47
Materiales de referencia y comparaciones interlaboratorios
Z Score
Points Assigned
<1.01
5
1.01 – 2.0
4
2.01 – 3.00
2
>3.00
0
Table 1. Points are then assigned for each
results of the Z score.
Ec/1/
THE EFFECT OF PROFICIENCY TESTING
PARTICIPATION ON LABORATORY PERFORMANCE
Ken Middlebrook and Andrew Morris
Canadian Association for Environmental Analytical Laboratories. (CAEAL)
Suite 300,265 Carling Avenue,Ottawa,Ontario K1S 2E1
Abstract.
Data accumulated through the proficiency testing (PT) program operated by the Canadian Association for Environmental Analytical Laboratories
(CAEAL) was evaluated to determine if proficiency testing participation results in improved laboratory performance.The data shows that,on average,PT per-
formancescores increaseoverthefirst fewrounds followedbyaplateau.Thiscorrespondstoadecreasedvariation inscores.It issuggestedthatthecombina-
tion of proficiency testing and conformance to ISO/IEC 17025 through accreditation is the best strategy to ensure consistently high quality analytical results.
Keywords.
Proficency testing program, laboratory performance, ISO/IEC 17025
EFECTO DE LA PARTICIPACIÓN EN PRUEBAS DE APTITUD EN EL DESEMPEÑO DE LOS LABORATORIOS
Resumen:
Se evaluaron los datos acumulados del programa de la prueba de aptitud (PA) operado por la Asociación Canadiense para los Laboratorios
Analíticos medioambientales (CAEAL) de modo de determinar si la participación en este programa da lugar a una mejora en el funcionamiento de los labo-
ratorios.Los datos demuestran que,en promedio, los resultados del puntaje en la participación en las PA se incrementan en las primeras rondas para luego
llegar a un valor constante. Lo que se corresponde a una disminución en la variación de los puntajes.Se sugiere que la combinación de la prueba de aptitud
y de la conformidad a ISO/IEC 17025 con la acreditación resulta ser la mejor estrategia para asegurar una constante y alta calidad de los resultados analíticos.
Palabras claves.
Programas de pruebas de aptitud,desempeño de los laboratorios, ISO/IEC 17025
I. Introduction
ISO/IEC Guide 43 (1997) [1], ILAC-G13:2000 [2] and CAN-P-1593 (2001) [3] detail the requirements for the develop-
ment and operation of proficiency testing (PT) schemes and the use of proficiency testing schemes by laboratory ac-
creditation bodies.The premise behind these standards is that participation in a properly designed and coordinated
PT program is an essential part of the quality assurance of an analytical laboratory.
The Canadian Association for Environmental Analytical Laboratories (CAEAL) operates a PT program for environmen-
tal laboratories.This program supports both accredited laboratories (through a partnership with the Standards Coun-
cil of Canada) and non-accredited laboratories.Participating laboratories include commercial,industrial,government,
and public health labs (~350 participating laboratories). Samples are generally shipped to laboratories twice per year
and the labs report on up to 200 parameters. Laboratories are evaluated on each parameter based on consensus
statistics (Standard Operating Procedure for the CAEAL Proficiency Testing Program:
.
An earlier study [4] indicated that accredited laboratories perform better than non-accredited. This conclusion is
not unexpected given that accreditation is dependent on conformance to a quality management standard (ISO/IEC
17025 [5], CAN-P-4D [6]). Although this study also suggested that performance for both accredited and non-accred-
ited laboratories improve over time, too little data was available, over too short a duration, to draw firm conclusions.
Although it is logical to assume that active participation in a PT program will improve the quality (i.e., accuracy and
precision) of analytical results, the few studies designed to measure this have typically been restricted to clinical test-
ing [7,8].
The current study uses ten years’ worth of data from the CAEAL PT program to examine the effect that participation
in a PT program has on laboratory performance.
II. Methods
CAEAL’s proficiency testing samples consist of four samples (with different concentrations) shipped to each labora-
tory twice per year. Reported results are checked for normality, outliers (Grubbs test), and the adjusted mean and
standard deviation estimated. Each result is assigned a z score based on the following;
Z
=
x
−
X
s
Where: x are a reported result; X is a adjusted mean and s is a adjusted standard deviation .
z
