ENV ISO 13843:2001

SLOVENSKI STANDARD
01-junij-2004
Kakovost vode – Navodilo za validacijo mikrobioloških metod (ISO/TR 13843:2000)
Water quality - Guidance on validation of microbiological methods (ISO/TR 13843:2000)
Wasserbeschaffenheit - Richtlinie zur Validierung mikrobiologischer Verfahren (ISO/TR
13843:2000)
Qualité de l'eau - Lignes directrices pour la validation des méthodes microbiologiques
(ISO/TR 13843:2000)
Ta slovenski standard je istoveten z: ENV ISO 13843:2001
ICS:
07.100.20 Mikrobiologija vode Microbiology of water
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.

EUROPEAN PRESTANDARD
ENV ISO 13843
PRÉNORME EUROPÉENNE
EUROPÄISCHE VORNORM
May 2001
ICS 07.100.20
English version
Water quality - Guidance on validation of microbiological
methods (ISO/TR 13843:2000)
Qualité de l'eau - Lignes directrices pour la validation des Wasserbeschaffenheit - Richtlinie zur Validierung
méthodes microbiologiques (ISO/TR 13843:2000) mikrobiologischer Verfahren (ISO/TR 13843:2000)
This European Prestandard (ENV) was approved by CEN on 7 April 2001 as a prospective standard for provisional application.
The period of validity of this ENV is limited initially to three years. After two years the members of CEN will be requested to submit their
comments, particularly on the question whether the ENV can be converted into a European Standard.
CEN members are required to announce the existence of this ENV in the same way as for an EN and to make the ENV available promptly
at national level in an appropriate form. It is permissible to keep conflicting national standards in force (in parallel to the ENV) until the final
decision about the possible conversion of the ENV into an EN is reached.
CEN members are the national standards bodies of Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Greece,
Iceland, Ireland, Italy, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland and United Kingdom.
EUROPEAN COMMITTEE FOR STANDARDIZATION
COMITÉ EUROPÉEN DE NORMALISATION
EUROPÄISCHES KOMITEE FÜR NORMUNG
Management Centre: rue de Stassart, 36  B-1050 Brussels
© 2001 CEN All rights of exploitation in any form and by any means reserved Ref. No. ENV ISO 13843:2001 E
worldwide for CEN national Members.

Page 2
CORRECTED 2001-08-15
Foreword
The text of the Technical Report from Technical Committee ISO/TC 147 "Water quality" of the
International Organization for Standardization (ISO) has been taken over as a European
Prestandard by Technical Committee CEN/TC 230 "Water analysis", the secretariat of which
is held by DIN.
This European Prestandard shall be given the status of a national standard, either by
publication of an identical text or by endorsement, at the latest by November 2001, and
conflicting national standards shall be withdrawn at the latest by November 2001.
According to the CEN/CENELEC Internal Regulations, the national standards organizations of
the following countries are bound to announce this European Prestandard: Austria, Belgium,
Czech Republic, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy,
Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland and the United
Kingdom.
Endorsement notice
The text of the Technical Report ISO/TR 13843:2000 has been approved by CEN as a
European Prestandard without any modifications.

TECHNICAL ISO/TR
REPORT 13843
First edition
2000-06-01
Water quality — Guidance on validation of
microbiological methods
Qualité de l'eau — Lignes directrices pour la validation des méthodes
microbiologiques
Reference number
ISO/TR 13843:2000(E)
©
ISO 2000
ISO/TR 13843:2000(E)
PDF disclaimer
This PDF file may contain embedded typefaces. In accordance with Adobe's licensing policy, this file may be printed or viewed but shall not
be edited unless the typefaces which are embedded are licensed to and installed on the computer performing the editing. In downloading this
file, parties accept therein the responsibility of not infringing Adobe's licensing policy. The ISO Central Secretariat accepts no liability in this
area.
Adobe is a trademark of Adobe Systems Incorporated.
Details of the software products used to create this PDF file can be found in the General Info relative to the file; the PDF-creation parameters
were optimized for printing. Every care has been taken to ensure that the file is suitable for use by ISO member bodies. In the unlikely event
that a problem relating to it is found, please inform the Central Secretariat at the address given below.
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means, electronic
or mechanical, including photocopying and microfilm, without permission in writing from either ISO at the address below or ISO's member body
in the country of the requester.
ISO copyright office
Case postale 56 � CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 734 10 79
E-mail copyright@iso.ch
Web www.iso.ch
Printed in Switzerland
ii © ISO 2000 – All rights reserved

ISO/TR 13843:2000(E)
Contents Page
Foreword.iv
1 Scope .1
2 Terms and definitions .1
3 Arrangement of the document .8
4 Basic concepts.8
4.1 General.8
4.2 Validation.8
4.3 Detectors .11
4.4 Performance characteristics .11
4.5 Specifications.11
5 Limitations and characteristic features of microbiological methods .12
5.1 Recovery of the analyte .12
5.2 Sample variance.12
5.3 Particle distribution and overdispersion.12
5.4 Interactions in the detector.12
5.5 Robustness .13
5.6 Spurious errors.13
5.7 Control and guidance charts.13
6 Mathematical models of variation.14
6.1 Unavoidable basic variation — The Poisson distribution .14
6.2 Overdispersion — The negative binomial model .17
6.3 Statistical and practical limits .20
6.4 General tests for randomness — Detection of overdispersion .21
7 Specifications — Current practice.21
8 Specifications — Recommended approach.22
9 Determination and expression of performance characteristics .23
9.1 General.23
9.2 Categorical characteristics related to specificity and selectivity.23
9.3 Working limits .24
9.4 Working range of MPN procedures.25
9.5 Precision.25
10 Procedures and steps of validation.26
10.1 General.26
10.2 Primary validation.26
10.3 Secondary validation.28
11 Designs for determining specifications .28
11.1 A general model for basic quantitative specifications .28
11.2 Precision of the entire analytical procedure.29
11.3 Categorical characteristics.29
11.4 Unplanned data.29
Annex A Statistical procedures and computer programs.30
Annex B Numerical examples .34
Annex C Example of a validation experiment.45
Bibliography.46
ISO/TR 13843:2000(E)
Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies (ISO
member bodies). The work of preparing International Standards is normally carried out through ISO technical
committees. Each member body interested in a subject for which a technical committee has been established has
the right to be represented on that committee. International organizations, governmental and non-governmental, in
liaison with ISO, also take part in the work. ISO collaborates closely with the International Electrotechnical
Commission (IEC) on all matters of electrotechnical standardization.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 3.
The main task of technical committees is to prepare International Standards. Draft International Standards adopted
by the technical committees are circulated to the member bodies for voting. Publication as an International
Standard requires approval by at least 75 % of the member bodies casting a vote.
In exceptional circumstances, when a technical committee has collected data of a different kind from that which is
normally published as an International Standard ("state of the art", for example), it may decide by a simple majority
vote of its participating members to publish a Technical Report. A Technical Report is entirely informative in nature
and does not have to be reviewed until the data it provides are considered to be no longer valid or useful.
Attention is drawn to the possibility that some of the elements of this Technical Report may be the subject of patent
rights. ISO shall not be held responsible for identifying any or all such patent rights.
ISO/TR 13843 was prepared by Technical Committee ISO/TC 147, Water quality, Subcommittee SC 4,
Microbiological methods.
iv © ISO 2000 – All rights reserved

TECHNICAL REPORT ISO/TR 13843:2000(E)
Water quality — Guidance on validation of microbiological
methods
1 Scope
This Technical Report deals with validation of microbiological methods, with particular emphasis on selective
quantitative methods in which the quantitative estimate is based on counting of particles either directly, with the aid
of a microscope, or indirectly, on the basis of growth (multiplication) into colonies or turbidity.
The principles and procedures within this scope are commonly known as the presence/absence (P/A), most
probable number (MPN), colony count and direct (microscopic) count.
This Technical Report does not apply to the validation of the so-called rapid or modern methods which mostly
depend on measuring products or changes due to microbial activity but do not address the detection of individual
particles.
2 Terms and definitions
For the purposes of this Technical Report, the following terms and definitions apply.
2.1
accuracy of measurement
closeness of the agreement between a test result and the accepted reference value
NOTE The term “accuracy”, when applied to a set of test results, involves a combination of random components and a
common systematic error or bias component.
[ISO 3534-1:1993, 3.11]
2.2
analyte
measurand
particular quantity subjected to measurement
NOTE 1 See reference [5].
NOTE 2 In microbiology the analyte is ideally defined as a list of taxonomically defined species. In many cases, in practice
the analyte can only be defined by group designations less accurate than taxonomic definitions.
2.3
analytical portion
test portion
volume of particle suspension inoculated into a detector unit
NOTE Examples of a detector unit are agar plate, membrane filter, test tube, microscopic grid square.
2.4
application range
range of particle concentrations routinely subjected to measurement by a method
ISO/TR 13843:2000(E)
2.5
categorical characteristic
method performance characteristic numerically expressed as a relative frequency based on P/A or +/� classification
2.6
CFU, deprecated
colony-forming unit, deprecated
CFP, deprecated
colony-forming particle, deprecated
NOTE The term was originally introduced to convey the idea that a colony may originate not only from a single cell but from
a solid chain or aggregate of cells, a cluster of spores, a piece of mycelium, etc. It mistakenly equates the number of colonies
observed to the number of living entities seeded on the medium. Growth unit, viable particle, propagule (2.27) and germ (2.13)
are terms with similar meanings but convey the original idea better and apply not only to colony-count methods but also to MPN
and P/A.
2.7
coefficient of variation
CV
relative standard deviation
for a non-negative characteristic, the ratio of the standard deviation to the average
NOTE 1 The ratio may be expressed as a percentage.
NOTE 2 The term "relative standard deviation" is sometimes used as an alternative to "coefficient of variation", but this use is
not recommended.
[ISO 3534-1:1993, 2.35]
NOTE 3 In this Technical Report the term coefficient of variation (CV) is used when the relative standard deviation is
expressed in percent (CV % = 100 RSD).
2.8
collaborative test
method or laboratory performance test where several laboratories join in an experiment planned and co-ordinated
by a leader laboratory
NOTE Collaborative tests are mainly of two types. Intercalibration exercises are made to allow laboratories to compare
their analytical results with those of other participating laboratories.
Method performance tests produce precision estimates (repeatability, reproducibility) out of data accumulated when several
participating laboratories study identical samples with a strictly standardized method.
2.9
confirmed [verified] colony count
x
presumptive colony count corrected for false positives
k
x= pc= c
n
where
c is the presumptive count;
p is the true positive rate;
n is the number of presumptive positives isolated for confirmation;
k is the number confirmed.
2 © ISO 2000 – All rights reserved

ISO/TR 13843:2000(E)
2.10
control chart
two-dimensional scattergram for monitoring method performance with control values obtained by a Type A study
NOTE In control charts the horizontal axis is usually in the time scale or ordinate scale and the control variable is the mean
or some precision measure (s, CV, RSD).
2.11
detector
particle detector
plate of solid matrix or a tube of liquid containing a nutrient medium for counting or detecting living microbial
particles
2.12
detection set
detector set
combination of plates or tubes on which quantitative estimation of microbial concentration in a sample is based
NOTE The detection set is the set of plates or tubes utilized for numerical estimation of a single value.
EXAMPLES Parallel plates of a suspension, plates from consecutive dilutions, 3 � 5 tube MPN system, microtitre plate.
2.13
germ
living entity capable of producing growth in a nutrient medium
cf. propagule (2.27)
2.14
guidance chart
two-dimensional scattergram for presenting method performance data (quantity or precision) with arbitrary guide
values or guide values obtained by Type B reasoning
NOTE In guidance charts, the horizontal axis is usually the colony count per detector.
2.15
heterogeneous Poisson distribution
distribution arising when the mean of a Poisson distribution varies randomly from occasion to occasion
NOTE 1 See reference [11].
NOTE 2 See also negative binomial distribution (2.19).
2.16
limit of detection
particle number x (per analytical portion) where the probability p of a negative result equals 5 %
NOTE 1 Probability of a positive result p(+) = 1 – p .
NOTE 2 a) Calculation of x via Poisson distribution:
��
11��
x= ln = ln = ln 20 = 3,00
� �
��
��
��
p 0,05
��
b) Calculation of x via negative binomial distribution:
��
�u
p � 1
��0 ��uu
�� 0,05��1 20 1
x�� �
22 2
uu u
ISO/TR 13843:2000(E)
2.17
limit of determination
lowest average particle concentration x per analytical portion where the expected relative standard uncertainty,
equals a specified value (RSD)
NOTE a) Calculation of x via Poisson distribution:
x �
RSD
� �
b) Calculation of x via negative binomial distribution:
x= , given overdispersion factor = u
RSD � u
� �
2.18
linearity
linear dependence of the signal on concentration of the analyte
cf. proportionality (2.28)
2.19
negative binomial distribution
a particular "overdispersed" statistical distribution of counts
NOTE 1 Its variance can be expressed as
=+� �
� u
where ��is the mean.
NOTE 2 In this Technical Report the square of the overdispersion factor (u) is substituted for the inverse of the exponent
(1/k) of the standard formula for the negative binomial distribution.
2.20
overdispersion
variation in excess of Poisson randomness
NOTE It is detected qualitatively by the Poisson index of dispersion, and measured quantitatively by estimating the
parameter u (overdispersion factor) of the negative binomial distribution.
2.21
overdispersion factor
u
additional random uncertainty of determination in excess of the Poisson distribution, measured in terms of relative
standard deviation
2.22
overlap error
crowding error
systematic depression of colony counts due to confluence of colonies
NOTE Quantitatively, overlap error depends primarily on the fraction of available growth space occupied by colonial
growth.
2.23
parallel counts
particle or colony numbers in equal analytical portions drawn from the same suspension
4 © ISO 2000 – All rights reserved

ISO/TR 13843:2000(E)
2.24
Poisson distribution
fully random distribution of particle numbers when sampling a perfectly mixed suspension
NOTE The probability P(k) of observing exactly k units in a test portion when the mean equals � is calculated from
k
�
��
Pk��=
e
k!
2.25
precision
closeness of agreement between independent test results obtained under stipulated conditions
NOTE Precision does not relate to the true value or the specified value. It is usually expressed in terms of imprecision and
computed as a standard deviation of the test results.
2.26
primary validation
full validation
establishment of the specifications for the performance of a new method and/or experimental verification that a
method meets theoretically derived quality criteria
2.27
propagule
a viable entity, vegetative cell, group of cells, spore, spore cluster, or a piece of fungal mycelium capable of growth
in a nutrient medium
cf. germ (2.13)
2.28
proportionality
agreement of observed particle counts with the volume (or dilution) of a series of analytical portions from a common
root suspension
NOTE Proportionality is computed for statistical evaluation as the log-likelihood ratio statistic G with n–1 degrees of
freedom.
2.29
qualitative method
method of analysis whose response is either the presence or absence of the analyte in a certain amount of sample
NOTE See reference [10].
2.30
recovery
general term for the number of particles estimated in a test portion or sample, with the understanding that there is a
true (although unknown) number of particles of which 100 % or less are "recovered" by the detector
2.31
relative accuracy
degree of correspondence between the response obtained by the reference method and the response obtained by
the alternative method on identical samples
NOTE See reference [10].
2.32
relative difference
d
difference between two measured values divided by their mean
ISO/TR 13843:2000(E)
2xx�
xx� ��
AB
AB
d= =
xx �x
AB
dd%1� 00
NOTE For all practical purposes, the same value results from the calculation d =ln(x )–ln(x ).
A B
2.33
relative recovery
ratio (A/B) of colony counts obtained by two methods tested on equal test portions of the same suspension, where
B is the reference (when applicable)
2.34
relative standard deviation
RSD
estimate of the standard deviation of a population from a sample of n results divided by the mean of that sample
s
RSD =
x
cf. coefficient of variation (2.7)
2.35
repeatability
closeness of the agreement between the results of successive measurements of the same measurand carried out
under the same conditions of measurement
NOTE 1 See Guide to the expression of uncertainty in measurement [6].
NOTE 2 Repeatability is computed as r =2,8s , where s is the repeatability standard deviation.
r r
2.36
reproducibility
closeness of the agreement between the results of measurements on the same measurand carried out under
changed conditions of measurement
NOTE 1 See Guide to the expression of uncertainty in measurement [6].
NOTE 2 Reproducibility is computed as R =2,8 s ,
R
where
s is the reproducibility standard deviation usually compounded from the between-laboratories standard deviation s and
R L
repeatability standard deviation s :
r
ss��s
RrL
2.37
robustness
insensitivity of an analytical method to small changes in procedure
NOTE 1 See reference [23].
NOTE 2 To examine the robustness it is advisable to “abuse” the method in a controlled way.
6 © ISO 2000 – All rights reserved

ISO/TR 13843:2000(E)
2.38
secondary validation
demonstration by experiment that an established method functions according to its specifications in the user's
hands
2.39
apparent selectivity
F
ratio of the number of target colonies to the total number of colonies in the same sample volume
F =lg(t/n)
where
t is the apparent concentration of presumptive target types estimated by counting colonies;
n is the concentration of total colonies.
2.40
sensitivity
fraction of the total number of positive cultures or colonies correctly assigned in the presumptive inspection
2.41
specificity
fraction of the total number of negative cultures or colonies correctly assigned in the presumptive inspection
2.42
standard uncertainty
uncertainty of the result of a measurement expressed as a standard deviation
NOTE See reference [5].
2.43
type A evaluation
�of uncertainty� method of evaluation of uncertainty by the statistical analysis of a series of observations
EXAMPLE Observations may be e.g. standard deviation, relative standard deviation.
NOTE 1 See references [5] and [6].
NOTE 2 Repeatability and reproducibility are often estimated by carrying out collaborative method performance tests where
several laboratories study "identical" samples provided by a central organizer [15].
2.44
type B evaluation
�of uncertainty� method of evaluation of uncertainty by means other than the statistical analysis of series of
observations e.g. from assumed probability distributions based on experience or other information
NOTE See references [5] and [6].
2.45
uncertainty
�of measurement� parameter, associated with the result of a measurement, that characterizes the dispersion of the
values that could reasonably be attributed to the measurand
NOTE See reference [6].
2.46
uncertainty
�of counting� relative standard deviation of results of repeated counting of the colonies or particles of the same
plate(s) or field(s) under stipulated conditions
ISO/TR 13843:2000(E)
EXAMPLE Stipulated conditions may be e.g. the same person or different persons in one laboratory, or different
laboratories.
2.47
validation range
range of mean number of particles per analytical portion for which obeyance of validation specifications (particularly
linearity) have been acceptably demonstrated
NOTE It is usually expressed as the range of "reliable" colony counts.
3 Arrangement of the document
The first part (clauses 4 to 8) of this Technical Report contains informative material on basic principles,
characteristics and limitations of microbiological methods, as well as on general aspects of validation. The second
half (clauses 9 to 11) is the actual validation document, containing specifications and recommended procedures for
their determination.
Old and new concepts and principles are not completely defined in the body of this Technical Report. Three
annexes are attached. Annex A details the statistical formulae most relevant to this document, annex B contains
numerical examples and annex C gives detailed plans for two validation experiments.
Statistical tests in the ordinary sense are not central to the ideas. Mathematical calculations are used mainly for the
purpose of providing convenient summaries of data and statistical distributions provide guidance values. A table of
the � distribution is the guide most frequently consulted.
The two BASIC programmes given in annex A are easily copied into desk-top computers or programmable pocket
calculators to help with the basic calculations most frequently needed.
4 Basic concepts
4.1 General
As far as particle statistics is concerned, microscopic counts obey the same laws as viable counts but they are, with
the exception of microcolony methods, free from the biological problems associated with growth. Differential stains,
specifically labelled complexes or other agents used for finding the target do not change the metrological principles.
The same validation principles as with selective colony methods can be applied.
Plaque counts of bacteriophages are in most respects similar to bacterial colony counts.
4.2 Validation
4.2.1 General
Validation means a process providing evidence that a method is capable of serving its intended purpose: to detect
or quantify a specified microbe or microbial group with adequate precision and accuracy. The total count methods
do not have a definable target group and can only be validated in relation to other methods or against theoretical
expectations of precision.
Validation is classified as primary or secondary according to its purpose.
4.2.2 Primary validation
Primary validation is an exploratory process with the aim of establishing the operational limits and performance
characteristics of a new, modified or otherwise inadequately characterized method. It should result in numerical and
descriptive specifications for the performance and include a detailed and unambiguous description of the target of
interest (positive colony, tube or plaque).
8 © ISO 2000 – All rights reserved

ISO/TR 13843:2000(E)
Primary validation characteristically proceeds by the use of specially designed test schemes.
A laboratory developing an in-house method or a variant of an existing standard should carry out the steps of
primary validation.
It is imperative that technicians involved in primary validation have considerable experience with other
microbiological methods.
4.2.3 Secondary validation
Secondary validation (also called verification) takes place when a laboratory proceeds to implement a method
developed elsewhere. Secondary validation focuses on gathering evidence that the laboratory is able to meet the
specifications established in primary validation. Presently, specifications are not available for most of the methods.
Results of external quality assurance (see 4.2.8) may have to be used as the first step towards complete secondary
validation.
Typically, secondary validation uses selected and simplified forms of the same procedures used in primary
validation, but possibly extended over a longer time. Natural samples are the optimal test materials and the work
need only address the procedure within the operational limits set by primary validation.
4.2.4 Analytical quality control (AQC)
Application of valid methods in their specified reliable limits does not automatically ensure valid results. Analytical
quality control (AQC) used in connection with daily routine analyses is necessary. It controls the ability to use a
method successfully.
AQC is a continuous process. Guidance charts, with limits derived from method specifications (from primary
validation) or from theoretical considerations are the principal tools.
The methods of AQC are extensions of the routine analytical process, e.g. replications at different levels, or simply
calculations not ordinarily performed on routine data. In addition, reference materials, intercalibrations and spiked
samples are used.
Analytical quality control is needed in connection with primary and secondary validation. Only results reliable in the
AQC sense should be used for derivation of validation criteria and performance characteristics.
International and national working groups have produced numerous documents on analytical quality control of
microbiological methods (e.g. references [1, 2, 3, 4, 7, 8, 20, 21, 24, 26]). Standards manuals also contain sections
on that subject. Although vital to validation, the methods of analytical quality control are not detailed in this
Technical Report. In everything that follows, it is assumed that laboratories have the appropriate analytical controls
and internal and external quality assurance systems in operation.
4.2.5 Equivalent methods
It is necessary to apply two methods in parallel on the same samples when developing an in-house method, and
also when collecting information to justify the use of an alternative method.
Method performance consists of many aspects. There is no single test of method equivalence, nor numerical
criteria for it. One method may be superior in specificity but inferior in recovery. All the collective information about
robustness, precision and specificity gained during validation tests can be used for method comparisons (examples
B.2, B.3, B.4 in annex B). The methods only need to be tested in parallel for recovery comparisons.
A method giving the highest recovery of confirmed target organisms is obviously the best, unless confirmation is
always required for routine use. A method giving somewhat lower recovery but not requiring confirmation may be
preferable. If high false negative rates or false positive rates observed in primary validation cannot be corrected by
more refined target colony definitions, the method should be deemed invalid.
ISO/TR 13843:2000(E)
4.2.6 Test materials
It is a popular notion that validation should simulate routine as much as possible. Natural samples with natural
concentrations of microbes should therefore be the main test materials. There are exceptions under some
circumstances.
Artificial materials (certified reference materials and spiked samples) are used in internal and external quality
assurance systems to ensure the basic proficiency of the laboratories participating in method validation excercises.
Spiking may be useful and even necessary in secondary validation whenever it is difficult to find natural samples
with target organisms. Laboratory personnel will be able to familiarize themselves with the target.
Negative samples (blanks) should be limited to internal quality assurance. Their inclusion among samples studied
for method equivalence may lead to a false impression of a good correlation between methods. If it were possible
to know in advance which natural samples contain no target organisms, they would be a suitable selection for
testing false positives in actual validation excercises.
The optimal concentration range for the validation of microbiological methods is narrower than the projected
application range. High concentrations are unnecessary. Such samples resemble pure cultures and do not put the
performance of the method or the laboratory to test.
Samples with very low bacterial content need to be studied for public health reasons, but are ill suited for method
comparisons and other validation exercises for statistical reasons. The problem is mostly avoidable, because
microbiological methods are generally not concentration-sensitive at the low end of the scale. Each individual germ
reacts with the nutrient medium almost independently of other particles in the sample. If a method has a low
recovery compared to another, the fact is more readily discovered with twenty or thirty colony-forming particles per
plate than with one or a few.
Methods found valid at concentrations sufficient for validation are trusted to work also at low analyte
concentrations.
4.2.7 Samples — Representativeness and sufficiency
Statistical theory provides solutions for calculating the number of samples required for different testing or estimation
situations [3, 13]. To be able to make use of the theory, the size of real effects of importance and the power for their
detection should be defined. An estimate of the uncertainty (precision) of the determination should be available and
random sampling should be practiced.
Many or all of the above requirements are difficult to meet in advance planning and execution of microbiological
method performance tests. Statistical techniques, if used at all, become rough guidelines.
The number and variety of samples examined ought to be sufficient to be convincing. Without the help of statistics,
there are no exact ways of deciding. In some instances, the first sample studied might give the answer that the
method is not good enough. Usually, however, more samples are needed. It may take a thousand samples to
"prove" that two P/A methods are not equivalent. Choice of too few examples may be a waste of time.
4.2.8 External quality assurance and other collaborative tests
Participation of several laboratories in studies of "homogeneous" material are considered essential tests of both
method and laboratory performance. (After outliers have been recognized and deleted, the remaining data are
thought to provide the necessary information on method performance and proficiency.)
Collaborative tests have been developed into a tool widely applied for testing precision characteristics of chemical
methods [34]. It seems somewhat premature to fully recommend the same in microbiology. It is assumed that all
the participating laboratories have several years of experience with the methods tested and a proven ability to use
them. The present experience is that collaborative experiments intended for method performance testing tend to
turn into laboratory proficiency tests and training excercises.
10 © ISO 2000 – All rights reserved

ISO/TR 13843:2000(E)
A number of microbiological methods have been in use for decades (e.g. Endo agar for total coliforms, mFC for
thermotolerant coliforms, m-Enterococcus agar for intestinal enterococci) by hundreds of laboratories. These
methods would therefore theoretically be suitable objects for collaborative method performance testing.
When making collaborative proficiency tests for specific target organisms with selective media, the samples almost
necessarily should be spiked with pure cultures or mixtures of organisms. Another solution is to use certified
reference materials. This is a simplified and artificial situation. Major difficulties experienced by different
laboratories in the routine use of methods on natural samples may be missed. As long as the detailed performance
characteristics of a microbiological method have not been expressed quantitatively, these types of external quality
assurance (EQA) schemes may nevertheless be the most satisfactory means towards secondary validation
(verification) of a method.
4.3 Detectors
4.3.1 General
It is often convenient to call the nutrient medium in its container a detector (2.11). Two types of detector, solid and
liquid, are employed in different microbiological method variants. They are also mostly associated with different
enumeration or detection principles: liquid with P/A and MPN, and solid with colony counts.
All forms of validation in microbiology focus on the performance of the detectors.
The set of tubes (MPN) or the series of (countable) plates used for analysis is called a detection set or a detector
set (2.12). Each individual MPN tube is a P/A detector.
EXAMPLE An individual well of a microtitre plate is a P/A detector. The whole plate when used as an MPN system is the
detection set.
4.3.2 Detector comparisons
For most colony-count methods a liquid counterpart with the same chemical composition but without the solid
matrix (agar, membrane filter) can be produced. The effect of the solid environment can be evaluated by comparing
colony counts with the equivalent MPN estimate provided that the reaction for target recognition is the same on
both types of detectors and the number of parallel tubes is large enough for adequate precision. Also the sensitivity
of P/A detectors can be evaluated by similar liquid-solid comparisons.
4.4 Performance characteristics
Performance characteristics should be quantifiable and testable to be of use in validation.
The terminology on performance characteristics in this Technical Report mostly follows the chemometric usage.
Because the original definitions of the terms do not always fit microbiological methods perfectly, they have been
modified and adapted as necessary.
The performance characteristics dealt with in this Technical Report are related to scope (list of situations and
sample types where the method is applicable), precision, linearity, recovery, working limits in terms of lowest and
highest recommendable colony number per plate, selectivity, specificity and robustness (ruggedness). Definitions
of these and other terms can be found in clause 2.
4.5 Specifications
Specifications are either numerical or qualitative expressions of performance characteristics or of working limits
derived from them. Primary validation should provide the following:
a) morphological identification of the (presumptive) target;
b) statements regarding incubation conditions (temperature, time, gas atmosphere, moisture) and media
characteristics (pH, stability);
ISO/TR 13843:2000(E)
c) a statement regarding reliable working limits in terms of colony or plaque numbers per detector (plate,
membrane filter) if possible;
d) expressions of uncertainty within the specified reliable limits;
e) scope and limitations.
5 Limitations and characteristic features of microbiological methods
5.1 Recovery of the analyte
The microbiological analyte consists of discrete living particles, variously called colony-forming units (CFU), colony-
forming particles (CFP), germs, propagules, etc. (see clause 2). The number of colonies observed is an
approximation of the number of living particles.
The useful arsenal of performance tests is limited by the near impossibility of knowing the true amount of the
analyte in a sample or in an analytical portion. Detectors cannot be challenged with an exactly known number of
germs.
Viability is defined by growth, i.e. by the method itself. Absolute recovery is undefinable and traceability is
impossible. As viability may appear different with different detectors and/or with different sample history, the relative
recovery (involving the new method and a reference) is a practicable performance characteristic even though the
true result is not known.
5.2 Sample variance
In the environment and even in laboratory samples, the distribution of particles is uneven. The sampling variance of
the environment is not a characteristic of the method, whereas the subsampling variance of a laboratory sample
may be considered so. It may not be possible to use mixing practices sufficient to ensure perfect mixing of sample
contents without some loss of viable cells. Within-sample variance often remains considerable and causes
problems in validation. Performance, especially precision and upper working limits, will need to be determined
separately for different matrices.
5.3 Particle distribution and overdispersion
Random variation due to uneven distribution of particles between parallel s
...