f/EPA
United States
Environmental Protection
Agency
Office of Radiation and Indoor Air
National Air and Radiation
Environmental Laboratory
EPA402-R-09-006
June 2009
www.epa.gov/narel
Method Validation Guide for Qualifying
Methods Used by Radiological
Laboratories Participating in Incident
Response Activities
H®hW
-------
-------
EPA 402-R-09-006
www.epa.gov
June 2009
Revision 0
Method Validation Guide for
Qualifying Methods Used by Radiological
Laboratories Participating in Incident
Response Activities
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Radiation and Indoor Air
National Air and Radiation Environmental Laboratory
Montgomery, AL 36115
Recycled/Recyclable
-------
This report was prepared for the National Air and Radiation Environmental Laboratory of the Office of Radiation
and Indoor Air, United States Environmental Protection Agency. It was prepared by Environmental Management
Support, Inc., of Silver Spring, Maryland, under contracts 68-W-03-038, work assignment 35, and EP-W-07-037,
work assignments B-33 and 1-33, all managed by David Carman. Mention of trade names or specific applications
does not imply endorsement or acceptance by EPA.
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Preface
This document describes proj ect method validation guidance that a radioanalytical laboratory should
comply with in order to validate methods used to process samples submitted during a radiological
or nuclear incident, such as that caused by a terrorist attack. EPA laboratories using radioanalytical
processes consistent with the guidance provided in the Radiological Laboratory Sample Analysis
Guide for Incidents of National Significance should first validate their methods according to the
guidance provided in this document. The use of the guidance in this document, as well as in the
Radiological Laboratory Sample Analysis Guide for Incidents of National Significance, will assist
in fulfilling EPA's responsibilities as outlined in the National Response Framework Nuclear/
Radiological Incident Annex. These responsibilities include response and recovery actions to detect
and identify radioactive substances, and to coordinate federal radiological monitoring and assessment
activities. Additionally this document identifies a formalized process for the development (Section
4.0) and testing (Section 5.0) of a new method so that there is confidence that radioanalytical results
meet project-specific data requirements.
The need to ensure adequate laboratory infrastructure to support response and recovery actions
following a major radiological incident has been recognized by a number of federal agencies. The
Integrated Consortium of Laboratory Networks (ICLN), created in 2005 by 10 federal agencies1,
consists of existing laboratory networks across the Federal Government. The ICLN is designed to
provide a national infrastructure with a coordinated and operational system of laboratory networks
that provide timely, high quality, and interpretable results for early detection and effective
consequence management of acts of terrorism and other events requiring an integrated laboratory
response. It also designates responsible federal agencies (RFAs) to provide laboratory support across
response phases for chemical, biological, and radiological agents. To meet its RFA responsibilities
for environmental samples, EPA has established the Environmental Response Laboratory Network
(ERLN) to address chemical, biological, and radiological threats. For radiological agents, EPA is the
RFA for monitoring, surveillance, and remediation, and will share responsibility for overall incident
response with the U.S. Department of Energy (DOE). As part of the ERLN, EPA's Office of
Radiation and Indoor Air is leading an initiative to ensure that sufficient environmental radioanalyti-
cal capability and competency exists across a core set of laboratories to carry out EPA's designated
RFA responsibilities.
Laboratories that support EPA's incident-response mission will undergo training and should adopt
the use of the material presented in this document, with emphasis on validating methods for expected
radionuclide and matrix combinations in the event of a terrorism incident involving radioactive
materials. As soon as reasonably possible, rapid radioanalytical methods expected to be used to
process anticipated radionuclide and matrix combinations from the early to intermediate phases of
a radiological incident should be validated according to the guidance of the document. During these
early phases of an incident response, there may be insufficient time to validate methods. Therefore,
it is prudent to validate the applicable radioanalytical methods for various sample matrices as part
of the preparatory actions that are necessary to respond properly to a possible radiological incident.
1 Departments of Agriculture, Commerce, Defense, Energy, Health and Human Services, Homeland Security, Interior,
Justice, and State, and the U.S. Environmental Protection Agency.
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Laboratories developing new methods and operational protocols should review the detailed guidance
on recommended radioanalytical practices found in the Multi-Agency Radiological Laboratory
Analytical Protocols Manual (MARLAP) referenced in this document. Familiarity with Chapters
6 and 7 of MARLAP will benefit readers of this document.
This document is one in a planned series designed to present radioanalytical laboratory personnel,
Incident Commanders (and their designees), and other field response personnel with key laboratory
operational considerations and likely radioanalytical requirements, decision paths, and default data
quality and measurement quality obj ectives for samples taken after a radiological or nuclear incident,
including incidents caused by a terrorist attack. Documents currently completed or in preparation
include:
• Radiological Laboratory Sample Analysis Guide for Incidents of National Significance -
Radionuclides in Water (EPA 402-R-07-007, January 2008)
• Radiological Laboratory Sample Analysis Guide for Incidents of National Significance -
Radionuclides in Air (EPA 402-R-09-007, June 2009)
• Radiological Laboratory Sample Screening Analysis Guide for Incidents of National
Significance (EPA 402-R-09-008, June 2009)
• Method Validation Guide for Qualifying Methods Used by Radiological Laboratories
Participating in Incident Response Activities (EPA 402-R-09-006, June 2009)
• Guide for Radiological Laboratories for the Identification, Preparation, and Implementation of
Core Operations for Radiological Incident Response (in preparation)
• Guide for Radiological Laboratories for the Control of Radioactive Contamination and
Radiation (in preparation)
• Radiological Laboratory Sample Analysis Guide for Incidents of National Significance -
Radionuclides in Soil (in preparation)
Comments on this document, or suggestions for future editions, should be addressed to:
Dr. John Griggs
U.S. Environmental Protection Agency
Office of Radiation and Indoor Air
National Air and Radiation Environmental Laboratory
540 South Morris Avenue
Montgomery, AL 36115-2601
(334) 270-3450
Griggs.John@epa.gov
11
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Acknowledgments
This manual was developed by the National Air and Radiation Environmental Laboratory (NAREL)
of EPA's Office of Radiation and Indoor Air (ORIA).
Dr. John Griggs served as project lead for this document. Several individuals provided valuable
support and input to this document throughout its development. Special acknowledgment and
appreciation are extended to Dr. Keith McCroan, ORIA/NAREL; Mr. Daniel Mackney for his
support of instrumental analysis, ORIA/NAREL; Dr. Lowell Ralston and Mr. Edward Tupin, CHP,
both of ORIA/Radiation Protection Division; Ms. Schatzi Fitz-James, Office of Emergency
Management, Homeland Security Laboratory Response Center; and Mr. David Garman,
ORIA/NAREL. We also wish to acknowledge the peer reviews conducted by Carolyn Wong, David
Burns, and Jack Bennett, whose thoughtful comments contributed greatly to the understanding and
quality of the report. Numerous other individuals both inside and outside of EPA provided peer
review of this document, and their suggestions contributed greatly to the quality and consistency of
the final document. Technical support was provided by Dr. N. Jay Bassin, Dr. Carl V. Gogolak, Dr.
Robert Litman, Dr. David McCurdy, Mr. Robert Shannon, and Dr. Anna Berne of Environmental
Management Support, Inc.
in
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Contents
Preface i
Acronyms, Abbreviations, Units, and Symbols vii
Radiometric and General Unit Conversions ix
1.0 Introduction 1
2.0 Method Validation Description 1
3.0 Method Description 2
4.0 Method Performance Characteristics 3
4.1 Method Uncertainty 3
4.2 Detection Capability 5
4.3 Bias and Trueness 5
4.4 Analyte Concentration Range 6
4.4.1 Derived Radionuclide Concentrations Corresponding to Established Action
Levels 6
4.4.2 Default Analytical Action Levels 7
4.5 Method Specificity 7
4.6 Method Ruggedness 8
5.0 Incident Response Method Validation Guidance, Tests, and Requirements 9
5.1 Method Specificity 9
5.2 Analyte Concentration Range 10
5.3 Matrix Considerations 12
5.4 Method Validation Levels for Testing the Required Method Uncertainty 13
5.4.1. Method Validation Requirements Based on MARLAP Concepts 13
5.4.2 Required Method Uncertainty Acceptance Criteria 15
5.4.2.1 Level B Method Validation: Same, Similar, or Slightly Different Matrix . . 16
5.4.2.2 Level C Method Validation: New Application of an Existing Method to a
Different Matrix 17
5.4.2.3 Level D Method Validation: Adapted or Newly Developed Methods, Including
Rapid Methods 17
5.4.2.4 Level E Method Validation: Adapted or Newly Developed Methods, Including
Rapid Methods, Using Method Validation Reference Materials 18
5.5 Verification of Required Detection Limit (MDC) Specification 18
5.5.1 Calculation of the Critical Net Concentration 20
5.5.2 Testing for the Required MDC 21
5.6 Method Bias Tests 22
5.6.1 Absolute Bias Testing 23
5.6.2 Relative Bias Testing 24
5.6.2.1 Test Level Samples with Same Known Value 24
iv
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
5.6.2.2 Test Level Samples with Slightly Different Known Values 25
6.0 Method Validation Documentation 25
7.0 References 26
Appendix A: Tables Summarizing the Derived Radionuclide Concentrations and Required Method
Uncertainties Corresponding to PAGs or Risks for the Water, Air, and Soil Matrices 27
Appendix B: Examples of the Method Validation Process for Required Method Uncertainty
Specifications 34
Appendix C: Example of the Method Validation Process for Verification of the Required MDC
Specification 37
Appendix D: Example of the Effect of Bias on the Probability of Failing the Method Validation
Acceptance Criteria for Required Method Uncertainty 39
Detecting Bias 40
Appendix E: An Alternative Method Validation Criterion 42
Introduction 42
Definitions 42
Alternative Method Validation Criterion 44
The Holst-Thyregod Test for Mean Squared Error 46
Example 46
Theoretical Comparison of Statistical Power 48
Reference 48
Appendix F: Glossary 51
Figures
Figure 1 - Method Validation Process for the Required Method Uncertainty MQO 19
Figure 2 - Validation Process for Verifying the Required MDC MQO 20
Figure Dl - Probability of a validation sample failing at concentration 100 pCi/L with and without
bias at various values of the method standard uncertainty 39
Figure D2 - Level D validation (21 samples) failing at test level as a function of relative method bias
for relative method uncertainties of 5%, 7.5%, 10%, and 12.5% 40
Figure Ela 49
Figure Elb 49
Figure Elc 50
Figure Eld 50
Tables
Table 1 - Default Analytical Action Levels for General Matrix Categories 11
Table 2 - Method Validation Test Concentrations 12
v
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Table 3 - Method Validation Requirements and Applicable to Required Method Uncertainty . 15
Table 4 - Required Method Uncertainty (WMR and #v) Values for Default AAL Test Levels . . 16
Table 5 - Method Validation Requirements Applicable to Required MDC Verification 22
Table Al - Alpha-Emitting Radionuclide Concentrations and Required Method Uncertainties in
Water Corresponding to 500- and 100-mrem AAL Derived Water Concentrations 27
Table A2 - Beta/Gamma-Emitting Radionuclide Concentrations in Water and Required Method
Uncertainties Corresponding to 500- and 100-mrem AAL Derived Water Concentrations . 28
Table A3 - Alpha-Emitting Radionuclide Concentrations in Air and Required Method Uncertainties
Corresponding to 2-rem and 500-mrem AAL Derived Air Concentrations 29
Table A4 - Beta/Gamma-Emitting Radionuclide Concentrations in Air and Required Method
Uncertainties Corresponding to 2-rem and 500-mrem AAL Derived Air Concentrations . . 30
Table A5 - Alpha-Emitting Radionuclide Concentrations in Air and Required Method Uncertainties
Corresponding to AAL Derived Air Concentrations 31
Table A6 - Beta/Gamma-Emitting Radionuclide Concentrations in Air and Required Method
Uncertainties Corresponding to AAL-Derived Air Concentrations (DACs) 32
Table A7 - Alpha and Beta/Gamma-Emitting Radionuclide Concentrations in Soil and Required
Method Uncertainties Corresponding to Derived Soil Concentrations 33
Table Bl - Required Method Uncertainty for Am-241 in Potable Water 35
Table B2 - Required Method Uncertainty for Am-241 in Street Runoff Water 36
Table Cl - Results of Blank Sample Analyses 37
Table C2 - Results of MDC Test Sample Analyses; Test Concentration = 2.0 pCi/L 38
Table El - Method Validation Measurement Results 47
Table E2 - Acceptance Limits, MARLAP Test 47
Table E3 - Method Validation Results, Alternative Test (fFTest) 48
VI
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Acronyms, Abbreviations, Units, and Symbols
(Excluding chemical symbols and formulas)
a alpha particle
a probability of a Type I decision error
AAL analytical action level
AL action level
APS analytical protocol specification
P beta particle
ft probability of a Type n decision error
Bq becquerel (1 dps)
CLNC critical net concentration level
CSU combined standard uncertainty
d day
DAC derived air concentration
DP decay product(s)
dpm disintegration per minute
dps disintegration per second
DQO data quality objective
DRC derived radionuclide concentration
DWC derived water concentration
EPA [United States] Environmental Protection Agency
y gamma ray
g gram
Gy gray [unit of absorbed radiation dose in materials; 1 gray =100 rad]
h hour
1C Incident Commander [or designee]
ISO International Organization for Standardization
keV thousand electron volts
L liter
m meter
MARLAP .... Multi-Agency Radiological Laboratory Analytical Protocols Manual
MARSSEVI . . . Multi-Agency Radiation Survey and Site Investigation Manual
MDB minimum detectable bias
MDC minimum detectable concentration
MeV million electron volts
min minute
mL milliliter(10"3L)
MQO measurement quality objective
mrem millirem (10~3 rem)
MSE mean squared error
MV method validation
MVRM method validation reference material
PAG protective action guide
pCi picocurie (10~12 Ci)
PE performance evaluation
vii
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
PT proficiency test/testing
QC quality control
rad unit of absorbed radiation dose in materials; 100 rad = gray
RDD radiological dispersal device (i.e., "dirty bomb")
rem roentgen equivalent man (traditional units; 1 rem = 0.01 Sv)
RSD relative standard deviation
s second
sBlanks standard deviation of blank sample net results
SI International System of Units
Sv sievert (1 sievert = 100 rem)
MMR required method uncertainty
#V relative required method uncertainty
y year
Vlll
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Radiometric and General Unit Conversions
To Convert
years (y)
disintegrations
per second (dps)
Bq
Bq/kg
Bq/m3
Bq/m3
microcuries per
milliliter
(nCi/mL)
disintegrations
per minute (dpm)
cubic feet (ft3)
gallons (gal)
gray (Gy)
roentgen equiva-
lent man (rem)
To
seconds (s)
minutes (min)
hours (h)
days (d)
becquerels (Bq)
picocuries (pCi)
pCi/g
pCi/L
Bq/L
pCi/L
uCi
pCi
cubic meters
(m3)
liters (L)
rad
si evert (Sv)
Multiply by
3.16x 107
5.26 x 105
8.77 x 103
3.65 x 102
1
27.0
2.70 x 1Q-2
2.70 x 1Q-2
io-3
109
4.50 x 1Q-7
4.50 x 1Q-1
2.83xlO~2
3.78
IO2
io-2
To Convert
s
min
h
d
Bq
pCi
pCi/g
pCi/L
Bq/L
pCi/L
pCi
cubic meters
(m3)
L
rad
Sv
To
y
dps
Bq
Bq/kg
Bq/m3
Bq/m3
uCi/mL
dpm
cubic feet
(ft3)
gal
Gy
rem
Multiply by
3.17 x 1Q-8
1.90 x 1Q-6
1.14 x 1Q-4
2.74 x 1Q-3
1
3.70 x 1Q-2
37.0
37.0
IO3
io-9
2.22
35.3
0.264
io-2
IO2
NOTE: Traditional units are used throughout this document instead of International System of Units
(SI) units. Protective Action Guides (PAGs) and their derived concentrations appear in official
documents in the traditional units and are in common usage. Conversion to SI units will be aided by
the unit conversions in this table. Conversions are exact to three significant figures, consistent with
their intended application.
IX
-------
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
1.0 Introduction
The United States Environmental Protection Agency (EPA) is responsible for assessing the extent
of environmental contamination and human health consequences in the event of a radiological
incident such as a terrorist incident involving radioactive materials. Although EPA will be mainly
involved in the intermediate and recovery phases of an incident response, there also may be
involvement in some activities in the early phase. For a terrorist event such as a radiological
dispersion device, the radionuclide(s) and the types and number of sample matrices that may be
collected and analyzed can vary dramatically depending on the type of device used and radioactive
material incorporated. The radioanalytical laboratories used to process the samples must not only be
capable of identifying and quantifying the radionuclide(s) in various matrices, but they must also
have the capacity to process a large number of samples in a short time (thousands of samples per
week). Sufficient laboratory capacity is a balance of adequate facility processing areas and nuclear
instrumentation, validated radioanalytical methods available, and trained staff.
In order to make proper assessments and decisions in the event of a radiological incident, EPA will
utilize only qualified radioanalytical laboratories that have the capability, capacity and quality needed
to process samples taken from affected areas. Analytical protocol specifications (APSs), including
measurement quality obj ectives (MQOs), will be preestablished to define the expected quality of the
data for incident situations. The objective of this document is to establish systematic and objective
methodologies and acceptance criteria for validating analytical methods, based on the stated quality
requirements of a specific incident-response project, such as recovery from a radiological dispersal
device. Laboratories developing new methods and operational protocols should review the detailed
guidance on recommended radioanalytical practices found in current editions of MARLAP and
MARS SIM.
Several radiological sample analysis guides for incident response have been developed that provide
information on the expected radionuclides of concern and MQOs to make decisions relative to
sample processing priorities for the water, air parti culate filter, and soil/solid matrices. As part of the
laboratory qualifying process, laboratories must demonstrate their ability to meet the APSs and
MQOs for the methods used to analyze each radionuclide and sample-matrix combination. EPA will
require an initial project method validation and a subsequent participation in a performance evalua-
tion (PE) program as a means to demonstrate that the methods used by a laboratory are capable of
meeting the MQOs for incident response applications. For incident-response applications, project
method validation will be required and applied to methods currently being used by the laboratories,
including EPA Safe Drinking Water Act required methods, as well as to newly developed methods
and methods that have been modified for incident response. Project method validation and
participation in a PE program will be required for gross alpha and beta screening methods as well.
In this document, the term "project method validation" is synonymous with "incident response
method validation."
2.0 Method Validation Description
The Multi-Agency Radiological Laboratory Analytical Protocols Manual (MARLAP) Chapter 6
discusses two distinct applications of method validation: general method validation and project
1
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
method validation. General method validation is the process of demonstrating that a method is
suitable for its general intended use, such as routine radioanalytical processing of samples for the
determination of environmental levels of a given radionuclide. For general method validation, the
methods would address internal measurement quality objectives, and typical sample matrix
constituents and nominally interfering concentrations of expected chemical and radionuclide
interferences. EPA has developed a draft general method validation process document (EPA 2006,
Validation and Peer Review of U.S. EPA Radiochemical Methods of Analysis) that covers the
method validation parameters for radioanalytical methods. That document provides guidance to
satisfy EPA requirements for general method validation for measurement uncertainty, method bias
and trueness, precision, detection capability, analyte concentration range, specificity and ruggedness.
In contrast, this document provides guidance on proj ect method validation applicable to methods for
processing samples during a response to a radiological incident, including radiological incidents of
national significance. Project method validation demonstrates that a method is capable of meeting
project-specific MQOs (in other words, a required method uncertainty at a specific radionuclide
concentration). The method selected for a project needs to address specific sample matrix
characteristics, chemical and radionuclide interferences, special sample preparation requirements,
sample-processing turnaround times, and MQOs defined in an analytical protocol specification
(APS). This document addresses the method validation expectations for an incident response for the
MQOs of the required method uncertainty and the required minimum detectable concentration
(MDC). The method validation procedures for the method uncertainty MQO follow the guidance
provided in MARLAP Chapter 6. As discussed in MARLAP, the principal MQO is the required
method uncertainty at an action level. Although the MDC MQO normally would not be specified as
an MQO for incidence response applications, this document provides method validation guidance
for a "required MDC" MQO.
Even though a laboratory has a method that has undergone general method validation, use of the
method for the incident response application will require project method validation. The degree of
effort and required level of project method validation will depend on the degree of method
development or use, and the MQOs of the project, as included in the APSs.
Proper planning is critical for successful method validation because many method validation
parameters must be considered, evaluated and documented. Method development and method
validation generally are not separate processes. The types of experiments conducted during method
development and the types of tests performed during method validation have many similarities.
3.0 Method Description
The components of a method or measurement process requiring validation should be clearly
described. Generally, a laboratory method includes all physical, chemical and radiometric processes
conducted at a laboratory in order to provide an analytical result. The processes for radiochemical
methods may include sample preparation or dissolution, chemical separations, preparation of sample
test sources, nuclear counting, analytical calculations, data review and qualification, and data
reporting (MARLAP Chapter 6). Method validation efforts should evaluate all process components
combined. Some radiochemical methods may also include procedures for sampling (e.g., methods
for radon in air analysis or for volatile radioactive organic compounds in soils and other solid
matrices), in which case the sampling procedures should be included in the validation tests. The
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
measurement process components validated, and the combination of procedures comprising a
method, must be clearly and completely stated.
The purpose of a method (i.e., measurement objectives) and the intended use of the data must be
clearly defined. In addition, method scope and applicability must be well defined and clearly
described and consistent with the documented performance of the method. These measures will help
minimize misapplication by the users. Method scope and applicability include the following:
• The measurement process components validated (e.g., sample preparation, dissolution, chemical
isolation, precipitation, final product for counting, radiation measurement process, etc.)
• The nature (chemical-physical form, type of radiation and quantity measured) of the radionuc-
lides and matrices (chemical and physical form) studied
• The range of analyte concentration levels for which the method is claimed to be suitable
• A description of any known limitations and any assumptions upon which a method is based (e.g.,
radiological and non-radiological interferences, minimum sample size, etc.)
• A description of how the method and analytical parameters chosen meet the measurement quality
objectives for the intended application, when applicable
• Aliquant sample size for processing
4.0 Method Performance Characteristics
The performance characteristics of a radiochemical method that may be evaluated in method
validation include:
• Method uncertainty at a specific radionuclide concentration (action level)
• Detection capability (minimum detectable concentration)
• Bias/trueness
• Analyte concentration range
• Method specificity
• Method ruggedness
A brief discussion of each of these performance characteristics will be covered in the following
sections. For more detailed information on a characteristic, the reader is referred to MARLAP
(Chapters 3 and 6); EPA (2006) Validation and Peer Review of U.S. EPA Radiochemical Methods
of Analysis; EURACHEM Guide (1998) The Fitness for Purpose of Analytical Methods, A
Laboratory Guide to Method Validation and Related Topics; ISO 17025; and ANSIN42.23.
4.1 Method Uncertainty
MARLAP defines method uncertainty as follows:
Method uncertainty refers to the predicted uncertainty of the result that would be measured if the
method were applied to a hypothetical laboratory sample with a specified analyte concentration.
Although individual measurement uncertainties will vary from one measured result to another, the
required method uncertainty is a target value for the individual measurement uncertainties, and is
an estimate of uncertainty (of measurement) before the sample is actually measured.
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Method uncertainty can be thought of as an estimate of the expected analytical standard deviation
at a specified radionuclide concentration. For certain proj ects, including incident response, a required
method uncertainty should be specified. An example of a required method uncertainty specification
would be "...at a 137Cs soil concentration of 10 pCi/g, the required method uncertainty is 1 pCi/g."
In many applications, including incident response laboratory analyses, the specified radionuclide
concentration is referred to as the analytical action level (AAL) and may be based on either incident-
specific, risk-based or regulatory mandated value, such as a protective action guide (PAG) as
presented in the Radiological Laboratory Sample Analysis Guides for Incidents of National
Significance. Radioanalytical results from an incident response are compared to action level
concentrations, and thus it is very important to have results that are of sufficient quality to support
decisions to be made. Specifying a required method uncertainty at the AAL ensures the data quality
needed to make decisions.
To be consistent with MARLAP, certain nomenclature for the required method uncertainty is used
for incident response applications. The notation "WMR" is specified for the absolute required method
uncertainty at or below the action level and has units of activity or activity concentration that match
the AAL value. Above the action level, a relative required method uncertainty #v, defined as the
MMR/AAL, is specified (
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
AALs and required method uncertainties. If so, the 1C should verify whether the laboratory can meet
the new method uncertainty requirements for the updated AALs.
4.2 Detection Capability
In some cases, the detection capability of a method, rather than the required method uncertainty, is
the important MQO of a proj ect. Detection capability for this guide uses the concept of the minimum
detectable concentration (MDC) or minimum detectable value. MARLAP defines the minimum
detectable value of the analyte concentration in a sample as:
An estimate of the smallest true value of the measurand that ensures a specified high probability,
1 - /3, of detection.3
For radioanalytical processes, the probability of detection (1 - /?) of 0.95 is commonly used. The
definition of the minimum detectable value presupposes that an appropriate detection criterion has
been specified, i.e., "critical net concentration" for this document. This approach assumes that the
measured radionuclide net concentration in a sample will be above the critical net concentration 95%
of the time if the true concentration is equal to the MDC.
MARLAP (Chapter 20) provides a detailed discussion on how to calculate the critical net
concentration and MDC using a number of equations for various applications. The equations
provided in MARLAP calculate estimates of these method detection parameters for a given method
based on either a measured signal response of a single blank sample or from a population of sample
blanks that have been processed by the method under evaluation. For those applications when a
required MDC for a method has been specified as an MQO, the detection capability of the method
should be evaluated during method validation.
4.3 Bias and Trueness
Bias refers to the overall magnitude of systematic errors associated with the use of an analytical
method. The presence of systematic errors can be determined only by comparison of the average of
many results with a reliable, accepted reference value. Method bias may be estimated by measuring
materials whose composition is reasonably well known, such as reference materials, by comparing
results to those from at least one alternate method or procedure, or by analyzing spiked materials.
ISO (1993a) defines bias as:
"[T]he mean value that would result from an infinite number of measurements of the same
measurand carried out under repeatability conditions minus a true value of the measurand."
According to MARLAP (Chapter 6), bias typically cannot be accurately determined from a single
result or a few results because of the uncertainty in the measurement process to determine the
measurand. Bias is normally expressed as the absolute or relative deviation of the average of a group
of samples from the "true" or "known" value. Since it is a calculated estimate, a bias should be
3 Here, ft means the probability of a Type II decision error.
5
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
reported with a combined standard uncertainty and include the number of data points used to
calculate the bias.
It is assumed that the mean response of the method is essentially a linear function of analyte
concentration over the useful range of the method. As defined in MARLAP, "this function can be
characterized by itsj/-intercept, which reflects the mean response at zero concentration, and its slope,
which reflects the ratio of the change in the mean response to a change in sample analyte
concentration." The "absolute bias" of a method can be thought of as the difference between the
average concentration of the radionuclide at the y-intercept and the true concentration of zero.
The 1C will specify a method bias limit as an APS when method bias is considered an important
method performance characteristic for the method or a quality parameter for the proj ect. Method bias
must be evaluated during method development, general and proj ect method validation processes, and
subsequently, the processing of batch quality control (QC) samples processed with the incident
response samples (MARLAP Chapter 7).
The method uncertainty acceptance criteria provided in MARLAP (Chapter 6), as well as for proj ect
method validation in Section 5.4, assume that laboratories would not use a method that has a
significant bias. When a method has excessive bias, the method validation test results for the
required method uncertainty will be unacceptable. Appendices D and E provide information on bias
evaluation methods as related to the method validation acceptance criteria.
4.4 Analyte Concentration Range
The analyte concentration range of a method is a method performance characteristic that defines the
span of radionuclide activity levels, as contained in a sample matrix, for which method performance
has been tested and data quality deemed acceptable for their intended use. However, not all sample
matrices encountered during an incident response will have preestablished analytical action levels
with corresponding required method uncertainty values or required MDCs. Therefore, incident
response method validation must be sufficiently flexible to address not only those typical sample
matrices (liquids, air filters, swipes, and soil/solids) for which there are action levels, but also those
matrices for which there are no specified action levels. The subsequent subsections discuss the
analytical concentration range options for method validation for both situations. For both options,
the method is to be tested at a low, mid and upper validation test concentration/activity except when
noted.
4.4.1 Derived Radionuclide Concentrations Corresponding to Established Action Levels
MARLAP (Chapter 6) recommends that a method be validated at the expected action level for a
radionuclide and matrix combination. Therefore, an analyte concentration range should include either
an established regulatory limit or a defined action level, typically near the midpoint of the
radionuclide activity (concentration) range for a project. For a radiological incident response
application, the established AAL would normally be a derived radionuclide concentration
corresponding to a PAG or a risk-based dose as designated by an agency representative. There may
be four or five action levels for the various matrices contaminated, with the range of concentrations
as great as four orders of magnitude. Derived radionuclide concentrations for the various established
AALs have been generated for water and air-filter matrices (AALs for soils/sediments and building
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
materials are being developed), and can be found in the Radiological Laboratory Sample Analysis
Guide for Incidents of National Significance series. Summary tables of the established AALs for
these three matrices can be found in Appendix A. AALs vary according to the matrix, phase of the
incident and applied PAG.
At most laboratories, samples that have been screened and found to contain very high radionuclide
activities probably will be subdivided (when possible) prior to specific radioanalytical processing.
Also, samples requiring multiple radionuclide analyses should be subdivided. Both situations will
result in lower-activity subsamples (aliquants) for processing. For example, when a sample that has
a very high radionuclide concentration or activity is received by a laboratory, the sample likely will
be subdivided (possibly into five parts) so that a different radioanalytical pathway for each
radionuclide may be performed in parallel. For aqueous and soil samples, the radionuclide
concentration of the radionuclides in the aliquants would be the same as the original aqueous and
soil concentrations, but the aliquant activity available for processing will be reduced proportionally
from the original sample size. For air-filters, the total activity on the filter matrix represents the
activity in a volume of air collected. For swipe samples, the activity on the sample represents the
activity removed from a surface area swiped. Air filter and swipe samples may be digested prior to
radiochemical processing and the digestate volume generated represents the total activity on the
original sample for the air volume collected or surface area swiped. Aliquanting these digestates to
obtain a lower subsample activity or for multiple analyses is also a common practice. Thus, it is
important to know the exact fraction of the original sample taken so that the analysts know that a
sufficient sub-sample quantity has been processed to ensure that the MQOs have been met.
When developing incident response methods for high activity samples, it is important to note that
the analytical concentration range and detection capability specifications will be significantly higher
than what is usually found in normal procedures for environmental monitoring sample processing.
This difference in concentration range should be emphasized in the procedure's scope.
4.4.2 Default Analytical Action Levels
Established AALs based on PAG-derived radionuclide concentrations may not be available for all
matrices encountered in an incident response, such as concrete or asphalt. In the absence of
established PAG action levels, default AALs may be used for the validation test concentration or
activity levels. Section 5.4 provides guidance on selecting default AALs applicable to method
validation for the three general matrix categories of liquids, air sampling media/swipes and solids.
The default AALs approximate the expected derived radionuclide activity level for a sample volume
or mass for a 100-mrem or 10"4 risk-based AAL. These AAL levels were chosen because they can
be conveniently scaled to other possible project-specific AALs for the various matrices. For example,
if a specific project had an AAL at 20 mrem (one-fifth of a 100 mrem AAL), the table values for the
AALs can be scaled down simply by dividing the listed values by five.
4.5 Method Specificity
MARLAP defines "method specificity" as "the ability of the method to measure the analyte of
concern in the presence of interferences." EURACFEM (1998) defines selectivity or specificity as
"the ability of a method to determine accurately and specifically the analyte of interest in the
presence of other components in a sample matrix under the stated conditions of the test."
7
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
By extension, this guide defines method specificity as:
The ability to correctly identify and quantify the radionuclide(s) of interest in the presence of other
interferences in a sample under stated conditions of the test.
Method specificity should be evaluated during method development and the general and project
method validation processes for the applicable matrices, radionuclide(s) of interest and known
interfering radionuclides. Method specificity may be evaluated during method validation by
analyzing:
• Matrix samples that have been characterized in terms of radionuclide and chemical constituent
content;
• Appropriate matrix blanks; and
• Matrix blanks spiked with interferences.
Each specific sample matrix should be tested for method specificity, e.g., concrete, asphalt, soil, etc.
Matrix samples and blanks should be chosen to be as representative of the target matrix as is
practical. When possible, matrix blanks should contain the chemical species and potential interfering
radionuclides, other than the radionuclide(s) of interest, at concentrations that are reasonably
expected to be present in an actual sample. Each of the three options to determine method specificity
may provide insight into the relative degree of expected quantitative effect that the interferences will
have on the identification and quantification of the radionuclide(s) of interest at different
concentrations.
Method specificity is typically expressed qualitatively and quantitatively. A radiochemical method
specificity statement would include descriptions of parameters, such as:
• Expected radionuclide and chemical interferences
• Effects of the interfering substances on the measurement process
• Measurement information that substantiates the identity of the analyte (e.g., half-life, or
decay emission and energy)
• Effects of oxidation or molecular state of the target or interfering radionuclides
• Chemical processes that can remove interfering materials (e.g., ion exchange, solvent
extraction)
• Summary of results from analysis of standards, reference materials and matrix blanks
4.6 Method Ruggedness
MARLAP defines "method ruggedness" as "the relative stability of method performance for small
variations in method parameter values." EURACHEM (1998) discusses the concept of method
ruggedness and robustness interchangeably. Ruggedness is a measure of how well a method's
performance stands up to less than perfect implementation. In any method there are certain steps
which if not carried out sufficiently, exactly or carefully may have a significant effect on method
performance and the reliability of the results. Typically, these critical steps are identified during the
method development process, and annotations are made in the method description that provide
limiting conditions and an allowable range of application. It is advantageous to identify the variables
in the method that have the most significant effect on the analytical results so that they are closely
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
controlled. Ruggedness or robustness tests have been developed which involve experimental designs
for examining method performance when minor changes are made in operating steps or in some
cases environmental conditions (EPA 2006, Validation and Peer Review of the U.S. EPA
Radiochemical Methods of Analysis). The tests involve making deliberate variations to the method,
and investigating the subsequent effect on performance.
An example of method ruggedness is the adjustment of pH during the separation of strontium from
calcium in the analysis of milk for 90Sr. The pH of the milk is buffered at 5.4 and disodium EDTA
is added prior to passing the solution through a cation exchange column. The calcium will effectively
complex with EDTA at this pH, forming an anion, while the strontium remains a cation. A pH lower
than about 5.2 will not provide enough EDTA anion to effectively complex calcium, and a pH
greater than about 5.5 will begin to effectively complex strontium. Thus for this analysis method,
ruggedness deals with pH control in the range of 5.2 to 5.5.
Method ruggedness is typically evaluated during method development and prior to method
validation. Therefore, no specific tests for ruggedness will be included in this document for project
method validation of the radioanalytical methods used for incident response.
5.0 Incident Response Method Validation Guidance, Tests, and Requirements
This section provides guidance, specific tests and minimum requirements for project method
validation for methods used to process samples from a radiological incident. This section addresses
the following selected method performance characteristics:
• Method specificity
• Analyte concentration range
• Method validation levels for testing the required method uncertainty
• Verification of required detection limit specification
Discussion of matrix considerations and method bias tests are also included in this section. Before
initiating the method validation process, a validation plan should be prepared that incorporates the
various guidance and requirements specified in this section and Section 6, Method Validation
Documentation.
5.1 Method Specificity
Method specificity is evaluated during general method validation for normal routine applications,
e.g., environmental surveillance programs. During general method validation, the method should be
evaluated for the applicable matrices and radionuclide(s) of interest and known interfering chemical
constituents and radionuclides over a typical expected range. For some incident response
applications, method validation testing for method specificity may be more focused than general
method validation. Incident response scenarios may involve one or many radionuclides and a
multitude of matrices. To ensure method specificity for the incident response application have been
met, the proficiency testing (PT) samples used for incident response method validation should
contain the known or expected concentration levels of the matrix chemical species and potential
interfering radionuclides. Adequate method specificity during project method validation should be
evaluated by analyzing:
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
• Matrix PT samples that have been characterized in terms of expected radionuclide and chemical
constituent content;
• Appropriate matrix blanks containing the applicable radionuclide and chemical interferences;
and
• Matrix blanks spiked with interferences.
During method development, decontamination factors4 should be evaluated for the more commonly
expected radionuclide interferences so that the final method can improve method performance and
adequately address radionuclide interferences. Also, the concentration of the interfering radionuc-
lides should be added during method development at their 100 mrem AAL-derived radionuclide
concentrations. Matrix blank results having no absolute bias would indicate adequate method
specificity. Excessive absolute or relative bias, erroneous chemical or radiotracer yields, or possibly
excessive method uncertainty may be indications of inadequate method specificity.
5.2 Analyte Concentration Range
The radionuclide concentration range applicable to method validation for radiological incident
response should extend from a lower bound (-0.5 AAL) to an upper bound (3 AAL) that are both
a multiple of an incident response action level. For radiological incident response applications, the
analyte concentration range for the method validation process and validation test levels should be
established based on the established PAG or risk-based derived radionuclide concentrations as
designated by a representative of the responsible government agency. If the laboratory has not been
provided with action levels by the 1C, default values listed in Table 1 may be used. Also, derived
radionuclide concentrations for the various AALs for water, air filter and soil matrices can be found
in the Radiological Laboratory Sample Analysis Guide for Incidents of National Significance.
Summary tables of the AAL-derived radionuclide concentrations for the water and air filter matrices
can be found in Appendix A5 of this document. These tables list the expected AALs for various
media mainly for the late intermediate and recovery phases, but also for the early phase. The AAL-
derived radionuclide concentrations will vary according to the matrix, phase of the incident and
applied PAG. For air filters, an activity per sample corresponding to an AAL concentration for an
assumed air volume sampled should be used.
The validation test concentration/activity values should be adjusted to reflect the typical sample
aliquant size that would be analyzed. In some cases, the original sample may be aliquanted directly,
but in other cases the sample must be completely digested before sample aliquanting. When the
radionuclide(s) identity is known, the number of aliquants may be small, but when the identity is not
known, the number of aliquants may be three or more depending on the decay particle emission type.
4 The term "decontamination factor" is defined as the amount of interferent in the sample before chemical separation
divided by the measured amount in the sample after chemical separation.
5 The 1C may develop and require other AALs and required method uncertainties. If so, the 1C should verify whether the
laboratory can meet the new method uncertainty requirements for the updated AALs. Calculating the test levels for
method validation should be consistent with Table 2.
10
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
In the absence of established PAG AALs, default AALs may be used. Default AALs (activity per
sample aliquant) for three general matrix categories and radionuclide emission type are provided in
Table 1. Default AALs can be used for similar matrix categories, as discussed in Section 5.3 (for all
solutions, use water; for swipes, use air-filter materials; and for pulverized concrete, use soil). It
should be noted that these default AALs and associated required method uncertainty values (Table
4) for the stated general matrix categories do not have a dosimetric basis but may be considered
adequate for method validation purposes.
TABLE 1 - Default Analytical Action Levels for General Matrix Categories
Matrix
Category
Liquids
Air Sampling
Media/Swipe
Solids - soil,
etc.
Size Assumptions for Values
5-mL Screen
100-mL Nuclide-Specific
68 m3 Screen
68 m3 (4 aliquants [2])
Nuclide-Specific
2g
100 g
500s
Default Test Level Activity in Each Sample
Aliquant (Total pCi) [1]
Alpha
(241Am)
2.0
40
22
5.5
Pure Beta
(90Sr)
12
240
1,900
480
Gamma
(60Co)
33
660
8,400
2,100
TBD [3]
[1] Test-level activity corresponds closely to 100-mremdose-derivedconcentrationvaluesforwaterand 10 4 risk-based
DAC for air (Appendix A). The table values were calculated for the noted radionuclides. To calculate air sampling
default AALs for lO^5 risk-based applications, the 10~4 risk-based values in the table can be scaled down by a factor
of 100. Table values for the solids and soil are pending.
[2] Test-level activity assumes that the air filter has been split into four aliquants after sample digestion.
[3 ] TBD: To be determined pending development of'Radiological Laboratory Sample Analysis Guide for Incidents of
National Significance-Radionuclides in Soils and Solids.
For each matrix category, default AALs are provided for both screening and specific radionuclide
methods requiring validation. The test level values stated in Table 1 were calculated using the DRC
for the 100-mrem AAL values for the gross alpha, beta, and gamma screening levels and for 241Am,
90Sr, and 60Co for the specific alpha and beta/gamma radionuclide categories. The default AALs have
been adjusted to reflect the typical sample aliquant size (column 2 of Table 1) that would be analyzed
by a laboratory.
For practical reasons and to prevent potential laboratory/instrumentation contamination and
radiological safety issues, the test levels for incident response method validation purposes are limited
to three levels related to the designated (established) PAG AALs (derived radionuclide
concentrations) or default AALs. The use of three test levels is consistent with the specifications
indicated in Table 3 and Section 5.4. For incident response method validation, the validation test
levels are denoted as lower, mid, and upper. For method validation levels B, C, D, and E (Section
5.4), the recommended three concentration test levels for the replicate PT samples are presented in
Table 2. The lower level test level of-0.5 AAL was chosen to avoid detectability issues that could
occur at lower test concentrations. The mid test level corresponds to the established PAG AAL or
default AAL test level. It is assumed that a laboratory will use the same sample aliquant size and
11
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
counting time to analyze the test samples for all three test-level concentrations (values typically
selected to meet the required method uncertainty at the mid test level, or AAL).
TABLE 2 - Method Validation Test Concentrations
Test Level
Upper
Mid
Lower
Relative Concentration
~3 AAL
~1 AAL
~ 0.5 AAL
5.3 Matrix Considerations
For method validation, the method under consideration shall address a specific radionuclide and
matrix combination. In many applications, a matrix may be described by a general name or type,
such as water or air particulate. However, when developing and documenting the applicability of a
method, a description of the sample matrix should be specific and address possible variations in the
matrix that may be encountered when such will impact method performance. In addition, validation
of a method applies only to the specifically defined matrix described in the method validation plan,
which must be consistent with the matrix description in the method applicability statement. Listed
below are some specific matrices that may be encountered for radioanalytical processing during an
incident:
• Liquids
- Fresh water
- Surface water
- Groundwater
- Rain
- Salt/brackish water
- Aqueous suspensions
- Aqueous solutions
- Sewer and water treatment effluents or discharges
- Collection of volatiles
- Organic liquids
- Liquids generated during decontamination activities
• Air sampling media
- Glass fiber, cellulose, acetate filters
- Charcoal canisters or loose particles
- Molecular sieve
- Silica gel
• Swipes
- Glass fiber, cellulose, acetate filter paper
• Solids
- Soil, sediment, stone, sod, vegetation, wood
- Manufactured/construction
12
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
- Concrete, asphalt, brick, ceramics, plaster, plastics, metals, clothes, paper, stone, wood, etc.
- Sludges
- Sewer and water treatment
- Solids generated during decontamination activities
5.4 Method Validation Levels for Testing the Required Method Uncertainty
The primary method validation approach used in this document follows the concepts presented in
MARLAP Chapter 6 for the required method uncertainty MQO. The MARLAP method validation
approach and validation acceptance criteria assume that the laboratory method being validated has
no significant bias. However, this may not always be the case. Appendix D provides an insight into
the effect of method bias on the probability of failing the MARLAP validation acceptance criteria.
An alternate approach that may be used to determine if a method has acceptable method validation
performance is presented in Appendix E. This approach is based on the mean squared error (MSB)
or root mean squared error concept and has a greater power to detect excessive imprecision or bias
in many cases.
If a method fails to meet the method validation acceptance criteria as presented in the subsequent
sections, the laboratory should:
• Evaluate the possible reasons for the failure;
• Identify the root causes for the failure; and
• Update the method with the appropriate corrections or additions to ensure the method will meet
the specified MQOs.
The updated method must go through another validation process using the same requirements
applied to the first attempt at method validation.
5.4.1. Method Validation Requirements Based on MARLAP Concepts
Similar to the MARLAP (Chapter 6) graded approach to project method validation, there are four
proposed tiers or "levels" of method validation (Levels B, C, D, E) to demonstrate a method's
capability of meeting the required method uncertainty MQO applicable to a radiological incident.
For this guide, the MARLAP method validation Level A for the same radionuclide and matrix
combination has been combined with validation Level B (see Table 3). The level(s) of method
validation needed should be designated by the 1C. The laboratory will select a method based on
various operational aspects and the status of it's existing methods to meet the required method
uncertainty UMR or 0v specification for a designated (established) AAL (Appendix A) or a required
method uncertainty for a default AAL (Section 5.2). The WMR is specified in the units of the AAL. The
0V is a fractional unitless value (e.g., 0.13) and is calculated by dividing the UMR by the AAL.
Appendix A contains tables listing the required method uncertainties for screening and nuclide-
specific methods for certain established AALs and sample matrices related to a potential radiological
incident.
The four levels (B-E) of method validation for testing compliance with the required method
uncertainty using specified PT samples cover the following:
13
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Level B - Existing methods for radionuclide and matrix combinations for same, similar or
slightly different matrices (internal PT samples);
Level C - Existing methods that require modification to accommodate matrix differences
(internal or external PT samples);
Level D - Adapted or newly developed methods (internal and external PT samples); and
Level E - Adapted or newly developed methods using method validation reference materials
(method validation reference materials).
During the method validation process, the laboratory shall evaluate the method as to the required
method uncertainty and relative bias for the three test concentrations (Table 2) for the specified
method validation level, as well as the absolute bias through the use of at least seven blanks (Section
5.6 and Appendix E). The acceptable performance of a method to meet the required method
uncertainty will vary according to the level of validation as described in subsequent subsections. It
should be noted that the probability of acceptable performance for meeting a required method
uncertainty specification is dependent on the magnitude of existing method bias. The greater the
magnitude of the method bias, the more likely the method will not meet the required method
uncertainty specification. If excessive bias is measured during the method validation process, the
method should be revised to eliminate the bias as much as possible.
For radiological incident response applications, the analyte concentration range for the method
validation process, and thus the validation test levels, should be established based on the established
PAG or risk-based derived radionuclide concentrations as designated by an agency representative.
Derived radionuclide concentrations corresponding to the various established AALs (PAG or risk-
based) for water, air particulate and soil matrices have been provided in the Radiological Laboratory
Sample Analysis Guide for Incidents of National Significance and summarized in the Appendix A.
In the absence of validation test activity levels based on established AALs (PAG or risk-based),
default AALs specified in Section 5.2 may be used. Validation test activity/sample levels (designated
established AALs or default AALs) are to be used in conjunction with all method validation levels
stated in Table 3 and the required method uncertainty values for the radionuclide and matrix
combinations provided in the Radiological Laboratory Sample Analysis Guide for Incidents of
National Significance (Appendix A) or in Section 5.2.
A method is considered validated for a project when it has met the required method uncertainty
acceptance criteria stated in Table 3 and the acceptance criteria for other method characteristics such
as bias and required MDC, as may be stated by the 1C. When the required method uncertainty
specifications in Table 3 are met for default AALs, it will be assumed that the method has met the
required method uncertainty acceptance criteria for all PAG or risk-based action levels above the
default AALs.
All method validation levels require replicate samples at three different validation test concentration/
activity levels below, at, and above the derived radionuclide concentration corresponding to an AAL
(designated, established, or default). To ensure testing for sufficient method specificity, the known
concentration levels of potentially interfering radionuclides should be included in the test samples.
14
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE 3 - Method Validation Requirements and Applicable to Required Method Uncertainty
Validation
Level [11
C
D
E
Application
Existing Method
Radionuclide - Same,
Similar or Slightly
Different Matrix
Similar Matrix:
New Application
Adapted, Newly
Developed, Rapid
Methods
Adapted, Newly
Developed, Rapid
Sample
Type
Internal
PT
Internal or
External
PT
Internal or
External
PT
Method
Validation
Reference
Materials
Acceptance
Criterion [2]
Measured Value
Within±2.8wMRor±
2.8 0v of
Validation Value
Measured Value
Within±2.9MMRor±
2.9 0v of
Validation Value
Measured Value
Within±3.0MMRor±
3 0 (o, of
Validation Value
Measured Value
Within±3.0ttMRor±
3. 00v of
Validation Value
Levels [4]
(Concentration)
3
3
3
3
RepUcates
3
5
7
7
#of
Analyses
9
15
21
21
[1]
[2]
[3]
[4]
MARLAP method validation Level A for the same radionuclide and matrix combination has been included into
validation Level B.
The acceptance criterion is applied to each analysis/test sample used for method validation, not the mean of the
analyses. MMR and 0v values are the required absolute and relative method uncertainty specifications stipulated in
the Radiological Laboratory Sample Analysis Guide for Incidents of National Significance for gross screening
concentrations and quantification of individual radionuclide concentrations in various matrices. The acceptance
criteria are chosen to give a false rejection rate of ~5% when the measurement process is unbiased, with a standard
deviation equal to the required method uncertainty (MMR or 0v)- The stated multiplier (k = 2.8, 2.9, 3.0) for the
required method uncertainty was calculated using the formula k = z.
iwwhere TV is the number of
'0.5+0.5(1-0)
measurements, a is the desired false rejection rate, and, for any p, zp denotes the />-quantile (0 < p < 1) of the
standard normal distribution (MARLAP Appendix G, Table G. 1). The MMR or 0v values are provided in Appendix
A or Table 4.
For certain matrices, not all samples in a given test level can be spiked with the same known radionuclide activity
or concentration. In such cases, the measured activity or concentration in the test sample should be compared to the
known value for that test sample.
At least seven blank samples should be analyzed as part of method validation but are not considered part of the three
required concentration test levels.
5.4.2 Required Method Uncertainty Acceptance Criteria
For all four method validation levels for method uncertainty, acceptable method validation is
determined by comparing each test sample result for a given test concentration or activity with the
required method uncertainty specification (k x UMR or k x pMR) provided in Table 3. The values for UMR
or #v are stipulated in the Radiological Laboratory Sample Analysis Guide for Incidents of National
Significance (Appendix A) for the three basic incident response matrices of water, air filter/swipes,
and soil. A "#" value can be either 2.8, 2.9, or 3.0. It should be noted that the required method
uncertainty specification and AAL-derived radionuclide concentrations may vary according to the
matrix, the phase of the incident response and applied PAG.
Appendix B provides examples for testing a method's acceptability to meet validation Level D for
an established PAG AAL and a default AAL test level in a water matrix.
15
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
5.4.2.1 Level B Method Validation: Same, Similar, or Slightly Different Matrix
Most qualified laboratories will have existing methods to analyze for the radionuclides of interest
in the three most common matrices of water, air parti culate filters and soil/sediment. Under method
validation Level B, a method that has been previously validated for a different project and used for
one matrix may be used for that same matrix or modified for use for a very similar matrix. An
example of a slightly different matrix might be a method used for water samples having low
dissolved solids, modified for water samples containing high dissolved solids. Level B requires the
laboratory to conduct a method validation study for the radionuclide and matrix combination where
three replicate samples from each of the three concentration levels are analyzed according to the
method. Table 2 is to be used to determine the lower, middle, and upper testing levels for the
replicate analyses. The test samples are internal PT samples prepared at the laboratory. In order to
determine if a proposed method meets the project MQO requirements for the required method
uncertainty, each internal PT sample result is compared with the method uncertainty acceptance
criteria in Table 3. The acceptance criteria in Table 3 for Level B validation stipulate that, for each
test sample analyzed, the measured value must be within ±2.8 UUR for test-level concentrations at or
less than the AAL or ± 2.8 f>MR for the test-level concentration above the AAL. These acceptance
criteria apply to either established AALs stated in Appendix A or default AALs (Section 5.2). The
values of WMR and (pUR for select radionuclide and matrix combinations are provided in Appendix A
for established PAG AALs or in Table 4 for default AALs. The required method uncertainty values
for the established PAG AALs are based on the Radiological Laboratory Sample Analysis Guide for
Incidents of National Significance for the three basic matrices addressed: water, air and soil. The
Table 4 values for UUR are base on the default AALs stated in Table 1, and the #v values are taken
from the air and water editions of the Radiological Laboratory Sample Analysis Guide for Incidents
of National Significance. For example, in Table 4 the UMR value of 5.2 pCi/sample for specific alpha-
emitting nuclides in a water matrix (column two, row two) is calculated by multiplying the relative
required method uncertainty (#v) of 0.13 (column three, row two) for this radionuclide and matrix
combination by the default AAL in Table 1 (column three, row two) of 40 pCi/sample value. The
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Notes:
[1] For each radionuclide category, the absolute required method uncertainty (MMR) is applied to the lower and mid test
levels. The relative required method uncertainty (#v) is to be used for the upper test level.
[2] Required method uncertainty values for water correspond to the 100-mrem dose-derived concentration values from
Scenario 1 of the Radiological Laboratory Sample Analysis Guide for Incidents of National Significance-
Radionuclides in Water (EPA 2008).
[3] Required method uncertainty values for air sampling media correspond to the 10~4 risk-based derived air concentra-
tion values from Scenario 1 of the Radiological Laboratory Sample Analysis Guide for Incidents of National
Significance-Radionuclides in Air (EPA 2009a).
[4] The default values stated in the specific emitting nuclide rows apply to all radionuclides in the designated emission
category. The reference to a radionuclide is presented only as information indicating the basis for the specific
emission category.
[5] TBD: To be determined pending development of Radiological Laboratory Sample Analysis Guide for Incidents of
National Significance-Radionuclides in Soils (In preparation).
[6] Table values have been rounded after calculations.
5.4.2.2 Level C Method Validation: New Application of an Existing Method to a Different
Matrix
When a laboratory has a validated method for a radionuclide in one matrix but not others, the method
may require modification to accommodate a completely different media/matrix (water versus soil)
or a very different similar matrix (two soils of different physicochemical compositions). The degree
of adaptation or modification needed will vary according to chemical and physical differences in the
two matrices. Because of the extent of these differences, a laboratory may choose to validate the
method for the radionuclide and matrix combination through method validation Level C. Level C
method validation requires the laboratory to conduct a method validation study wherein five replicate
samples from each of the three concentration levels are analyzed according to the method. Table 2
is to be used to determine the lower, mid and upper testing levels for the replicate analyses. The test
samples are internal PT samples prepared at the laboratory. In order to determine if a proposed
method meets the project MQO requirements for the required method uncertainty, each internal PT
sample result is compared with the method uncertainty acceptance criteria of Table 3. The
acceptance criteria stated in Table 3 for Level C for new applications stipulate that, for each test
sample analyzed, the measured value must be within ±2.9 WMR for test level concentrations at or less
than the AAL or ± 2.9f>MR for the test level concentration above the AAL. These acceptance criteria
apply to either established PAG AALs stated in Appendix A or default AALs (see Section 5.2). The
values of z/MR and f>MR for select radionuclide and matrix combinations are provided in Appendix A
for established PAG AALs or in Table 4 for default AALs.
5.4.2.3 Level D Method Validation: Adapted or Newly Developed Methods, Including Rapid
Methods
In some cases, a laboratory may not have a method for a certain radionuclide and matrix combina-
tion. For such situations, the laboratory may either develop a new method internally or adapt a
method from the literature. In this case, the new method should undergo general method validation
first and then incident response method validation. A laboratory would validate the new method for
the radionuclide and matrix combination through method validation Level D or E (Section 5.4.2.4).
Level D method validation requires the laboratory to conduct a method validation study wherein
seven replicate samples from each of the three concentration levels are analyzed according to the
method. Table 2 is to be used to determine the lower, mid and upper testing levels for the replicate
17
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
analyses. For validation Level D, the test samples are internal PT samples prepared at the laboratory.
In order to determine if a proposed method meets the project MQO requirements for the required
method uncertainty, each internal PT sample result is compared with the method uncertainty
acceptance criteria in Table 3. The acceptance criteria stated in Table 3 for Level D validation
stipulate that, for each test sample analyzed, the measured value must be within ±3.0 UMR for test
level concentrations at or less than the AAL or ± 3.0
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
validation process for the required method uncertainty MQO. General method validation require-
ments are to be met prior to the initiation of this verification process. The specifications given in this
section are distinct from those given for the required method uncertainty MQO method-validation
process.
/ 1 v
/ • \
/ Radionuclide/
\ mitrixAAL
\specified /
vYes
13. Develop method
for radionuclide and
specified matrix
/Method "*
exists for
radionuclide
/
No
14. Complete general
method validation
No
19. „
/Can existing
method be
modified for /
matrix
18, Level D required
(Sections 5.4.1 & 5.4.2.3)
/ Method
exists for same/
.similar matrix
7
Yes
Yes
20. Modify method and
complete general MV
5. Level B required
(Sections 5.4.1 & 5.4.2.1)
21. Level C required
(Sections 5.4.1 & 5.4.2.2)
Documentation
meets intent of
Sections
No
19. Complete Sections
5 and 6 requirements
for the method
17. Level E required
(Sections 5.4.1 8,5.4.2.4)
Modify Method
Go to 5 for Level B
Goto Step 21 for Level C
Goto StsplBforLevelD
Go to Step 17 for Level E
7.
\M(
Yes
ratory X,
specifies \. No^
nd method /
ables /
f
9. Revise method to
include MQOs, MV
documentation, and
references
i
/
' •,/
10. Use calculations
cited in Section 5.6
dto
tjiv/
j
11. MV for this
radionuclids/matrix
combination
completed
Figure 1 - Method Validation Process for the Required Method Uncertainty MQO
19
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Required MDC is
the MQO
Has method
been established
and tested for
radionuclide-matrix
combination
Evaluate method
for matrix effects
Revise or develop method
chemical-separations or
counting techniques
Analyze at least 7 blanks to
establish critical net
concentration, CLNC
Analyze 10 replicate samples
spiked at the required MDC
Are the
number of
replicate results
below. CLNC
<2
MDC MQO
has been
verified
Figure 2 - Validation Process for Verifying the Required MDC MQO
5.5.1 Calculation of the Critical Net Concentration
The critical net concentration shall be calculated for the required MDC method validation process.
The calculation of the critical net concentration for the method is based on the analytical results of
the blank matrix samples used in the MDC validation process. A minimum of seven blank samples
is required. To ensure testing for sufficient method specificity, the matrix blanks should contain the
anticipated concentration levels of chemical interferences and the potential interfering radionuclides
(naturally occurring and incident response-related).
The critical net concentration (CLNC), with a Type I error probability of a = 0.05, is calculated using
the following equation (consistent with MARLAP, Chapter 20, Equation 20.35):
(pCi/unit) = ^ _a(n -1) x
(1)
20
-------
_ Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities _
where %anks is the standard deviation of the n blank-sample net results (corrected for instrument
background) in radionuclide concentration units of pCi/sample, and t^Jn- 1) is the (l-a)-quantile
of the ^-distribution with n- 1 degrees of freedom (see MARLAP Table G.2 in Appendix G).
Although the Type I error rate of 0.05 is routinely used and accepted, it is possible that other error
rates may be used in incident response situations.
For seven (minimum) blank results (six degrees of freedom) and a Type I error probability of 0.05,
Equation (1) reduces to:
(pCi/unit) =1.94x*Blanks (2)
If the number of blank samples is different than the minimum value of seven, refer to MARLAP
Chapter 20, Attachment 20A for appropriate guidance. Care must be taken to ensure that all samples
and blanks are analyzed under conditions that are typical of those used for routine analyses using the
same sample weight or volume and with the same instruments with representative counting efficien-
cies and background levels. The calculated critical net concentration will be used in the verification
process to determine if a method is capable of meeting the required MDC specification as described
in Section 5.5.2.
5.5.2 Testing for the Required MDC
When a required MDC specification for a radionuclide and matrix combination is given as an MQO
rather than the required method uncertainty, the method should be validated by verifying that the
method can meet the required MDC. As noted in Table 5, method validation for the required MDC
specifies that ten replicate samples, each spiked at the required MDC, should be analyzed and
evaluated. In addition, the results of at least seven blank samples are used to determine the critical
net concentration of the method (Section 5.5.1). The ten replicate spiked samples and seven blanks
should contain the chemical species and potential interfering radionuclides which are reasonably
expected to be present in an actual sample. To ensure the testing for sufficient method specificity,
the expected concentration levels of the chemical species and potential interfering radionuclides
should be used during testing. Figure 2 (page 20) provides an overview of the method validation
process for verifying the required MDC MQO.
Testing for the required MDC verification is based on the null hypothesis that the true MDC for the
method is at or below the required MDC. If the true MDC of the method has been calculated
properly and is equal to or less than the required MDC, the probability of failing to detect the
radionuclide at or above the critical net concentration is at most/?. For project method validation
related to incident response, ft is assumed to be 0.05. The number of "non-detects" (sample results
below the critical net concentration) for a set of n samples spiked at the required MDC is assumed
to have a binomial distribution with parameters ft and n. For a set of ten samples spiked at the
required MDC, the number of Y sample results expected to be below the critical net concentration
is not more than two (2) for a/? of 0.05. If Y is greater than two, the null hypothesis is rejected.
The following protocol should be used to verify a method's capability to meet the required method
MDC for each radionuclide (including gross screening)-matrix combination:
21
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
1. Analyze a minimum of seven blank samples (representing the matrix of interest) for the
radionuclide under consideration.
2. From the blank sample net results, calculate the estimated Critical Net Concentration
(Section 5.5.1), CZNC.
3. Analyze ten replicate samples (representing the matrix of interest) spiked at the required
MDC for the radionuclide under consideration.
4. From the results of the ten replicate samples spiked at the required MDC, determine the
number (Y) of sample results at or below the estimated Critical Net Concentration.
5. If Y < 2, the method evaluated at the required MDC passes the test for the required MDC
specification.
6. If Y > 2, the method evaluated at the required MDC fails the test for the required MDC
specification.
Appendix C provides an example for testing a method's capability to meet a required MDC
specification.
TABLE 5 - Method Validation Requirements Applicable to Required MDC Verification
Method
Characteristic
Detection
Capability
Application
Required
MDC
Specification
Sample
Type
Internal
PT
Acceptance
Criterion
Number of Sample
Results Below
Critical Net
Concentration Value
< 2
Levels
(Concentrations)
Single Concentration
at the Required MDC
Value
Replicates
10
#of
Analyses
10
Note: At least seven blank samples should be analyzed to estimate the critical net concentration as part of the required
MDC verification.
5.6 Method Bias Tests
In order to provide quality data, a method should not have a significant bias. Depending on the
radiological incident, acceptable absolute and relative bias criteria for a method may be specified by
the 1C. Since the degree of acceptability of method bias depends on many parameters and circum-
stances, specific acceptance criteria for method bias have not been included for this method
validation process. However, because the acceptance criteria for method uncertainty and required
MDC verification will not tolerate a significant method bias or measurement uncertainty, acceptable
method bias is indirectly evaluated when evaluating method uncertainty and the required MDC.
Appendix D provides an example of the effect of bias on the probability of failing the required
method uncertainty validation acceptance criteria for method validation Level D.
Method bias is initially evaluated during method development, general and proj ect method validation
processes, and then continuously during the processing of incident response samples using batch QC
samples (MARLAP Chapter 7). Tests for absolute and relative biases shall be made for the method
22
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
validation level specified by the 1C. The absolute bias shall be evaluated using the blank sample
results and the relative bias evaluated for each test level (lower, mid and upper) using the results of
the replicates.
When there is a significant absolute or relative bias, the probability of failing the required method
uncertainty acceptance criteria of Section 5.4.2 may become significant depending on the magnitude
of the actual method uncertainty. Appendix D provides an example of the probability of failing
method validation Level D for three actual method uncertainty values (as compared to the required
method uncertainty) as a function of relative bias up to 20%. In general, to avoid failure to meet the
method validation acceptance criteria, it is best to have an actual method uncertainty at the AAL that
is a fraction of the required method uncertainty.
The following equations, taken from MARLAP Chapter 6 (Attachment 6A) and other statistical
references, are used to test for absolute and relative biases.
5.6.1 Absolute Bias Testing
The protocol for testing for absolute bias is the following:
1. Calculate the mean (X) for "N" (at least seven) blank sample net results using Equation 3.
X = — > X. (3)
N ^'=1 '
where N should be at least seven blank sample results.
2. Calculate the experimental standard deviation (sx) of the same results6 using Equation 4.
(4)
3. Use Equation 5 to calculate the |T| value:
6 Notice that the sum under the radical in equation 4 is divided by the number of degrees of freedom, N- I, not the
number of results, N. When calculated in this manner, sx2 is an unbiased estimator for the variance of the results. If the
1 v->"
true mean of the results, UY, were known, a better estimate of the variance would be — > (X - u ), but because the
>/••*>> N Z_!l=l ^ • t*,J ••
mean is estimated from the data, the number of degrees of freedom is reduced by 1. Notice also that the expression in
the denominator of the right-hand side of Equation 5 gives the experimental standard deviation of the mean, more
commonly known as the "standard error of the mean." The division by -JW in this case accounts for the effect of
averaging TV independent results.
23
-------
_ Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities _
4. An absolute bias in the measurement process is indicated if
\T\>tl_aJ2(N-l) (6)
where, t^^ (N-i) represents the (1 - a/2)-quantile of the ^-distribution with N-l degrees of
freedom. For seven blanks, an absolute bias is identified at a significance level of 0.05, when
|T| > 2.447.
5.6.2 Relative Bias Testing
5.6.2.1 Test Level Samples with Same Known Value
When the samples for a test level have the same concentration (e.g., water) or activity, the protocol
for testing relative bias for each method validation test level is the following:
1 . Calculate the mean (X) and estimated standard deviation (sx) of the replicate results for each
method validation test level using Equations 3 and 4, respectively.
2. Use Equation 7 to calculate the |T| value
\T\= \X~K\ - (7)
where:
X is the average measured value
sx is the experimental standard deviation of the measured values
N is the number of replicates
K is the reference value
u(K) is the standard uncertainty of the reference value
3. A relative bias in the measurement process is indicated if
m^-a^ff) (8)
The number of effective degrees of freedom for the T statistic is calculated as follows:
(9)
veff as calculated by the equation generally is not an integer so veff should be truncated (rounded
down) to an integer. Then, given the significance level of 0.05, the critical value for \T\ is defined
tobe Vo/2(veffX the (1 - a/2)-quantile ofthe ^-distribution with veffdegrees of freedom (seeMARLAP
Appendix G, Table G.2 ).
24
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
5.6.2.2 Test Level Samples with Slightly Different Known Values
When the PT samples for a test level have slightly different concentrations or activities (e.g.,
independently prepared7 water samples, air filters, or swipes), the following protocol (paired Mest)
for testing relative bias for each method validation test level is:
1. Calculate the average difference (D) between the measured value and the known spiked
value using Equation 10:
(10)
/v ^"^/=1
where
Xt is the measured value for the ith sample at a particular test level
Kt is the known value for the same sample
N is the number of samples at that test level
2. Calculate the standard deviation of the differences, SD, as:
(11)
where D,=X,-K,.
3. Calculate the absolute value of the t statistic as:
\T\=
4. A relative bias in the measurement process for a test level is indicated if
m^i-a/2^-1) (13)
6.0 Method Validation Documentation
The information and data to be retained should be specified in the method validation plan for each
radionuclide and matrix combination. When the laboratory conducts project method validation for
incident response applications, the detailed analytical method and all records, laboratory workbooks,
and matrix spike data used to validate the analytical method should be retained on file and be
retrievable for a specified length of time after the method has been discontinued. Data evaluations
such as comparison of individual results to the validation acceptance criteria and absolute bias in
7 During the preparation of the proficiency test samples for a test level, the spread in activity deposited on the samples
of the test level should be controlled so that the coefficient of variation of the test-sample activities does not exceed 3%.
25
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
blanks and, when available, method precision and bias, should be part of the data validation package
retained as part of the documentation related to the laboratory's quality system. In addition, for each
radionuclide and matrix combination, a synoptic method validation report containing the analytical
method identification, method validation acceptance criteria, test levels, validation results and a
method acceptability decision should be generated and retained.
7.0 References
American National Standards Institute (ANSI) N42.23. Reapproved 2003. Measurement and
Associated Instrumentation Quality Assurance for Radioassay Laboratories.
U.S. Environmental Protection Agency (EPA). 2006. Validation and Peer Review of U.S.
Environmental Protection Agency Radiochemical Methods of Analysis. FEM Document Number
2006-01, November 8.
U.S. Environmental Protection Agency (EPA). 2008. Radiological Laboratory Sample Analysis
Guide for Incidents of National Significance-Radionuclides in Water. Revision 0. Office of Air
and Radiation, Washington, DC. EPA 402-R-07-007, January. Available at: www.epa.gov/
narel/recent_info.html.
U.S. Environmental Protection Agency (EPA). 2009a. Radiological Laboratory Sample Analysis
Guide for Incidents of National Significance-Radionuclides in Air. Revision 0. Office of Air and
Radiation, Washington, DC. EPA402-R-09-007, June. Available at: www.epa.gov/narel/recent_
info.html.
U.S. Environmental Protection Agency (EPA). 2009b. Radiological Laboratory Sample Screening
Analysis Guide for Incidents of National Significance. Revision 0. Office of Air and Radiation,
Washington, DC. EPA402-R-09-008, June. Available at: www.epa.gov/narel/recent_info.html.
EURACHEMGuide. 1998 The Fitness for Purpose of Analytical Methods, A Laboratory Guide to
Method Validation and Related Topics. Available at: http://www.eurachem.org/.
International StandardISO/IEC17025. 2005. General Requirements for the Competence of Testing
and Calibration Laboratories. ISO, Geneva, Switzerland.
International Organization for Standardization (ISO). 1993. International Vocabulary of Basic and
General Terms in Metrology. 2nd Edition. ISO, Geneva, Switzerland.
Multi-Agency Radiological Laboratory Analytical Protocols Manual (MARLAP). 2004. EPA 402-
B-04-001A, July. Volume I, Chapters 6, 7, 20, Glossary; Volume U and Volume IE, Appendix
G. Available at: www.epa.gov/radiation/marlap.
Multi-Agency Radiation Survey and Site Investigation Manual (MARSSEVI), Revision 1. 2000.
NUREG-1575 Rev 1, EPA 402-R-97-016 Revl, DOE/EH-0624 Revl. August. Available from
www.epa.gov/radiation/marssim/.
26
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Appendix A:
Tables Summarizing the Derived Radionuclide Concentrations and Required
Method Uncertainties Corresponding to PAGs or Risks for the Water, Air, and Soil
Matrices
TABLE Al - Alphs
Water Corres
Radionuclide
Gross a Screen [5]
Am-241
Cm-242
Cm-243
Cm-244
Np-237
Po-210
Pu-238
Pu-239
Pu-240
Ra-226 [4]
Th-228 [4]
Th-230
Th-232
U-234
U-235
U-238
i-Emitting Radionuclide Concentrations and Required Method Uncertainties in
jonding to 500- and 100-mrem AAL Derived Water Concentrations (DWCs)
pCi/L
500 mrem
AAL
DWC
[i]Pl
2.0xl03
2.0xl03
1.4xl04
2.5xl03
2.9xl03
3.9xl03
130
1.8xl03
l.VxlO3
1.7xl03
910
2.6xl03
1.8xl03
1.6xl03
6.3xlQ3
6.6xl03
7.0xl03
Screening
Methods
Required
Method
Uncertainty
(u \ [6]
\UMR)
610
610
4.3xlQ3
760
880
1.2xl03
40
550
520
520
280
790
550
490
1.9xl03
2.0xl03
2.1xl03
100 mrem
AAL
DWC
[11P1P1
400
400
2.8xl03
500
580
780
26
360
340
340
180
520
360
320
1.3xl03
1.3xl03
1.4xl03
Screening
Methods
Required
Method
Uncertainty
(u \ [6]
\UMR)
120
120
850
150
180
240
7.9
110
100
100
55
160
110
97
400
400
430
Nuclide-
Specific
Required
Method
Uncertainty
(u } [6]
\UMR)
—
50
350
63
73
98
3.3
45
43
43
23
65
45
40
160
160
180
Notes:
[1] Values are based on the dose conversion factors in Federal Guidance Report No. 13, CD Supplement, 5-
year-old child and the 50thpercentile of water consumption.
[2] 365-day intake.
[3] Values obtained by dividing 500-mrem DWC values by 5.
[4] Includes the dose from the decay products originating from the 226Ra or 228Th in the body.
[5] Values for gross alpha screening are based on 241Am.
[6] The required relative method uncertainty (#v) for values greater than the AALs in this table is obtained
by dividing the «MR value by the corresponding AAL.
27
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE A2 - Beta/Gamma-Emitting Radionuclide Concentrations in Water and Required Method
Uncertainties Corresponding to 500- and 100-mrem AAL Derived Water Concentrations (DWCs)
Radionuclide
Gross P/Y
Screen[4]
Ac-227DP[6]
Ce-141
Ce-144
Co-57
Co-60
Cs-134
Cs-137
H-3
1-125
1-129
1-131
fr- 192
Mo-99
P-32
Pd-103
Pu-241
Ra-228[6]
Ru-103
Ru-106
Se-75
Sr-89
Sr-90
Tc-99
pCi/L
500 mrem
AAL DWC
[i]Pl
5.8xl04
l.lxlQ3
2.2xl05
2.9xl04
6.3xl05
3.3xl04
4.3xl04
5.8xl04
7.7xl06
1.3xl04
3.3xl03
5.4xl03
1.2xl05
3.2xl05
5.9xl04
7.8xl05
l.OxlO5
160
2.3xl05
2.2xl04
6.7xl04
6.3xl04
1.2xl04
2.4xl05
Screening
Methods
Required
Method
Uncertainty
(«MR)[5]
1.8xl04
330
6.7xlQ4
8.8xl03
1.9xl05
l.OxlO4
1.3xl04
1.8xl04
2.3xlQ6
4.0xl03
l.OxlO3
1.6xl03
3.6xl04
9.7xl04
1.8xl04
2.4xl05
3.0xl04
49
7.0xl04
6.7xl03
2.0xl04
1.9xl04
3.6xl03
7.3xl04
100 mrem
AAL DWC
[11P1P1
1.2xl04
220
4.4xl04
5.8xl03
1.3xl05
6.6xl03
8.6xl03
1.2xl04
l.SxlO6
2.6xl03
660
l.lxlO3
2.4xl04
6.4xl04
1.2xl04
1.6xl05
2.0xl04
32
4.6xl04
4.4xl03
1.3xl04
1.3xl04
2.4xl03
4.8xl04
Screening
Methods
Required
Method
Uncertainty
(«MR)[5]
3.6xl03
67
1.3xl04
1.8xl03
4.0xl04
2.0xl03
2.6xl03
3.6xl03
4.6xl05
790
200
330
7.3xl03
1.9xl04
3.6xl03
4.9xl04
6.1xl03
9.7
1.4xl04
1.3xl03
4.0xl03
4.0xl03
730
l.SxlO4
Nuclide-
Specific
Required
Method
Uncertainty
(«MR)[5]
—
28
5.5xl03
730
1.6xl04
830
l.lxlO3
l.SxlO3
1.9xl05
320
83
140
3.0xl03
S.lxlO3
l.SxlO3
2.0xl04
2.5xl03
4.0
5.8xl03
550
1.6xl03
1.6xl03
300
6.0xl03
Notes:
[1] Values are based on the dose conversion factors in Federal Guidance Report No. 13, CD Supplement, 5-
year-old child and the 50thpercentile of water consumption.
[2] 365-day intake.
[3] Values obtained by dividing 500-mrem DWC values by 5.
[4] Gross beta screening values are based on 137Cs.
[5] The required relative method uncertainty (#v) for values greater than the AALs is obtained by dividing
the um value in this table by the corresponding AAL value.
[6] Includes the dose from the decay products originating from the 228Ra or 227Ac in the body.
28
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE A3 - Alpha-Emitting Radionuclide Concentrations in Air and Required Method
Uncertainties Corresponding to 2-rem and 500-mrem AAL Derived Air Concentrations (DACs)
Radionuclide
Gross a
Screen™
Am-241
Cm-242
Cm-243
Cm-244
Np-237
Po-210
Pu-238
Pu-239
Pu-240
Ra-226 [2]
Th-228 [2]
Th-230
Th-232
U-234
U-235
U-238
pCi/m3
2 rem
AAL
DAC
[i]
0.70
0.70
11
0.97
1.2
1.3
16
0.62
0.56
0.56
7.0
1.7
0.66
0.61
7.1
7.9
8.3
Screening
Method
Required
Method
Uncertainty
(«MR)
[3]
0.21
0.21
3.3
0.29
0.36
0.40
4.9
0.19
0.17
0.17
2.1
0.52
0.20
0.19
2.2
2.4
2.5
Nuclide-
Specific
Required
Method
Uncertainty
(«MR)
[3]
0.088
1.4
0.12
0.15
0.16
2.0
0.081
0.071
0.071
0.88
0.21
0.083
0.077
0.89
0.99
1.0
500 mrem
AAL
DAC
[i]
0.17
0.17
2.8
0.24
0.29
0.34
3.9
0.15
0.14
0.14
1.8
0.42
0.17
0.15
1.8
2.0
2.1
Screening
Method
Required
Method
Uncertainty
(«MR)
[3]
0.052
0.052
0.85
0.073
0.088
0.10
1.2
0.046
0.043
0.043
0.55
0.13
0.052
0.046
0.55
0.61
0.64
Nuclide-
Specific
Required
Method
Uncertainty
(«MR)
[3]
0.021
0.35
0.030
0.037
0.043
0.49
0.020
0.018
0.018
0.23
0.053
0.021
0.019
0.23
0.25
0.26
Notes:
[1] Morbidity for long-term inhalation. Child as receptor. Value corresponds to solubility class having
lowest value.
[2] Includes the dose from the decay products originating from the 226Ra or 228Th in the body.
[3] Required method uncertainty values are based on a sampled aerosol volume of 68 m3 at the 2 rem or 500-
mrem DAC. The required relative method uncertainty (#v) for values greaterthan the AALs in this table
is obtained by dividing the «MR value in this table by the corresponding AAL value.
[4] The gross a screening values are not related to a specific radionuclide.
29
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE A4 - Beta/Gamma-Emitting Radionuclide Concentrations in Air and Required Method
Uncertainties Corresponding to 2-rem and 500-mrem AAL Derived Air Concentrations (DACs)
Radionuclide
Gross (3
Screen[5]
Ac-227+DP [2]
Ce-141
Ce-144
Co-57
Co-60
Cs-134
Cs-137
H-3
1-125 [6]
1-129 [6]
1-131 [6]
IT- 192
Mo-99
P-32
Pd-103
Pu-241
Ra-228 [2]
Ru-103
Ru-106
Se-75
Sr-89
Sr-90
Tc-99
pCi/m3
2 rem
AAL
DAC
1?]
420
0.43
1.8xl04
1.3xl03
6.7xl04
2.2xl03
3.3xl03
1.7xl03
2.6xl05
1.3xl04
1.9xl03
9.1xl03
l.OxlO4
6.8xl04
1.7xl04
l.SxlO5
29
4.2
2.3xl04
l.OxlO3
S.OxlO4
8.4xl03
420
S.OxlO3
Screening
Method
Required
Method
Uncertainty
(«MR)
[3]
130
0.13
5.5xl03
400
2.0xl04
670
l.OxlO3
520
7.9xl04
4.0xl03
580
2.8xl03
3.0xl03
2.1xl04
5.2xl03
4.6xl04
8.8
1.3
7.0xl03
300
1.5xl04
2.6xl03
130
1.5xl03
Nuclide-
Specific
Required
Method
Uncertainty
(«MR)
[3]
0.054
2.3xl03
160
8.4xl03
280
420
210
3.3xl04
1.6xl03
240
l.lxlO3
1.3xl03
8.6xl03
2.1xl03
1.9xl04
3.7
0.53
2.9xl03
130
6.3xl03
l.lxlO3
53
630
500 mrem
AAL
DAC
[1,4]
110
0.11
4.5xl03
320
1.7xl04
540
820
430
6.4xl04
3.2xl03
470
2.3xl03
2.5xl03
1.7xl04
4.3xl03
3.8xl04
7.3
1.0
5.7xl03
250
l.SxlO4
2.1xl03
110
l.SxlO3
Screening
Method
Required
Method
Uncertainty
(«MR)
[3]
33
0.033
1.4xl03
97
5.2xl03
170
250
130
1.9xl04
970
140
700
760
5.2xl03
1.3xl03
1.2xl04
2.2
0.30
1.7xl03
76
4.0xl03
640
33
400
Nuclide-
Specific
Required
Method
Uncertainty
(«MR)
[3]
0.014
570
40
2.1xl03
69
100
54
S.lxlO3
400
59
290
310
2.1xl03
540
4.8xl03
0.92
0.13
720
31
1.6xl03
260
14
160
Notes:
[1] Derived air concentration yielding stated committed effective dose assuming a 365-day year. Child as receptor.
Value corresponds to solubility class having lowest value.
[2] Includes the dose from the decay products originating from the 228Ra or 227Ac in the body. DP refers to "decay
products."
[3] Required method uncertainty values are based on a sampled aerosol volume of 68 m3 at the 2 rem or 500-mrem
DAC. The required relative method uncertainty (#v) for values greater than the AALs in this table is obtained by
dividing the WMR value in this table by the corresponding AAL value.
[4] All nuclides can be collected on a fibrous or membrane air filter media except 3H, 125I, 129I, and 131I in the vapor
states.
[5] Gross beta screening values are based on 90Sr.
[6] These values are based on the vapor plus particulate dose rate.
30
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE AS - Alpha-Emitting Radionuclide Concentrations in Air and Required Method
Uncertainties Corresponding to AAL Derived Air Concentrations (DACs)
Radionuclide
Gross a Screen [4]
Am-241
Cm-242
Cm-243
Cm-244
Np-237
Po-210
Pu-238
Pu-239
Pu-240
Ra-226 [2]
Th-228 [2]
Th-230
Th-232
U-234
U-235
U-238
pCi/m3
ID'4 Risk
AAL DAC
[i]
0.33
0.33
0.62
0.34
0.35
0.43
0.86
0.24
0.22
0.22
0.44
0.094
0.36
0.30
0.45
0.49
0.52
ID'4 Risk AAL
Required
Method
Uncertainty (HMR)
[3]
0.042
0.042
0.078
0.043
0.044
0.054
0.11
0.030
0.028
0.028
0.055
0.012
0.045
0.038
0.057
0.062
0.065
ID'6 Risk
AAL DAC
[i]
3.3xlO~3
3.3xlO~3
6.2xlO~3
3.4xlO~3
3.5xlO~3
4.3 xlO~3
8.6xlO-3
2.4xlO~3
2.2xlO~3
2.2xlO~3
4.4xlO~3
9.4xlO-4
3.6xlO~3
3.0xlO-3
4.5 xlO~3
4.9xlO~3
5.2xlO~3
ID'6 Risk AAL
Required Method
Uncertainty (HMR)
[3]
4.2xlO-4
4.2xlO-4
7.8xlQ-4
4.3 xlQ-4
4.4xlO-4
5.4xlO-4
l.lxlQ-3
3.0xlQ-4
2.8xlQ-4
2.8xlQ-4
5.5xlQ-4
1.2xlQ-4
4.5 xlQ-4
3.8xlQ-4
5.7xlQ-4
6.2xlQ-4
6.5 xlQ-4
Notes:
[1] Morbidity for long-term inhalation. Value corresponds to solubility class having lowest value.
[2] Includes the dose from the decay products originating from the 226Ra or 228Th in the body.
[3] Required method uncertainty values are based on a sampled aerosol volume of 1,600 m3 at the 10~4 and
10~6 risk DACs, respectively. The required relative method uncertainty (0«R) for values greater than the
AALs in the table is obtained by dividing the u^ value by the corresponding AAL value.
[4] The gross a screening values are not related to a specific radionuclide.
31
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE A6 - Beta/Gamma-Emitting Radionuclide Concentrations in Air and Required
Method Uncertainties Corresponding to AAL-Derived Air Concentrations (DACs)
Radionuclide
Gross |3 Screen
(Sr-90)
Ac-227+DP [2]
Ce-141
Ce-144
Co-57
Co-60
Cs-134
Cs-137
H-3 Vapor
1-125
1-129
1-131
Ir-192
Mo-99
P-32
Pd-103
Pu-241
Ra-228 [2]
Ru-103
Ru-106
Se-75
Sr-89
Sr-90
Tc-99
pCi/m3
ID'4 Risk
AAL DAC
[1,4]
29
0.083
920
69
3.3xl03
120
180
110
l.SxlO4
1.2xl03
200
640
510
2.6xl03
890
7.0xl03
14
0.28
1.2xl03
56
2.5xl03
410
29
330
ID'4 Risk AAL
Required Method
Uncertainty (HMR)
[3]
3.8
0.010
120
8.7
420
15
23
14
1.9xl03
150
25
81
64
330
110
880
1.8
3.5xlO~2
150
7.1
310
52
3.7
42
ID'6 Risk
AAL DAC
[1,4]
0.29
8.3 xlO~4
9.2
0.69
33
1.2
1.8
1.1
150
12
2
6.4
5.1
26
8.9
70
0.14
2.8xlO~3
12
0.56
25
4.1
0.29
3.3
ID'6 Risk AAL
Required
Method
Uncertainty (HMR)
[3]
0.038
i.oxio-4
1.2
0.087
4.2
0.15
0.23
0.14
19
1.5
0.25
0.81
0.64
3.3
1.1
8.8
0.018
3.5xlQ-4
1.5
0.071
3.1
0.52
0.037
0.42
Notes:
[1] Morbidity for long-term inhalation. Value corresponds to solubility class having lowest value.
[2] Includes the dose from the decay products originating from the 228Ra or 227Ac in the body.
[3] Required method uncertainty values are based on a sampled aerosol volume of 1,600 m3 at the 10~4 and
10~6 risk DAC, respectively. The required relative method uncertainty (cpMR) for values greater than the
AALs in the table is obtained by dividing the u^ value by the corresponding AAL value.
[4] All nuclides can be collected on a fibrous or membrane air filter media except 3H, 1251,129I, and 131I in the
vapor states.
32
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE A7 - Alpha and Beta/Gamma-Emitting Radionuclide Concentrations in Soil and
Required Method Uncertainties Corresponding to Derived Soil Concentrations
Table to be determined following publication o/"Radiological Laboratory Sample Analysis Guide
for Incidents of National Significance-Radionuclides in Soil.
33
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Appendix B:
Examples of the Method Validation Process for
Required Method Uncertainty Specifications
Two examples are provided to demonstrate the method validation process when the MQO involves
a required method uncertainty (WMR or MR (± 39%) of known value > AAL.
Test levels (Table 2): Lower (0.5 x AAL = 200 pCi/L; Mid (AAL) = 400 pCi/L; Upper (3 x AAL)
= 1,200 pCi/L
8 EPA 2008. Radiological Laboratory Sample Analysis Guide for Incidents of National Significance - Radionuclides
in Water, Table 9A.
34
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Data Evaluation:
TABLE Bl - Required Method Uncertainty for Am-241 in Potable Water
Test
Sample
1
2
3
4
5
6
7
Lower Test Level
Concentration
200 pCi/L
Acceptable Range:
50 to 350 pCi/L
Measured
Value
± 1 CSU[3]
221 ±27
179 ± 24
210 ±26
190 ±25
169 ±25
225 ± 27
213 ±26
Acceptable
Value
(Y/N)
Y
Y
Y
Y
Y
Y
Y
Mid Test Level
Concentration'11
400 pCi/L[2]
Acceptable Range:
250 to 550 pCi/L
Measured
Value
± 1 CSUP1
429 ± 40
381 ±37
405 ± 39
304 ± 32
362 ±36
458 ± 42
390 ±38
Acceptable
Value
(Y/N)
Y
Y
Y
Y
Y
Y
Y
Upper Test Level
Concentration
1,200 pCi/L
Acceptable Range:
732 to 1,670 pCi/L
Measured
Value
± 1 CSU[3]
1,283 ± 87
1,117±78
1,241 ± 85
1,159 ±80
1,262 ± 86
1,138 ±79
994 ± 72
Acceptable
Value (Y/N)
Y
Y
Y
Y
Y
Y
Y
Notes:
[1] Mid test level is at the AAL (Table 2)
[2] AAL taken from Table 9A, Radiological Laboratory Sample Analysis Guide for Incidents of National
Significance-Radionuclides in Water (EPA 2008).
[3] Approximate combined standard uncertainty for a 100-minute count on an alpha detector having atypical
detector efficiency plus another 5 % uncertainty for other method parameters at the action level. Counting
time was estimated so that the required method uncertainty would be met at the AAL. All samples would
be counted for the same length of time regardless of the test level.
Example 2. Method Validation for Am-241 in Street Runoff Water - Default AAL and
Required Method Uncertainty
Nuclide. 241Am
Matrix. Street runoff water
Default AAL 40 pCi/sample (Table 1, liquid, specific nuclide)
Proposed Method. Radiochemistry with alpha spectrometry. Specific nuclide measurement.
Required Method Validation Level. D, new matrix
Required Method Uncertainty. 5.2 pCi/test sample at AAL or below; 13% above AAL
(Table 4)
Acceptance Criteria (Table 2): Measured Value within ±3 WMR (±15.6 pCi/sample) of known value
MR (± 39%) of known value > AAL
35
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Method Validation Test Levels (Table 2): Lower (0.5 x AAL) = 20 pCi/sample; Mid (AAL) = 40
pCi/sample; Upper (3 x AAL) =120 pCi/sample
Data Evaluation:
TABLE B2 - Required Method Uncertainty for Am-241 in Street Runoff Water
Test
Sample
1
2
3
4
5
6
7
Lower Test Level Concentration
20 pCi/sample
Acceptable Range:
4.4 to 35.6 pCi
Measured Value
± 1 CSU[3]
22.3 ±2.3
17.6 ±2.0
20.9 ±2.2
23.4 ±2.4
15.8±1.9
21.7 ±2.2
18.8±2.1
Acceptable
Value (Y/N)
Y
Y
Y
Y
Y
Y
Y
Mid Test Level Concentration'11
40 pCi/samplepl
Acceptable Range:
24.4 to 55.6 pCi
Measured Value
± 1 CSU[3]
44.2 ±3.6
36.7 ±3.2
42.4 ±3. 5
38.1 ±3.2
50.5 ±3. 9
41. 5 ±3.4
31.1±2.8
Acceptable
Value (Y/N)
Y
Y
Y
Y
Y
Y
Y
Upper Test Level Concentration
120 pCi/sample
Acceptable Range:
73.2 to 167 pCi
Measured
Value
± 1 CSU[3]
128.6 ±8.0
112.2 ±7.2
124.7 ±7.8
117.0 ±7.4
140.0 ±8.6
122.0 ±7.7
113.4 ±7.2
Acceptable
Value (Y/N)
Y
Y
Y
Y
Y
Y
Y
Notes:
[1] Mid test level is at the AAL, (Table 2)
[2] Table 1, liquid, specific nuclide
[3 ] Approximate combined standard uncertainty for a 15 -minute count on an alpha detector having a typical
detector efficiency plus another 5 % uncertainty for other method parameters at the action level. Counting
time was estimated so that the required method uncertainty would be met at the AAL. All samples would
be counted for the same length of time regardless of the test level. Sample volume ~ 100 mL.
36
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Appendix C:
Example of the Method Validation Process
for Verification of the Required MDC Specification
Refer to Section 5.5.2 (page 21) for the protocol to follow for verifying that a method's MDC meets
the required MDC specification.
Nuclide. 90Sr
Matrix. Street runoff water
Required MDC = 2 pCi/L (MQO designated by Incident Commander)
Proposed Method. Radiochemistry with beta counting on gas proportional counter. Sample volume
= 1 L, counting time = 240 minutes. Analytical result calculations to include detector efficiency,
detector background (cpm) and 90Y ingrowth factor.
Number of Blanks. 1
Number of Spiked Test Samples. 10
Testing Level. 2 pCi/L of 90Sr
Calculations:
a) 90Sr concentration and associated combined standard uncertainty for the blanks and test samples.
b) Critical Net Concentration = 1.94 x standard deviation of the seven blank results.
c) Number (Y) of sample results at or below the estimated Critical Net Concentration
Test. Does the number (Y) of sample results at or below the estimated Critical Net Concentration
exceed 2?
• If Y < 2, the method tested at the required MDC passes the test for the required MDC
specification.
• If Y > 2, the method tested at the required MDC fails the test for the required MDC specification.
TABLE Cl - Results of Blank Sample Analyses
Blank Number
1
2
3
4
5
6
7
Average
Standard Deviation of Results
Critical Net Concentration
Result (pCi/L)
-0.21 ±0.44
0.10 ±0.45
0.44 ±0.46
0.82 ±0.46
-0.40 ±0.44
-0.75 ±0.44
0.61 ±0.46
0.09
0.57
1.11
37
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE C2 - Results of MDC Test Sample Analyses; Test Concentration = 2.0 pCi/L
Test Sample Number
1
2
3
4
5
6
7
8
9
10
Average
Standard Deviation of Results
Y - Number of Results < Critical Net
Concentration
Result (pCi/L)
2.57 ±0.50
1.00 ±0.47
2.43 ±0.50
1.57 ±0.48
2.29 ±0.50
1.71 ±0.48
2.01 ±0.49
3. 14 ±0.52
0.86 ±0.46
1.43 ±0.48
1.90
0.72
Result < Critical Net
Concentration (1.11 pCi/L)
N
Y
N
N
N
N
N
N
Y
N
2
Conclusion: The hypothesis that the true MDC for the method is at or below the required MDC
cannot be rejected. Therefore, the method is assumed to be capable of meeting the required MDC
specification.
38
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Appendix D:
Example of the Effect of Bias on the Probability of Failing the Method Validation
Acceptance Criteria for Required Method Uncertainty
Suppose one is validating a method for water using Level D acceptance criteria, so tests should be
made at three concentration levels with seven samples at each level. Consider that the action level
is 100 pCi/L, and that this is one of the test levels. Also suppose that the required method uncertainty
at 100 pCi/L is uUR = 10 pCi/L, i.e. the relative required method uncertainty is
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Figure D2 shows the overall
probability of failing the Level D
validation as a function of bias and
relative method uncertainty. It is clear
that if the method just meets the
required relative method uncertainty,
then there is not much room to
accommodate bias. However, when
the relative method uncertainty is
7.5%, about an equal amount of bias
might be tolerated. If the relative
method uncertainty is half that
required, three times as much bias can
exist without bringing the probability
of failure over 5%.
• When the actual relative method
uncertainty is one-half or less than
the relative required method un-
certainty, biases up to 15% may
be tolerated without substantially
increasing the probability of
failing the acceptance criteria.
Bias (%)
Figure D2 - Level D validation (21 samples) failing at test
level as a function of relative method bias for relative
method uncertainties of 5%, 7.5%, 10%, and 12.5%.
• When the actual relative method uncertainty is equal to the relative required method uncertainty,
it is best not to have a bias in order to maintain a reasonable probability of passing the acceptance
criteria.
• When the actual relative method uncertainty is greater than the relative required method
uncertainty, the probability of failing the acceptance criteria is extremely high, regardless of the
magnitude of the bias.
Detecting Bias
Testing for a bias smaller than the standard uncertainty is difficult. There must be at least 16 replicate
measurements to make the minimum detectable bias (MDB) less than a, the true method standard
deviation, and at least 54 replicate measurements to make the MDB less than oil (see MARLAP
Chapter 6A). This problem seems to be inescapable for absolute bias tests based on method-blank
analyses. For relative bias tests based on spiked samples, statistics can be improved if there are high-
activity samples whose reference values have small uncertainties.
The mean squared error (MSB) is the sum of the squared differences between the measurements and
the true values. The MSB is the sum of the variance, a2, and the square of the bias, b2. The root MSB
= -vcr2 + b2 . If the root mean squared error is kept below the required method uncertainty, the
MQOs are likely to be met. If the bias is less than one third the relative method standard uncertainty,
bias will only contribute 10% to the MSB.
40
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
The validation criteria in MARLAP were developed with the presumption that any known biases in
the method will be corrected, and that any remaining bias will be small compared to the method
uncertainty. Thus, the primary focus was placed on detecting an unacceptably high method
uncertainty. Note that if the reverse is true, namely that the method uncertainty is much smaller than
the bias, the method may pass the acceptance criteria while having what might be considered
unacceptably high bias. In the extreme case of zero method uncertainty, a method with bias up to
30% might still pass the criteria. In this case different validation criteria may be desirable. Appendix
E contains alternative method validation criteria which treat the detection of excessive bias and
imprecision more equally using the concept of MSB.
41
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Appendix E:
An Alternative Method Validation Criterion
Introduction
The method validation process and acceptance criteria described in Section 5.4 are based on the
criteria recommended in Chapter 6 of the Multi-Agency Radiological Laboratory Analytical
Protocols Manual (MARLAP). This appendix presents alternative method performance acceptance
criteria that may have greater power to detect large imprecision and bias in some situations.
However, the number of test levels and replicates for the appropriate method validation level (B, C,
D, E) as presented in Section 5.4.2 are still to be used.
Every measurement process involves both bias and imprecision to some degree. The MARLAP
method validation criterion is predicated on the assumption that the laboratory has eliminated any
substantial bias in the measurement process, so that measurement results are likely to be evenly
distributed about the true value. If this assumption is not true, use of the MARLAP test alone may
in some cases allow a method with a substantial bias to be accepted for use. Although MARLAP
recommends that the candidate method be evaluated for bias, it does not recommend an objective
criterion for determining whether a detected bias is tolerable. Furthermore, as noted in MARLAP
and in Appendix D of this document, testing for bias tends to be difficult in any case, because of the
number of measurements required to detect a bias that is comparable in magnitude to the standard
deviation.
The assumption of this appendix is that a measurement process may be considered adequate for its
intended use if a certain combination of bias and imprecision, called the "root mean squared error,"
does not exceed the required uncertainty. According to this view, the fact that bias is hard to quantify
is less troublesome, because what one cares about most is not bias alone or imprecision alone but
a combination of the two.
Definitions
/^ /v /v 2
Suppose X is an estimator for some parameter^. The variance of X, denoted by V(X~) or O'% ,
/v
is defined as the expected value of the square of the deviation of X from its mean.
<4 = E[(X- E(X))2} (El)
where £'(•) denotes the expected value (mean) of the operand within the brackets or parentheses. The
square root of the variance, denoted by
-------
_ Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities _
,«, ^
The bias of X is defined as the difference between the expected value (or mean) of X and the
value of the parameter^. In symbols,
K (E3)
^ ~ ~
The bias of X also equals its mean error. So Bias(X) = E(X — K) .
/^ /\
If X is an unbiased estimator (i.e., if E(X) = K\ then the standard deviation is a good measure of
the overall quality of X as an estimator. However, in the context of laboratory analyses, the
estimator X is typically the result of a measurement made using a specified method and
measurement process, and in this situation, X is usually biased to some extent. It is common to
evaluate a laboratory method or measurement process in terms of both the bias and imprecision
(standard deviation) of the estimator x. During method validation, separate limits may be set for the
maximum allowable bias and for the maximum allowable standard deviation; however, since the
overall quality of a measurement process is affected by both bias and imprecision, one may instead
choose to specify a limit for some combination of the bias and imprecision. If this is done, then a
biased but precise method may be considered to be as good as an essentially unbiased but less precise
method.
Note that although neither the bias nor the standard deviation is ever known exactly, it is possible
to use statistical methods to test hypotheses about their magnitudes or to determine likely bounds for
their values. Note also that acknowledging the existence of bias in a measurement process does not
mean that one should cease trying to find and eliminate the causes of any significant bias.
The "mean squared error" or the "root mean squared error" of an estimator is often used as a measure
/v
of the estimator's overall quality. The mean squared error of X , as the name implies, is the
s\.
expected value of the squared error of X . So:
MSE(X) = E[(X -K)2] (E4)
/^ 9
Notice that the definition of A/fSE(X) resembles that of the variance &£ , but with K substituted
^ ^
for the mean E(X~) . It can be shown mathematically that the mean squared error ofX is equal to
the sum of its squared bias and its variance.
MSE(X) = Bias(X)2 + cr2- (E5)
-A
^ A.
The root mean squared error ofX is simply the positive square root of MSE(X~) .
^MSE(X) = ^E[(X-Kf] = Bias(X)2 + a\ (E6)
So the root mean squared error can be viewed as a mathematical combination of bias and
imprecision. For an unbiased estimator, the root mean squared error is exactly equal to the standard
deviation, but for a biased estimator, the root mean squared error is always larger than the standard
deviation.
43
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
The approach to method validation described in this document is based on the concept of a required
uncertainty at each activity level. If one interprets this required uncertainty, w, as a required bound
for JMSE\ X\ , then an unbiased method can have a standard deviation 0% as large as w, or a
perfectly precise method can have a bias as large as u. In general, both the bias and standard
deviation may be nonzero, but in principle, neither the bias nor the standard deviation is allowed to
exceed the required uncertainty, w, at any level of activity.
Alternative Method Validation Criterion
The validation procedure of Section 5.4 involves making several measurements of samples spiked
at known activity levels. Let L denote the number of activity levels and N the number of
measurements made at each level. Then the test described in Section 5.4 compares each result JQ
(where /' denotes the number of the activity level andy denotes the number of the measurement) to
acceptance limits:
Kt ± kut (E7)
where
Kt = target value at the /'th activity level (1 < i < L)
k = uncertainty multiplier (from Table 3)
ut = required uncertainty at the /'th analyte level
The method is judged acceptable if every result^ falls within the appropriate acceptance limits for
its activity level.
As noted under Table 3, the uncertainty multiplier, k, may be calculated as follows:
where a is the chosen significance level, or the probability of a false rejection (a = 0.05), and for any
p, zp denotes the/>-quantile of the standard normal distribution. (Note that &is rounded to two figures
in Table 3 .) The multiplier k also equals the square root of the (1 - a)llm -quantile of the chi-squared
distribution with one degree of freedom, and for purposes of exposition, it will be convenient to use
the latter interpretation here.
The required uncertainty, ut, at each activity level equals the required method uncertainty, WMR, if Kt <
AL, and it equals (pMRKf ifKf > AL.
„={"-' ifK'~AL (E9)
' \<,Kl,ifKl>AL
44
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
A more traditional presentation of the same statistical test would define a test statistic and a critical
value for that statistic. For this test, the statistic can be defined as:
M= maxZ^ (E10)
l
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
TABLE E3 - Method Validation Results, Alternative Test (WTesi)
Measurement
(/)
1
2
3
4
5
6
7
Activity Level (i)
1
K{ = 50 pCi/L
H; = 10.0 pCi/L
^
36.1
39
42.2
44.4
47.5
40.2
44
^• = 2>2 =
r - £,
v y i
«,
-1.39
-1.10
-0.78
-0.56
-0.25
-0.98
-0.60
5.45
2
^ = 100 pCi/L
H; = 10.0 pCi/L
^
83.2
83.7
84.8
75.4
82.3
94.7
88.4
^ = M =
X,J~K.
Z
u,
-1.68
-1.63
-1.52
-2.46
-1.77
-0.53
-1.16
18.6
3
Ki = 300 pCi/L
M; = 30.0 pCi/L
*,
256.1
235.2
249
258.5
265.2
255.7
254.5
^• = 2>2 =
X -K
7 U i
ui
-1.4633
-2.1600
-1.7000
-1.3833
-1.1600
-1.4767
-1.5167
17.4
Because the statistics W2 and W3 both exceed the critical value wc = 17.1, the method is judged to
be unacceptable.
Theoretical Comparison of Statistical Power
The following set of four figures graphically illustrates the power of the MARLAP test and the W
test for the same conditions assumed in Figure D2 of Appendix D. The power of the Holst-Thyregod
(H-T) test is also graphed for comparison. The scenario (as above) involves a Level D validation of
a method for a project where the required method uncertainty is 10 pCi/L at an action level of
100 pCi/L. Each of the following figures assumes a different value for the ratio of the relative
standard deviation (RSD) to the required uncertainty at each activity level. In Figure Ela, the ratio
is 0.5, so that the RSD at the action level equals 5 %. The ratios for Figures Elb, Elc, and Eld are
0.75, 1, and 1.25, respectively. In each graph, the horizontal axis represents possible values for the
relative bias of the method ranging from 0 to 20 %. The vertical axis, labeled P, represents the
probability that a method with the given relative standard deviation and relative bias will be rejected.
In every case, the Wtest outperforms the MARLAP test, although the differences are most noticeable
when the precision of the method is good but the bias is large. Also note that the power of the Holst-
Thyregod test exceeds that of the Wtest in Figures Ela and Elb but not in Elc and Eld.
Reference
Hoist, Erik and Poul Thyregod. 1999. "A statistical test for the mean squared error," Journal of
Statistical Computation and Simulation, 63:4, 321-347.
48
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
5 % RSD at 100 pCi/L
5 10 15
Relative Bias (%)
20
MARLAP
W
H-T
Figure E1a
RSD = 7.5 %
5 10 15
Relative Bias (%)
20
•MARLAP
•W
•H-T
Figure E1b
49
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
0
RSD = 10 %
•MARLAP
•W
•H-T
5 10 15 20
Relative Bias (%)
Figure E1c
RSD = 12.5%
5 10 15
Relative Bias (%)
•MARLAP
•W
•H-T
20
Figure E1d
50
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
Appendix F:
Glossary
accuracy: The closeness of a measured result to the true value of the quantity being measured.
Various recognized authorities have given the word "accuracy" different technical definitions,
expressed in terms of bias and imprecision. Following MARLAP, this document avoids all of
these technical definitions and uses the term "accuracy" in its common, ordinary sense.
aerosol: A suspension of fine solid or liquid particles within a gaseous matrix (usually air).
aliquant: A representative portion of a homogeneous sample removed for the purpose of analysis
or other chemical treatment. The quantity removed is not an evenly divisible part of the whole
sample. An aliquot, by contrast, is an evenly divisible part of the whole.
analyte: For this document, an analyte is a specific radionuclide or a category of radionuclides that
comprise gross alpha or beta analyses. An analyte may be on the list of radionuclides of interest
or a radionuclide of concern for a project. See target analyte.
analyte concentration range: (1) Method validation definition - the radionuclide concentration
range corresponding to three test levels (lower, mid and upper) that are used during method
validation. The mid level concentration corresponds to the action level. (2) MQO definition -
the expected concentration range (minimum to maximum) of an analyte expected to be present
in a sample for a given proj ect. While most analytical protocols are applicable over a fairly large
range of concentration for the radionuclide of interest, performance over a required concen-
tration range can serve as a measurement quality objective for the protocol selection process, and
some analytical protocols may be eliminated if they cannot accommodate the expected range of
concentration.
analytical action level (AAL): The value of a quantity that will cause the decision maker to choose
one of the alternative actions. The action level may be a derived concentration level (such as the
derived water concentration in this document), background level, release criteria, regulatory
decision limit, etc. The AAL is often associated with the type of media, target analyte, and
concentration limit. Some AALs are expressed in terms of a derived radionuclide concentration
corresponding to a dose or risk, such as a protective action guide. MARLAP uses the term
"action level."
analytical decision level (ADL). The minimum measured value for the radionuclide concentration
in a sample that indicates the amount of radionuclide present is equal to or greater than the
analytical action level at a specified Type II error rate. (Assumes that method uncertainty
requirements have been met.) Any measurement result equal to or greater than the applicable
ADL is considered to have exceeded the corresponding analytical action level. MARLAP uses
the term "critical level."
analytical protocol specification (APS): The output of a directed planning process that contains the
project's analytical data needs and requirements in an organized, concise form. The level of
51
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
specificity in the APSs should be limited to those requirements that are considered essential to
meeting the project's analytical data requirements to allow the laboratory the flexibility of
selecting the protocols or methods that meet the analytical requirements.
background (instrument): Radiation detected by an instrument when no source is present. The
background radiation that is detected may come from radionuclides in the materials of construc-
tion of the detector, its housing, its electronics, and the building, as well as the environment and
natural radiation.
background level: A term that usually refers to the presence of radioactivity or radiation in the
environment. From an analytical perspective, the presence of background radioactivity in
samples needs to be considered when clarifying the radioanalytical aspects of the decision or
study question. Many radionuclides are present in measurable quantities in the environment.
bias (of a measurement process): A persistent deviation of the mean measured result from the true
or accepted reference value of the quantity being measured, which does not vary if a measure-
ment is repeated.
blank (analytical or method): A sample that is assumed to be essentially free of the target analyte
(the "unknown"), which is carried through the radiochemical preparation, analysis, mounting,
and measurement process in the same manner as a routine sample of a given matrix.
calibration: The set of operations that establish, under specified conditions, the relationship between
values indicated by a measuring instrument or measuring system, or values represented by a
material measure, and the corresponding known value of a parameter of interest.
calibration source: A prepared source, made from a certified reference material (standard), that is
used for calibrating instruments.
carrier: (1) A stable isotopic form of a tracer element or nonisotopic material added to effectively
increase the quantity of a tracer element during radiochemical procedures, ensuring conventional
behavior of the element in solution. (2) A substance in appreciable amount that, when associated
with a tracer of a specified substance, will carry the tracer with it through a chemical or physical
process, or prevent the tracer from undergoing non-specific processes due to its low
concentration (IUPAC, 1995). A stable isotope of a radionuclide (usually the analyte) added to
increase the total amount of that element so that a measurable mass of the element is present.
chain of custody: Procedures that provide the means to trace the possession and handling of a
sample from collection to data reporting.
check source: A material used to validate the operability of a radiation measurement device,
sometimes used for instrument quality control. See source, radioactive.
52
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
combined standard uncertainty: Standard uncertainty of an output estimate calculated by com-
bining the standard uncertainties of the input estimates. The combined standard uncertainty of
y is denoted by uc(y). See uncertainty (of measurement).
critical level. Termed analytical decision level in this document in the context of evaluating sample
results relative to an analytical action level. In the context of analyte detection, critical level
means the minimum measured value (e.g., of the instrument signal or the radionuclide concentra-
tion) that indicates a positive (nonzero) amount of a radionuclide is present in the material within
a specified probable error. The critical level is sometimes called the critical value or decision
level.
critical net concentration. Similar in concept as the "critical level."
data quality objective (DQO) Qualitative and quantitative statements that clarify the study
objectives, define the most appropriate type of data to collect, determine the most appropriate
conditions from which to collect the data, and specify tolerable limits on decision error rates.
Because DQOs will be used to establish the quality and quantity of data needed to support
decisions, they should encompass the total uncertainty resulting from all data collection
activities, including analytical and sampling activities.
default AAL test level: Radionuclide test concentration for a given general matrix category to be
used in the method validation process in the absences of PAG or risk-based AALs.
derived air concentration (DA C): The concentration of a radionuclide, in pCi/m3, that would result
in exposure to a specified dose level. Generally refers to a protective action guide or other
specified dose- or risk-based factor related to an analytical action level. In this document, for
example, the "500-mrem DAC for 239Pu" is the concentration of 239Pu, in pCi/m3, that would
result in an exposure of 500 mrem and would refer to the 500-mrem PAG. The DAC is
radionuclide-specific.
derived radionuclide concentration (DRC): General application term used in discussions involving
both of the terms DAC and DWC.
derived water concentration (DWC): The concentration of a radionuclide, in pCi/L, that would
result in exposure to a specified dose level. Generally refers to ^.protective action guide or other
specified dose- or risk-based factor related to an analytical action level.
detection capability: The capability of a measurement process to distinguish small amounts of
analyte from zero.
detection limit: The smallest value of the amount or concentration of analyte that ensures a specified
high probability of detection. Also called "minimum detectable value."
53
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
discrimination limit (DL). The DL is the point where it is important to be able to distinguish
expected signal from the analytical action level. The boundaries of the gray region.
dose equivalent: Quantity that expresses all radiations on a common scale for calculating the
effective absorbed dose. This quantity is the product of absorbed dose (grays (Gy) or rads)
multiplied by a quality factor and any other modifying factors (MARSSEVI, 2000). The quality
factor adjusts the absorbed dose because not all types of ionizing radiation create the same effect
on human tissue. For example, a dose equivalent of one sievert (Sv) requires 1 Gy of beta or
gamma radiation, but only 0.05 Gy of alpha radiation or 0.1 Gy of neutron radiation. Because the
sievert is a large unit, radiation doses often are expressed in millisieverts (mSv). See total
effective dose equivalent and roentgen.
gray (Gy): The International System of Units (SI) unit for absorbed radiation dose. One gray is 1
joule of energy absorbed per kilogram of matter, equal to 100 rad. See sievert.
gray region: The range of possible values in which the consequences of decision errors are relatively
minor. Specifying a gray region is necessary because variability in the analyte in a population and
imprecision in the measurement system combine to produce variability in the data such that the
decision may be "too close to call" when the true value is very near the analytical action level.
The gray region establishes the minimum distance from the analytical action level where it is
most important to control Type II decision errors.
hypothesis testing: The use of statistical procedures to decide whether a null hypothesis should be
rejected in favor of an alternative hypothesis or not rejected.
incident response method validation: Proj ect method validation for incident response applications.
See project method validation and method validation.
interferences: The presence of other chemicals or radionuclides in a sample that hinder the ability
to analyze for the radionuclide of interest.
MARLAP Process: A performance-based approach that develops Analytical Protocol Specifications,
and uses these requirements as criteria for the analytical protocol selection, development, and
evaluation processes, and as criteria for the evaluation of the resulting laboratory data. This
process, which spans the three phases of the data life cycle for a proj ect, is the basis for achieving
MARLAP's basic goal of ensuring that radioanalytical data will meet a project's or program's
data requirements or needs.
measurand: "Particular quantity subject to measurement" (ISO, 1993a).
measurement quality objective (MQO): The analytical data requirements of the data quality
objectives, which are project- or program-specific and can be quantitative or qualitative. These
analytical data requirements serve as measurement performance criteria or objectives of the
analytical process. MARLAP refers to these performance objectives as MQOs. Examples of
quantitative MQOs include statements of required analyte detectability and the uncertainty of
54
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
the analytical protocol at a specified radionuclide concentration, such as the action level.
Examples of qualitative MQOs include statements of the required specificity of the analytical
protocol (e.g., the ability to analyze for the radionuclide of interest [or target analyte] given the
presence of interferences).
measurement uncertainty: See uncertainty.
method blank: A sample assumed to be essentially target analyte-free that is carried through the
radiochemical preparation, analysis, mounting and measurement process in the same manner as
a routine sample of a given matrix.
method performance characteristics: The characteristics of a specific analytical method such as
method uncertainty., method range., method specificity, and method ruggedness. MARLAP
recommends developing measurement quality objectives for select method performance
characteristics., particularly for the uncertainty (of measurement) at a specified concentration
(typically the action level).
method ruggedness: The relative stability of method performance for small variations in method
parameter values.
method specificity: The ability of the method to measure the analyte of concern in the presence of
interferences.
method uncertainty: Refers to the predicted uncertainty of the result that would be measured if the
method were applied to a hypothetical laboratory sample with a specified analyte concentration.
Although individual measurement uncertainties will vary from one measured result to another,
the required method uncertainty is a target value for the individual measurement uncertainties
and is an estimate of uncertainty before the sample is actually measured.
method validation (MV): The demonstration thatthe method selected for the analysis of a particular
analyte in a given matrix is capable of providing analytical results to meet the proj ecf s measure-
ment quality objectives and any other requirements in the analytical protocol specifications.
minimum detectable concentration (MDC): An estimate of the smallest true value of the analyte
concentration that gives a specified high probability of detection.
nuclide-specific analysis: Radiochemical analysis performed to isolate and measure a specific
radionuclide.
null hypothesis (H^: One of two mutually exclusive statements tested in a statistical hypothesis test
(compare with alternative hypothesis). The null hypothesis is presumed to be true unless the test
provides sufficient evidence to the contrary, in which case the null hypothesis is rejected and the
alternative hypothesis (Hj) is accepted.
55
-------
_ Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities _
performance evaluation (PE) program: A laboratory's participation in an internal or external
program of analyzing proficiency-testing samples appropriate for the analytes and matrices under
consideration (i.e., PE program traceable to a national standards body, such as NIST). Reference-
material samples used to evaluate the performance of the laboratory may be called performance-
evaluation, performance or proficiency-testing samples or materials. See proficiency test
samples.
precision: The closeness of agreement between independent test results obtained by applying the
experimental procedure under stipulated conditions. Precision may be expressed as the standard
deviation. Conversely, imprecision is the variation of the results in a set of replicate
measurements.
proficiency test (PT) samples: Samples having a known radionuclide concentration used in a PE
program or internally at the laboratory for method validation and for the measurement of bias.
project method validation: The demonstrated method applicability for a particular project.
protective action guide (PAG). The radiation dose to individuals in the general population that
warrants protective action following a radiological event. In this document, PAGs limit the
proj ected radiation doses for different exposure periods: not to exceed 2-rem total effective dose
equivalent (TEDE) during the first year, 500-mrem TEDE during the second year, or 5 rem over
the next 50 years (including the first and second years of the incident). See derived water
concentration and analytical action level.
quality control (QC): The overall system of technical activities that measures the attributes and
performance of a process, item, or service against defined standards to verify that they meet the
stated requirements established by the proj ect; operational techniques and activities that are used
to fulfill requirements for quality. This system of activities and checks is used to ensure that
measurement systems are maintained within prescribed limits, providing protection against out-
of-control conditions and ensuring that the results are of acceptable quality.
radiochemical analysis: The analysis of a sample matrix for its radionuclide content, both
qualitatively and quantitatively.
radionuclide: A nuclide that is radioactive (capable of undergoing radioactive decay).
relative required method uncertainty ((p^): The required method uncertainty divided by the
analytical action level. The relative required method uncertainty is applied to radionuclide
concentrations above the analytical action level. A key measurement quality objective.
rem: The common unit for the effective or equivalent dose of radiation received by a living
organism, equal to the actual dose (in rads) multiplied by a factor representing the danger of the
radiation. Rem is an abbreviation for "roentgen equivalent man," meaning that it measures the
biological effects of ionizing radiation in humans. One rem is equal to 0.01 Sv. See sievert.
56
-------
Method Validation Guide for Radiological Laboratories Participating in Incident Response Activities
replicates: Two or more aliquants of a homogeneous sample whose independent measurements are
used to determine the precision of laboratory preparation and analytical procedures.
required method uncertainty (UMB): Method uncertainty at a specified concentration. A key
measurement quality objective. See relative required method uncertainty.
required minimum detectable concentration (RMDC): An upper limit for the minimum detectable
concentration required by some projects.
sample: (1) A portion of material selected from a larger quantity of material. (2) A set of individual
samples or measurements drawn from a population whose properties are studied to gain informa-
tion about the entire population.
screening method: An economical gross measurement (alpha, beta, gamma) used in a tiered
approach to method selection that can be applied to analyte concentrations below an analyte
level in the analytical protocol specifications or below a fraction of the specified action level.
sievert (Sv): The SI unit for the effective dose of radiation received by a living organism. It is the
actual dose received (grays in SI or rads in traditional units) times a factor that is larger for more
dangerous forms of radiation. One Sv is 100 rem. Radiation doses are often measured in mSv.
An effective dose of 1 Sv requires 1 gray of beta or gamma radiation, but only 0.05 Gy of alpha
radiation or 0.1 Gy of neutron radiation.
swipe: A filter pad used to determine the level of general radioactive contamination when it is wiped
over a specific area, about 100 cm2 in area. Also called "smear" or "wipe."
target analyte: A radionuclide on the list of radionuclides of interest or a radionuclide of concern
for a project. For incident response applications, typical radionuclides of interest are provided
in Appendix A.
total effective dose equivalent: The sum of the effective dose equivalent (for external exposure) and
the committed effective dose equivalent (for internal exposure), expressed in units of Sv or rem.
Type I decision error: In a hypothesis test, the error made by rejecting the null hypothesis when it
is true. A Type I decision error is sometimes called a "false rejection" or a "false positive."
Type II decision error: In a hypothesis test, the error made by failing to reject the null hypothesis
when it is false. A Type II decision error is sometimes called a "false acceptance" or a "false
negative."
uncertainty: A parameter, usually associated with the result of a measurement, that characterizes the
dispersion of the values that could reasonably be attributed to the measurand.
57
------- |