E-Mail karger@karger.com
Gynecologic Cytopathology
Acta Cytologica 2013;57:585–590
DOI: 10.1159/000353843
External Quality Control of Cervical
Cytopathology:
Interlaboratory Variability
Cinara Zago Silveira Ázara Edna Joana Cláudio Manrique
Nádja Lindany Alves de Souza Andryne Rego Rodrigues
Suelene Brito do Nascimento Tavares Rita Goreti Amaral
Unit for External Quality Control, School of Pharmacy, Federal University of Goiás, Goiânia, Brazil
ization of diagnostic criteria, and enhance the accuracy of
screening and improve the quality of cytopathology test re-
sults, it is necessary to perform external quality control.
© 2013 S. Karger AG, Basel
Introduction
The cervical cancer screening programs based on
screening tests has proven its efficiency in many countries
where such programs are organized and mortality rates
have been reduced [1]. However, this screening test must
present a good diagnostic accuracy [1, 2].
The screening tests show a high percentage of false
negatives (FN), with a range of 2–62%. The main causes
of the low sensitivity are related to mistakes during the
collection of material, the examination of the smear or the
interpretation of results [3–5]. Programs of quality con-
trol in cytopathology are an alternative that attempts to
minimize these mistakes [6, 7].
Internal quality control is a necessary tool in the rou-
tine of laboratories in order to reduce mistakes in the
assessment and interpretation of screening tests [8, 9].
However, the evaluation of the performance of different
laboratories can only be achieved through external qual-
ity control [10].
Key Words
Quality control · Cytology · Cervical screening
Abstract
Objective: To compare the variability of screening tests held
at laboratories with the Unit for External Quality Control
(UEQC), checking the frequency of cases that were discor-
dant, false-positive, false-negative, unsatisfactory or that had
a delay in clinical management and diagnostic agreement.
Materials and Methods: The study analyzed 10,053 screen-
ing tests from January 2007 to December 2008, including all
positive cases, all those that fall under unsatisfactory and at
least 10% of negative screening tests. The magnitude of the
agreement was analyzed using the kappa coefficient. Re-
sults: Out of the 10,053 cases analyzed, 7.59% were consid-
ered disagreeing, and it was estimated that 1.1% were false-
negative. There was a delay in the clinical procedure regard-
ing 2.44% cases. There were 2.82% of cases identified as
false-positive and 1.24% as unsatisfactory. The diagnostic
agreement was excellent (kappa = 0.81). The agreement of
most laboratories concerning screening tests was classified
as very good. The agreement of the sample adequacy was
reasonable (kappa = 0.30) and the agreement regarding the
representation of epithelia was considered excellent. Con-
clusion: Most laboratories showed very good agreement;
however, it is worthy of note that to establish the standard-
Received: January 14, 2013
Accepted after revision: June 19, 2013
Published online: October 1, 2013
Correspondence to: Dr. Cinara Zago Silveira Ázara
Avenida Engenheiro Eurico Viana, Quadra 1, Lotes 1/2/3
nº 30 Apartamento 1106, Bloco A, Residencial Spazzio Gran Ville Bairro Alto da Glória
Goiânia, Goiás 74815-725 (Brazil)
E-Mail cinarazago @ hotmail.com
© 2013 S. Karger AG, Basel
0001–5547/13/0576–0585$38.00/0
www.karger.com/acy