Clsi Ila21 A

I/LA21-A Replaces I/LA21-P
Vol. 22 No. 9 Vol. 19 No. 19

Clinical Evaluation of Immunoassays; Approved Guideline
This document addresses the need for clinical evaluation of new immunoassays and new applications of
existing assays. As a guide to designing and executing a clinical evaluation, this document will aid
clinical and regulatory personnel responsible for commercializing products, developers of “in-house”
assays for institutional use, and developers of assays used for monitoring pharmacologic effects of new
drugs or biologics.
A guideline for global application developed through the NCCLS consensus process.
NCCLS...
Serving the World’s Medical Science Community Through Voluntary Consensus
NCCLS is an international, interdisciplinary, nonprofit, the need for field evaluation or data collection, documents
standards-developing, and educational organization that may also be made available for review at an intermediate
promotes the development and use of voluntary (i.e., “tentative”) consensus level.
consensus standards and guidelines within the healthcare
Proposed An NCCLS consensus document undergoes the
community. It is recognized worldwide for the
first stage of review by the healthcare community as a
application of its unique consensus process in the
proposed standard or guideline. The document should
development of standards and guidelines for patient
receive a wide and thorough technical review, including an
testing and related healthcare issues. NCCLS is based on
overall review of its scope, approach, and utility, and a line-
the principle that consensus is an effective and cost-
by-line review of its technical and editorial content.
effective way to improve patient testing and healthcare
services. Tentative A tentative standard or guideline is made
available for review and comment only when a
In addition to developing and promoting the use of
recommended method has a well-defined need for a field
voluntary consensus standards and guidelines, NCCLS
evaluation or when a recommended protocol requires that
provides an open and unbiased forum to address critical
specific data be collected. It should be reviewed to ensure its
issues affecting the quality of patient testing and health
utility.
care.
Approved An approved standard or guideline has achieved
PUBLICATIONS
consensus within the healthcare community. It should be
An NCCLS document is published as a standard, reviewed to assess the utility of the final document, to
guideline, or committee report. ensure attainment of consensus (i.e., that comments on
earlier versions have been satisfactorily addressed), and to
Standard A document developed through the consensus
identify the need for additional consensus documents.
process that clearly identifies specific, essential
requirements for materials, methods, or practices for use NCCLS standards and guidelines represent a consensus
in an unmodified form. A standard may, in addition, opinion on good practices and reflect the substantial
contain discretionary elements, which are clearly agreement by materially affected, competent, and interested
identified. parties obtained by following NCCLS’s established
consensus procedures. Provisions in NCCLS standards and
Guideline A document developed through the
guidelines may be more or less stringent than applicable
consensus process describing criteria for a general
regulations. Consequently, conformance to this voluntary
operating practice, procedure, or material for voluntary
consensus document does not relieve the user of
use. A guideline may be used as written or modified by
responsibility for compliance with applicable regulations.
the user to fit specific needs.
COMMENTS
Report A document that has not been subjected to
consensus review and is released by the Board of The comments of users are essential to the consensus
Directors. process. Anyone may submit a comment, and all comments
are addressed, according to the consensus process, by the
CONSENSUS PROCESS
NCCLS committee that wrote the document. All comments,
The NCCLS voluntary consensus process is a protocol including those that result in a change to the document when
establishing formal criteria for: published at the next consensus level and those that do not
result in a change, are responded to by the committee in an
• the authorization of a project appendix to the document. Readers are strongly encouraged
• the development and open review of documents to comment in any form and at any time on any NCCLS
document. Address comments to the NCCLS Executive
• the revision of documents in response to comments Offices, 940 West Valley Road, Suite 1400, Wayne, PA
by users 19087, USA.
• the acceptance of a document as a consensus VOLUNTEER PARTICIPATION
standard or guideline.
Healthcare professionals in all specialties are urged to
Most NCCLS documents are subject to two levels of volunteer for participation in NCCLS projects. Please
consensus—“proposed” and “approved.” Depending on contact the NCCLS Executive Offices for additional
information on committee participation.
Volume 22 I/LA21-A

Abstract
This document addresses the need for the clinical evaluation of new immunoassays and new applications
of existing assays. Existing NCCLS documents provide guidance for assessing analytical performance,
methods comparison, and clinical accuracy of laboratory tests. However, none of the documents define
the elements that are integral to generating clinical data. As a guide to designing, executing, and
analyzing a clinical evaluation, this document will aid clinical and regulatory personnel responsible for
commercializing products, developers of “in-house” assays for institutional use, and developers of assays
used for monitoring pharmacologic effects of new drugs or biologics.
The elements of this guideline include (1) a brief review of the analytical performance measures that must
be in place prior to testing clinical specimens; (2) a thorough discussion of the planning and design
considerations that are necessary for a successful evaluation; (3) a description of requirements for
conducting the evaluation through monitoring and database management; and (4) a development plan for
an effective analysis and evaluation.
NCCLS. Clinical Evaluation of Immunoassays; Approved Guideline. NCCLS document I/LA21-A (ISBN
1-56238-455-4). NCCLS, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania 19087-1898 USA,
2002.
THE NCCLS consensus process, which is the mechanism for moving a document through two or more
levels of review by the healthcare community, is an ongoing process. Users should expect revised
editions of any given document. Because rapid changes in technology may affect the procedures,
methods, and protocols in a standard or guideline, users should replace outdated editions with the
current editions of NCCLS documents. Current editions are listed in the NCCLS Catalog, which is
distributed to member organizations, and to nonmembers on request. If your organization is not a
member and would like to become one, and to request a copy of the NCCLS Catalog, contact the
NCCLS Executive Offices. Telephone: 610.688.0100; Fax: 610.688.0700; E-Mail: exoffice@nccls.org;
Website: www.nccls.org
i
Number 9 NCCLS
ii
I/LA21-A
ISBN 1-56238-455-4
ISSN 0273-3099
Volume 22 Number 9
Linda Ivor, Chairholder
Peter Maxim, Ph.D.
Donald R. Parker, Ph.D.
Gregory P. Payne, RAC
John Ridderhof, Dr.P.H.
Carol Ryerson, Ph.D.
Lawrence E. Schaefer
Mark H. Zweig, M.D.
Number 9 NCCLS
This publication is protected by copyright. No part of it may be reproduced, stored in a retrieval system,
or transmitted or made available in any form or by any means (electronic, mechanical, photocopying,
recording, or otherwise) without prior written permission from NCCLS, except as stated below.
NCCLS hereby grants permission to reproduce limited portions of this publication for use in laboratory
procedure manuals at a single site, for interlibrary loan, or for use in educational programs provided that
multiple copies of such reproduction shall include the following notice, be distributed without charge,
and, in no event, contain more than 20% of the document’s text.
Reproduced with permission, from NCCLS publication I/LA21-A—Clinical Evaluation

of Immunoassays; Approved Guideline (ISBN 1-56238-455-4). Copies of the current
edition may be obtained from NCCLS, 940 West Valley Road, Suite 1400, Wayne,
Pennsylvania 19087-1898, USA.
Permission to reproduce or otherwise use the text of this document to an extent that exceeds the
exemptions granted here or under the Copyright Law must be obtained from NCCLS by written request.
To request such permission, address inquiries to the Executive Director, NCCLS, 940 West Valley Road,
Suite 1400, Wayne, Pennsylvania 19087-1898, USA.
Copyright ©2002. The National Committee for Clinical Laboratory Standards.
Suggested Citation
(NCCLS. Clinical Evaluation of Immunoassays; Approved Guideline. NCCLS document I/LA21-A

[ISBN 1-56238-455-4]. NCCLS, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania 19087-1898
USA, 2002.)
Proposed Guideline
September 1999
Approved Guideline
June 2002
ISBN 1-56238-455-4
ISSN 0273-3099
iv
Volume 22 I/LA21-A
Committee Membership
Area Committee on Immunology and Ligand Assay
W. Harry Hannon, Ph.D. Centers for Disease Control and Prevention

Chairholder Atlanta, Georgia
Jeffrey Warren, M.D. University of Michigan

Vice-Chairholder Ann Arbor, Michigan
Subcommittee on Clinical Evaluations of Immunoassays
Linda Ivor Gen-Probe Incorporated

Chairholder San Diego, California
Peter Maxim, Ph.D. FDA Center for Devices/Rad. Health

Rockville, Maryland
Donald R. Parker, Ph.D. Bayer Corporation

Elkhart, Indiana
Gregory P. Payne, RAC Hybritech

San Diego, California
John C. Ridderhof, Dr.P.H. Centers for Disease Control and Prevention

Atlanta, Georgia
Carol Ryerson, Ph.D. Eastman Kodak Co.

Rochester, New York
Lawrence E. Schaefer Abbott Laboratories

Abbott Park, Illinois
Mark H. Zweig, M.D. National Institutes of Health

Bethesda, Maryland
Advisors
John A. Boffa, B.L.D. American Association of Bioanalysts

Howell, New Jersey
Prof. Aw Tar Choon National University Hospital

Singapore
Anna J. DeMarinis, M.T.(ASCP) bioMerieux, Inc.

Rockland, Massachusetts
Steven I. Gutman, M.D., MBA FDA Center for Devices/Rad. Health

Rockville, Maryland
Sheri L. Hall Beckman Coulter, Inc.

Brea, California
v
Number 9 NCCLS
Advisors (Continued)
Lorin Hills, Ph.D. LifeScan, Inc.

Milpitas, California
Hassan Khayam-Bashi, Ph.D. San Francisco General Hospital

San Francisco, California
Edward M. Levine, Ph.D. Diagnostic Products Corporation

Los Angeles, California
Terry McGovern Rockland, Massachusetts
Ronald W. McLawhon, M.D., Ph.D. University of Chicago Medical Center

Chicago, Illinois
William F. Skogen, Ph.D. Dade Behring Inc. Glasgow

Newark, Delaware
Leonard Wilson Food and Drug Administration

Rockville, Maryland
Patrice E. Polgar NCCLS

Project Manager/Editor Wayne, Pennsylvania
Donna M. Wilhelm NCCLS

Assistant Editor Wayne, Pennsylvania
vi
Volume 22 I/LA21-A
Active Membership
(as of 1 April 2002)
Sustaining Members College of Medical Laboratory Deutsches Institut für Normung

Technologists of Ontario (DIN)
Abbott Laboratories College of Physicians and FDA Center for Devices and
American Association for Surgeons of Saskatchewan Radiological Health
Clinical Chemistry ESCMID FDA Center for Veterinary
Bayer Corporation Fundación Bioquímica Argentina Medicine
Beckman Coulter, Inc. International Association of Medical FDA Division of Anti-Infective
BD and Company Laboratory Technologists Drug Products
bioMérieux, Inc. International Council for Health Care Financing
CLMA Standardization in Haematology Administration/CLIA Program
College of American Pathologists International Federation of Health Care Financing
GlaxoSmithKline Clinical Chemistry Administration
Nippon Becton Dickinson Co., Ltd. Italian Society of Clinical Iowa State Hygienic Laboratory
Ortho-Clinical Diagnostics, Inc. Biochemistry and Clinical Massachusetts Department of
Roche Diagnostics, Inc. Molecular Biology Public Health Laboratories
Japan Society of Clinical Chemistry National Association of Testing
Professional Members Japanese Committee for Clinical Authorities – Australia
Laboratory Standards National Center of Infectious
AISAR-Associazione Italiana per lo Joint Commission on Accreditation and Parasitic Diseases (Bulgaria)
Studio degli of Healthcare Organizations National Health Laboratory Service
American Academy of Family National Academy of Clinical (South Africa)
Physicians Biochemistry National Institute of Standards
American Association for National Society for and Technology
Clinical Chemistry Histotechnology, Inc. New York State Department of
American Association for Ontario Medical Association Health
Respiratory Care Quality Management Program- Ohio Department of Health
American Chemical Society Laboratory Service Ontario Ministry of Health
American Medical Technologists RCPA Quality Assurance Programs Pennsylvania Dept. of Health
American Public Health Association PTY Limited Saskatchewan Health-Provincial
American Society for Clinical Sociedade Brasileira de Analises Laboratory
Laboratory Science Clinicas Scientific Institute of Public Health;
American Society of Hematology Sociedade Brasileira de Belgium Ministry of Social
American Society for Microbiology Patologia Clinica Affairs, Public Health and the
American Type Culture Sociedad Espanola de Bioquimica Environment
Collection, Inc. Clinica y Patologia Molecular Swedish Institute for Infectious
Asociación Española Primera de Turkish Society of Microbiology Disease Control
Socorros (Uruguay) Thailand Department of Medical
Asociacion Mexicana de Government Members Sciences
Bioquimica Clinica A.C.
Assn. of Public Health Laboratories Association of Public Health Industry Members
Assoc. Micro. Clinici Italiani- Laboratories
A.M.C.L.I. Armed Forces Institute of Pathology AB Biodisk
British Society for Antimicrobial BC Centre for Disease Control Abbott Laboratories
Chemotherapy Centers for Disease Control and Abbott Laboratories, MediSense
CADIME-Camara De Instituciones Prevention Products
De Diagnostico Medico Centers for Medicare & Medicaid Acrometrix Corporation
Canadian Society for Medical Services/CLIA Program Advanced Medicine, Inc.
Laboratory Science—Société Centers for Medicare & Medicaid Agilent Technologies, Inc.
Canadienne de Science de Services Ammirati Regulatory Consulting
Laboratoire Médical Chinese Committee for Clinical Anaerobe Systems
Clinical Laboratory Management Laboratory Standards Asséssor
Association Commonwealth of Pennsylvania AstraZeneca
COLA Bureau of Laboratories Aventis
College of American Pathologists Department of Veterans Affairs Axis-Shield POC AS
Bayer Corporation – Elkhart, IN
vii
Number 9 NCCLS
Bayer Corporation – Tarrytown, NY Essential Therapeutics, Inc. Roche Diagnostics GmbH

Bayer Corporation – West Haven, EXPERTech Associates, Inc. Roche Diagnostics, Inc.
CT F. Hoffman-La Roche AG Roche Laboratories (Div.
Bayer Medical Ltd. Fort Dodge Animal Health Hoffmann-La Roche Inc.)
BD General Hospital Vienna (Austria) Sarstedt, Inc.
BD Biosciences – San Jose, CA Gen-Probe SARL Laboratoire Carron (France)
BD Consumer Products GlaxoSmithKline Schering Corporation
BD Diagnostic Systems Greiner Bio-One Inc. Schleicher & Schuell, Inc.
BD Italia S.P.A. Helena Laboratories Second Opinion
BD VACUTAINER Systems Home Diagnostics, Inc. Showa Yakuhin Kako Company,
Beckman Coulter, Inc. Immunicon Corporation Ltd.
Beckman Coulter, Inc. Primary Care Instrumentation Laboratory Streck Laboratories, Inc.
Diagnostics International Technidyne SurroMed, Inc.
Beckman Coulter K.K. (Japan) Corporation Sysmex Corporation (Japan)
Bio-Development SRL IntraBiotics Pharmaceuticals, Inc. Sysmex Corporation
Bio-Inova Life Sciences I-STAT Corporation (Long Grove, IL)
International Johnson and Johnson Pharmaceutical The Clinical Microbiology Institute
Bio-Inova Life Sciences North Research and Development, The Toledo Hospital (OH)
America L.L.C. Trek Diagnostic Systems, Inc.
BioMedia Laboratories Sdn Bhd Kendall Sherwood-Davis & Geck Versicor, Inc.
BioMérieux (NC) LAB-Interlink, Inc. Vetoquinol S.A.
bioMérieux, Inc. (MO) Laboratory Specialists, Inc. Visible Genetics, Inc.
Biometrology Consultants Labtest Diagnostica S.A. Vysis, Inc.
Bio-Rad Laboratories, Inc. LifeScan, Inc. (a Johnson & Wallac Oy
Bio-Rad Laboratories, Inc. - France Johnson Company) Wyeth-Ayerst
Biotest AG Lilly Research Laboratories Xyletech Systems, Inc.
Blaine Healthcare Associates, Inc. Macemon Consultants YD Consultant
Bristol-Myers Squibb Company Medical Device Consultants, Inc. YD Diagnostics (Seoul, Korea)
Canadian External Quality Merck & Company, Inc.
Assessment Laboratory Minigrip/Zip-Pak Trade Associations
Capital Management Consulting, Molecular Diagnostics, Inc.
Inc. mvi Sciences (MA) AdvaMed
Carl Schaper Nabi Association of Medical
Checkpoint Development Inc. Nichols Institute Diagnostics Diagnostic Manufacturers
Chiron Corporation (Div. of Quest Diagnostics, Inc.) Japan Association Clinical
ChromaVision Medical Systems, Nissui Pharmaceutical Co., Ltd. Reagents Ind. (Tokyo, Japan)
Inc. Nippon Becton Dickinson Co., Ltd. Medical Industry Association
Chronolab Ag Norfolk Associates, Inc. of Australia
Clinical Design Group Inc. Novartis Pharmaceuticals
Clinical Laboratory Improvement Corporation Associate Active Members
Consultants Ortho-Clinical Diagnostics, Inc.
Cognigen (Raritan, NJ) 20th Medical Group (SC)
Community Medical Center (NJ) Ortho-Clinical Diagnostics, Inc. 31st Medical Group/SGSL (APO,
Control Lab (Brazil) (Rochester, NY) AE)
Copan Diagnostics Inc. Oxoid Inc. 67th CSH Wuerzburg, GE (NY)
Cosmetic Ingredient Review Paratek Pharmaceuticals 121st General Hospital (CA)
Cubist Pharmaceuticals Pfizer Inc Academisch Ziekenhuis-VUB
Dade Behring Inc. - Deerfield, IL Pharmacia Corporation (Belgium)
Dade Behring Inc. - Glasgow, DE Powers Consulting Services Acadiana Medical Laboratories,
Dade Behring Inc. - Marburg, Premier Inc. LTD (LA)
Germany Procter & Gamble Adena Regional Medical Center
Dade Behring Inc. - Sacramento, CA Pharmaceuticals, Inc. (OH)
Dade Behring Inc. - San Jose, CA The Product Development Group Advocate Healthcare Lutheran
Diagnostic Products Corporation QSE Consulting General (IL)
Eiken Chemical Company, Ltd. Quintiles, Inc. Akershus Central Hospital and AFA
Electa Lab s.r.l. Radiometer America, Inc. (Norway)
Enterprise Analysis Corporation Radiometer Medical A/S Albemarle Hospital (NC)
David G. Rhoads Associates, Inc. Allegheny General Hospital (PA)
viii
Volume 22 I/LA21-A
Allegheny University of the Clinical Laboratory Partners, LLC Harris Methodist Erath County
Health Sciences (PA) (CT) (TX)
Allina Health System (MN) CLSI Laboratories (PA) Harris Methodist Fort Worth (TX)
Alton Ochsner Medical Columbia Regional Hospital (MO) Hartford Hospital (CT)
Foundation (LA) Commonwealth of Kentucky Headwaters Health Authority
American Medical Laboratories Community Hospital of Lancaster (Alberta, Canada)
(VA) (PA) Health Network Lab (PA)
Arkansas Department of Health CompuNet Clinical Laboratories Health Partners Laboratories (VA)
ARUP at University Hospital (UT) (OH) Heartland Regional Medical Center
Armed Forces Research Institute of Cook Children’s Medical Center (MO)
Medical Science (APO, AP) (TX) Highlands Regional Medical Center
Associated Regional & Covance Central Laboratory (FL)
University Pathologists (UT) Services (IN) Hoag Memorial Hospital
Aurora Consolidated Danish Veterinary Laboratory Presbyterian (CA)
Laboratories (WI) (Denmark) Holmes Regional Medical Center
Azienda Ospedale Di Lecco (Italy) Danville Regional Medical Center (FL)
Bay Medical Center (MI) (VA) Holzer Medical Center (OH)
Baystate Medical Center (MA) Delaware Public Health Laboratory Hospital for Sick Children
Bbaguas Duzen Laboratories Department of Health & (Toronto, ON, Canada)
(Turkey) Community Services (New Hospital Sousa Martins (Portugal)
Bermuda Hospitals Board Brunswick, Canada) Hotel Dieu Hospital (Windsor, ON,
Bo Ali Hospital (Iran) DesPeres Hospital (MO) Canada)
British Columbia Cancer Agency DeTar Hospital (TX) Houston Medical Center (GA)
(Vancouver, BC, Canada) Detroit Health Department (MI) Huddinge University Hospital
Broward General Medical Center Dr. Everett Chalmers Hospital (Sweden)
(FL) (New Brunswick, Canada) Hurley Medical Center (MI)
Calgary Laboratory Services Doctors Hospital (Bahamas) Indiana State Board of Health
Carilion Consolidated Laboratory Duke University Medical Center Indiana University
(VA) (NC) Institute of Medical and Veterinary
Cathay General Hospital (Taiwan) E.A. Conway Medical Center (LA) Science (Australia)
CB Healthcare Complex Eastern Maine Medical Center Instituto Scientifico HS. Raffaele
(Sydney, NS, Canada) East Side Clinical Laboratory (RI) (Italy)
Central Peninsula General Hospital Elyria Memorial Hospital (OH) International Health Management
(AK) Emory University Hospital (GA) Associates, Inc. (IL)
Central Texas Veterans Health Care Esoterix Center for Infectious Jackson Memorial Hospital (FL)
System Disease (TX) Jersey Shore Medical Center (NJ)
Centre Hospitalier Regional del la Fairview-University Medical John C. Lincoln Hospital (AZ)
Citadelle (Belgium) Center (MN) John F. Kennedy Medical Center
Centro Diagnostico Italiano Federal Medical Center (MN) (NJ)
(Milano, Italy) Florida Hospital East Orlando John Peter Smith Hospital (TX)
Champlain Valley Physicians Foothills Hospital (Calgary, AB, Kadlec Medical Center (WA)
Hospital (NY) Canada) Kaiser Permanente Medical Care
Changi General Hospital Fort St. John General Hospital (CA)
(Singapore) (Fort St. John, BC, Canada) Kaiser Permanente (MD)
Children’s Hospital (NE) Fox Chase Cancer Center (PA) Kantonsspital (Switzerland)
Children’s Hospital & Clinics (MN) Fresenius Medical Care/Spectra Keller Army Community Hospital
Children’s Hospital King's East (NJ) (NY)
Daughters (VA) Fresno Community Hospital and Kenora-Rainy River Regional
Children’s Hospital Medical Center Medical Center Laboratory Program (Ontario,
(Akron, OH) Frye Regional Medical Center (NC) Canada)
Children’s Hospital of Gambro Healthcare Laboratory Kern Medical Center (CA)
Philadelphia (PA) Services (FL) Kimball Medical Center (NJ)
Children’s Medical Center of Dallas Gateway Medical Center (TN) King Faisal Specialist Hospital
(TX) Geisinger Medical Center (PA) (Saudi Arabia)
Clarian Health–Methodist Hospital Grady Memorial Hospital (GA) King Khalid National Guard Hospital
(IN) Guthrie Clinic Laboratories (PA) (Saudi Arabia)
Clendo Lab (Puerto Rico) Hahnemann University Hospital King’s Daughter Medical Center
(PA) (KY)
ix
Number 9 NCCLS
Klinični Center (Slovenia) Montreal General Hospital Royal Columbian Hospital (New
Laboratories at Bonfils (CO) (Canada) Westminster, BC, Canada)
Laboratoire de Santé Publique du MRL Pharmaceutical Services, Inc. Sacred Heart Hospital (MD)
Quebec (Canada) (VA) Saint Mary’s Regional Medical
Laboratório Fleury S/C Ltda. MRL Reference Laboratory (CA) Center (NV)
(Brazil) Nassau County Medical Center St. Alexius Medical Center (ND)
Laboratory Corporation of America (NY) St. Anthony Hospital (CO)
(NJ) National Institutes of Health (MD) St. Anthony’s Hospital (FL)
Laboratory Corporation of Naval Surface Warfare Center (IN) St. Barnabas Medical Center (NJ)
America (MO) Nebraska Health System St-Eustache Hospital (Quebec,
LAC and USC Healthcare New Britain General Hospital (CT) Canada)
Network (CA) New England Fertility Institute St. Francis Medical Ctr. (CA)
Lakeland Regional Medical Center (CT) St. John Hospital and Medical
(FL) North Carolina State Laboratory of Center (MI)
Lancaster General Hospital (PA) Public Health St. John Regional Hospital (St.
Langley Air Force Base (VA) Northern Indiana Education John, NB, Canada)
LeBonheur Children’s Foundation St. Joseph Hospital (NE)
Medical Center (TN) North Kansas City Hospital (MO) St. Joseph’s Hospital – Marshfield
Libero Instituto Univ. Campus North Shore – Long Island Jewish Clinic (WI)
BioMedico (Italy) Health System Laboratories (NY) St. Joseph Mercy Hospital (MI)
Long Beach Memorial Medical Northwestern Memorial Hospital St. Luke’s Regional Medical
Center (CA) (IL) Center (IA)
Louisiana State University O.L. Vrouwziekenhuis (Belgium) St. Mary Medical Center (IN)
Medical Center Ordre professionnel des St. Mary of the Plains Hospital
Maccabi Medical Care and Health technologists médicaux du (TX)
Fund (Israel) Québec St. Mary’s Hospital & Medical
Magee Womens Hospital (PA) Ospedali Riuniti (Italy) Center (CO)
Malcolm Grow USAF Medical The Ottawa Hospital St. Paul’s Hospital (Vancouver, BC,
Center (MD) (Ottawa, ON, Canada) Montreal)
Manitoba Health (Winnipeg, Our Lady of Lourdes Hospital (NJ) St. Vincent Medical Center (CA)
Canada) Our Lady of the Resurrection Ste. Justine Hospital (Montreal, PQ,
Martin Luther King/Drew Medical Medical Center (IL) Canada)
Center (CA) Pathology and Cytology Salina Regional Health Center (KS)
Massachusetts General Hospital Laboratories, Inc. (KY) San Francisco General Hospital
(Microbiology Laboratory) The Permanente Medical Group (CA)
MDS Metro Laboratory Services (CA) Santa Clara Valley Medical Center
(Burnaby, BC, Canada) Piedmont Hospital (GA) (CA)
Medical College of Virginia Pikeville Methodist Hospital (KY) Seoul Nat’l University Hospital
Hospital Pocono Hospital (PA) (Korea)
Medicare/Medicaid Certification, Presbyterian Hospital of Dallas Shanghai Center for the
State of North Carolina (TX) Clinical Laboratory (China)
Memorial Medical Center (IL) Queen Elizabeth Hospital (Prince South Bend Medical Foundation
Memorial Medical Center (LA) Edward Island, Canada) (IN)
Jefferson Davis Hwy Queensland Health Pathology Southwest Texas Methodist Hospital
Memorial Medical Center (LA) Services (Australia) (TX)
Napoleon Avenue Quest Diagnostics Incorporated South Western Area Pathology
Mescalero Indian Hospital (NM) (CA) Service (Australia)
Methodist Hospital (TX) Quintiles Laboratories, Ltd. (GA) Southern Maine Medical Center
Methodist Hospitals of Memphis Regions Hospital Specialty Laboratories, Inc. (CA)
(TN) Reid Hospital & Health Care Stanford Hospital and Clinics (CA)
MetroHealth Medical Center (OH) Services (IN) State of Washington Department of
Michigan Department of Research Medical Center (MO) Health
Community Health Rex Healthcare (NC) Stony Brook University Hospital
Mississippi Baptist Medical Center Rhode Island Department of Health (NY)
Monte Tabor – Centro Italo - Laboratories Stormont-Vail Regional Medical
Brazileiro de Promocao (Brazil) Riyadh Armed Forces Hospital Center (KS)
Montreal Children’s Hospital (Saudi Arabia) Sun Health-Boswell Hospital (AZ)
(Canada)
x
Volume 22 I/LA21-A
Sunrise Hospital and Medical University of Alabama-Birmingham VA (San Diego) Medical Center
Center (NV) Hospital (CA)
Swedish Medical Center – University of Alberta Hospitals VA (Tuskegee) Medical Center
Providence Campus (WA) (Canada) (AL)
Tampa General Hospital (FL) University of Colorado Health VA Outpatient Clinic (OH)
Temple University Hospital (PA) Science Center Vejle Hospital (Denmark)
Tenet Odessa Regional Hospital University of Chicago Hospitals Washington Adventist Hospital
(TX) (IL) (MD)
The Toledo Hospital (OH) University of Illinois Medical Center Washoe Medical Center
Touro Infirmary (LA) University of the Ryukyus (Japan) Laboratory (NV)
Trident Regional Medical Center University of Texas M.D. Anderson West Jefferson Medical Center
(SC) Cancer Center (LA)
Tripler Army Medical Center (HI) University of Virginia Medical Wilford Hall Medical Center (TX)
Truman Medical Center (MO) Center William Beaumont Hospital (MI)
UCSF Medical Center (CA) University of Washington Williamsburg Community Hospital
UNC Hospitals (NC) UZ-KUL Medical Center (Belgium) (VA)
University Hospital (Gent) VA (Denver) Medical Center (CO) Winn Army Community Hospital
(Belgium) VA (Kansas City) Medical Center (GA)
University Hospitals of Cleveland (MO) Winnipeg Regional Health
(OH) VA (Western NY) Healthcare Authority (Winnipeg, Canada)
The University Hospitals (OK) System Wishard Memorial Hospital (IN)
Yonsei University College of
Medicine (Korea)
York Hospital (PA)
OFFICERS BOARD OF DIRECTORS
F. Alan Andersen, Ph.D., Susan Blonshine, RRT, RPFT, Tadashi Kawai, M.D., Ph.D.
President FAARC International Clinical Pathology
Cosmetic Ingredient Review TechEd Center
Donna M. Meyer, Ph.D., Wayne Brinster J. Stephen Kroger, M.D., FACP

President Elect BD COLA
CHRISTUS Health
Kurt H. Davis, FCSMLS, CAE Gary L. Myers, Ph.D.
Emil Voelkert, Ph.D. Canadian Society for Medical Centers for Disease Control and
Secretary Laboratory Science Prevention
Roche Diagnostics GmbH
Robert L. Habig, Ph.D. Barbara G. Painter, Ph.D.
Gerald A. Hoeltge, M.D. Ampersand Medical Bayer Corporation
Treasurer
The Cleveland Clinic Foundation Thomas L. Hearn, Ph.D. Ann M. Willey, Ph.D., J.D.
Centers for Disease Control and New York State Department of
William F. Koch, Ph.D., Prevention Health
Immediate Past President
National Institute of Standards Carolyn D. Jones, J.D., M.P.H. Judith A. Yost, M.A., M.T.(ASCP)
and Technology AdvaMed Centers for Medicare & Medicaid
Services
John V. Bergen, Ph.D.,
Executive Director
xi
Number 9 NCCLS
xii
Volume 22 I/LA21-A
Contents
Abstract ................................................................................................................................................... i
Committee Membership ......................................................................................................................... v
Active Membership .............................................................................................................................. vii
Foreword .............................................................................................................................................. xv
1 Introduction ............................................................................................................................... 1
1.1 Scope............................................................................................................................. 1
1.2 Definitions .................................................................................................................... 1
2 Establishment of Analytical Performance Prior to Clinical Evaluations .................................. 3
2.1 Assay Components........................................................................................................ 3
2.2 Specimen Requirements................................................................................................ 5
2.3 Procedural Design......................................................................................................... 6
2.4 Analytical Measures...................................................................................................... 6
3 Clinical Evaluation: Planning and Design................................................................................. 7
3.1 Investigator’s Manual ................................................................................................... 7
3.2 Ethical Considerations .................................................................................................. 9
3.3 Clinical Evaluation Protocol ....................................................................................... 11
3.4 Clinical Evaluation Objectives.................................................................................... 12
3.5 Selection of Investigator and Evaluation Site ............................................................. 13
3.6 Evaluating Performance Characteristics ..................................................................... 14
3.7 Classification of Subjects............................................................................................ 17
3.8 Evaluation Population................................................................................................. 18
3.9 Considerations for Masking ........................................................................................ 19
4 Conducting the Clinical Evaluation ........................................................................................ 20
4.1 Monitoring Clinical Evaluations................................................................................. 20
4.2 Management of Database............................................................................................ 21
4.3 Quality Assurance of Data Integrity ........................................................................... 22
4.4 Retention of Records................................................................................................... 22
5 Analysis of Clinical Evaluation Data ...................................................................................... 22
5.1 Performance of Statistical Tests.................................................................................. 23
5.2 Documentation of Performance Characteristics ......................................................... 23
5.3 Clinical Evaluation Summary ..................................................................................... 24
References ............................................................................................................................................ 26
Additional References .......................................................................................................................... 27
Summary of Comments and Subcommittee Responses ....................................................................... 28
Summary of Delegate Comments and Subcommittee Responses ........................................................ 36
Related NCCLS Publications ............................................................................................................... 37
xiii
Number 9 NCCLS
xiv
Volume 22 I/LA21-A
Foreword
Currently, no uniform guidelines exist that adequately describe the requirements for the clinical
evaluation of immunoassays. Historically, assay developers—primarily in vitro diagnostic
manufacturers—have based their approach to designing and conducting clinical evaluations on what may
be required by government regulatory agencies, who review the demonstration of safety and
effectiveness. Increasingly, assays developed by nongovernment-regulated entities are being used as
clinical measures. These assays may include those developed for the purpose of measuring end points of
drugs under development. While these end-point assays may provide appropriate analysis of analytical
performance that is consistent with laboratory requirements and regulations, evaluation of clinical
performance may be incomplete or lacking.
In preparing this guideline, the subcommittee considered three areas of need regarding the clinical use of
immunoassays. First, for manufacturers of in vitro diagnostic assays, this guideline will provide a
checklist to review against their approach to addressing regulatory requirements for commercialization of
products. Second, for laboratories engaged in the development of immunoassays for use within their
institutions, this guideline will provide direction in designing an evaluation of the assay’s clinical
performance. And third, for those scientists involved in evaluating new therapeutic agents, this guideline
will provide direction in establishing immunoassays as reliable clinical end points.
For the purposes of this document, clinical performance refers to accuracy— correct classification, i.e.,
clinical sensitivity and specificity—and does not refer to clinical utility, which may include the effects of
environment, economy, and patient outcomes. While there is mention of an assay’s analytical
performance, users should refer to existing NCCLS documents (see Related NCCLS Publications) and to
other sources for more detailed information.
Because the scope of this document does not limit its application to industry or to the clinical or research
laboratory, the subcommittee has used the term “clinical evaluation” in place of “clinical study” or
“clinical trial.” While considered interchangeable from the subcommittee’s perspective, the reader should
use the term that is appropriate for the reader’s institution.
It should also be acknowledged that there are different types of evaluations for new assays, including
comparative and clinical. Comparative evaluations are typically performed when the laboratorian is
considering substituting an assay from one manufacturer with another from a different manufacturer.
While having its own unique forms of execution and analyses, this evaluation is simply comparing one
assay to another without the postulation of any clinical questions. A reference for comparative evaluations
is NCCLS document EP9—Method Comparison and Bias Estimation Using Patient Samples. While it
may involve a comparative approach, the clinical evaluation is required for the application of a new assay,
a new analyte, or for a new, intended use of an existing analyte.
In the assay development to implementation/commercialization continuum, this guideline addresses the

activities associated with preclinical testing and clinical evaluation requirements, evaluation design, and
analysis. While written for immunoassay developers, the information has broad application to other assay
formats. It is the subcommittee’s intent that this document will be expanded to ensure that the full range
of in vitro diagnostic assays is addressed.
A Note on Terminology
NCCLS, as a global leader in standardization and harmonization, is firmly committed to achieving global
harmonization wherever possible. Harmonization is a process of recognizing, understanding, and
explaining differences while taking steps to achieve worldwide uniformity. NCCLS recognizes that
medical conventions in the global metrological community have evolved differently in the United States,
Europe, and elsewhere; that these differences are reflected in NCCLS, ISO, and CEN documents; and that
xv
Number 9 NCCLS
legally required use of terms, regional usage, and different consensus timelines are all obstacles to
harmonization. In light of this, NCCLS recognizes that harmonization of terms facilitates the global
application of standards and is an area of immediate attention. Implementation of this policy must be an
evolutionary and educational process that begins with new projects and revisions of existing documents.
In the context of this document, it is necessary to point out that several terms are used differently in the
USA and other countries, notably those in Europe.
In Europe the term "performance evaluation" is used for the assessment of quality of in vitro diagnostic
products both with their analytical and medical (diagnostic) characteristics. "Clinical evaluation" in
European terms is applied mostly to the evaluation of medical products, which are used on or in patients
or when it refers to clinical studies of drugs, under much more stringent conditions. Appropriately, the
USA term "clinical evaluation" in the context of this document would translate into "diagnostic
evaluation" in Europe. Consequently, the terms "diagnostic sensitivity" and "diagnostic specificity" are
used in Europe, with the corresponding expressions "clinical sensitivity" and "clinical specificity" in the
USA, as they are applied in this document.
Also, in order to align the usage of terms to ISO, the term "trueness" is used in this document when
referring to the closeness of the agreement between the average value from a large series of measurements
and to a true value of a measurand. The term "accuracy," in its metrological sense, refers to the closeness
of the agreement between the result of a (single) measurement and a true value of a measurand, thus
comprising both random and systematic effects.
All terms and definitions will be reviewed again for consistency with international use, and revised
appropriately during the next scheduled revision of this document.
Standard Precautions
Because it is often impossible to know what might be infectious, all human blood specimens are to be
treated as infectious and handled according to “standard precautions.” Standard precautions are new
guidelines that combine the major features of “universal precautions and body substance isolation”
practices. Standard precautions cover the transmission of any pathogen and thus are more comprehensive
than universal precautions which are intended to apply only to transmission of blood-borne pathogens.
Standard precaution and universal precaution guidelines are available from the U.S. Centers for Disease
Control and Prevention (Guideline for Isolation Precautions in Hospitals. Infection Control and Hospital
Epidemiology. CDC. 1996;Vol 17;1:53-80), (MMWR 1987;36[suppl 2S]2S-18S), and (MMWR
1988;37:377-382, 387-388). For specific precautions for preventing the laboratory transmission of blood-
borne infection from laboratory instruments and materials and for recommendations for the management
of blood-borne exposure, refer to NCCLS document M29—Protection of Laboratory Workers from
Occupationally Acquired Infections.
Key Words
Clinical evaluation, diagnostic evaluation, clinical evaluation investigator, clinical performance

characteristics, diagnostic performance characteristics, database management, evaluation population,
informed consent, institutional review board, pilot evaluation, sponsor, statistical tests
xvi
Volume 22 I/LA21-A
The Quality System Approach

NCCLS subscribes to a quality system approach in the development of standards and guidelines, which
facilitates project management; defines a document structure via a template; and provides a process to
identify needed documents through a gap analysis. The approach is based on the model presented in the
most current edition of NCCLS HS1— A Quality System Model for Health Care. The quality system
approach applies a core set of “quality system essentials (QSEs),” basic to any organization, to all
operations in any healthcare service’s path of workflow. The QSEs provide the framework for delivery of
any type of product or service, serving as a manager’s guide. The quality system essentials (QSEs) are:
QSEs
Documents & Records Information Management
Organization Occurrence Management
Personnel Assessment
Equipment Process Improvement
Purchasing & Inventory Service & Satisfaction
Process Control Facilities & Safety
NCCLS document I/LA21-A—Clinical Evaluation of Immunoassays addresses the following Quality

System Essentials (QSEs)
Purchasing &
&
&
Improvement
Organization
Management
Management
Information
Satisfaction
Assessment
Occurrence
Documents
Equipment
& Records
Personnel
Inventory
Facilities
Control
Process
Process
Service
Safety
X X X X
Adapted from NCCLS document HS1—A Quality System Model for Health Care.
xvii
Number 9 NCCLS
xviii
Volume 22 I/LA21-A
1 Introduction
In vitro diagnostic assay development occurs in multiple environments, each with different requirements
for the actual use of the assay. Developers that manufacture and market their assays are required to
obtain governmental approval or clearance in some countries, such as in the U.S., prior to commercial
sales. Others may develop assays that do not require the same extent of testing and review. For example,
clinical research laboratories may develop assays for use within their institutions, thereby limiting the
assay’s use to a select patient population. Another example is the development of research assays for use
in monitoring pharmacologic effects or unfavorable reactions. Whatever the intended use, the assay can
only provide medical benefit when it has undergone adequate clinical evaluation.
This document provides uniform guidelines for clinical evaluation of in vitro diagnostic immunoassays.
Separated into four basic sections, this guidance is intended to provide all assay developers with a
consistent approach to establishing clinical performance characteristics. First, the preclinical evaluation is
described, including recommendations for establishing analytical performance. Second, the planning and
design of the clinical evaluation is discussed. Third, a description of conducting the evaluation is given,
and fourth, analysis and summary documentation is reviewed.
1.1 Scope
The scope of this document is to provide guidance for evaluating the clinical performance characteristics
of an immunoassay, with no regard to the status or environment of the assay developer. The information
provided will have broad applications to multiple assay formats and their uses.
1.2 Definitionsa
Accuracy// Measurement accuracy// Accuracy of measurement, n - Closeness of the agreement

between the result of a measurement and a true value of the measurand {/analyte}; See Trueness.
Analytical specificity, n - The ability of a measurement procedure to determine solely the measurable
quantity it purports to measure.
Clinical evaluation, n - An investigation of the clinical performance characteristics of a new (or new
indication for use) in vitro diagnostic assay in controlled clinical settings; NOTE: The term "clinical
evaluation" is equivalent to the term "diagnostic evaluation."
Clinical feasibility/Pilot evaluation, n – An evaluation performed using patient specimens to assess the
potential application of a new assay to some clinical use; NOTE: Typically conducted by the sponsor, the
evaluation may take place in a clinical setting or in the sponsor’s laboratory.
Clinical investigator, n – A person under whose direction a clinical evaluation is conducted.
Clinical sensitivity, n - The proportion of patients with a well-defined clinical disorder whose test values
are positive or exceed a defined decision limit (i.e., a positive result and identification of the patients who
have a disease); NOTES: a) It is the fraction of clinically true positive classifications divided by the sum
a
Some of these definitions are found in NCCLS document NRSCL8—Terminology and Definitions for Use in NCCLS
Documents. For complete definitions and detailed source information, please refer to the most current edition of that document.
An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 1

Number 9 NCCLS
of clinically true positive plus clinically false negative classifications; b) The term "Clinical sensitivity" is
equivalent to "Diagnostic sensitivity."
Clinical specificity, n - The proportion of subjects who do not have a specified clinical disorder and
whose test results are negative or within the defined decision limit; NOTES: a) It is the fraction of
clinically true negative classifications divided by the sum of clinically true negative plus clinically false
positive classifications; b) The term "clinical specificity" is equivalent to "diagnostic specificity."
Confidence interval, n - An interval estimate of a population parameter computed so that the statement
“the population parameter lies in this interval” will be true ... in a stated proportion of the times such
statements are made.
Interim analysis, n - Any data analysis that is performed during the clinical evaluation and before the
evaluation is completed; NOTES: a) Analyses performed at intervals throughout the evaluation provide
the sponsor with estimates of the new assay’s performance; b) These estimates are of limited validity due
to insufficient sample size.
Minimal detectable concentration (MDC), n – The lowest concentration that can be accurately
estimated by a given test system; NOTE: Sometimes this is defined as the lowest concentration that can
be statistically distinguished from zero, but other definitions may be used.
Negative predictive value, NPV, n – The likelihood that an individual with a negative test result does not
have the disease, or other characteristic which the test is designed to detect. NOTE: This varies with
prevalence of the disease unless the test is 100% sensitive.
Population//(study group), n - The totality of items under consideration; NOTE: The study group is the
sample subset of the population which is actually studied.
Positive predictive value, PPV, n – The likelihood that an individual with a positive test result has a
particular disease, or characteristic, which the test is designed to detect; NOTE: This varies with
prevalence of the disease unless the test is 100% specific.
Preclinical evaluation, n - A preliminary evaluation to determine if the new assay’s analytical

performance characteristics meet the desired design goals; NOTE: It serves as a determinant for
proceeding to the clinical evaluation.
Precision, n – The closeness of agreement between independent test results obtained under prescribed
{/stipulated} conditions; NOTE: Precision of a given measurement procedure is usually subdivided
according to specific conditions into Repeatability and Reproducibility. Please refer to these terms
below.
Repeatability of measurements {/Measurement repeatability}, n - The closeness of the agreement

between the results of successive measurements of the same measurand carried out under the same
conditions of measurement; NOTE: Repeatability is often termed in laboratory medicine as "within-run
precision."
Reportable range, n - The range of test values over which the relationship between the instrument, kit, or
system's measurement response is shown to be valid; NOTES: a) For this document, the range of values
(in units appropriate for the analyte) over which the acceptability criteria for the method have been met;
that is, where errors due to nonlinearity, imprecision, or other sources are within defined limits; b) This is
similar to the VIM definition for "measuring range" or "working range," i.e., a set of values of
measurands for which the error of a measuring instrument is intended to lie within specified limits; c) The
reportable range of the assay should be established prior to beginning the clinical evaluation.
2 An NCCLS global consensus guideline. ©NCCLS. All rights reserved.

Volume 22 I/LA21-A
Reproducibility, of measurements {/Measurement reproducibility}, n - The closeness of the

agreement between the results of measurements of the same measurand, where the measurements are
carried out under changed conditions; NOTE: The changed conditions may refer different lots, runs, time
(day), technician, etc.
Subject, n - A person, with disease or without disease, enrolled in an evaluation; NOTE: An equivalent
term is "Proband."
Traceability, n - A property of the result of a measurement or the value of a standard whereby it can be
related to stated references, usually national or international standards, through an unbroken chain of
comparisons, all having stated uncertainties; NOTES: a) The concept is often expressed by the adjective
“traceable”; b) The unbroken chain of comparisons is called a traceability chain.
Trueness/Trueness of measurement, n - Closeness of agreement between the average value obtained

from a large series of results of measurements and a true value; NOTE: The difference between the
average value from a large series of measurements and the true is also called “bias” or systematic error.
See Accuracy.
2 Establishment of Analytical Performance Prior to Clinical Evaluations

Although the focus of this document is on the clinical evaluation of immunoassays, the information
provided will have broad applications to multiple assay formats and their uses. It is imperative that the
assay be well designed and optimized prior to starting a clinical evaluation. Therefore, before beginning a
clinical evaluation, the assay should be produced as pilot lots according to prepared documentation, and
not in a research and development modality. Modifications of the product design, materials, or operating
conditions during the course of a clinical evaluation may invalidate the entire evaluation, due to the fact
that the analytical performance characteristics of the assay may have changed while the evaluation was in
progress. Such a change can significantly affect the variability of test results and may result in an inability
to validate clinical performance. The key elements in the analytical characterization of an assay are
briefly described below with references provided for more detailed information.
2.1 Assay Components
Immunoassays refer to any laboratory method for detecting a substance, or antigen, by using an antibody
that is reactive with it. It is important to recognize that some immunoassays claiming to measure the
same analyte may in fact give different results when applied to a particular sample or reference material
because respective reagents recognize a different repertoire of epitopes in the substance thus leading to
results for different although related quantities.1 In some cases the substances are mixtures of molecular
species with clinically relevant properties in common, but with different structures in varying proportions,
e.g., substances involving analytes such as antibodies toward antigens, glycoproteins, tumor antigens, and
protein hormones.1 For many analytes measured by immunoassays, there is no traceability for trueness
above the manufacturer’s selected measurement procedure or working calibrator until internationally
agreed upon reference measurements and materials become available.1 The following section addresses
the characterization of antigen and antibody, and calibration, as well as preparation and handling of other
assay components.
2.1.1 Antigen and Antibody Characterization
Whether the antigen and antibody are produced internally or purchased from a vendor, it is important to
ensure that an adequate supply of these critical reagents is available to avoid frequent material changes to
new lots. Antigen characterization using biochemical, immunochemical, or immunological techniques is
necessary to demonstrate purity, the nature and possibly quantity of potentially cross-reacting antigens,
and the presence of all essential reactants in the antigen preparation. For example, it may be important
Number 9 NCCLS
that proteins of 34, 69, and 157 kD all be present in the antigen. If the antigen is prepared from cell
culture or by recombinant procedures, appropriate documentation should be available. Antibody should
be tested by the most sensitive procedure available to determine the specificity and to characterize cross-
reactivity with related antigens. A popular technique for establishing antibody specificity is western
immunoblotting using a crude antigen mixture. In some cases, additional information, such as calculation
of binding constants, etc., may be needed. If monoclonal antibodies are used, there should be adequate
documentation of the procedures used for production of the hybridoma, as well as selection and expansion
of the monoclonal antibody. If these reagents are purchased, sufficient characterization may have already
been done, and the needed documentation can be supplied by the vendor. To use these reagents, the assay
developer must establish acceptance criteria that will be used to release each new lot of antigen or
antibody.
2.1.2 Calibrators, Standards, and Controls
Calibrators are materials with established analyte values that are used to standardize the instrument or
assay method. Prepared at specific concentration intervals, the calibrators establish the assay dose
response curve and should encompass the critical determination points. For example, if the key medical
decision point for the assay is 2.0 ng/mL, there should be a calibrator close to that value.
Assay controls are materials that contain assay-specific analyte in an assay-compatible matrix and are
tested with patient samples to ascertain the reliability of the assay. Rather than a preset value, controls
may have a range of values and variability that should be consistent with the required accuracy for the
assay working range. Generally, positive and negative controls are used for qualitative assays, and low-
and high-level controls are used for the quantitative assays. In quantitative assays, there should be one
control with analyte concentration near the decision point. A narrative description of these materials can
be found in the most current edition of NCCLS document I/LA18—Specifications for Immunological
Testing for Infectious Diseases.
In addition to assay controls provided by the assay manufacturer or by the laboratory, commercial
controls are available for assay developers and assay users. Typically, these controls contain different
concentrations of one or several different analytes and may be used in addition to the assay controls to
validate each assay run. In addition to use in assay validation, commercial controls may be used by the
assay developer to challenge laboratory testing personnel’s skills in the performance of an assay. These
challenge controls are not used to aid directly in clinical diagnosis. Commercial controls are assigned
specified control ranges that may be determined by the manufacturer of the testing methodology. These
ranges are specified in the commercial controls package insert for each level of control. Availability of
commercial controls during the assay development process provides the developer with a measure of the
product’s trueness and precision. If using unassayed controls, developers should establish acceptability
limits prior to beginning the evaluation.b
National or international standards or reference materials provide the basis for calibration and may be
used by assay developers and research or clinical laboratories to ensure harmonization across assays.
When such suitable materials are available, a developer should use these materials to set values (establish
traceability) for the assay calibrators, and an assay user can confirm that the product is in fact
standardized to a particular reference material. In some cases, reference materials and procedures may
not exist. For these assays, the developer must establish a means of ensuring purity of the calibration
material and its preservation in a stable matrix, as well as a means of establishing its concentration.
b
Refer to FDA guideline, “Points to Consider Guidance Document on Assayed and Unassayed Quality Control
Material.” Draft February 1999.

Volume 22 I/LA21-A
The trueness of the measurement of a value to a defined calibrator depends upon the metrological
traceability of the value.1 When possible, the calibration should be traced to a method and/or material of a
higher metrological order/hierarchy.1 The goals of calibration hierarchies are to improve traceability of
results. The hierarchy of traceability depends upon on the availability of components for each of the
various metrological levels of measurement procedures and calibrators; five levels of this traceability
chain can be identified. 1 The metrological traceability chain ranges from the first level of availability of
primary reference methods and primary reference materials for value assignment through several different
iterations of these in combinations to the last level where no reference sources are available. 1 When
neither reference measurement procedure nor reference materials for calibration are available, the
manufacturer can establish “in-house” measurement procedures and calibrators to support their value
assignment to their product’s calibrator. This condition may occur frequently for immunoassays and
traceability may be particularly difficult with these immunoprocedures.
2.1.3 Other Reagents
Reagents that will be provided with the assay, such as buffers, enzyme substrates, and stopping reagents,
will need to be provided in the most practical and stable form (liquid or liquid concentrate, lyophilized,
powder). Specifications will need to be established for those materials that are needed to run the assay but
will not be provided with the assay (i.e., microplates, spectrophotometers, and water baths).
2.1.4 Stability, Handling, and Storage
Generally, real-time stability studies are performed to establish the expiration dating for the entire kit and
its individual components under recommended storage conditions, and this information is provided on the
label. Alternatively, accelerated stability studies can be conducted at elevated temperatures for materials
that are suitable for testing by these methods. Assay components with unique storage requirements, such
as reduced shelf-life after reconstitution, aliquoting and freezing after reconstitution, or storage in reduced
light, should have labeling that indicates any special handling.
Warnings and precautions may be required for some reagents. Materials containing metallic azides have
to be identified and proper disposal conditions described. Special warnings are required for any kit that
contains radioactive materials. Kits that contain reagents made with human source materials must contain
a statement that the source materials were tested and found negative for the presence of HIV and hepatitis
antigens. Additional warnings and handling recommendations may be necessary depending on the kit
contents. Material Safety Data Sheets should be available for hazardous materials.
2.2 Specimen Requirements
2.2.1 Type
Serum is the most commonly recommended specimen for immunoassay. It should be clearly stated if
hemolysis, lipemia, cloudiness, or any other sample conditions are to be avoided. When plasma is
included as a specimen, the recommended anticoagulants should also be listed, as some may interfere in
assay performance. Comparison studies with paired samples should be performed to show that the same
results are obtained with each sample type. Similar studies are needed to show that whole blood or urine
or another specimen type is acceptable. Any special caveats regarding sample collection should be
provided also. NCCLS document H3— Procedures for the Collection of Diagnostic Blood Specimens by
Venipuncture offers additional details.
2.2.2 Stability, Handling, and Storage
It is incumbent on the assay developer or manufacturer to establish the conditions and time frame
whereby a sample or specimen can be stored. These data to support stability claims can be generated

Number 9 NCCLS
from real-time studies or might come from the literature or historical references, e.g. Tietz.2 All patient
specimens are to be treated with standard precautions, because it is impossible to know whether they are
infectious. NCCLS document M29— Protection of Laboratory Workers from Occupationally Acquired
Infections deals specifically with this issue.
2.3 Procedural Design
2.3.1 Assay Method Optimization
Optimization of the immunoassay can be considered to encompass all aspects of the test system. First,
reproducibility should be established for the reaction conditions, linearity, working range, analytical
specificity, minimal detectable concentration, etc. These operating conditions are then incorporated into a
well-defined, step-by-step procedure. Second, the reference interval (normal range) of analyte levels, the
expected values in appropriate test populations for which the assay is intended to be used, the key medical
decision points, and/or cutoff are estimated. These values will be validated and challenged in the clinical
studies. NCCLS document EP10— Preliminary Evaluation of Quantitative Clinical Laboratory Methods
gives additional guidance on experimental design and data analysis for the preliminary evaluation of
analytical device performance.
2.3.2 QA/QC
Ensuring that appropriate quality assurance and quality control procedures are followed requires the
availability of the materials and protocols necessary for proper and accurate test performance. These
include control materials and concentrations for positive and negative controls; any additional control
parameters that need to be addressed; established QA criteria; directions for performing QC;
interpretation of QC; and recommendations for action regarding discrepant QC results. If the device
contains internal or procedural controls, these should be described with an explanation of the assay
performance parameters that they control.
2.3.3 Training
The magnitude of training and skill level required to perform the assay should be delineated. Depending
on the complexity of the test, training can range from a detailed, step-by-step procedure that is provided
as a part of the product labeling, to an in-depth user manual, or to a formal on-site training session with
practice specimens and certification requirements.
2.3.4 Ancillary Equipment
If the assay is not being developed as a part of an instrument system, the specifications for ancillary
equipment, such as washers, diluters, readers, data management, and processing hardware and software,
must be provided.
2.4 Analytical Measures
2.4.1 Performance Characteristics
The nonclinical laboratory testing is typically conducted prior to initiating a clinical study in order to
establish the performance characteristics of the new assay. The specific experiments performed may
vary, depending on whether the assay is quantitative, semiquantitative, or qualitative. Some of these
studies are described in detail in other NCCLS evaluation protocols: EP5—Evaluation of Precision
Performance of Clinical Chemistry Devices; EP6—Evaluation of the Linearity of Quantitative Analytical
Methods; EP9—Method Comparison and Bias Estimation Using Patient Samples; and EP14—Evaluation
of Matrix Effects.

Volume 22 I/LA21-A
The following is a general list of studies which have been used to establish immunoassay performance
characteristics:
• validation of cutoffs and/or calibration curve for a qualitative assay;
• determination of lower and upper limits of detection (dynamic range) and linearity for a quantitative
or semiquantitative assay; and
• precision, anticoagulant study, cross-reactivity, interference testing, minimal detectable concentration,

and bias relative to a comparator method (or traceability to a reference measurement procedure1).
It is important to conduct evaluations of trueness and precision (repeatability and reproducibility) at each
clinical evaluation site to ensure the comparability of clinical test data prior to testing clinical specimens.
2.4.2 Interference Studies
The common interfering substances tested during assay development include hemoglobin, lipids,
bilirubin, and protein concentration. Other substances that have been investigated include therapeutic
drugs, over-the-counter medications, and dietary supplements. Additional potential interferences will be
highly dependent on the assay being developed, the specimen type, and the proposed test population. For
additional guidance, the reader is referred to NCCLS documents EP7—Interference Testing in Clinical
Chemistry and I/LA18—Specifications for Immunological Testing for Infectious Diseases.
2.4.3 Cross-Reactivity Studies
Cross-reactivity studies should be conducted to define the analytical specificity of the immunoassay for
the specific antibody or antigen the assay is designed to detect. The experiment is conducted with a large
excess of a variety of natural substances and chemical analogs to challenge the assay’s specificity. The
test substances examined vary greatly depending on the assay design. Substances that have been
investigated include patient sera from subjects with related viruses or diseases, microorganisms that could
be encountered in test specimens, and microorganisms which could be introduced when handling test
specimens. If the antigen or antisera utilized in the assay is a recombinant protein, consider testing sera
containing antibodies against the organism in which the vectors were induced.
3 Clinical Evaluation: Planning and Design

Adequate organization and communication are key elements to any clinical evaluation. Initial face-to-face
interaction between the clinical evaluation sponsor or assay developer and the clinical investigator prior to
beginning the evaluation establishes a productive working relationship which will prove invaluable
throughout the entire clinical evaluation period. As a part of this initial interaction, the clinical
investigator should be provided with the appropriate materials necessary to plan for, implement, and
complete a proper clinical evaluation. The evaluation sponsor has the responsibility for organizing and
providing such materials in an “investigator’s manual.” In addition, the sponsor must be familiar with
and must fully inform the investigator of the appropriate ethical considerations relative to the evaluation
being undertaken.
3.1 Investigator’s Manual
It is recommended that the evaluation sponsor prepare an investigator’s manual and provide it to each
evaluation site. It should provide a set of clear and concise instructions which will serve as a reference
guide during the clinical evaluation period. This manual may be used in conjunction with frequent person-
to-person communications between the sponsor and investigator as the evaluation proceeds. It is

Number 9 NCCLS
recommended that the manual be titled with the name of the immunoassay being evaluated, the names of
the site and investigator, and the name and phone/fax numbers of the sponsor. Optional sponsor names
and numbers may be provided for evening and weekend emergencies. The manual shall include, at a
minimum, the following sections:
(1) Clinical protocol. A dated and revision-number-controlled copy of the clinical evaluation protocol.
(See Section 3.3.)
(2) Draft package insert. A draft-written description of the immunoassay performance instructions,
including equipment required, specimen collection/handling, and precautions, as well as a technical
description of the new assay method and its principles of performance.
(3) Institutional review board (IRB). Where appropriate, document that IRB review has occurred,
along with a copy of an approved, blank, patient informed consent form.
(4) Data collection. Instructions describing the manner in which data (including patient histories, if
appropriate) are to be collected, documented, and transmitted to the sponsor.
(5) Forms management. Copies of forms to be used during the evaluation, with examples of each
properly filled out.
(6) Daily work organization. Suggestions for organization of daily work load related to the clinical
evaluation (e.g., equipment maintenance required daily/weekly, recommendations for appropriate
number of patient specimen runs per day, daily paperwork completion schedule, and preparation
schedule for data/materials shipments back to the sponsor).
(7) Supply inventory control. Suggestions for maintaining adequate levels of supplies required for
the evaluation, including instructions for ordering additional supplies/materials from either the
sponsor or outside vendors, are key to timely completion of the evaluation without interruption.
(8) Shipment instructions. Complete instructions regarding the shipment of supplies, equipment,
data, and/or patient specimens, if applicable, from the evaluation site to the sponsor. It is
recommended, if commercial couriers are used for this purpose, that preprinted courier forms be
provided to the evaluation site. If specimens are to be shipped, instructions must include the proper
labeling and packaging of these specimens in compliance with good laboratory practices and
governmental regulations.
(9) Safety. A complete analysis of any possible safety hazards which could confront the clinical
investigator, including those associated with specimen collection/handling, reagent
preparation/testing, or instrument/equipment operation. Refer to NCCLS document M29−
Protection of Laboratory Workers from Occupationally Acquired Infections for additional
information.
(10) Environment. Instructions for the safe and environmentally sound handling of reagent waste,
residual patient specimens, packaging, reagent vapors/aerosols, or other disposable supplies
associated with the clinical evaluation.
(11) Responsibilities. A description of responsibilities of each participant: the monitor, evaluation

investigator, and additional personnel involved with the performance of the clinical evaluation.
Other items to include in the investigator’s manual include:
• key site personnel curriculum vitae;

Volume 22 I/LA21-A
• monitor site visit log;

• subject enrollment list form;
• correspondence log (includes phone calls, letters, faxes, and email);
• laboratory certification; and
• investigational materials inventory accountability log.
3.2 Ethical Considerations
The protection of human subjects is critical to the conduct of medical research worldwide. Each
evaluation investigator must obtain approval from the institutional review board (IRB) or the independent
ethics committee (IEC), groups that are responsible for the implementation and oversight of methods to
ensure that protections are appropriate and adequate.
To adequately protect the subjects involved in medical research, the investigator should consider: risk
analysis and management; institutional review board approval; informed consent; linkage and patient
confidentiality; and use of data in patient management. These issues are discussed in further detail below.
3.2.1 Risk Analysis and Management
Risks to the subjects must be minimized whether or not the evaluation requires review by an institutional
review board. Risks may include physical harm, psychological harm, and violation of privacy. Any
possible risks must be outweighed by the benefits from participation in the evaluation. These benefits
might include improved medical treatment resulting from the evaluation, or possible benefits to society or
future patients resulting from the introduction of a new assay or a new application of an existing assay.
3.2.2 Institutional Review Board
An institutional review board (IRB) is a group formally designated by an institution to review, approve,
and monitor any research involving human subjects. Outside of the U.S., an IRB is often called an “ethics
committee.” Although most institutions have their own IRB, in some cases an investigator or sponsor
may find it appropriate or necessary to contact a commercial IRB for evaluation review. Commercial
IRBs generally charge a fee for their services. The purpose of the review board is to protect the welfare
and rights of human subjects in any evaluation.
In reviewing any evaluation, the IRB will consider the scientific merit (including sound scientific design),
the ratio of the risks to the benefits, and any bias in subject selection. The IRB will also consider the
balance between ethical and social benefits and is responsible for the protection of any vulnerable
populations (e.g., children, mentally infirm, prisoners).
Sufficient time should be allowed in the project plan for IRB review. Some protocol reviews can take up
to three months depending on the issues raised by the IRB. The clinical evaluation protocol and the
patient informed consent form are prepared by the sponsor and principal investigator. The IRB may
request that a specific format be used for submitting evaluation information. The principal investigator
submits these evaluation documents to the IRB for review. The IRB may waive the requirement for
review, or conduct a full board review. The IRB determines the requirement for patient consent and will
approve the specific language in the patient consent form. An original copy of the patient consent form, if
required for the evaluation, must bear a dated stamp from the IRB. The portion of clinical evaluation
using subjects or subject specimens cannot begin prior to receipt of an IRB waiver or approval
documentation.

Number 9 NCCLS
3.2.3 Informed Consent
If indicated as a requirement by the IRB or the IEC, the investigator must ensure that a patient consent
form is signed by each evaluation subject. The consent form needs to be written using nontechnical
language that is understandable to the subjects. In some cases the clinical investigators or their designee
must provide oral explanations of all pertinent aspects of the evaluation. In signing the consent form, the
subject acknowledges that his/her consent was given freely without coercion. The basic elements of
informed consent include:c
(1) explanation of the purpose of the research, the duration of the evaluation, and a description of the
procedures to be used (procedures are either experimental or a part of established medical practice);
(2) description of any reasonably foreseeable risks or discomforts to the patients;
(3) description of any benefits to the patients or to others;
(4) disclosure of appropriate alternative procedures;
(5) statement regarding the confidentiality of any records;
(6) explanation of any compensation and available treatments in the event of complications;
(7) contact name for answers to questions related to the research and patients’ rights in the event of a
research-related injury; and
(8) statement that participation is voluntary and may be discontinued at any time without penalty.
Many IRBs or IECs will not require informed consent for clinical evaluations using surplus specimens,
provided that there is no link to the patient’s identity. Informed consent can be required when clinical
evaluations include, for example, access to patient medical records, specifications for patient treatment, or
collection of specimens for the specific purpose of the evaluation.
3.2.4 Patient Confidentiality
Although clinical evaluations can require that the evaluation subject’s specimen be linked to the patient’s
medical records, extreme care must be taken to protect the confidentiality of the patient. Identification
numbers for subjects and/or clinical specimens can be recorded using a coded system to protect individual
identity. Social security numbers and patient names should not be recorded as identifiers on data
collection forms.
In clinical evaluations where a link between the specimen and patient is not required by the clinical
protocol, information should be recorded such that subjects cannot be identified, directly or through
identifiers linked to the subjects.
3.2.5 Use of Data in Patient Management
Use of the immunoassay data for patient management (e.g., treatment decisions) generated during a
clinical evaluation of an in vitro diagnostic assay (IVD) is not generally accepted practice. This is
particularly applicable to the use of IVD tests for which the performance characteristics have not yet been
established. Exceptions include conditions in which the evaluation design requires disclosure of test
results, or in which the researcher would be medically negligent by not informing the subjects. An
c
In the U.S., refer to 21 CFR Part 50.25.

Volume 22 I/LA21-A
example of the need for disclosure of test results is found in the evaluation of volunteer blood donors with
a new HIV-1/2 immunoassay. In this case, disclosure of positive results by the new assay is required to
allow for donor deferral until donor suitability can be resolved.
3.3 Clinical Evaluation Protocol
The clinical protocol is the most important document associated with the clinical evaluation. It is the set
of instructions that each investigator will follow to conduct the evaluation. It must be accurate, clinically
correct, and written in clear, concise language. During protocol development, an evaluation sponsor can
gain valuable insight into its effectiveness by reviewing the protocol with several key investigators. In
addition, review by a biostatistician may be beneficial as the protocol is being developed. Prior to
initiating the evaluation, it is important to ensure that the evaluation is feasible and meets the stated
objectives. In addition to review by the investigators, it may be possible to review the protocol with the
appropriate regulatory body. For example, in the U.S., the FDA will review draft clinical protocols for
new assay systems requiring approval or clearance. The major elements of a clinical protocol are outlined
below:
(1) Table of Contents. A well-indexed table of contents makes it easier for the laboratory staff to
quickly find important information any time during the evaluation. All the information needed to
perform the clinical evaluation can be found in this document.
(2) Introduction. Provides background information about the new assay, including proposed
indications for use; describes clinical significance for which the assay is being evaluated; and
describes or references previous research.
(3) Description of the New Assay Method or Device. Provides details on the method and procedure.
(4) Objectives/End Points. Specific, measurable objectives for the clinical evaluation that define the
purpose of the evaluation, with reference to particular performance characteristics being evaluated,
(e.g., clinical sensitivity, clinical specificity, etc.). Considers the need for well-defined clinical end
points.
(5) Design End Points. Summary of the evaluation plan; defines measures for clinical diagnosis
evaluation (if appropriate to evaluation design), duration of evaluation, type and estimated number
of specimens or patients to be tested, statistical justification for sample size estimate, information
on interpretation of results, plan for discrepancy resolution, and testing format, including test
methods and specimen coding system.
(6) Study Population. Inclusion and exclusion criteria for patients and/or specimens; pertinent
population demographics; logistics of recruiting and selecting subjects.
(7) Description of Assay Methods. Describes all other methods to be used in the clinical evaluation,
including all methods for discrepancy analysis (e.g., new assay method, other immunoassays,
confirmatory tests).
(8) Site Training. Describes training needed for the clinical evaluation, including any requirement for
demonstrating proficiency with the new assay method, as well as any other methods to be used in
the evaluation.
(9) Specimens. Acquisition, storage, handling, and shipping.
(10) Warnings and Precautions. Cautions for handling infectious material.

Number 9 NCCLS
(11) Limitations of the New Assay System. Describes any stress conditions that might lead to product
failure; lists any limitations of the new assay system.
(12) Data Management. Provides a description of methods of data collection, data collection forms,
subject enrollment log, and procedures for transferring data to the sponsor.
(13) Data Analysis. Provides a description of statistical methods to be used for determining clinical and
analytical performance characteristics.
(14) Monitoring Schedule. Identifies the evaluation monitor and primary method of communication
with the site; indicates a schedule for monitoring visits.
(15) Principal Investigator Responsibilities. Some general obligations include: ensuring the
investigation is conducted according to the protocol and the agreement with the study sponsor;
protecting the rights of subjects; obtaining appropriate informed consent; controlling distribution of
investigational devices; retaining records as specified by the protocol; and assuring that an IRB is
informed for initial and continuing review of the study. Additional study specific responsibilities
may be listed in the study protocol.
(16) Materials. Lists materials the sponsor will provide and those the investigator is expected to
provide.
(17) Key Contacts. Lists contact information for both the evaluation site and the sponsor, complete with
name, address, phone, FAX, and E-mail.
(18) Appendixes. May include: sample data collection forms, sample informed consent, draft product
instructions, evaluation outline for quick reference, and flow chart for handling specimens through
all possible result algorithms.
Additional elements to be considered for inclusion in the clinical protocol are:
• number of sites;
• masking (if appropriate);
• Subject discontinuation procedures;
• source documents and record retention;
• risk/benefit analysis, risk management;
• IRB and informed consent requirements; and
• subject confidentiality and subject ID log description.
(19) Termination of evaluation. May include closeout procedures and early rules/clauses for
consideration of termination.
3.4 Clinical Evaluation Objectives
A successful clinical evaluation must have clearly established objectives and clinical end points. The
objectives should be specific and measurable. For example, “to establish the clinical sensitivity and
specificity of a new infectious disease assay” is too general. The objective is better stated in specific
terms, identifying the clinical management question and the target population. Hence, a more complete
statement of the objective might be “to estimate the clinical sensitivity and specificity of a new serum
immunoassay for analyte X to be used to identify clinically important coronary artery disease (CAD) in
women presenting with signs and/or symptoms suggestive of CAD.”

Volume 22 I/LA21-A
The objective is to distinguish those women with important CAD from those without; the target
population would be symptomatic women, and the end point might be coronary angiography. A clinical
evaluation may be designed with multiple objectives. A second objective might be to screen
asymptomatic males to identify those with occult CAD. This is a separate question, requiring different
subjects, and might even use a different specimen. Different clinical applications for a given test require
separate evaluations, whether or not they are conducted simultaneously.
3.4.1 Clinical Feasibility/Pilot Evaluation
There may be a number of unknowns in the evaluation design that could be established with a pilot
evaluation. Generally, a pilot evaluation has a similar design but may not have a statistically significant
sample size relative to a full clinical evaluation. The outcome of the pilot evaluation can help to confirm
the evaluation design for the larger evaluation. Some of the reasons for conducting a pilot evaluation
include:
(1) to evaluate a new sample preparation method;
(2) to ensure that the evaluation design is feasible in the laboratory;
(3) to evaluate new data retrieval and data management methods;
(4) to better establish appropriate clinical end points;
(5) to confirm patient selection criteria, thus better defining the evaluation population;
(6) to evaluate the performance of the new assay with fresh (unfrozen) clinical specimens;
(7) to evaluate a new assay system in the laboratory setting (for instance, combining the reagents and
instruments for the first time);
(8) to determine expected variability and means in assay values for different patient populations and/or
proposed diagnostic cutoffs to better estimate the sample size for the clinical evaluation;
(9) to modify and optimize (improve) the evaluation protocol; and
(10) to validate the performance of the investigator’s method or the use of the reference assay system.
3.5 Selection of Investigator and Evaluation Site
Selection of qualified investigators is an essential part of any successful clinical evaluation. The
following is a list of attributes to look for when selecting an investigator and clinical evaluation site:
(1) The investigator has demonstrated expertise in the clinical field of interest.
(2) The investigator has sufficient time from other commitments for the evaluation.
(3) The investigator/institution is unconstrained by other commitments.
(4) Consider the laboratory’s experience with particular technology associated with the new assay.
(5) Consider the laboratory’s experience with other methods to be used during the clinical evaluation.
(6) The investigator has previous experience with clinical studies.

Number 9 NCCLS
(7) The investigator/institution demonstrates willingness to participate.
(8) The investigator/staff is committed to complying with good clinical practice (GCP) standards.
(9) The investigator/institution understands the need for informed consent.
(10) IRB reviews are conducted on a timely basis.
(11) The investigator/staff are committed to following the evaluation protocol.
(12) The facilities are adequate), and there is secure storage for records and investigational product.
(13) Convenient access to patient records is provided.
(14) There is a well-organized approach to documentation and record keeping.
(15) Sufficiently qualified staff can be assigned to the evaluation, especially an evaluation coordinator.
(16) The staff turnover rate is low.
(17) The staff are given sufficient time from other duties to accomplish the evaluation.
(18) Patients/subjects who are willing to participate in research are accessible.
(19) The patient population/demographics are consistent with the evaluation design.
(20) Adequate number of subjects/specimens meeting inclusion/exclusions criteria exists, as well as the
ability to enroll enough subjects consistent with evaluation timelines and total specimen targets.
A visit to the evaluation site prior to final selection of investigators can be valuable in evaluating the
above criteria. A face-to-face interview with the investigator can sometimes be more revealing than a
phone conversation. Good clinical practice standards refer to the following types of site visits: 1) a pre-
evaluation visit to confirm staff and facilities are adequate; 2) an evaluation initiation visit, in order to
train the site personnel regarding the protocol and evaluation requirements, as well as confirming that the
new assay method is working well; 3) a visit to monitor evaluation progress and review raw data and
patient charts; and 4) a close-out visit to conclude all evaluation activities and confirm that all evaluation
documents are in order.
3.6 Evaluating Performance Characteristics
3.6.1 Clinical Performance
What is “clinical performance?” Clinical performance is the performance of a diagnostic test in a clinical
context. Clearly defined objectives and end points naturally describe the intended diagnostic use of the
assay in its medical context. This use indicates what the relevant performance characteristics under
investigation are, and therefore, what should be evaluated. What is the clinical management issue? Who
are the subjects relevant to this objective? What are the end points? An example might be as follows:
Objective: To identify, among individuals who are over a certain age and have positive results for fecal
occult blood (FOB), those who are candidates for colonoscopic exam. Subjects: In this case, persons over
the age of interest and having a positive FOB. End point: Presence or absence of cancer based on
histopathologic examination of lesions identified and removed in a direct exam of the colon.

Volume 22 I/LA21-A
Assume an immunoassay for a serum marker for colon cancer is being evaluated for this purpose. The
subjects have some probability of colon cancer as a result of their age and positive FOB. The issue is
whether the test can provide new information to revise that pretest probability (prior probability) for each
individual to a posttest probability of cancer. This posttest probability would be used to distinguish
between those who are candidates for some further exam (such as colonoscopy) and those that need not
have it. This ability to separate the subjects is the clinical performance of interest.
Possible uses of an assay include: (1) screening for disease or conditions in the absence of indications; (2)
pursuing a differential diagnostic workup in response to signs/symptoms or some other clinical
indications; (3) following a known condition to determine a change in clinical status (with or without
interventions); (4) choosing among intervention options; or (5) assessing the status of an individual, such
as whether they have been immunized or exposed to some infectious agent. In each of the possible
applications, a specific question is being addressed about a defined target group of subjects, in an attempt
to make some clinically useful distinction among the subjects. The ability of the assay to make this
distinction, that is, to discriminate among the relevant subgroups (for example, healthy versus diabetic;
normal risk versus increased risk; prostate cancer versus benign hypertrophy in subjects with difficulty
urinating; relapse of cancer following therapy; subjects with estrogen receptor positive versus negative
breast cancers; potential blood donors with and without hepatitis) is often termed “clinical accuracy” and
is the performance characteristic at issue.
How can this discrimination ability be described and evaluated? One common approach is by determining
clinical sensitivity and specificity. Generally sensitivity and specificity vary as the decision threshold
(cutoff) changes, and they are reciprocal. As one increases, the other decreases. A test, then, exhibits a
spectrum of trade-offs between sensitivity and specificity as the cutoff changes. For continuous variables,
the most comprehensive description of this spectrum of clinical sensitivity and specificity is provided by
the receiver operating characteristic (ROC) analysis, because it examines all possible decision thresholds
and all the corresponding combinations of sensitivity and specificity. This makes it a global assessment. It
is independent of prevalence (prior probability) as well. The ROC plot itself is a qualitative representation
of clinical sensitivity and specificity; the area under the ROC plot is a quantitative one. (See the most
current edition of NCCLS document GP10—Assessment of the Clinical Accuracy of Laboratory Tests
Using Receiver Operating Characteristic (ROC) Plots.) In the example mentioned above, ROC of the
results for a new serum marker of colon cancer would indicate the ability of this new test to discriminate
between those individuals who were found to have an unknown cancer and those who did not. The ROC
data could also be used to compare this new marker to an existing one. And further, the ROC data could
be combined with other relevant data on prevalence and on the cost of false results to evaluate and select
particular decision thresholds for clinical use. For assays with semiquantitative results such as “negative,”
1+, 2+, etc., ROC analysis can be employed, though the smaller number of different results will yield
plots with fewer points.
Once particular decision thresholds are identified, performance characteristics, such as positive and
negative predictive value (PV), may be useful descriptors of the meaning of individual test results. PV
analysis, unlike ROC analysis, is prevalence-dependent. ROC describes how well the test discriminates
between two categories of subjects, while PV describes the probable meaning of normal or abnormal
results once the cutoff and prevalence are defined.
For purely qualitative data, test results expressed simply as “positive” or “negative” in 2 x 2 tables as
illustrated below, may be used to characterize the clinical performance of each test.

Number 9 NCCLS
# Subjects
Affected Unaffected
# Test Positive A B
Results
Negative C D
Using these tables, from two or more different tests, assays can be compared to one another using the
appropriate statistics.
3.6.2 Analytical Performance
The analytical performance characteristics of a diagnostic test should be well defined before instituting a
clinical evaluation. The clinical evaluation phase should include verification components for
maintenance of the analytical performance characteristics. This verification of performance characteristics
is also an important series of measurements to ensure that the analytical performance characteristics of a
particular test are reproducible at different test sites and with different staff performing and interpreting
the test. For many IVDs, the analytical performance characteristics may be established by the
manufacturer or a central laboratory/researcher within an institution. In these instances, the analytical
performance characteristics, including traceability, trueness, and precision (repeatability and repro-
ducibility), must be established/verified at each clinical evaluation site to ensure the comparability of
clinical test data prior to testing clinical specimens. The manufacturer or central laboratory should
provide each test site with selected control materials that can verify the trueness and reproducibility of
analytical performance.
3.6.3 Statistical Power and Sample Size Considerations
Determining the sample size for a clinical evaluation may require the assistance of a biostatistician to
define the evaluation objectives, selection of evaluation populations, and statistical methods. If the intent
of the clinical evaluation is to determine test efficacy or compare different tests using ROC plots, then the
sample size may require sufficient numbers to provide a valid estimate of the ROC plot for the IVD. The
clinical performance of an IVD may be characterized by determination of clinical sensitivity and
specificity, as well as the confidence intervals associated with each of these performance characteristics at
selected decision thresholds. Since the confidence intervals for either the sensitivity or the specificity will
decrease as the sample size increases, the sample size calculations for a clinical evaluation may also
include prior considerations about the desired confidence intervals for sensitivity and specificity.
Sample size considerations will also depend on the intended use of the test, the evaluation population(s),
and the prevalence of disease in the evaluation population. For many diseases, a clinical evaluation may
require multiple sites to provide sufficient numbers of diseased individuals and to analyze test data in
different populations.
In studies with the intent to determine reference values and reference intervals for quantitative tests, the
sample size is determined by the reference interval the investigator wishes to present. More samples are
needed if a wider reference interval estimate (such as central 95% as compared to 90%) is targeted. In any
case, a minimum of 120 subjects is recommended for such a study. For more information on determining
the sample size in a reference intervals study, see NCCLS document C28—How to Define and Determine
Reference Intervals in the Clinical Laboratory.

Volume 22 I/LA21-A
3.6.4 Methods of Statistical Analysis (Confidence Intervals and Considerations for Pooling Data)
Techniques for plotting ROC curves, calculating areas under the plots, and determining relevant
confidence intervals have been described.3,4 Computer software is available. In order to pool data from
multiple evaluation sites, the evaluation must be designed and executed to ensure consistency from site to
site in all aspects, particularly selection of subjects, sample collection and storage, and analytical
techniques. This consistency must be verified before the data from the various sites are actually merged.
Consultation with biostatisticians is recommended while the evaluation is being designed.
Any subjects, samples, or data which are lost, deleted, or otherwise do not conform to the protocol should
be accounted for and described.
3.6.5 Reproducibility Testing
Reproducibility testing may be assessed on the basis of lot-to-lot, day-to-day, run-to-run, and
technologist-to-technologist variation of the new assay method. This portion of the evaluation is
generally conducted with a panel of specimens provided to each site participating in the evaluation (See
the most current version of NCCLS document EP5—Evaluation of Precision Performance of Clinical
Chemistry Devices.
3.7 Classification of Subjects
For most clinical evaluations that require patient information, definitive diagnoses of enrolled subjects are
required for accurate classification. Ensuring that each subject meets certain criteria may require
extensive workup and/or access to medical records. Clear classification of some subjects may be difficult
or impossible (see Section 3.7.2). Excluding subjects with equivocal diagnoses from the data analysis
must be explicitly acknowledged. Such exclusions introduce bias in the estimates of clinical sensitivity
and specificity.
3.7.1 Algorithm
It is important that the actual state of health of each subject be established as definitively as possible and
without knowledge of the test results under evaluation. It is this definitive classification against which the
classification provided by the assays under evaluation will be assessed. In deriving this algorithm for
classification, it is desirable to go beyond routine means of clinical classification and instead design as
thorough a clinical analysis as feasible. This analysis may include, for example, histopathology, state-of-
the-art imaging, nonroutine laboratory testing, surgical information, autopsy data, and short- or long-term
clinical outcome. The assays under evaluation should not be included in the algorithm for establishing the
definitive classification of subjects.
3.7.2 Strengths and Limitations of Classification
The strongest evaluation will be based on definitive classification of the subjects. Any error in the
primary classification will result in distortion (either upward or downward) of the apparent clinical
accuracy of the assays under evaluation.3,5,6 This is why a rigorous workup is desirable. It is not unusual,
however, that a truly definitive classification cannot be established for a particular disease state
(myocardial infarction, for example), or for particular individuals with equivocal findings. Alternative
approaches have been reported and discussed.3,7,8,9 In the case of the former problem, the validity of the
entire evaluation is limited and should be recognized as such. Sometimes, clinical follow-up and
determination of outcome can provide a useful alternative to a real-time diagnosis or categorization. In
fact, outcome may be the preferred means to classify subjects. For example, in the case of a prenatal assay
for a genetic abnormality such as Down’s syndrome, waiting until parturition and then examining the
newborn child, rather than trying to classify the fetus in utero would provide a truly definitive

Number 9 NCCLS
classification. Likewise, waiting to observe the behavior of a tumor may provide a more relevant clinical
classification which may be more definitive than imaging or even histopathology.
3.7.3 Analysis of Discrepant Specimens
When a predicate device (a currently available diagnostic test) is used in place of a definitive diagnosis to
assess the clinical performance of a new test (or new clinical application), discrepancies may occur
between or among the results of the various test devices. This is anticipated, since rarely would results
match perfectly among two or more devices. Even when results do match perfectly, both the predicate
and the new test may be incorrect. It is often assumed that the percentage of incorrect but matching results
is zero or small. Even if this assumption is not true, comparative information about the clinical accuracy
of the predicate device versus the new test can still be obtained by investigating only the discrepant
specimens using a definitive diagnosis to see whether the new test or the predicate is correct. This
investigation provides information on whether the predicate or the new test makes fewer classification
errors, and the results from the discrepancy analysis can be provided to the user. However, the results
should not be provided in the performance calculations, since the investigation does not provide data
needed to calculate statistically unbiased estimates of performance for either test. If the additional testing
used to further characterize discrepant specimens is not 100% accurate, then discrepancy resolution may
not be useful at all. The only information obtained is whether the predicate and the new test agree, but it
provides no information about the clinical accuracy of the two tests. It is appropriate to calculate
performance characteristics if the protocol involves a definitive diagnosis on all specimens, or on all
discrepant specimens and a random sample of specimens where initial results agreed. The latter requires
statistical and practical planning in the protocol stage, and may not be simple or cost effective. It is
recommended that planning for discrepant analysis be included in the study design. A scheme for
evaluating the relevant subjects to determine their actual clinical classification can be established in the
study design and then implemented after the results are obtained.
The frequency of discrepancies can be described by using a 2 x 2 table as shown below.
# Results Test 2
Positive Negative
# Results Positive A B B and C

Test 1 are discrepant
Negative C D
Cases falling in the lower left and upper right boxes are discrepant. Note that changing the decision
threshold, or cutoff, for any given assay will change the numbers appearing in the boxes. Thus
discrepancy results can increase or decrease as the decision thresholds vary. While varying the cutoff and
examining the change in the discrepancies can provide insight into the nature of the discrepancies, the
actual choice of the new test cutoff should be made prior to the start of the comparison study.
3.8 Evaluation Population
3.8.1 Establish Clinically Relevant Population
Once the clinical question being addressed by the clinical evaluation is clearly and fully defined (see
Section 4.1), the targeted population will be evident. Nevertheless, an explicit and detailed description of
this relevant clinical population may be developed, so that relevant samples can be defined by rigorous,

Volume 22 I/LA21-A
unambiguous inclusion/exclusion criteria. Careful attention is required to ensure an appropriate spectrum

of individuals.3,5,6,10,11 The subjects actually selected for evaluation in a clinical evaluation should be
representative of that larger subject population for whom the test is ultimately intended. Selection
procedures which use only the most convenient or readily accessible subjects are likely to introduce bias
into the evaluation by obtaining a non-representative sample. For example, a selection bias may be
created by enrolling employees, only those patients seen early in the day on weekdays, only those subjects
for whom the laboratory has leftover samples, or only those subjects whose true classification is clearly
and easily established. The conclusions of such an evaluation may not be applicable to the larger
population.
As mentioned above and in the literature, healthy controls or other so-called “normals” may not be
relevant to the evaluation, unless the clinical question being tested by the evaluation involves using the
test for screening such apparently healthy individuals. If this were the case, then some defined group of
healthy or asymptomatic individuals would comprise the entire evaluation sample.
3.8.2 Subject Exclusion/Inclusion Criteria
The inclusion/exclusion criteria uniquely characterize the evaluation population and relate this population
to the intended use of the test. Care should be taken to avoid biases in the criteria and in the actual
selection process.10,11 Criteria for selection of subjects should be clearly established before selection
begins and then adhered to strictly in conducting the evaluation. Examples of inclusion/exclusion criteria
include medication use, presence of a disease state or stage, and fasting or dietary requirements. In
addition to patient criteria, immunoassay evaluations must also consider specimen criteria, such as storage
conditions, time from procurement to testing, and specimen volume and quality. Carefully designed and
written criteria are more likely to provide data that can be pooled from multiple sites for statistical
analysis. All sites involved must be able and willing to agree to the criteria and to operate in strict
conformity.
3.8.3 Patient History Documentation
After establishing subject information requirements for the evaluation, procedures for collecting and
recording the data should be established and documented. The sponsor should consider that access to
patient charts and data may be facilitated if the physician evaluating the patient is the principle
investigator or coinvestigator. Depending upon institutional or sponsor requirements, each documented
history must be signed by the clinical personnel taking the information. All sites should agree to conform
to these procedures and maintain confidentiality of all data.
3.9 Considerations for Masking
The use of masking, formerly termed “blinding,” may be necessary to prevent bias that may be introduced
through knowledge of results of an ongoing evaluation. Evaluator bias is a type of investigator bias in
which the person interpreting or recording test results may inadvertently bias the measurements through
subtle changes that may favor the test method being evaluated. When evaluating a test method, the
investigators should not have access to either the clinical information or the reference test method result.
The test method results should be recorded for analyses and comparison with the reference methods and
clinical data after completion of the evaluation. Each patient specimen should be separately coded so that
only the principle investigator or sponsor can link patient clinical information; however, the clinical
investigator or the sponsor must be able to link results from the test undergoing evaluation with reference
test results. Although the use of different test personnel and instrumentation with objective quantitation
would seem to minimize bias, there are many sources of bias that cannot be retrospectively measured or
controlled. The use of coded patient samples, with no access to patient identifiers or patient clinical data,
is the only method to ensure that evaluator bias has not altered the clinical evaluation data.

Number 9 NCCLS
4 Conducting the Clinical Evaluation
4.1 Monitoring Clinical Evaluations
The responsibilities of the monitor are to ensure the safety, rights, and welfare of the patients in the
evaluation, to ensure the quality and integrity of the evaluation data, and to ensure adherence to the
evaluation protocol. It is recommended that the clinical monitor be an individual who possesses the
education, experience, and communication skills appropriate to monitor the progress of the clinical
evaluation. In the case of a clinical evaluation sponsored by a manufacturer, the monitor is typically
identified within the sponsoring organization. In the case of an evaluation initiated by an end-user
laboratory, the monitor may be identified within the end user’s organization. Alternatively, a contract
research organization (CRO) may be contracted to monitor an evaluation. (In the U.S., refer to the FDA
Guideline for the Monitoring of Clinical Investigations; 1988). The duties of the clinical monitor include,
but are not limited to:
(1) assay procedure training for clinical laboratory staff who will perform the testing;
(2) providing training on the study protocol and study forms for all appropriate personnel, including the
study coordinator, principle investigator, study nurses, and clinical laboratory staff;
(3) acting as the primary communication link between the sponsor and the evaluation site;
(4) verifying records/data;
(5) verifying forms completion;
(6) verifying compliance with the evaluation protocol;
(7) reviewing IRB correspondence;
(8) reviewing original patient informed consent forms;
(9) maintaining device accountability;
(10) documenting monitoring activities, such as phone logs and site visit reports;
(11) conducting site visits; and
(12) ensuring that sites have sufficient supplies of forms, evaluation materials, and investigational kits
and reagents to complete the evaluation.
Training the clinical laboratory staff who perform the testing is an important aspect of the clinical
evaluation. Verification that training was effective can be easily accomplished by providing a panel of
test specimens having known values. Using these panels, laboratory personnel from each site can
demonstrate proficiency with the new assay method prior to commencing the evaluation. This can be an
effective test of the training materials and new assay instructions, as well as the capability of the
laboratory personnel. If identical performance of a complex method is important to the evaluation design,
demonstration of proficiency at all evaluation sites is essential prior to commencing the evaluation. It is
well worth incurring a short delay while additional training is provided or personnel are reassigned to
ensure the quality of the data.
Close monitoring of the data as it is collected can help to quickly identify problems which can then be
addressed in a timely manner. Periodic site-monitoring visits may be required to manage the evaluation.

Volume 22 I/LA21-A
For simple evaluation designs, phone communication may be sufficient. For studies of short duration,
daily phone calls may be necessary; for longer studies, weekly calls may be more appropriate. Electronic
transfer of data can facilitate monitoring from a distance. Alternatively, data sent to the sponsor by fax on
a regular schedule may also suffice.
It is important to keep a record of all phone and on-site monitor communications. The record should
include information reviewed, findings, conclusions, and action taken to correct any deficiencies. A pre-
evaluation visit is conducted to ensure that the investigator and the laboratory facility qualify for the
evaluation. Monitoring visits during the evaluation may be necessary to ensure data integrity and to
verify protocol compliance. A closeout visit at the end of the evaluation is essential to complete the
account of all investigational devices, and to ensure that all evaluation records are in order.
4.2 Management of Database
The evaluation sponsor is responsible for designing, verifying, and maintaining a database management
system. Critical to the successful completion of an evaluation, the database management system design
must meet the requirements of the clinical protocol for ensuring accurate and reliable data. Optimally, the
design should be verified well in advance of the start of testing to allow for modifications. A clinical
feasibility or pilot evaluation may serve as a means of verifying the management system. The system
should be well documented with information that may include:
(1) the names and roles of each site’s principal investigators, coinvestigators, and technical
staff;
(2) a schedule of reagent/kit shipments and lot numbers;
(3) the mode of data transmission, such as paper copies, fax, floppy disk, or by modem;
(4) an example and explanation of a data reporting form and case report form;
(5) a description of data entry into the sponsor’s database and means of ensuring clerical accuracy;
(6) a brief description of data availability for statistical analysis;
(7) a means of identifying and analyzing discrepant results, as well as missing data;
(8) procedures for conducting data audits and for general compliance;
(9) general troubleshooting information;
(10) procedures for data review; and
(11) a description of procedures for archiving data.
In addition to defining these activities, the sponsor should clarify the general flow of data. For example,
most clinical sites will transfer their data to the monitor, who ensures delivery to the database manager.
Once the data is entered into the database, the manager may submit it for statistical analysis. The
statistician may then send his/her analysis to the monitor for incorporation into a regulatory submission,
internal documentation, or publication. Consideration must also be given to the site procedures for
documenting supporting data, such as the clinical history or ongoing testing, including x-rays, body scans,
or biopsies.

Number 9 NCCLS
4.3 Quality Assurance of Data Integrity
Creating conditions to minimize mistakes should be a primary concern of the sponsor and the
investigator. Responsibility for efforts to reduce mistakes lies with the sponsor and with the investigative
team. The clinical monitor should plan an ongoing data review for 100% accuracy of records and total
compliance by the site personnel. This review is considered part of the monitoring process. Data audits
are conducted by sampling the evaluation population. Typically, audits involve 5 to 10% randomly
selected records. If problem records are detected, an additional 10% may be audited to determine the
extent of the errors. Ideally, audits are conducted by internal QA data auditors or by external contractors.
Because these resources may not be available, audits may be completed by the monitor or his/her
designee.
An initial step in reducing mistakes is ensuring adequate training of the investigative team. Prior to the
start of the evaluation, the technical personnel at the site should be trained in all aspects of testing using
the new immunoassay and any comparative tests that may be part of the evaluation design. The site
should be thoroughly familiar with the evaluation protocol and inclusion/exclusion criteria for subjects or
specimens. In addition, the site personnel should be instructed on the proper means of using the data
forms or other data applications, and should have a thorough review of troubleshooting. The monitor
should develop a means of verifying that the site personnel have received this training and have
demonstrated a proficiency in testing and transferring data.
An audit of the database management system provides the sponsor with a direct estimate of the database
quality. Assuming a 100% data entry check is completed as data is transferred to the sponsor, audits or
data queries can be scheduled as often as needed to complete the evaluation. The audits should follow the
data flow, with special attention given to the data transfer points. For example, the auditor should check
that any information extracted from the patient’s chart (source document) is accurate and that data
reporting forms accurately reflect instrument printer results. In addition, the auditor should review the
handling of warning flags, missing data, and outliers, as well as any changes in data entries. All data
queries that are detected during an audit should be resolved and documented.
4.4 Retention of Records
Data from clinical evaluations shall be retained in accordance with regulations in countries where the
product is being registered or used.d For those laboratories conducting a methods comparison for a new
assay or a new clinical application, the retention time begins when the assay is placed into clinical
practice. Records should be retained in an easily accessible location that would allow for review by
agency or laboratory surveyor personnel.
For any evaluation, the records may take the form of hard copy (paper) documents and report forms,
computer disks, computer mainframe tapes, or other forms of electronic transmission. The monitor is
responsible for a plan that ensures retention of original data that cannot be modified without clear notation
and justification.
5 Analysis of Clinical Evaluation Data

The analysis should clearly state the hypothesis (clinical issue) tested, the statistical tests used to describe
the data obtained, and the assumptions behind the tests. All statistical procedures should be referenced.
d
In the case of assays requiring agency review, such as in the U.S., the investigator and sponsor are required to retain records for
two years from the time of regulatory clearance or approval or from the time withdrawn by the sponsor.

Volume 22 I/LA21-A
5.1 Performance of Statistical Tests
Processing of data should be determined in advance, including the statistics to be calculated and the
method of calculation (see Sections 3.6.1 and 3.7.3). Computer software is generally available for
analyzing data generated in the evaluation. Appropriate statistical approaches should be identified at the
design stage through consultation with a statistician, and relevant software should then be selected,
implemented, and validated. Before pooling data from multiple sites or multiple evaluations, it is
important to establish the criteria for pooling data.
5.1.1 Continuous Quantitative Data
Tests providing continuous data should be evaluated (and compared to one another) by ROC analysis as
already mentioned (see Section 3.6.1). Complete plots of false-positive fraction versus true-positive
fraction (or an equivalent alternative plot, such as true-negative fraction versus true-positive fraction)
should be generated for each test, along with confidence intervals. Areas under the plot, along with their
confidence intervals, should also be obtained. The significance of any differences observed among plots
and/or among areas under the plots can be determined.
5.1.2 Discontinuous (i.e., discrete or ratings) Quantitative Data
For discontinuous data with more than a few classes of results, ROC analysis can be used to evaluate
results and to compare tests, though some variations may be required.
5.1.3 Qualitative Data
As discussed in Section 3.6.1, qualitative data is best characterized by application to a 2 x 2 discrepancy

table in which the positive or negative test results are compared to reference results. The reference may be
represented by the actual “affected status” of the subject or by the positive or negative results of a
predicate assay. Because a perfect reference capable of identifying true-positive specimens and true-
negative specimens is rarely available, the investigator should have a good understanding of the reference
assay true performance. In some cases, a more precise comparison may be made using two references,
where their combined efficiencies provide greater clinical accuracy.
ROC analysis may be appropriate if the positive or negative test result is dependent upon a measured
signal that is compared to a preassigned cutoff.
5.2 Documentation of Performance Characteristics
Major outcomes of a well-planned and well-executed immunoassay clinical evaluation are the data used
to calculate performance characteristics of the assay under evaluation. Performance characteristics are
generated for inclusion in the package insert which accompanies the product as it is sold to the end-user
laboratory or, if for in-house use only, to appear in assay documentation records. These characteristics
represent the performance that is to be expected when testing a population(s) similar to that evaluated
during the assay clinical evaluation.
Ideally, clinical sensitivity and clinical specificity should be established by comparing the new test to a
definitive patient classification (diagnosis) or other measure of clinical or operational truth, i.e., use of a
well-established diagnostic reference method or combination of methods. When estimates of clinical
sensitivity or, specificity are not possible, agreement of the new device in comparison to a predicate or
comparator assay should be clearly reported. For each performance characteristic that is claimed, there
should be a calculation of the exact 95% confidence interval. In addition, a description of the
population(s) evaluated and the number of specimens tested, as well as a complete description of the
discrepant resolution algorithm, should be included in any data summary.

Number 9 NCCLS
Intra-assay and interassay measures of repeatability and overall reproducibility should be presented when
describing the immunoassay performance characteristics. Where appropriate, the presentation of
performance data in tabular format is recommended.
5.3 Clinical Evaluation Summary
The format of the clinical evaluation summary will be dictated by the purpose for writing it. A clinical
evaluation summary may be intended for regulatory submissions, literature publications, clinical
laboratory reports, or inclusion into manufacturer product inserts.
A summary of studies for regulatory submission might be expected to include:
(1) Clinical studies:
• Evaluation objectives.
• Overview (a summary of the evaluation design and evaluation plan).
• Clinical evaluation sites and qualification.
• Interpretation of results.
• Testing algorithms (detailed plan for resolving discrepant results for a methods comparison
evaluation).
• Methods of statistical analysis (including statistical justification for sample size).
• Evaluation summary. In this portion of the report, the data to address each evaluation objective is
described separately, along with a description of the patient population (inclusion/exclusion
criteria and demographics). Headings for this section might include: clinical sensitivity, clinical
specificity, calibration, reproducibility testing, etc.
• Conclusions.
(2) Preclinical studies (nonclinical):
• Interfering substances/cross-reactivity.
• Factors influencing specimen collection, handling, and processing.
• Specimen requirements and storage conditions.
• Analytical performance characteristics.
• Microbial contamination.
• Antibody/antigen characteristics.
A document describing the assay performance characteristics may include:
(1) Intended uses of the test.

(2) Summary and explanation of the test.
(3) Principles of the procedure.
(4) Warnings and precautions. (May include cautions for handling infectious materials and/or special
handling for specific chemical components, or a statement “for investigational use only” or “for
research use only”).
(5) Materials provided.
(6) Materials required but not provided.
(7) Description of the procedure.
(8) Quality control.
(9) Limitations of the procedure. (Describe any stress conditions that might lead to product failure; list
any limitations of the new assay system).
(10) Expected results and their interpretation (if applicable).
(11) Performance characteristics:

Volume 22 I/LA21-A
• Clinical performance (e.g., clinical sensitivity, clinical specificity, NPV, PPV).

• Analytical performance (minimal detectable concentration, analytical specificity, calibration,
reproducibility, etc).
(12) References (bibliography).

Number 9 NCCLS
References
1
PrEN ISO 17511. (Draft) In vitro diagnostic medical devices—Measurement of quantities in
biological samples—Metrological traceability of values assigned to calibrators and control materials.
Geneva: International Organization for Standardization; 2000.
2
Tietz NW, ed. Clinical Guide to Laboratory Tests. 3rd ed. Philadelphia: WB Saunders; 1995:358.
3
Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: A fundamental tool in
clinical medicine. Clin Chem. 1993;39:561-577.
4
McNeil BJ, Hanley JA. Statistical approaches to the analysis of receiver-operating characteristic
(ROC) curves. Med Decis Making. 1984;4:137-150.
5
Robertson EA, Zweig MH, Van Steirteghem AC. Evaluating the clinical efficacy of laboratory tests.
Am J Clin Path. 1983;79:78-86.
6
Zweig MH, Robertson EA. Why we need better test evaluations. Clin Chem. 1982;28:1272-1276.
7
Valenstein PN. Evaluating diagnostic tests with imperfect standards. Am J Clin Pathol. 1990;93:252-
258.
8
Revesz G, Kundel HL, Bonitatibus M. The effect of verification on the assessment of imaging
techniques. Invest Radiol. 1983;18:194-198.
9
Henkelman RM, Kay I, Bronskill MJ. Receiver operator characteristic (ROC) analysis without truth.
Med Decis Making. 1990;10:24-29.
10
Ransohoff DF, Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic
tests. N Engl J Med. 1978;299:926-930.
11
Reid MC, Lachs MS, Feinstein AR. Use of methodological standards in diagnostic test research.
JAMA. 1995;274:645-651.

Volume 22 I/LA21-A
Additional References
Bross I. Misclassification in 2 x 2 tables. Biometrics. 1954;10:478-486.
Hagdu A. Bias in the evaluation of DNA-amplification tests for detecting Chlamydia trachomatis. Stat
Med. 1997;16:1391-1399.
Lipman HB, Astles JR. Quantifying the bias associated with use of discrepant analysis. Clin Chem.
1998;44;1:108-115.
Miller WC. Bias in discrepant analysis: When two wrongs don’t make a right. J Clin Epidemiol.
1998;51:219-231.
Peters T, Westgard JO. Evaluation of methods. In: Tietz. Textbook of Clinical Chemistry. Burtis and
Ashwood, eds. 1995;225-237.
Galen RSA. Gambino SR. Beyond Normality: The Predictive Value and Efficiency of Medical Diagnoses.
New York: Wiley; 1975.
International and European Standards
PrEN 13612. Performance evaluation of in vitro diagnostic medical devices. Brussels: European
Committee for Standardization (CEN); 1999.
Food and Drug Administration (FDA) Guidelines
“Guideline for the Monitoring of Clinical Investigations,” published January 1988.

PMA Review Statistical Checklist
“Points to Consider Guidance Document on Assayed and Unassayed Quality Control

Material,” draft February 1999.
Federal Regulations
FDA Center for Research and Radiological Health (CDRH) Internet web site: www.fda.gov/cdrh/
21 CFR 50 - Protection of Human Subjects
21 CFR 54 – Financial Disclosure by Clinical Investigators
21 CFR 56 - Institutional Review Boards
21 CFR 812 - Investigational Device Exemptions (includes responsibilities for sponsors, monitors, and
investigators)
21 CFR 814 - Premarket Approval (PMA) of Medical Devices
International Conference on Harmonisation of Technical Requirements for Registration of
Pharmaceuticals for Human Use. ICH Harmonised Tripartite Guideline. ICH Steering Committee. May
1, 1996.

Number 9 NCCLS
NCCLS consensus procedures include an appeals process that is described in detail in Section 9
of the Administrative Procedures. For further information, contact the Executive Offices or visit
our website at www.nccls.org.
Summary of Comments and Subcommittee Responses

I/LA21-P: Clinical Evaluation of Immunoassays; Proposed Guideline
General
1. The Introduction and Section 2 should be made consistent with the concepts and requirements of
prEN/ISO 17511 (Metrological traceability of values assigned to calibrators and control materials).
Comments were limited to those items involving inconsistencies with terms and definitions used in
this NCCLS document. These terms and definitions are also currently inconsistent with well-assessed
ISO Standards, like VIM (1993), GUM (1993), ISO 3534-1 (1993), ISO 5725-1 (1994), and others
which are well known at ISO/TC 212 level. In addition, EDMA has prepared a terminology draft
document on the subject based on the definitions reported in ISO and CEN standards and in some
relevant publications presently available.
• Sections 2.1 and 2.1.2 have been revised to be consistent with the traceability concepts in prEN
ISO 17511 and, where appropriate, the subcommittee has modified terminology to be consistent
with international use. Please see the "Note on Terminology" in the Foreword and Section 1.2.
Also see the responses to Comments 21, 22, and 36.
2. Should it be necessary to compare different vendors’ antigens and/or antisera in the preclinical stages
of establishment of analytical performance of an immunoassay (i.e., antisera may have been produced
slightly differently, or to different epitopes of the analyte)?
• This detail was not intended to be part of the scope or intent of this document.
3. During the course of the clinical evaluation, are patients’ symptoms followed or communicated to the
investigator (i.e., remission of disease state; exuberation of symptoms; development of side effects
from therapies, etc.)?
• Where appropriate, patients’ signs and symptoms are followed for the study design. For
example, an integral part of the evaluation of the assay could be to predict clinical end points
such as the response to therapy, or remission or exacerbation of disease.
4. In the case of discrepant specimen results between the existing method and the new test, could the
third test results be biased if there are limited methodologies to choose from for the analyte in
question (i.e., it may have already been evaluated against the existing method)?
• Section 3.7.3 has been revised for clarification.
5. Reference to commercial controls does not consider unassayed controls or controls that don't have
values for specific methods. Did you consider providing information about the utility of these?

Volume 22 I/LA21-A
• As stated in Section 2.1.2, if using unassayed controls, developers should establish acceptability
limits prior to beginning the evaluation. The subcommittee has provided an additional
reference to support the text.
6. Discussion on specimen stability and storage does not mention doing studies that evaluate the
intended use/environment for the test, e.g., specimens for tests intended for use in an outpatient
testing environment should have extended stability studies to sufficiently cover the typical amount of
time required for specimen transport. This may be particularly important for laboratories that receive
large numbers of specimens from off-site.
• There is adequate discussion of this topic in Sections 2.1.2, 2.2.1, 2.2.2 and 3.1; but the onus of
responsibility to address, establish, and explain any caveats for the assay (e.g., testing
environment, stability requirements, transport, etc.) lies with the assay developer or the
manufacturer.
7. The description of reportable range does not address assays with single point calibration. Should it?
• See the response to Comment 23.
8. The daily work organization seems a little sketchy. Have you considered adding other important
details, such as run size, calibration and control frequency, etc.?
• This document provides a representative list of daily tasks. Each study is expected to have a
unique list, which should be delineated in the study protocol.
9. This document, although a good discussion document, does not address specific issues regarding how
to carry out clinicals. Perhaps more detail regarding evaluation that is needed for quantitative vs.
qualitative assays or screening vs. diagnostic assay would be appropriate.
• This is not the intent of the document. The subcommittee recommends that the reader refer to
other reference materials such as NCCLS document GP10—Assessment of the Clinical Accuracy
of Laboratory Tests Using Receiver Operating Characteristic (ROC) Plots; 510K summaries;
PMA summaries of safety and effectiveness; manufacturers’ product package inserts; ICH
guidelines with harmonized tripartite guidelines; or literature references for specific analytes
and intended uses.
10. There is minimal discussion regarding the depth of evaluation that must be conducted for different
assays with various intended uses. Perhaps a table could be developed that would give specific
analytes and the relative degree of evaluation that must occur, e.g., what must be done for a TSH
assay vs. a cancer marker assay.
• The subcommittee refers readers to NCCLS documents GP10—Assessment of the Clinical

Accuracy of Laboratory Tests Using Receiver Operating Characteristic (ROC) Plots and I/LA18—
Specifications for Immunological Testing for Infectious Disease (see Related NCCLS
Publications).
Foreword
11. Clinical evaluation, clinical study, clinical trial. This paper should distinguish carefully between
“performance evaluation” and “clinical evaluation.” In Europe the IVD directive requires
performance evaluation of all commercial medical devices including immunoassay kits. Performance

Number 9 NCCLS
evaluation is further described in prEN 13612. “Clinical study” or “clinical trial” is not required and
is not necessary to ensure safe and proper use.
Whether an immunoassay has clinical relevance in medical utility is a question of the practice of
laboratory medicine and not a question of product safety. Even if an immunoassay performs exactly,
its clinical performance will vary from study to study because of the heterogeneity of patient samples
and the fact that the assay can only be defined by the reactivity of the antibody used.
• Please see the Note on Terminology in the Foreword and the Definitions section. Also, prEN
13162 has been added as an additional reference.
Section 1.2
12. The first definition given for analytical sensitivity. “Quantitative testing” is the correct ISO definition
for analytical sensitivity (i.e., slope of the calibration curve). By contrast, the second definition for
“qualitative testing” cannot fit the concept of analytical sensitivity. Indeed, it is called “clinical
sensitivity” or “diagnostic sensitivity,” as correctly indicated later in the same page 1, under “clinical
sensitivity.”
• After reviewing the definitions for “analytical sensitivity,” the subcommittee decided to remove
this term from the document and replace it with the term “minimal detectable concentration
(MDC).” The subcommittee agreed that this is more appropriate for the document.
Section 2
13. “Modifications of the product design, materials, or operating conditions during the course of a clinical
evaluation may invalidate the entire study, due to the fact that the analytical performance
characteristics of the assay may have changed while the study was in progress.”
Add the following sentence: “Changes should be evaluated for possible impact to the study, and those
portions of the study repeated if necessary.”
• The subcommittee believes that consideration of any changes to the assay and the effect on the
evaluation are implicit in the paragraph. The text remains unchanged.
Section 2.1.2
14. Last sentence. “For these assays, the developer must establish a means of ensuring purity of the
calibration material and its preservation in a stable matrix, as well as a means of validating its
concentration.” Change “validating” to “establishing.”
• The text has been revised.
15. Section 2.1.2: "Calibrators, Standards, and Controls." There is also information on that topic in
Sections 2.3.2, 2.4.1, and 2.4.3.
• Sections 2.3.2, 2.4.1, and 2.4.3 have been revised.

Volume 22 I/LA21-A
Section 2.1.4
16. 2nd Paragraph. The statement “source materials were tested for the presence of HIV and hepatitis
antigens” should continue to state that they were found negative. Perhaps the testing method should
also be mentioned, as it may become relevant for retrospective correlation of aberrant cases.
• The text has been revised by adding, “and found negative” to the fourth sentence of the second
paragraph.
Section 2.2.1
17. Section 2.2.1: "Type." There is also information regarding interfering substances in Section 2.4.2.
Also in Section 2.2.1, would you want to include that the investigator needs to know not only
specimen type, but also the minimum amount of specimen necessary to perform the test accurately?
• Section 2.2.1 and 2.4.2 are not redundant for interfering substances but provide synergistic
information. The amount of required specimen is defined by default in the assay instructions,
and by the evaluation protocol provided in Section 3.3.
Section 2.4.1
18. You may wish to combine the documentation of this data in this section rather than in a separate
section (now Section 2.5) much further down within the document, or you can create a separate
section for ALL documentation required.
• Section 2.5 has been deleted.
19. “The analytical performance characteristics of the assay that need to be established may include
accuracy, precision ... analytical sensitivity (minimal detectable level), recovery of analyte, and
method comparison...” Delete “minimal detectable level” and add “limit of quantitation.”
Rationale: Analytical sensitivity is already defined in the Definitions section and “minimal detectable
level” is not complete.
The limit of quantitation (LOQ) or functional sensitivity is evaluated in immunoassay clinical studies.
For example, transplant drug assays often evaluate LOQ based on a predetermined maximum %CV
that is acceptable to end users. Similar studies have been done for PSA and TSH.
• The subcommittee has replaced the term "analytical sensitivity" with the term “minimal
detectable concentration (MDC)” throughout the text. In addition, the text in Section 2.4.1 has
been revised.
20. Here the “analytical sensitivity” is not defined as the ratio between response and stimulus (i.e., the
slope of the calibration curve), in agreement with the previous definition of Section 1.2 (page 1). Here
the same term is defined, in brackets, as the “minimal detectable limit” (i.e., the detection limit),
which is not a synonym for "analytical sensitivity" (see EDMA Terminology paper).
• The term “analytical sensitivity” has been replaced with “minimal detectable concentration
(MDC)” and a definition for "minimal detectable concentration" has been added to the
Definitions section, as this term is more appropriate.

Number 9 NCCLS
21. There is an evident contradiction between the definitions in Sections 1.2 and 2.4.1. Moreover,
“accuracy” is not defined, and it is not clear whether it means the combination of “trueness” and
“precision,” according to ISO/ 5725-1, or only the component of systematic error, called “trueness.”
This should be clarified, because, in U.S. literature, “accuracy” is too often used synonymously with
“trueness,” which is not correct, according to ISO standards mentioned above.
• A definition for "accuracy" has been added to the Definitions section. Moreover, the definitions
and terminology throughout the document have been changed to maintain consistency with the
ISO standards mentioned above. Please see the "Note on Terminology" in the Foreword and
the responses to Comments 1, 22, and 36.
22. At last, “precision” is here correctly defined as intra-assay, now called “repeatability,” and interassay
now called “reproducibility.” The two new terms (according to VIM: 1993) should appear instead of
intra- or interassay, lot-to-lot, or other similar terms.
• The definitions for “repeatability” and “reproducibility” have been added to the Definitions
section and they are used appropriately throughout the text. Descriptions such as lot-to-lot are
mentioned in an effort to explain the terms mentioned above. Please see the responses to
Comments 1, 21, and 36.
Section 2.4.3
23. Delete the second and fourth sentences, and revise the third and fifth sentences to read as follows:
• Samples that read high and outside of the reportable range should be diluted and rerun and those
reading lower than the reportable range limit should be reported as below the level of the
sensitivity of the assay.
• In most circumstances, any reading obtained above the analytical sensitivity of the assay or above
the minimal detectable concentration is considered within the reportable range of the assay.
Rationale: The first sentence does not hold true for all assays. The rest of the paragraph is
clarification.
• The section on "Reportable Range" has been deleted, and a definition for "reportable range"
has been added to Section 1.2.
24. What is the difference between analytical sensitivity and MDC? Why mention both? If different,
define?
• Please see the response to Comment 20.
Section 2.5
25. This doesn’t say anything, except that “you should” in the document.
• The subcommittee agrees and has removed Section 2.5 from the document.

Volume 22 I/LA21-A
Section 3.1
26. Safety. The statement should include a requirement for inclusion of detailed instructions for
handling/cleaning up spills/accidents. The details should be specific to the infectious risks and
reagents involved.
• This subcommittee believes the current text falls under the scope of good laboratory practice.
The subcommittee refers the reader to the current version of NCCLS document M29—
Protection of Laboratory Workers from Occupationally Acquired Infections for additional
information.
27. Add a section on “Supply inventory control.”
Rationale: Accountability of IUO reagents needs to be added to this section. Governing bodies are
concerned that IUO reagents be used for a specific purpose and not diverted and used elsewhere. It is
increasingly important to be able to account for all materials that are shipped, used, and destroyed or
returned during a clinical evaluation.
• The text has been revised by adding the bullet, “investigational materials inventory
accountability log.”
Section 3.2.1
28. This section should perhaps be worded more strongly regarding the importance of potential risks to
the human subjects and the critical nature of that decision. The present language is somewhat weak;
however, I qualify my concerns with the fact that I have limited research exposure.
• The subcommittee appreciates the commenter’s concern, but notes that these issues are covered
by Institutional Review Board (IRB) review.
Section 3.2.2
29. Last line. Change “clinical evaluation” to read “portion of clinical evaluation using subjects or subject
specimens.”
Rationale: This change in wording allows the site to set up the assay, train, and do initial performance
evaluations prior to or concurrent with the IRB review.
• The subcommittee has revised the text as indicated.
Section 3.2.5
30. When research results are returned to the subject or a healthcare provider, the facility where the test is
performed must have an effective, appropriate CLIA certificate.
• The subcommittee directs the commenter to Section 3.1, which addresses the need for
laboratory certification.

Number 9 NCCLS
Section 3.3
31. Page 11, (15) Principal Investigator Responsibilities: The statement should perhaps be a general,
inclusive statement of broad responsibility rather than a mere caution.
Section 3.4
32. Last two sentences. Be more specific and clear, and give examples.
Section 3.5
33. Delete (18).
Rationale: This seems to be saying the same as item 21.
Section 3.6.1
34. First line, “The objectives of a clinical evaluation will depend on the proposed clinical use and
indications for use of the immunoassay.” Change “proposed clinical” to “intended.”
Rationale: This change puts this sentence more in line with the scope of the document and narrows
the requirements of providing clinical use studies for every assay in development, which is sometimes
unnecessary and would be burdensome.
• The first paragraph of Section 3.6.1 has been rewritten.
35. Third paragraph: Delete the sentence beginning “Use of ROC analysis...”
Rationale: If one were using cancer diagnostic assays, for example, and only tested patients with
cancer, it is unclear how ROC analysis would be effective, since there would be no way to calculate
clinical specificity (the x-axis is plotted as 1-specificity).
Section 3.6.2
36. In the middle of this chapter there is a sentence, reading: "the analytical performance characteristics,
including reproducibility, accuracy, and precision must be...."
In fact, as stated in Section 2.4.1, precision may be expressed as reproducibility and repeatability.
Thus, the sentence should mention precision (i.e., repeatability and reproducibility) or only precision
or only repeatability and reproducibility. Otherwise, the reader gets confused, as in this case, where it
seems that reproducibility is different from precision, whilst it is one component of precision.
Moreover, in this chapter, "accuracy" is again not defined and it is not clear what it means (see
comment on Section 2.4.1).

Volume 22 I/LA21-A
It is absolutely not sufficient to check accuracy and reproducibility (by contrast, it is even wrong) in
order to verify the analytical performance of a system.
All the specific performance characteristics have to be checked, including analytical sensitivity,
analytical specificity, precision (including repeatability and reproducibility), trueness, and accuracy.
In certain cases, also diagnostic (or clinical) sensitivity and specificity have to be checked.
• The ISO definition for “accuracy” has been added to the Definitions section and it is used
appropriately throughout the text. Furthermore, the Definitions section and the terminology
have been revised to be consistent with international use. Please see the "Note on Terminology"
in the Foreword and the responses to Comments 1, 21, and 22.
Section 3.8.2
37. Revise the fourth sentence as follows: “Examples of inclusion/exclusion criteria include medication
use, presence of a disease state or stage, fasting or dietary requirements, and adequate specimen
volume for initial as well as discordant or discrepant retesting.”
Rationale: It is important to allow this type of exclusion within the protocol.
• The text has been revised to read, “specimen volume and quality” in the fifth sentence.
Section 3.9
38. Delete the sentence beginning with “Although the use of different test personnel…” and replace it
with the following: “Care should be taken by the study sponsor to provide a protocol that allows for
elimination of evaluator bias. This may mean that only coded patient samples may be used throughout
the protocol.”
Rationale: The requirement to use coded samples in a study using automated instruments is of
questionable utility in controlling bias and would add a degree of complexity that may not be
necessary.
• The subcommittee does not agree with the rationale. Not all bias is eliminated by the use of
automated instruments.
Section 4.1
39. This section states the importance of training, but training is also addressed in Section 2.3.4.
• Section 4.1 has been revised.
Section 4.4
40. Delete the first sentence and replace it with the following: “Data from clinical evaluations shall be
retained in accordance with regulations in countries where the product is being registered or used.”
Rationale: There are various regulations surrounding retention of data which change depending on
which country(ies) the product is marketed in. Leave this statement more open.
• The subcommittee agrees, and the text has been revised.

Number 9 NCCLS
Summary of Delegate Comments and Subcommittee Responses

I/LA21-A: Clinical Evaluation of Immunoassays; Approved Guideline
General
I saw the proposed version (September 1999) early in the year 2000; and I communicated my comments
to various persons of the EDMA Standardization Working Party. My main concerns at that time were that
the guideline was definitely NOT in line with current concepts of traceability especially for the Type B
analytes, i.e., all the glycoproteins measured by immunoprocedures (cf. prEN ISO 17511). The type B
analytes comprise some 600 entries including tumor markers, hormones, virology markers etc.
With the new version, I had hoped to find that in Section 1.2, Section 2 as a whole, Section 3 (particularly
3.6.2 and 3.7.3.) adequate discussion was given to the many problems in this most important area of
laboratory medicine. I felt disappointed to find only some vague allusions to them or no discussion at all.
I believe that NCCLS has been fully informed during the years about the new concepts; in my modest
opinion, I think they should have been brought to the attention of the subcommittee.
Therefore, I can only come to the conclusion that the I/LA21 guideline is strictly for regional purposes,
i.e., within the USA, where many of us work for a global market world. I must confess that it surprises me
that so many members of the subcommittee, coming from the U.S.-IVD Industry, have not made
objections at an earlier stage of development of this guideline.
As a suggestion, I propose that a number of persons knowledgeable in the new concepts meet with a
selection (emphasis on the US IVD Industry) of the members of the subcommittee in order to try to sort
out the problems. I have made the same suggestion more than 1.5 years ago, but have never received an
appropriate answer.
I am sorry to say that I cannot be more positive about this guideline; perhaps the ideas and the perceptions
will change?
• Please see the response to Comment 1 in the Summary of Comments section regarding the
proposed-level version of this guideline.

Volume 22 I/LA21-A
Related NCCLS Publications*

EP5-A Evaluation of Precision Performance of Clinical Chemistry Devices; Approved
Guideline (1999). This document provides guidance for designing an experiment to evaluate
the precision performance of clinical chemistry devices; recommendations on comparing the
resulting precision estimates with manufacturer precision performance claims and
determining the validity of such comparisons; and manufacturer guidelines for establishing
claims.
EP6-P2 Evaluation of the Linearity of Quantitative Analytical Methods; Proposed Guideline

(2001). This document provides guidelines for characterizing the linearity of a method during
a method evaluation; for checking linearity as part of routine quality assurance; and for
determining and stating a manufacturer's claim for linear range.
EP7-P Interference Testing in Clinical Chemistry; Proposed Guideline (1986). This document
provides background information and procedures for characterizing the effects of interfering
substances on test results.
EP9-A Method Comparison and Bias Estimation Using Patient Samples; Approved Guideline
(1995). This document addresses procedures for determining the bias between two clinical
methods or devices; and design of a method comparison experiment using split patient
samples and data analysis.
EP10-A Preliminary Evaluation of Quantitative Clinical Laboratory Methods; Approved

Guideline (1998). This guideline addresses experimental design and data analysis for
preliminary performance evaluation of an analytical method or device.
EP14-A Evaluation of Matrix Effects; Approved Guideline (2000). This document provides
guidance for evaluating the error or bias in analyte measurements that results from the sample
matrix (physiological or artificial) when two analytical methods are compared.
GP10-A Assessment of the Clinical Accuracy of Laboratory Tests Using Receiver Operating
Characteristic (ROC) Plots; Approved Guideline (1995). This document describes the
design of a study to evaluate clinical accuracy of laboratory tests; procedures for preparing
ROC curves; glossary of terms; and computer software program information.
H3-A4 Procedures for the Collection of Diagnostic Blood Specimens by Venipuncture;

Approved Standard—Fourth Edition (1998). This document provides procedures for the
collection of diagnostic specimens by venipuncture, including line draws, blood culture
collection, and venipuncture in children.
I/LA18-A Specifications for Immunological Testing for Infectious Diseases; Approved Guideline
(1994). This guideline outlines specimen requirements; performance criteria; algorithms for
the potential use of sequential or duplicate testing; recommendations for intermethod
comparisons of immunological test kits that detect infectious diseases; and specifications for
development of reference materials.
*
Proposed- and tentative-level documents are being advanced through the NCCLS consensus process; therefore, readers should refer to the most
recent editions.

Number 9 NCCLS
Related NCCLS Publications (Continued)

M29-A2 Protection of Laboratory Workers from Occupationally Acquired Infections;
Approved Guideline – Second Edition (2001). This document provides guidance on
the risk of transmission of hepatitis viruses and human immunodeficiency viruses in any
laboratory setting; specific precautions for preventing the laboratory transmission of
blood-borne infection from laboratory instruments and materials; and recommendations
for the management of blood-borne exposure.
NRSCL8-A Terminology and Definitions for Use in NCCLS Documents; Approved Standard
(1998). This document provides standard definitions for use in NCCLS standards and
guidelines, and for submitting candidate reference methods and materials to the National
Reference System for the Clinical Laboratory (NRSCL).

NCCLS T 940 West Valley Road T Suite 1400 T Wayne, PA 19087 T USA T PHONE 610.688.0100
FAX 610.688.0700 T E-MAIL: exoffice@nccls.org T WEBSITE: www.nccls.org T ISBN 1-56238-455-4

Clsi Ila21 A

Uploaded by

Copyright:

Available Formats

Clsi Ila21 A

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Clsi Ila21 A

Uploaded by

Copyright:

Available Formats

I/LA21-A Replaces I/LA21-P

Vol. 22 No. 9 Vol. 19 No. 19

Clinical Evaluation of Immunoassays; Approved Guideline

Reproduced with permission, from NCCLS publication I/LA21-A—Clinical Evaluation

Copyright ©2002. The National Committee for Clinical Laboratory Standards.

(NCCLS. Clinical Evaluation of Immunoassays; Approved Guideline. NCCLS document I/LA21-A

Area Committee on Immunology and Ligand Assay

W. Harry Hannon, Ph.D. Centers for Disease Control and Prevention

Jeffrey Warren, M.D. University of Michigan

Subcommittee on Clinical Evaluations of Immunoassays

Linda Ivor Gen-Probe Incorporated

Peter Maxim, Ph.D. FDA Center for Devices/Rad. Health

Donald R. Parker, Ph.D. Bayer Corporation

Gregory P. Payne, RAC Hybritech

John C. Ridderhof, Dr.P.H. Centers for Disease Control and Prevention

Carol Ryerson, Ph.D. Eastman Kodak Co.

Lawrence E. Schaefer Abbott Laboratories

Mark H. Zweig, M.D. National Institutes of Health

John A. Boffa, B.L.D. American Association of Bioanalysts

Prof. Aw Tar Choon National University Hospital

Anna J. DeMarinis, M.T.(ASCP) bioMerieux, Inc.

Steven I. Gutman, M.D., MBA FDA Center for Devices/Rad. Health

Sheri L. Hall Beckman Coulter, Inc.

Lorin Hills, Ph.D. LifeScan, Inc.

Hassan Khayam-Bashi, Ph.D. San Francisco General Hospital

Edward M. Levine, Ph.D. Diagnostic Products Corporation

Terry McGovern Rockland, Massachusetts

Ronald W. McLawhon, M.D., Ph.D. University of Chicago Medical Center

William F. Skogen, Ph.D. Dade Behring Inc. Glasgow

Leonard Wilson Food and Drug Administration

Patrice E. Polgar NCCLS

Donna M. Wilhelm NCCLS

Sustaining Members College of Medical Laboratory Deutsches Institut für Normung

Bayer Corporation – Tarrytown, NY Essential Therapeutics, Inc. Roche Diagnostics GmbH

OFFICERS BOARD OF DIRECTORS

Donna M. Meyer, Ph.D., Wayne Brinster J. Stephen Kroger, M.D., FACP

Committee Membership ......................................................................................................................... v

Active Membership .............................................................................................................................. vii

Additional References .......................................................................................................................... 27

Summary of Comments and Subcommittee Responses ....................................................................... 28

Summary of Delegate Comments and Subcommittee Responses ........................................................ 36

Related NCCLS Publications ............................................................................................................... 37

In the assay development to implementation/commercialization continuum, this guideline addresses the

Clinical evaluation, diagnostic evaluation, clinical evaluation investigator, clinical performance

The Quality System Approach

NCCLS document I/LA21-A—Clinical Evaluation of Immunoassays addresses the following Quality

Clinical Evaluation of Immunoassays; Approved Guideline

Accuracy// Measurement accuracy// Accuracy of measurement, n - Closeness of the agreement

Clinical investigator, n – A person under whose direction a clinical evaluation is conducted.

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 1

Preclinical evaluation, n - A preliminary evaluation to determine if the new assay’s analytical

Repeatability of measurements {/Measurement repeatability}, n - The closeness of the agreement

2 An NCCLS global consensus guideline. ©NCCLS. All rights reserved.

Reproducibility, of measurements {/Measurement reproducibility}, n - The closeness of the

Trueness/Trueness of measurement, n - Closeness of agreement between the average value obtained

2 Establishment of Analytical Performance Prior to Clinical Evaluations

2.1 Assay Components

2.1.1 Antigen and Antibody Characterization

2.1.2 Calibrators, Standards, and Controls

4 An NCCLS global consensus guideline. ©NCCLS. All rights reserved.