate measurement uncertainty
, it may also be composed of multiple systematic errors. The larger the bias value, the more it indicates that this standard uses the overall method to evaluate measurement uncertainty. It is based on the total variables that affect the test result analysis procedure. This total variable includes precision (arbitrary components) and bias (systematic components). In practical applications in the field of food microbiology, bias is usually not considered (4.2).
The overall method evaluation of measurement uncertainty in this standard is obtained by experimentally evaluating the standard deviation of the reproducibility of the final result of the entire measurement procedure. This standard deviation is equivalent to the combined standard uncertainty (5.1). The overall approach can be viewed as a “black box” system (as shown in Figure 1), where the main sources of uncertainty in food microbiology are identified. This diagram helps to identify the sources of uncertainty, whether or not they are included in the experimental plan2
Laboratory
Culture medium
Random error
Secondary sampling/
Primary dilution
Operator/time
RB/T151—2016
Figure 1 Schematic diagram of the main sources of uncertainty in food microbiology (assessment of measurement uncertainty) Sampling is an important component of the total error, but it is not part of the uncertainty of the measurement itself. Secondary sampling refers to the part of the sample that is taken out for testing. The preparation of the initial suspension in the bacterial counting technique belongs to secondary sampling. The main sources of uncertainty in the analysis process are operator, time, equipment, culture medium and reagents. Finally, the remaining random error is caused by unexplained factors and is often considered only when the laboratory is evaluated under repeatable conditions. At the same time, the use of the holistic method to assess uncertainty can control 4.2 Bias
According to experience, the assessment of the uncertainty of measurement of food microbial counts generally does not take bias into account. That is, the analytical steps directly determine the measurement results, such as the number of colonies per unit sample, and it is actually impossible to obtain the true value. Even if the materials used are approved and the values ​​obtained have been verified by laboratories, only part of the total bias can be assessed. At the same time, some bias can be assessed through inter-laboratory studies, which are used in the two options for evaluating the reproducibility standard deviation in this standard (6 and 7). Although this standard does not consider the uncertainty bias component, the laboratory bias that occurs in practice can be controlled by multiple parties, such as inter-laboratory proficiency tests or standard material controls. 5 General
5.1 Combined standard uncertainty
In this standard, the combined standard uncertainty is calculated by the reproducibility standard deviation (5.2) of the final measurement result. 5.2 Reproducibility standard deviation
The following are three possible situations for estimating the reproducibility standard deviation (s), in the following order of priority: The first: actual Intra-laboratory reproducibility standard deviation, i.e., intra-laboratory intermediate precision. It refers to the precision of the measurement results in the same laboratory, which is determined by factors such as matrix, time and operator due to changes in laboratory internal conditions; Note: Because it refers to the original English version of ISO/TS19036:2006, the name "intra-laboratory reproducibility standard deviation" is continued to be used for expression. The second type: inter-laboratory collaborative reproducibility standard deviation; the third type: inter-laboratory level test reproducibility standard deviation. 3
RB/T151—2016
The first case is the most important.
The basic principles for estimating the reproducibility standard deviation are listed in 5.4. The above three cases will be explained in detail in Chapters 6 to 8.
5.3 Expanded uncertainty
According to ISO/IEC Guide 98-3: 2008, the expanded uncertainty U is the product of the combined standard uncertainty u(y) (3.3) and the inclusion factor table (3.5). The k value in this standard is 2 (corresponding to a confidence level of approximately 95%), calculated according to formula (1): U=ku.(y)=2u .(y)=2sR
Where:
Inclusion factor;
Expanded uncertaintybZxz.net
u(y)
Combined standard uncertainty
Reproducibility standard deviation
Basic principles for estimating reproducibility standard deviation·（1）
The black concept cited in this standard actually advocates that as many sources of uncertainty as possible in Figure 1 should be considered, especially the laboratory should understand the distribution of microorganisms in the test matrix and take it into account in the secondary sampling component of the measurement uncertainty ( 1.1) The reproducibility standard deviation should be estimated for each standard microorganism (or a group consistent with the standard microorganism) and preferably for each matrix (or a group consistent with the matrix). Note 1: “Reproducibility” is the value of measurement uncertainty that is determined for a specified laboratory method used to obtain routine test results. Note 2: The main factors affecting the estimation of measurement uncertainty are the laboratory’s own specific test conditions (e.g. different operators, operating procedures, instrumentation, etc.). The influence of the experimental space analysis method on the measurement uncertainty will be evaluated and the unspecific test results will be obtained in accordance with the standards of B/127025-2008. For example, the key factors that may affect the measurement results from the method or laboratory should be determined, such as the source and type of culture medium or other reagents, calculation method (manual or automatic operator, etc.), and it should be proved that these key factors are under effective control. In order to ensure that the estimation can reflect the actual situation and the test results can be effectively controlled, necessary process monitoring should be carried out on the estimation of measurement uncertainty. When any key factors in the measurement change, the measurement uncertainty needs to be re-estimated. 6 In-laboratory reproducibility standard deviation 6.1 Overview The in-laboratory reproducibility standard deviation is the first choice for obtaining measurement uncertainty. Laboratory Based on this, the measurement uncertainty value can be given in the test report. The theoretical disadvantage of this method is that it cannot take into account bias. In food microbiology, the influence of the matrix on the measurement uncertainty is inevitable, so the experimental plan needs to consider the influence of secondary sampling from the test part of the sample (such as the tested food sample) on the measurement uncertainty. For each target microorganism (or a group consistent with the target microorganism) and a specified matrix, the implementation of the experimental plan (6.2) is carried out on at least 10 samples in the same matrix. In order to cover the changes in operating conditions caused by different times as much as possible, repeated experiments need to be carried out on different days, so that cumulative data over a period of time can be obtained. The number of matrix types tested depends on the diversity of matrices routinely analyzed in the laboratory. The selected matrix should have a good influence on the uncertainty value and on the type of matrix used in previous laboratory analyses. It should be representative in terms of type and suitable for the microorganisms being tested. Appendix A provides us with a guide to matrix selection by listing the experimental results of a large number of international-level laboratories. Its purpose is to evaluate the measurement uncertainty components related to secondary sampling and preparation of initial suspensions from the test portion of the laboratory sample. For more guidance, please refer to Appendix B of ISO16140:2003.
RB/T151—2016
Without considering low-level contamination, convert the data into logarithmic form and then calculate the standard deviation, which can stabilize the reproducibility changes of the standard deviation under different contamination levels. Therefore, it is not necessary to estimate the reproducibility standard deviation of each contamination level one by one, but the selected samples and (or) dilution concentrations should still cover the concentration range of routine testing as much as possible. When selecting samples, use as much white natural as possible. Contaminated samples are used to give a more realistic estimate of the measurement uncertainty, which can better characterize the experimental results of naturally contaminated samples. If inoculation of the target bacteria is required, it must be strictly controlled to prevent other influences on the results. The inoculation needs to be designed to simulate the real contaminants as much as possible.
6.2 Experimental plan
The experimental plan is described in Figure 2.
Food samples
Operator 1 (Case A)
Primary suspension
Artificial contaminants
(if required)
Different cases
Operator 2 (Case B)
Primary suspension
Figure 2 Experimental plan for calculating the standard deviation of intra-laboratory reproducibility For each sample, each operator takes a test dose, uses it to prepare a primary suspension, and performs one analysis. The analysis is performed as a normal test (e.g., a series of 10-fold serial dilutions, each dilution is inoculated into 12 blood cells). In the test, the "operator" can be a group of people (technicians), each responsible for a designated part. In this case, the entire team is considered as one operator. When the task allocation of a member changes, the team member is considered as another operator. This approach is derived from the black box approach mentioned in 4.1. Different sources of uncertainty, such as subsampling, matrix properties, residual random errors, operators, time, etc., are considered at the same time. Case A and Case B should be as different as possible and should include as many different situations encountered in the laboratory on different testing dates as possible, such as technicians, batches of culture media and reagents, vortex mixers, pH meters, incubators, analysis times, etc. If the contamination of food samples is relatively stable (this situation is relatively rare in food microbiology), Case A and Case B should include different analysis dates. 6.3 Application
Figure 3 shows the main sources of uncertainty in this protocol, as well as the excluded items (sampling and bias). 5
RB/T151—2016
Note: Excluded sources are indicated by a cross.
Culture medium
Secondary sampling/
Primary dilution
Random error
Operator/time
Figure 3 Main sources of uncertainty covered and excluded in the intra-laboratory reproducibility experiment Results
This protocol incorporates the effects of sampling of the test part into the total uncertainty assessment. In addition, in food microbiology, the natural contamination level of foods (especially processed, cooked solid foods, etc.) often varies greatly. Therefore, this protocol takes into account the differences in results caused by this heterogeneity, which is very important when evaluating sample analysis results that exceed a certain range (e.g., as specified in microbiological standards). Note: If the uncertainty is estimated for an artificially contaminated initial suspension (Figure 2), the influence of matrix differences on the uncertainty can be ignored. However, this approach is not feasible in some cases. The distribution of natural contaminants is closely related to the type of matrix, which is why the experimental protocol is to repeat the analysis routinely using each matrix (or a consistent group of matrices). Of course, the more matrices a laboratory analyzes, the greater the workload. This protocol does not include the possible influence of bias on the measurement uncertainty. 6.4 Calculations
By convention, data (microbial count results) units are converted from CFU/g or CFU/mL to log1 (CFU/g) or log1e (CFU/mL) before calculations
Note: According to ISO, the symbol for decimal logarithms is "1g". However, in this standard, it is recommended to use the symbol \logio\
widely used in the field of food microbiology laboratories to calculate the reproducibility standard deviation 5R of n given matrix samples, which is calculated according to formula (2): SR
Where:
Reproducibility standard deviation;
(yiA-YB)
Logarithmic transformation value, unit is log1. (CFU/g) or logh (CFU/mL); number of given matrix samples;
Order number of samples, i=1～n (n≥10); Number of reproducibility conditions, j=A or j=B. Table 1 lists examples of aerobic mesophilic bacteria in mixed poultry meat. 6
-（2)
Calculation of reproducibility standard deviation
RB/T151—2016
Example of counting aerobic mesophilic bacteria in mixed poultry meat YiAlogIo(TA)
For y; the value is taken as logarithm, then the reproducibility standard deviation is: SR
（Yi—y）\/2
Reproducibility standard deviation for inter-laboratory collaboration
7.1 Overview||tt ||0.0064+0.0017+.+0.0453
yBloglo（rB）
(YiA-B)\
V0.02191-0.15(log1o)CFU/g
If the method for routine testing in the laboratory has been submitted to the laboratory for confirmation, the laboratory can use this reproducibility standard deviation calculation method to calculate the measurement uncertainty under the following prerequisites (see below). The standard deviation from the inter-laboratory reproducibility study is only related to the method and has nothing to do with the laboratory reporting the uncertainty results. These prerequisites are:
The laboratory deviation must be consistent with the repeatability and reproducibility expected to be obtained in the inter-laboratory collaborative study; the precision obtained from the laboratory measurement must be consistent with the repeatability and reproducibility expected to be obtained in the inter-laboratory collaborative study; the inter-laboratory collaborative study must correctly include various sources of uncertainty (especially sample preparation and homogenization). GB/Z225532010 describes these situations in detail. 7.2 Application in food microbiology
Figure 4 shows the main sources of (sampling) uncertainty included and excluded in this protocol. 7
RB/T151—2016
Note: Excluded sources are marked with a cross.
Media
Subsampling/primary dilution
Repeatability standard deviation
Operator/time
Figure 4 Main sources of uncertainty included and excluded in interlaboratory collaborative studies Subsampling, initial suspension preparation and matrix have an influence on the uncertainty depending on the experimental design. Results
This method enables a laboratory participating in an interlaboratory collaborative study to assess its experimental bias, a partial bias component of the measurement uncertainty, which is not covered in detail in this standard. The calculation method is given in Chapter 6. However, in food microbiology, this method has certain limitations, which are listed below. This is why it is listed as the second case in the order of priority in 5.2.
Except for checking that interlaboratory precision and bias values ​​agree with those obtained in interlaboratory studies using the method, few reproducibility parameters have actually been derived from interlaboratory standardized reference method studies. Moreover, it is difficult to generalize from the analysis of a particular sample to the analysis of routine samples. The precision derived from interlaboratory studies is obtained under certain constraints that incorporate matrix, microbial species, contamination level, and a given background microbial flora (if present). Furthermore, the homogeneity and stability of the samples during delivery to the laboratory space reduces the chance of sample contamination compared to reality, resulting in an underestimation of uncertainty, as the requirement for sample homogeneity in collaborative studies is met. 8 Standard Deviation of Reproducibility in Interlaboratory Proficiency Tests If a laboratory participates in an interlaboratory proficiency test, the standard deviation of reproducibility derived from the test may be used to deduce its measurement uncertainty if:
The method used in the interlaboratory test is the same as that used in routine analysis; The samples used in the test are similar (in terms of matrix and contamination level) to the samples used in routine analysis; The participating laboratories do not use different test methods or a sufficient number of participants use the same method to ensure that the standard deviation of reproducibility is correctly estimated.
Figure 4 illustrates the main sources of uncertainty and one excluded factor in this approach: sampling. One of the purposes of this approach is to enable the laboratories participating in the interlaboratory proficiency test to assess some of the bias components of the measurement uncertainty, which are not covered in detail in this standard approach. The calculation method is given in Chapter 6. 9 Representation of Measurement Uncertainty in Test Reports Once the measurement uncertainty has been derived in accordance with the above method, it may be reported together with the test results, if necessary, in the report as an interval in log1. units, a natural value (CFU per gram or per milliliter) or a percentage. Examples are given below. The reported result 9
or the value of the uncertainty interval shall not exceed two significant figures RB/T151—2016
The test result is expressed as y=log10, and the reproducibility standard deviation is expressed as sR. When the coverage factor k is 2 (confidence level 95%), the expanded uncertainty U is equal to 25R.
The test results can be reported in the following form: y±2sr (logio);
y·logioLy-2sR·y+2s;
xCFU/g or CFU/mL[10%-2R, 10+2k], %, 10%+10%+2start
αCFU/g or CFU/mL10%_
10%-2R
Example: The reproducibility standard deviation 8k is ±0.15log19, so the expanded uncertainty U with a coverage factor of 2 (95% confidence level) is 0.15×2-0.3log10: the test result is 5.0log1oCFU/g. The test results may be reported in the following form: -5.0log1o±0.3logie
5.0logio[4.7,5.3];
10°CFU/g[5X10/,2X105]
-10CFU/g[10550%,105+100%
RB/T151—2016
Annex A
(Informative)
Test results of uncertainty components related to secondary sampling tests and initial suspension preparation A.1 Introduction and experimental plan
The tests were organized by AFSSA (France) on behalf of the ISO/TC34/SC9 Secretariat and were conducted in 2003 and 2004. The aim was to determine the influence of the secondary sampling and initial suspension preparation components on the measurement uncertainty in different product matrices. One sample can produce 8 sets of counting results under different conditions, as shown in Figure A.1. Laboratory ()
Initial suspension (A)
Condition (1)
Description:
Initial suspension (A), initial suspension (B)
Condition (1), condition (2)
No. A11, A12, etc.
Condition (2)
Condition (1)
Initial suspension (B)
Condition (2)
Two sets of initial suspensions prepared separately, the conditions of the two sets of suspensions are as different as possible (such as different operators, different balance levels, different batches of dilutions, etc.); two sets of conditions are as different as possible (different operators, different batches of culture media, different incubators, etc.); two operations under repeated conditions (for example, two sets of dilutions of the initial suspension under each condition). Note: The experimental scheme in Figure 2 (6.2) is a simplified version of Figure A.1. Only A11 and B21 in Figure A.1 are listed in Figure 2. Figure A.1
Experimental plan
A.2 Results
A.2.1 Overview
A total of 79 laboratories participated in the test. The laboratories were numbered from 1 to 79. Each laboratory used one or more methods to test one or more types of matrices. Therefore, a total of 122 data files were collected (one laboratory, one method and one matrix). Among them, 28 data files were not used because they did not meet the test standards. In this way, 94 reproducibility standard deviations (sR) were obtained, each from one laboratory, one method and one matrix. In addition, according to the variation component theory, each reproducibility standard deviation can be identified from the following three sources of uncertainty: a) related to the matrix, subsampling of the test part and initial dilution; b)
related to the reproducibility situation (operator/time); c
related to random errors under repeatability conditions.
Tip: This standard content only shows part of the intercepted content of the complete standard. If you need the complete standard, please go to the top to download the complete standard document for free.