624

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

How to validate analytical methods

Roger Wood*

Food Contaminants Division, Joint Food Safety and Standards Group, Ministry of Agriculture,
Fisheries and Food, c / o Institute of Food Research, Norwich Research Park, Colney,
Norwich NR4 7UQ, UK
The requirement for laboratories to use a `fully
validated' method of analysis is now accepted
or required in many sectors of analysis. Fully
validated means that a method must have
been assessed in a collaborative trial. The signicance of these requirements is described
because analysts will increasingly be
required to justify their choice of method in
the light of them. In addition, the requirements
and procedures that may be used to obtain
methods that have been validated `in-house',
without full validation through collaborative
trial, are also outlined in the article, these
pointing the way to international acceptance
of such methods in future as the cost of carrying out full validation of methods through
collaborative
trials
becomes
prohibitory. z1999 Published by Elsevier Science
B.V. All rights reserved.
Keywords: Analytical method; Validation; Collaborative trial;
Study; Legislation

1. Introduction
The selection and development of methods of analysis has traditionally been a subject of importance to
laboratories working in analytical laboratories, often
to the extent that the practical application of the
method is neglected. Possibly, this greater emphasis
is because most organisations, be they governmental
or one of the International Standardising Organisations working in some areas, develop methods of analysis, incorporate them into legislation or International
Standards but then do not have any mechanism to
assess how well such methods are being applied. However, this approach is becoming superseded with the
need to demonstrate that the application of the method
*Corresponding author. Tel.: +44 (0) 1603-255000;
Fax: +44 (0) 1603-507723.
E-mail: [email protected]
0165-9936/99/$ ^ see front matter
PII: S 0 1 6 5 - 9 9 3 6 ( 9 9 ) 0 0 1 5 0 - 8

is also being successfully achieved. In addition analysts will be allowed a greater freedom of choice of
analytical method provided the method chosen meets
certain pre-dened criteria. This so-called criteria
approach or performance-based approach towards
methods of analysis is being progressively adopted
by legislative authorities. Thus the emphasis on the
methods of analysis, particularly for foodstuffs
areas, is now changing with the formal legislative
requirement for and introduction of accreditation,
prociency testing and dened internal quality control
procedures into the laboratory. Nevertheless it is
essential that the quality of the method of analysis is
fully recognised and appreciated, this particularly with
the introduction of the criteria approach as analysts
will increasing have to justify their choice of method.
Notwithstanding other quality assurance requirements, laboratories must ensure the quality of all
their analytical methods in use; they must be validated
and veried. The various means of ensuring the quality of methods vary from one analytical discipline to
the other and the objective of the analysis inuences
the extent of the quality assurance work. The requirements on methods used in process control are, for
example, generally less demanding than the requirements on methods for end product control and ofcial
food control.
The necessity for laboratories to use a `fully validated' method of analysis is now universally accepted
or required within many sectors of analysis. A description of these requirements is therefore given below as
analysts will increasingly be required to justify their
choice of method in the light of these `fully validated'
requirements. The meaning of `fully validated' within
the food sector is illustrated by the requirements of the
various international organisations which develop or
adopt standard methods of Analysis; these are
described in the rst part of this paper.
Most method validation guides start with discussions on how criteria such as specicity, accuracy
and precision of the method shall be established. The
analytical problem, requirements of the customers and
choices of analytical principles are seldom mentioned
1999 Published by Elsevier Science B.V. All rights reserved.

TRAC 2572 2-9-99

625

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

in this context. The introduction of a `criteria

approach' can also be seen as a way of introducing a
`customer approach' into method validation discussion. Quality can be dened as `fullling the customer
requirement', something that is possible only when the
method to be used is suitable for solving the analytical
problem.
The rst step in a `full validation procedure' therefore should be to identify and document `customer
requirements' and the analytical problem, what is analytically and economically possible and other specic
requirements on sampling, laboratory environment,
external environment, etc. A `validation plan' should
be written that indicates the method criteria needed
and addresses questions such as:
is the method going to be used (e.g. ofcial
when
food control and in-house process control methods

may have to full different criteria on e.g. precision

and accuracy),
what type of answer is required ^ qualitative or
quantitative, and
in what state is the analyte, i.e. whether is it bound,
free, etc.

This article describes the requirements for methods

of analysis, and in particular the extent of the validation required and the considerations to be made in their
selection. It is divided into three parts, these being:
and the acceptability of methods of
Introduction
analysis, types of validation, the requirements of

ofcial organisations and the criteria approach for

methods of analysis.
The requirements for and procedures to obtain
methods that have been fully validated through
collaborative trials.

requirements for and procedures to obtain

The
methods that have been validated `in-house'.

2. Introduction and the acceptability of

methods of analysis, types of validation
and the criteria approach for methods of
analysis
Analytical methods must be selected on the basis of
customer needs. In many cases this is possible only if
the laboratory assists the customer to select the
method, since they often lack specic analytical competence. Prerequisite for a successful selection is that
the purpose of the analysis or examination, together
with other specic needs, is well documented in an
analytical order / request or in a project or study plan.
Laboratories carrying out analytical work for
ofcial control do not always have the freedom to
select the method. In particular microbiological methods are often laid down in the legislation, e.g. both EU
and national legislation frequently specify which
reference methods are to be used in cases of dispute.
In some cases laboratories use internally ( in-house )
developed methods or signicantly modify standard
methods. Such methods must be validated at the relevant concentration ranges, before being taken into routine use.
If the method of choice is an established standard
reference method from, for example, the AOAC
INTERNATIONAL, the laboratory usually only
needs to verify that it can achieve the performance
characteristics given in the method, especially trueness and precision, and demonstrate that the method
is suitable for the intended use ( see Table 1 ). The
extent and nature of such verication work depend

Table 1
Existing validation

Laboratory requirement

Fully validated method ( has been studied in

a collaborative trial )
Fully validated method, but new matrix or
new instruments used
Well-established, but not collaboratively
studied method
Method published in the scientic literature;
characteristics given
Method published in the scientic literature;
no characteristics given
Method developed in-house

Verication that the laboratory is capable of achieving the performance characteristics of the method ( or is able to full the requirements of the analytical task )
Verication of trueness and precision; possibly also the detection limit
Verication, supplemented with limited validation ( e.g. with regard to reproducibility )
Verication, supplemented with limited validation ( e.g. with regard to repeatability
and reproducibility )
Full validation and verication
Full validation and verication

TRAC 2572 2-9-99

626

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

on the needs of the customer. If, for example, the

method can give 1% precision whilst only 5% is
needed, it is usually sufcient if the laboratory is
able to demonstrate that 5% precision is achieved.
It has to be appreciated that far from all the methods
which have been adopted and published by the international standardising organisations have been validated
in full method-performance interlaboratory ( collaborative ) trials. Further information on validation and
verication of analytical methods is given in Sections
3 and 4 of this paper.
2.1. The introduction of a new or unfamiliar
method into the laboratory
2.1.1. Responsibility for carrying out validation
and verication
When a laboratory intends to use a method with
which it is unfamiliar it is the responsibility of the
laboratory to verify that it is competent to use the
method. Usually national or international organisations, such AOAC INTERNATIONAL, International
Organisation for Standardization ( ISO ) etc., have
undertaken the interlaboratory validation of the
method in a method performance ( collaborative )
trial. The extent of laboratory internal validation and
verication depends on the context in which the
method is to be used. Some suggestions as to the extent
of validation and verication measures are given in
Table 1.
2.1.2. Competency requirements
The introduction of new analytical methods
requires that both the parties developing and validating them, and the laboratory that will subsequently
verify its ability to use them routinely are sufciently
competent, i.e. are sufciently knowledgeable, experienced and procient. When new analytical techniques based on, for example, immunological or
DNA properties are considered, the laboratory may
have to employ new staff with specic competence.
It is not unusual for new rapid test kits to be marketed with the comment that they are ``so simple that
anyone can use them''. Often this is not correct and it is
necessary for the laboratory to acquire new knowledge
or otherwise widen its competence before starting to
use the kit. Manufacturers of rapid analysis kits sometimes make minor changes both in the documentation
and in the composition of the kit; it is therefore necessary for the user to be knowledgeable and observant
but also that any instructions accompanying the kits
etc. are carefully read. Knowledge of components

included and inter-batch variation may be of critical

importance when, for example, polyclonal antibodies
are used in ELISA analyses, or Tris buffers are used in
PCR analyses.
2.1.3. Evaluation of published validation and
verication data
Validation results, i.e. the performance characteristics of analytical methods, are often given as repeatability and reproducibility data, normally as standard
deviations or relative standard deviations.
The value of the published data varies depending on
the extent of the validation or verication. In some
cases results are based on experiments carried out
using synthetic aqueous standard solutions. The
results in such cases are naturally not applicable to
the same analyte present in a complex matrix. In
other cases method studies have been appropriately
carried out in accordance with internationally established protocols ( see Section 3 ). In order to give the
user the possibility of evaluating and utilising published results it must be clearly described how the
performance characteristics were estimated or determined.
Methods published in the scientic literature are
often accompanied by insufcient information on the
nature and extent of quality assurance measures
applied during the validation of the method, such as
a description of how important analytical parameters
( temperature and pH ) were controlled. It is also
important to know whether the originating laboratory
has participated in prociency testing schemes or has
used reference materials to support the development
work, and indeed whether it had been subjected to any
third party evaluation, e.g. through formal accreditation.
The absence of performance characteristics or lack
of information on the general competence of the originating laboratory makes it difcult to assess the credibility of incomplete validation data published in
methods.
2.1.4. Using a new or unfamiliar method
Before taking new or unfamiliar methods into routine use the appropriate staff must decide if the requirements of the customer and of the analytical problem
are being met. The judgement should be based both
on published validation data and on experimental
data resulting from the laboratory's own verication work. Methods should be formally approved in
writing, and their elds of application clearly
described.

TRAC 2572 2-9-99

627

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

The following aspects should be considered:

standard operating procedure, and its avail the
ability,
information the laboratory has on the method:
what
is it based on a standard or reference method, or has

it been developed in-house?,

whether any deviation in a method as compared to
the `reference method' is fully described and the
effects of the deviation have been investigated,
whether the method has been veried, e.g. by
analysing spiked samples of relevant matrices,
whether the method has been in use in the
laboratory for a time period of a minimum of,
say, 3 months during which a number of `real'
samples of relevant types have been analysed,
whether quality control procedures are in place, e.g.
analysis of reference or control materials, or control
strains,
whether the laboratory participates in prociency
testing schemes and evaluates, on a continuous
basis, the results,
whether the measurement uncertainty has been
estimated, and
if a sensory laboratory, whether it monitors the
performance of individual sensory assessors and of
panels.

2.2. Requirements of legislation and international

standardising agents for methods of analysis

The development of methods of analysis for incorporation into international standards or into foodstuff
legislation was, until comparatively recently, not systematic. As a consequence not all the present standard
methods have been collaboratively validated; even if
collaboratively validated, the method performance
characteristics are not always published as part of
the method thus making it difcult for the analyst to
verify his or her capability to use the method in an
appropriate way.
In general, it is more common for chemical standard methods to be `fully validated' and supplied
with information on e.g. precision ( reproducibility
and repeatability ) than is the case for microbiological methods. One reason why there are few
`fully validated' microbiological methods is that
there are no dedicated guidelines on method validation
for microbiological methods; in practice the chemical
protocols are adapted by the microbiological
sector but this is not being carried out in a systematic
manner.

Most international organisations now develop

their own methods in a dened way or stipulate
conditions to which their methods should comply. In
the food sector the most important of these, i.e. the
Codex Alimentarius Commission, European Union,
the European Committee for Standardisation ( CEN )
and the AOAC INTERNATIONAL ( AOACI ) are
very similar. Although these organisations are targeted towards the food sector, they may be taken
as indicative of the direction that all sectors are moving.
The essential requirement of all these organisations
is that the method has been subjected to a collaborative
trial before it is adopted or recommended by that
organisation. There are other possibilities for the
standardisation of a method, i.e. through the results
of prociency testing schemes or through an inhouse validation, but the rst and `easiest' procedure
for a laboratory seeking a method for any particular
analyte / matrix combination is to use a method published by one of the international standardising organisations.
The requirements of the four main non-commodity
specic methods standardising organisations in the
food sector are very similar so only the EU requirements are described below.
2.2.1. Requirements of the European Union
The European Union is attempting to harmonise
sampling and analysis procedures to meet the current
demands of the national and inter-national enforcement agencies and the likely increased problems
that the open market will bring. To aid this the EU
issued a Directive on Sampling and Methods of Analysis [ 1 ]. The Directive contains a technical annex, in
which the need to carry out a collaborative trial on a
method before it can be adopted by the Community is
emphasised.
The criteria to which Community methods of analysis for foodstuffs should now conform are as stringent as those recommended by any International
Organisation following adoption of the Directive.
The requirements are given in the Annex to the Directive. They are:

1.

Methods of analysis which are to be considered

for adoption under the provisions of the Directive
shall be examined with respect to the following
criteria:

1.1. specicity
1.2. accuracy

TRAC 2572 2-9-99

628

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

1.3. precision; repeatability intra-laboratory ( within

laboratory ), reproducibility inter-laboratory
( within laboratory and between laboratories )
1.4. limit of detection
1.5. sensitivity
1.6. practicability and applicability under normal laboratory conditions
1.7. other criteria which may be selected as required.
2.

3.

4.

The precision values referred to in 1.3 shall be

obtained from a collaborative trial which has
been conducted in accordance with an internationally recognised protocol on collaborative trials ( e.g. International Organisation for Standardisation Precision of Test Methods ( ISO 5725 /
1981 ) [ 2 ]). The repeatability and reproducibility
values shall be expressed in an internationally
recognised form ( e.g. the 95% condence intervals as dened by ISO 5725 / 1981 ). The results
from the collaborative trial shall be published or
be freely available. [Note: this reference has
since been updated and only the latest reference
is now cited.]
Methods of analysis which are applicable uniformly to various groups of commodities should
be given preference over methods which apply to
individual commodities.
Methods of analysis adopted under this Directive
should be edited in the standard layout for methods of analysis recommended by the International Organisation for Standardisation.

2.2.2. Requirements of ofcial bodies ^ the valid

analytical method
Consideration of the above requirements means that
all legislative methods from the European Union must
be fully validated, i.e. have been subjected to a collaborative trial conforming to an internationally recognised protocol. In the food sector this is also followed
by the Codex Alimentarius Commission as well as
methods of analysis stemming from the main ISOs.
Indeed, these requirements are now adopted by most
other international bodies. Thus it is essential that laboratories use methods which comply to the requirements in order to be able to ensure acceptance of
their analytical methodology by their customers.
The concept of the valid analytical method is such
that the accuracy and precision ( now trueness ) of the
method are determined.

2.3. Future requirements for methods of analysis

^ criteria of methods of analysis

Notwithstanding the above there is now an increasing tendency for the laboratory to be allowed freedom
of choice of the analytical method. This is because it is
now recognised that for a laboratory to report satisfactory ( and acceptable ) analytical results it must undertake a number of measures, only one of which is to
choose and use an appropriate or prescribed method of
analysis. This is best illustrated by consideration of
activities in:
1.

2.

3.

GATT, where the agreement on the Application

of Sanitary and Phytosanitary Measures recommends mutual recognition between governments. For the Codex Alimentarius Commission,
this means that the concept of equivalence should
be adopted,
the Codex Alimentarius Commission, which has
endorsed the Codex Alimentarius Commission
the IUPAC / ISO / AOAC Harmonised Protocol
for the Prociency Testing of ( Chemical ) Analytical Laboratories [ 3 ], and the CCMAS which
is recommending for endorsement by the Codex
Alimentarius Commission the IUPAC / ISO /
AOAC Harmonised Guidelines for Internal
Quality Control in Analytical Chemistry Laboratories [ 4 ] and which, at its Twenty-rst Session, recommended laboratory quality standards
for laboratories involved in import / export certication work [ 5 ],
the EU, where as a result of the adoption of the
Additional Measures Food Control Directive
[ 6 ], food control analytical laboratories will be
required to become accredited ( to the European
equivalent of ISO Guide 25 ), participate in prociency testing schemes and to use fully validated
methods whenever such methods are available.
Such validation for methods equates to the
requirements for methods of analysis outlined
in the Codex Alimentarius Commission's Procedural Manual [ 7 ].

As a result of such activities the focus of attention is

moving towards how well the analyst ( laboratory )
carries out his analytical procedures and away from
just ensuring that he is using a prescribed analytical
method.

TRAC 2572 2-9-99

629

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

3. The requirements for and procedures

to obtain methods that have been fully
validated through collaborative trials
As seen from the above, all `ofcial' methods
of analysis are required to include precision data;
such data can only be obtained through a collaborative trial and hence the stress that is given to collaboratively tested and validated methods in the food
sector.
3.1. What is a collaborative trial?

A collaborative trial is a procedure whereby the

precision of a method of analysis may be assessed
and quantied. The precision of a method is usually
expressed in terms of repeatability and reproducibility
values. Accuracy is not the objective. Because of the
importance of collaborative trials they are described in
greater detail below using the Outline and Recommendations from the IUPAC / ISO / AOAC Harmonised
Protocols.
3.2. IUPAC / ISO / AOAC INTERNATIONAL
harmonisation protocol

Recently there has been progress towards a universal acceptance of collaboratively tested methods and
collaborative trial results and methods, no matter by
whom these trials are organised. This has been aided
by the publication of the IUPAC / ISO / AOAC
INTERNATIONAL Harmonisation Protocols on Collaborative Studies [ 8,9 ]. These Protocols were developed under the auspices of the International Union of
Pure and Applied Chemists ( IUPAC ) aided by representatives from the major organisations interested in
conducting collaborative studies. In particular, from
the food sector, the AOAC INTERNATIONAL, the
International Organization for Standardization ( ISO ),
the International Dairy Federation ( IDF ), the Collaborative International Analytical Council for Pesticides
( CIPAC ), the Nordic Analytical Committee
(NMKL ), the Codex Committee on Methods of Analysis and Sampling and the International Ofce of
Cocoa and Chocolate were involved.
The Protocols give a series of recommendations;
they address the topics listed below:
components that make up a collaborative trial
The
Participants
Sample type
Sample homogeneity

plan ^ number of materials (material^

Sample
analyte/concentration level/matrix combination)
of replicates
Number
Split
level
or double)
Combination(single
blind
Blind replicates replicates and split level
Known replicates
Independent replicate analyses
The method(s) to be tested
Pilot trial/pre-trial
The trial proper
Statistical analysis
Valid data
One-way analyses of variance
Initial estimation
Outlier treatment
Precision parameters
Final report

3.3. Other procedures for the validation of a

method of analysis

International organisations are now considering

other procedures for the validation of methods of analysis besides a complete collaborative trial. Instructions on the use of results from prociency testing
schemes have been developed by the UK Ministry of
Agriculture, Fisheries and Food [ 10 ] and there is now
a procedure for the validation of test kit methods
developed under the `MicroVal' project.
3.4. Assessment of the acceptability of the
precision characteristics of a method of analysis:
calculation of HORRAT values

There is no formal requirement in the European

Union or in Codex as to the acceptability of the precision characteristics of any particular method. However, the calculated repeatability and reproducibility
values can be compared with existing methods and a
comparison made. If these are satisfactory then the
method can used as a validated method. If there is no
method with which to compare the precision parameters then theoretical repeatability and reproducibility
values can be calculated from the Horwitz equation
[ 11 ]. This is best achieved by the use of HORRAT
values to give a measure of the acceptability of the
precision characteristics of a method.
The HORRAT value is: (RSDR derived from the
collaborative trial )/(RSDR predicted from the Horwitz
equation ).

TRAC 2572 2-9-99

630

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

Thus, HoR , the HORRAT value for reproducibility,

is the observed RSDR value divided by the RSDR
value calculated from the Horwitz equation at the concentration of interest.
3.5. Interpretation

If the HORRAT values are 2 or less, then the method

may be assumed to have satisfactory reproducibility
values. Laboratories should ensure that the methods
which they employ meet this criterion.
3.6. Calculation of the Horwitz value

The Horwitz value is derived from the Horwitz

equation, which states that for any method:
RSDR 2130:5logC
and that the value is independent of matrix / analyte.
The major values are:
Concentration ratio
1 ( 100%)
1031
1032 ( 1%)
1033
1034
1035
1036 ( ppm )
1037
1038
1039 ( ppb )

RSDR
2
2.8
4
5.6
8
11
16
23
32
45

Horwitz has derived the equation after assessing the

results from many (V3000 ) collaborative trials.
Although it represents the average RSDR values and
is an approximation of the possible precision that can
be achieved, the data points from `acceptable' collaborative trials are less than twice the predicted RSDR
values at the concentrations of interest. This idealised
smoothed curve is found to be independent of the
nature of the analyte or of the analytical technique
that was used to make the measurement. In general
the values taken from this curve are indicative of the
precision that is achievable and acceptable of an analytical method by different laboratories. Its use provides a satisfactory and simple means of assessing
method precision acceptability.
This procedure is increasingly being used by organisations to assess the acceptability of the precision
characteristics of their methods.

4. The requirements for and procedures

to obtain methods that have been
validated `in-house'
It is a requirement that in many sectors that methods
of analysis should, wherever possible, be fully validated, i.e. have been subjected to a collaborative trial.
There are, however, many situations where this is not
feasible or such methods are not available, and indeed
many sectors do not take this approach. As a result the
need for laboratories to develop and use their own inhouse methods of analysis is well recognised in the
analytical community. Until recently such in-house
method validation has been undertaken on an ad hoc
basis. It has become recognised that such validation
should be carried out on a more formal basis and a
number of organisations have developed procedures
and protocols which meet such needs. The Codex Alimentarius Commission has agreed that the topic
should be included in its formal work programme
[ 12 ]. In particular, it is recognised within the Codex
system that there are some sectors of food analysis,
e.g. the veterinary residues in food sector, where it is
very unlikely that fully validated methods are, or
likely to become, available.
Validation can be dened as the process of determining the suitability of a measurement system for
providing useful analytical data. The term interlaboratory comparison ( i.e. collaborative trial ) is often
taken to be synonymous with method validation in
the food sector. According to ISO / IEC Guide 25,
this is but one of a number of ways of validating analytical methods. The others include one or more of the
following:
using references or reference materials;
calibration
comparison
of results achieved with other methods;
systemic assessment
of the factors inuencing the
result; and
of the uncertainty of the results based
assessment
on scientic knowledge and practical experience.
As stated previously ISO, IUPAC and AOAC
INTERNATIONAL have co-operated to produce
agreed protocols or guidelines on the `Design, Conduct and Interpretation of Method Performance Studies' [ 8,9 ] on the `Prociency Testing of ( Chemical )
Analytical Laboratories' [ 3 ] on `Internal Quality
Control in Analytical Chemistry Laboratories' [ 4 ]
and on the use the use of recovery information in analytical measurement [ 13 ]. The working group that

TRAC 2572 2-9-99

631

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

produced these protocols / guidelines has been mandated by IUPAC to prepare guidelines on the inhouse validation of methods of analysis. These guidelines will outline minimum recommendations to laboratories producing analytical data on procedures that
should be employed to ensure adequate validation of
their methods before undertaking further validation
involving other peer laboratories. The Guidelines are
intended to be generally applicable across the analytical community. Appendix A gives an indication of the
topics that will be addressed in the Guidelines.
A draft of the guidelines will be discussed at an
International Symposium on the Harmonisation of
Quality Assurance Systems in Chemical Laboratory,
the proceedings from which will be published by the
UK Royal Society of Chemistry.
4.1. Protocols for the in-house validation of
analytical methods

There are a number of authoritative texts which

have been developed for the in-house validation of
analytical methods, and in particular:

accepted protocols. However, with the requirement

for laboratories to comply with dened quality standards, and because of the difculty of obtaining sufcient participants to enable collaborative trials to be
carried out in every situation ( i.e. for every combination of analyte / matrix ) there is now a requirement for
an internationally accepted set of guidelines dealing
with in-house methods to be developed. This is being
addressed by IUPAC.

Appendix

Contents of IUPAC guidelines on inhouse method validation

The contents of the Guidelines will be such that the
following will be addressed:
Introduction
Denitions and terminology used in guidelines
Concepts and general principles
The introduction of a new or unfamiliar method
into the laboratory
Requirements for individual parameters

A Protocol on the Validation of Chemical Analytical Methods developed by the Nordic Committee on Food Analysis [ 14 ],
A generic laboratory guide developed by EURACHEM produced by the UK Laboratory of the
Government Chemist with the support of the UK
Department of Trade and Industry Valid Analytical Measurement Initiative [ 15 ],
An Interlaboratory Analytical Method Validation Short Course developed by the AOAC
INTERNATIONAL [ 16 ],
A Guide to the Validation of Methods developed
by the Dutch Inspectorate for Health Protection
[ 17 ], and
A guide to Analytical Quality Assurance in Public Analyst Laboratories prepared by the UK
Association of Public Analysts [ 18 ].

Applicability
Selectivity
Calibration
Accuracy
Precision
Range
of quantication
Limit
Limit
of detection
Sensitivity
Ruggedness
Practicability

Conclusions
Recommendations
References
Appendix I: Extract from Report `Validation of
Analytical Methods for Food Control' prepared
by the FAO / IAEA

They will be considered by the IUPAC Group preparing its Guidelines.

5. Conclusions
It is now recognised that methods must be validated
before use. Ideally methods should be `fully validated', i.e. have been subjected to a collaborative
trial which itself complies with internationally

References
[ 1 ] Council Directive 85 / 591 / EEC Concerning the Introduction of Community Methods of Sampling and Analysis

TRAC 2572 2-9-99

632

trends in analytical chemistry, vol. 18, nos. 9+10, 1999

for the Monitoring of Foodstuffs Intended for Human

Consumption, O.J. L372 of 31.12.1985.
[ 2 ] Precision of Test Methods, Geneva, 1994, ISO 5725 ( previous editions were issued in 1981 and 1986 ).
[ 3 ] M. Thompson, R. Wood ( Editors ), The International
Harmonised Protocol for the Prociency Testing of
( Chemical ) Analytical Laboratories, Pure Appl. Chem.
65 ( 1993 ) 2123 ( also published in J. AOAC Int. 76
( 1993 ) 926 ).
[ 4 ] M. Thompson, R. Wood ( Editors ), Guidelines on internal
quality control in analytical chemistry laboratories, Pure
Appl. Chem. 67 ( 1995 ) 649.
[ 5 ] Report of the 21st Session of the Codex Committee on
Methods of Analysis and Sampling, FAO, Rome, 1997,
ALINORM 97 / 23A.
[ 6 ] Council Directive 93 / 99EEC on the Subject of Additional Measures Concerning the Ofcial Control of Foodstuffs, O.J. L290 of 24.11.1993.
[ 7 ] Procedural Manual of the Codex Alimentarius Commission, 10th Edition, FAO, Rome, 1997.
[ 8 ] W. Horwitz ( Editor ), Protocol for the design, conduct and
interpretation of method performance studies, Pure Appl.
Chem. 70 ( 1998 ) 855.
[ 9 ] W. Horwitz ( Editor ), Protocol for the design, conduct and
interpretation of method performance studies, Pure Appl.
Chem. 67 ( 1995 ) 331.

[ 10 ] MAFF Validated Methods for the Analysis of Foodstuffs:

VOA ^ Introduction, General Considerations and Analytical Quality Control, J. Assoc. Publ. Anal. ( in preparation ).
[ 11 ] W. Horwitz, Evaluation of analytical methods used for
regulation of foods and drugs, Anal. Chem. 54 ( 1982 )
67A.
[ 12 ] Report of the 22nd Session of the Codex Alimentarius
Commission, FAO, Rome, 1997, ALINORM 97 / 37.
[ 13 ] S. Ellison, M. Thompson, P. Willetts, R. Wood ( Editors ),
Harmonised Guidelines for the Use of Recovery Information in Analytical Measurement, Pure Appl. Chem. ( in
press ).
[ 14 ] Validation of Chemical Analytical Methods, NMKL Secretariat, Helsinki, 1996, NMKL Procedure No. 4.
[ 15 ] Method Validation ^ A Laboratory Guide, EURACHEM
Secretariat, Laboratory of the Government Chemist, Teddington, 1996.
[ 16 ] An Interlaboratory Analytical Method Validation Short
Course developed by the AOAC INTERNATIONAL,
AOAC INTERNATIONAL, Gaithersburg, MD, 1996.
[ 17 ] Validation of Methods, Inspectorate for Health Protection, Rijswijk, Report 95-001.
[ 18 ] A Protocol for Analytical Quality Assurance in Public
Analysts' Laboratories, Assoc. Publ. Anal., Shefeld,
1986.

Step-by-step implementation of a quality

system in the laboratory
Maria Joa o Benoliel

Empresa Portuguesa das Aguas Livres S.A., Laboratorio Central da EPAL, Rua do Alviela 12,
P-1170-012 Lisboa codex, Portugal
In recent years laboratories have undergone
huge transformations due to the technological development of inspection and testing
equipment; the introduction of computerised
and automated systems; keen competitiveness between companies / laboratories as a
result of demand within Europe and on the
international market; and greater consumer
awareness of the quality of the products available. Laboratory accreditation, though a voluntary process, is formal recognition by an
accreditation body of the laboratory's competence to carry out certain tests. This article
presents those aspects which should be
taken into account in the step-by-step implementation of a quality system and also makes
reference to the requirements for the opera-

tion of accredited laboratories in accordance

with European Standard EN 45001. z1999
Elsevier Science B.V. All rights reserved.
Keywords: Accreditation; Quality assurance system;
Validation

1. Introduction
Quality is a daily, on-going challenge contributing
to technological and scientic development. Quality
has also become a factor of competitiveness between
laboratories and companies. In recent years there has
been growing concern about quality-related aspects as
a result of the creation of the internal European market
accompanied by the trend toward globalisation of the

0165-9936/99/$ ^ see front matter

PII: S 0 1 6 5 - 9 9 3 6 ( 9 9 ) 0 0 1 6 8 - 5

1999 Elsevier Science B.V. All rights reserved.

TRAC 2572 2-9-99