Revista Romn de Medicin de Laborator Vol. 18, Nr.

1/4, Martie 2010

65

COURSE NOTES
A model for calculating measurement uncertainty in
medical laboratories
Model pentru calcularea incertitudinii de msurare n
laboratoarele medicale
Irina L. Dumitriu1, Bogdan Gurzu1, Simona M. Sltineanu1, Liliana Foia2,
Traian Mutiu3, Corina chiriac4, Mioara Achirecesei4, Maria Enea4
1. Department of Functional Sciences, Faculty of Medicine, Gr. T. Popa University of Medicine and
Pharmacy, Iasi
2. Department of Semiology and Prevention, Faculty of Dental Medicine, Gr. T. Popa University of
Medicine and Pharmacy, Iasi
3. Politehnica University of Timisoara
4. Medical Investigations PRAXIS Iai.

Abstract
Introduction: All medical laboratories that require recognition for competency assessment have to estimate the uncertainty of measurement of assay test results where relevant and possible (ISO 15189:2007
Medical laboratories - Particular requirements for quality and competence). The repeated quantitative examination of an analyte with the same method will offer more or less different results. This is happening because the
outcome of an assay depends not only upon the analyte itself, but also upon few error factors that could yield
doubts about the obtained result. The mathematical, quantitative expression of this doubt is known as uncertainty
of measurement (UM). Methods: It is the responsibility of each medical laboratory to identify all error sources
that can be quantified and converted in standard deviations that could be used to estimate the type A or B of uncertainty. In the case of Romanian medical laboratories, the European Accreditation (EA) accepted as reference
documents for UM estimation the Guide to the Expression of Uncertainty in Measurement (GUM) and Romanian
Standard SR ENV 13005. Discussion and conclusion: In this paper, authors present and discuss the modalities
of UM estimation in two different situations: when the used reference materials (calibrators) are or are not
traceable to certified reference materials (CRM). Complete and informative UM reporting can only lead to better
decisions in healthcare.
Keywords: precision, accuracy, uncertainty of measurement, calibrator.

Rezumat
Introducere. Toate laboratoarele medicale care doresc recunoatere de competen trebuie s-i estimeze incertitudinea de msurare acolo unde este relevant i posibil (SR EN ISO 15189, Standard Romn, Laboratoare medicale, Cerine particulare pentru calitate i competen, 2007). n cazul unor msurari repetate
ale unui analit obinem rezultate diferite, mai mult sau mai puin apropiate ntre ele, dei valoarea analitului

66

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

este aceeai. Valoarea unui rezultat msurat nu depinde numai de valoarea nsi, ci i de o serie de factori de
eroare, care aduc o nencredere, un dubiu asupra rezultatului obinut. Exprimarea matematic, cantitativ a
acestei nencrederi se numete incertitudine de msurare (UM). Material i metod. Este responsabilitatea fiecrui laborator s-i indentifice toate sursele de eroare care pot fi cuantificate i convertite n deviaii standard
pe baza crora s-i estimeze UM de tip A i de tip B. Pentru laboratoarele medicale din Romnia, European
Acreditation (EA) recunoate ca documente de baz pentru estimarea UM, Guide to the Expression of Uncertainty in Measurement (GUM) i Standardul Romn SR ENV 13005 (Ghid pentru exprimarea incertitudinii de
msurare). Discuii i concluzii. Autorii prezint modalitile de estimare a UM n laboratoarele medicale din
ara noastr cnd calibratorii utilizai sunt trasabili sau nu la materiale de referin certificate (MRC). Raportarea corect i complet a UM influeneaz decizia terapeutic.
Cuvinte-cheie: precizie, acuratee, incertitudine de msurare, calibrator.

All medical laboratories that require recognition for competency assessment have to estimate the uncertainty of measurement of assay
test results where relevant and possible (1).
The term uncertainty means a doubt
and uncertainty of measurement (UM) is a
doubt on the validity of outcome measurements.
UM is only applicable to results of numerical
quantitative measurement.
According to GUM the result of a
measurement is only an approximation or estim-

ate of the value of the measurand and thus is

complete only when accompanied by a statement of the uncertainty of that estimate (2). Indeed, due to measurement uncertainty, a true
value of measurement can never be known.
There are several definitions of the UM:
estimation of the range of values within which
the true value of a measurand lies (2);
parameter characterizing the dispersion of the
quantity values being attributed to a measurand, based on the information used (3).
From daily practice it is well
known that when a measurement is repeated n times we obtain n different
results, although the analyte is the
same. This happens because the results
of a measurement are depending not
only upon the analyte itself, but also
upon a number of error factors that
could yield doubts about the estimate.
It is important not to confuse
the terms error of measurement and
uncertainty of measurement (2, 3).
An error of measurement is the difference between the measured value and
the true value of the analyte being
measured. As such, the error is a
single value that can be applied as a
correction to the result. The UM is a
range of values and this range cannot
be used to correct a measurements
Figure 1. Precision and accuracy of measurements (modified after result (4).
Mark Martinec http://www.ijs.si/time/).

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

67

X =

xi

x =1

n
n

(Xi X )2

SD = i = 1
n 1
SD
CV =
100 , where
X
X = average of measured values
Xi = individual measured values
SD = standard deviation
CV = coefficient of variation
n = number of measurements

Figure 2. Precision of measurements on control serum.

Results obtained by repeated measurements on a control material have a Gaussian distribution.

The UM is the number after sign

and it is expressed as standard deviation (or a
given multiple of it), or the half-width of an interval having a stated level of confidence (2).
Uncertainty of measurement comprises,
in general, many components grouped in 2 types
(A and B) derived from method for estimation
of their numerical value (2, 5, 6).
Type A uncertainty is obtained by calculation from a series of repeated measurements, using statistical methods. Type A uncertainty comprises random errors which arise
from random effects. Random errors cannot be
eliminated but the uncertainty due to their effect
may be reduced by increasing the number of
measurements and applying statistical analysis.
Type B uncertainty appears by means
other than those used for a Type A evaluation.
For example, by using data from outside
sources: calibration certificates, manufacturers
specifications, previous measurement data, experience with the behavior of the instruments,
intercomparison scheme and all other relevant

information. Type B uncertainty encompasses

systematic errors.
Each component of uncertainty is represented by an estimated standard deviation
named standard uncertainty (Ui) equal to the
positive square root of the estimated variance.
The performance of quantitative tests
from medical laboratory is attested by two components: precision and accuracy of measurements
(Figure 1) (7). These two components should be
taken into account for a reasonable estimation of
measurement uncertainty in medical laboratories.
Precision is characterized by the dispersion of values obtained by repeated measurements of an analyte and it is expressed by standard deviation (SD) and coefficient of variation
(CV) (Figure 2).
Precision of measurements is evaluated
by type A uncertainty (8).
Type A uncertainty is based on a statistical analysis of n different values obtained from
repeated measurements of an analyte using the
same method.

68

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

long period of time. For that reason,

medical laboratories estimate type A
measurement uncertainty by using
control material (in which the analytes
are stable for a long period of time).
Control material is homologue to patient sample. Control material is measured by the same measurement system, with the same calibrator and reagents as patient sample. Therefore, by
extrapolation, the precision of measurements using the control material is
identical with the precision of measurements using patient sample in
Figure 3. The accuracy of the measurement refers to closeness of which the analytes have only shortthe agreement between the result of a measurand and a true
term stability.
value.
Accuracy represents the difference
between
the result (the average
The mean is determined by adding a
of results) obtained by measurement and the true
group of measured values, then dividing the
value of a measurand which is expressed by Bias
total by the number of measurements in the
(Figure 3).
group as equation [1]:
Accuracy of measurements is evaluated
n
x
by
type
B
uncertainty. Type B uncertainty is de
x1 + x 2 + ... + x + ... + x n
i [1],
i
=
i
1
termined
by
other means than type A uncerX=
=
n
n
tainty, namely non statistical methods from a
where:
priori data:
uncertainty of the value assigned to the calibrators
X = mean of the measured values,
the data from external control (intercomparisons)
Xi = individual measurements,
traceable certificate report of equipment
n = number of the values Xi in the group.
data from validation process (2, 9).
Then the standard deviation is determThe estimation of the UM increases the
ined by equation [2].
reliability of the results obtained in medical
This standard deviation (SD) represented
laboratories. For a higher credibility, every single
by a statistically estimate (experimental standard
measured result must be accompanied by UM.
deviation) constitutes type A uncertainty.
The component standard uncertainties
There are many possibilities for assessing
are
combined
according to the law of propagatype A uncertainty. Type A uncertainty is based on
tion of uncertainty to produce an overall value
repeated measurements of the same analyte over a

(x1 x ) + (x 2 x )
2

Equation [2]:

SD =

+ ... + x x
n

n 1

Xi = individual values obtained from repeated measurements

X = mean of the measured values
n = number of measurements (2, 6).

n
2
Xi X
, where
= i =1
n 1

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

of uncertainty, known as the combined standard uncertainty (UMc). UMc is given by the
square root of the sum of the squares:

UM c = U2 + U2B [3].
A
Expanded uncertainty (U) is calculated by multiplying the standard uncertainty
UMc with an appropriate coverage factor k:
U = k UMc [4], for a desired level
of confidence of 95%, k is approximately 2.
Uncertainty reported with the result is
expanded uncertainty and is then conveniently
expressed as:
Y U [5], where:
Y = measured result,
U = expanded uncertainty.

The Guide to the expression of uncertainty in measurement (GUM) (2) was developed for estimating UM in all laboratories
types. In non-medical laboratories the potential
sources of uncertainty are usually readily identifiable, quantifiable and converted in standard
deviations in fields such as physical and chemical measurements (e.g. electrical, materials, optics, etc). These standard deviations are used for
calculation of UMc.
The Working Group for the Medical
Testing Laboratory recognizes that the implementation of the uncertainty of measurement requirement offers opportunities for pathology
laboratories to add value to their diagnostic services, particularly in educating users to better
understand the limitations of tests, and in recognizing when clinically significant changes in patient results have or have not occurred. ISO
15189 also recognizes that the rigor of estimating uncertainty of measurement may be based
on the needs of the client.
In medical testing there are many potential uncertainties that can significantly affect
test results (for example: poor specimen collection or transport, patient related factors such as

69

biological variation and the presence of drugs,

clerical and reporting errors etc). Not all these
factors can be quantified in standard deviations
and therefore such factors are excluded from the
estimation of uncertainty of measurement (7). It
is important to identify such factors and keep
them under an adequate management. In medical laboratories is difficult to strictly respect
GUM regulations. However, medical laboratory
measurements are strictly monitored by an internal and external control system, and therefore
the data generated can be used to estimate UM.
A basic requirement of GUM is to establish a mathematical expression which has to
include all the input quantities that significantly influences the test result. In medical
laboratories rigorously implementation of
GUM is difficult, but the above requirement of
GUM must be respected.
Whatever approach is used, estimation
of UM should consider the basic requirements
of GUM and the following steps (2, 4):
1. complete definition (or specification) of
the measurand;
2. careful consideration and understanding of
each phase of the measurement process;
3. mathematical expression of the equation
relating measurand (Y) and input quantities (Xi)
using a functional relationship in the form
Y = f ( X1, X2, X3...Xn). Functional relationship f
should enclose all input quantities that contribute
to the final result;
4. estimation of a certain value for all input
quantities in functional relationship f using statistical methods or other means;
5. evaluation of standard uncertainty for
every quantity Xi;
6. calculation of combined standard uncertainty;
7. calculation of expanded standard uncertainty with a chosen coverage factor;
8. reporting uncertainty of measurement.
1. Defining the measurand as fully as possible
in terms of designation, matrix, unit of measure,
because lack of measurand definition brings a

70

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

major uncertainty to the final test result. This

will yield major discrepancies between the
measurement results in different laboratories for
the same analyte.
Analyte is a term used to identify the
substance or constituent of interest (e.g. creatinine) that is the subject of measurement. However, a substance can have a number of properties (e.g. concentration, different matrix, different unit of measure, different measurement
methods), some or all of which can be utilized
to quantify the substance in an appropriate
measuring system. The particular quantifiable
property of the analyte used in the measuring
system is called the measurand (7).
For example, the analyte is creatinine
and the measurand is creatinine concentration in
serum, plasma or urine expressed in mg/dl or
mol/l. Therefore it is essential to define as
completely as possible the quantity that is measured (i.e. the measurand) by a given procedure.
2. Establishing the measurement principle of
the specified technical procedure for performing
a test concerning of all reaction phases is exemplified by the principle of measurement of creatinine in a biological sample. Picric acid reacts in
alkaline conditions with creatinine from analysed
biological sample to form picramic acid. This is a
colored product with a color intensity which is
measured at a certain wave length (510 nm) as
optic density directly proportional with creatinine
concentration from analysed biological sample.
3. Establishing a mathematical formula that
determines the functional relationship between
measured quantity Y and measured quantities
(Xi, input quantities upon which final test result
depends). In order to convert optical density
into a test result a reference material (calibrator)
with a known value of creatinine (measurand) is
used. The calibrator is measured in the same
conditions as the analyzed biological sample. At
the end of the reaction, optical density (OD)
read by the equipment for reference material has
the same value as the value assigned to the cal-

ibrator from the specifications of the manufacturer. This process it is called calibration and
represents a set of operations that establish, under specified conditions, the relationship
between values of quantities indicated by a
measuring instrument or measuring system, or
values represented by a material measure or a
reference material, and the corresponding values
realized by standards (3). Therefore, the result
y of creatinine from patient sample could be calculated by rule of three:
2500 OD................. 3.6 (calibrator value)

1800 OD.................. y

y=

ODsample
ODcalibrator

valuecalibrator [6],

where OD is optical density (of calibrator, respectively to patient sample).

Figure 4 represents the calibration
curve for creatinine in 2 points, respectively 0
and 3.6 mg/dl. If the calibrator value of 3.6
mg/dl includes UM expressed as range value
(see the dotted lines of Figure 4), the optical
density of final product reaction of patient
sample, 1800, will lead to different results according to the calibration curve used.
Since the reference materials (calibrators, etalons and standards) are very homogenous it is admitted that value distribution is rectangular (9). Thus, calibration curve can be situated anywhere in the range marked with dotted
lines, not necessarily in the middle of the range.
The constructed calibration curve could
be validated and used for carrying out the measurements on the patients samples only after is
verified using the control materials of known
values because a lot of error sources from used
equipment used, reagents, medium conditions,
laboratory technicians etc. may appear. The
measurement system (equipment, reagents, calibration curve and analyst) is verified in terms
of correctness by testing control materials on
different levels. These control materials have
predetermined values and could alert the analyst
regarding the errors of measurement system.

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

Figure 4. Calibration curve in 2 points.

The results obtained by measuring control materials are analyzed considering Westgard rules and the acceptance / rejection criteria
of each laboratory. Internal control tests with
control materials are similar to diagnostic tests
of patient samples. Similarly with diagnostic
tests in patients, which identify the health problems of the patient, the tests of internal control
identify the health problems (errors) of the
measurement system (10). Rapid detection and
adequate treatment is depending upon precise
and quick identification of the health
problem.
The reference materials (calibrators, etalons, standards) are considered Xi quantities
upon which final test result depends, while control material is used for verifying and validating
the correctness of calibration but does not have
a direct influence on final result.
4. Estimation of a value for all input quantities
The result of a quantitative measurement Y is a functional relationship f (Xi), where
Xi represents all the factors upon which the
measurement depends (X1, X2, X3 ... Xn) (4, 7).
All relevant sources of uncertainties that could

71

influence the test result should be identified.

Consideration should be given to all relevant
factors that may contribute to the overall uncertainty of a measurement and could influence the
test result. Some examples are given below:
X1 = factors affecting pre-analytical phase
cannot be usually quantified, but must be
controlled by an adequate management;
X2 = measurement procedure (method, reagents, measuring instruments, laboratory
hardware etc);
X3 = reference material (calibrator);
X4 = factors affecting post-analytical phase;
....
Xn = others factors.
In medical laboratories, after identification of all input quantities Xi, it could be observed that the measurement procedure used
(method, reagents, instruments, laboratory stuff
etc.) and the reference materials (calibrators) are
the only quantifiable factors that could be converted in standard deviations.
The best estimation for the measurement
procedure used is the average of the measured
values obtained in n repeated measurements (the
precision) between-day and within-day measuring technique (from data of internal control).
The estimation of the value of reference
material (calibrator) is offered by the manufacturer.
5. Estimation of standard uncertainty for
each quantity Xi
At least 30 values of control serum on
at least 30 different consecutive days were used
in the measurement procedure. In this case the
Xi input estimate is usually the average value.
All values from the internal control have Gaussian (normal) distribution. Standard uncertainty
U(xi) is the estimated standard deviation of the
mean calculated by formula [2]. This standard
uncertainty is estimated by statistical calculation
of measurement values and therefore it is named
Type A uncertainty.
When the medical laboratory uses reference materials which are not traceable to CRM,
the laboratory does not know the UM of the cal-

72

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

6. Calculation of combined standard

uncertainty (UMc)
Calculation of combined standard
uncertainty is done following equation
[5], in which the medical laboratory has
the responsibility to choose many UBi:

UMc = U 2 + U 2 + ... + U 2 [7]

A
B
B
1
n
where:
UA = U measurement procedure (from internal
quality control),
UB1 = Ucalibrator (from manufacturer specifications),
UB2 = UBias (from external quality control),
Uother factors (from speciality literature
data) (4).
Under these conditions, equation
[5] becomes equation [8].

Figure 5. UM of used reference material (calibrator)

represents type B uncertainty (UB) because it is taken from
manufacturer specifications ( X = average values measured). 7. Calculation of expanded standard un-

ibrator although this is a measured value and is

accompanied by UM. When used calibrators are
traceable to certified reference materials
(CRM), their values are accompanied by extended uncertainty. Uncertainty of the values assigned to the calibrator can be obtained by request from the manufacturer usually as extended
uncertainty (for 95% confidence interval of the
calibrator assigned value, k=2) (Figure 5).
Combined uncertainty of the values assigned to
the calibrator is used in medical laboratories, so
Uextended should be divided by 2.
Uncertainty of the values assigned to
the calibrator indicates the accuracy measurement as type B uncertainty because it is taken
from the manufacturers specifications and it is
not a result of statistical calculations performed
by the medical laboratory (11, 12).

certainty (UMexpanded) with a chosen coverage factor

Expanded uncertainty, usually shown
by the symbol U, is obtained by multiplying the
UMc by a coverage factor, denoted by symbol
k = 2, for a level of confidence of approximately
95% (2).
UM expanded = U = UMc  k [9].
8. Reporting uncertainty of measurement
Most laboratories have chosen until
now not to state measurement uncertainty in
their test reports. Instead, such information has
been given only when the physicians have specifically required it. It is important for laboratories to understand the clinical implications of the
results of the measurements they report and to
be aware of those where UM could affect clinical interpretations and patient management.

Equation [8]:
2
2
2
2
UM c = U measurementprocedure + Ucalibrator + UBias(external control) ... + Uother factors .

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

73

Figure 6. Combined standard uncertainty comprises type A standard uncertainty (UA, precision of
measurements) and type B standard uncertainty (UB, accuracy of measurements).
UA is statistically calculated from values of independent measurements with Gaussian distribution. Regarding
type B uncertainty, there are many kinds of UB according to literature specifications: UB1, UB2, .. ,UBn. When a
medical laboratory chooses to estimate its accuracy of measurements from Bias intercomparisons, the standard
deviation of Bias (Bias values from many participations to external control have rectangular distribution) has to
be calculated following the formula SD Bias =

a
, where a = Bias (adapted after SR ENV 13005:2005).
3

Uncertainty reported with the result is

expanded uncertainty and is then conveniently
expressed as:
Y U [10],
where:
Y = measured result,
U = expanded uncertainty.
The laboratory should offer detailed
data to users (physician, patient) about the pattern estimation used for UM.
Each laboratory has the responsibility to
identify all quantifiable error sources, which can

be converted in standard deviations used for estimating type A and type B standard uncertainty.
Type A standard uncertainty is assessed
from internal quality control data (statistical calculation on at least 30 values of control serum
from at least 30 different consecutive days).
Type A standard uncertainty illustrates the precision of laboratory measurements.
When the laboratory uses traceable calibrators to certified reference materials (CRM)
with attributed UM, this can be treated as type B
standard uncertainty. Type B standard uncertainty shows the accuracy of measurements.

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

74

Table 1. Estimation of UM when the medical laboratory has data about UMcalibrator
Quantity
Measurement
Units
Reference intervals
Test principle
Calibrator
traceability
Precision /
Imprecision
(Type A uncertainty)

CREATININE
Creatinine concentration in human serum
mg/dL, mol/L
Men 0,6 1,1 mg/dl
Women 0,5 0,9 mg/dl
53 97 mol/L
44 80 mol/L
Creatinine in alkaline solution reacts with picrate to form a colored complex. The rate of formation of
the complex is measured.
Is traceable to the NIST reference material 909b level 2 and by the reference method IC-GC/MS
Internal quality control data on multiparameter control serum for: 1.01.2009 31.03.2009
Number of measurements
72

Quality control MEAN

3.55 mg/dl

SD
0.17 mg/dl

CV (%)
4.84 %

Calibrator traceability
Uncertainty of
calibrator
(Type B uncertainty)

UMcombined (UMc)
UA

UB

Calibrator

Target value

Creatinine

3,6 mg/dl

Total Uncertainty
(Uexpanded)
0,26 mg/dl

Per cent Uncertainty

7,1%

Uncertainty is calculated as the half of the 95% confidence interval of the calibrator assigned value. The
true value should fall in the range (assigned valueuncertainty) with 95% probability.
The medical laboratory is responsable for taking account other types
2
2
UB from apriori data beside UMcalibrator.
UM c = U + U

Umeasurement procedure = standard deviation of quality control results for period 1.01.2009 31.03.2009
UA = 0,17 mg/dL
UMcombined (calibrator) obtained from manufacturer as standard deviation. The manufacturer offers
UMexpanded (for a level of confidence of 95%, k=2), so that the laboratory should divide by 2 the
UMexpanded offered by manufacturer.
UM
expanded 0, 26
UB = UMcombined (calibrator) =
=
= 0,13 mg/dL
k
2
UMc = UMcombined

2
2
UMc = SD (precision
from internal control) + UM combined (calibrator) =
= 0,17 2 + 0,132 = 0,30 2 = 0,30
UMc

Reporting UM

Note:
the traceability of calibrators to CRM provides the accuracy of measurements;
in this case, the usage of other types of UB is supplementary, but not mandatory because the traceability of calibrators to CRM covers the accuracy of measurements (Figure 6);
the proportion brought by other types of UB to the UMc is lower. The medical laboratory could
choose whether to take into account the Ubias (from external control) beside U calibrator for calculation of UMc.
U = UMexpanded = UMc  k = 0,30  2 = 0,60
k =2, for level of confidence of 95%
Y U, where
Y = measured result
U = expanded uncertainty
Mean value of creatinine 3.55 0,60 mg/dl

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

Unfortunately, in most medical laboratories from our country calibrators without traceable values are used. If an appropriate reference
material or reference procedure is unavailable,
then alternative approaches may be used, e.g.
external quality assessment data or inter-laboratory comparisons (13).
In this case, accuracy of measurements
is estimated by type B uncertainty calculated as
standard deviation of Bias values from external
control. Bias is the difference between measured result and average results of participant
medical laboratories based on an intercomparisons scheme (external control). Bias values from
many external control participations have a rectangular distribution (variance) given by:
(max min )2 = 2a 2 = 4a 2 [11] (6).
U 2 (x ) =
i
12
12
12
For this reason, standard deviation of
Bias is calculated by formula:

( )

U = SD =

4a 2
=
12

4a 2
2a
a
=
=
43 2 3
3

[12],

(Figure 6).
In this paper the authors summarize the
estimation of UM in two situations: when the
assigned values of traceable calibrators to CRM
are or not accompanied by UM as shown in
Tables 1 and 2.
The assigned values of traceable calibrators to CRM are accompanied by UM as
shown in Table 1.
Manufacturers do not always provide
the uncertainty of the values assigned to calibrators. In this case UM is estimated by following the model shown in Table 2.

Conclusions
1. Uncertainty Measurement of a measured
result increases its reliability.
2. A reasonable estimation of UM includes precision (UA) and accuracy of measurements (UB):

75

a. Precision is expressed as standard

deviation of the values obtained in internal quality control using a large number of measurements on at least 30 days
period.
b. Measurements results in medical
laboratories are used for clinical decisions, and therefore it is very important
to include accuracy of measurement in
UMcombined. Each medical laboratory has
the responsibility to guarantee the accuracy of measurements by using reference
materials (calibrators) with traceability to
CRM and / or participations of the laboratory to intercomparison schemes.
3. The laboratories should request data about
the traceability of calibrators from the manufacturers. Uncertainty of the value assigned to the
calibrator has a major importance in assessing
measurement accuracy.
4. When the medical laboratory does not have
data about the calibrator values UM, the accuracy of measurements could be estimated by
standard deviation of Bias from external quality
control (intercomparison schemes).
5. It is important for laboratories to understand
the clinical implications of the results of the measurements they report and to be aware of those results where UM could affect clinical interpretations
and patient management. The clinician needs to be
aware of the UM together with the result to make
a correct diagnosis. The uncertainty of the result is
important, e.g. when looking at the limits of reference interval or cut off. It is necessary to estimate
the UM when comparing results to allowable values, e.g. tolerance limits or allowable (legal) concentrations (doping control).
6. Where uncertainty of measurement information, if reported, could significantly affect clinical interpretations and patient management
(e.g. tumor marker monitoring), such information should be readily available on request and
comprehensive about the laboratorys method of
assessment.

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

76

Table 2. Estimation of UM when the medical laboratory does not have data about UMcalibrator
Quantity
Measurement
Units
Reference intervals
Test principle
Calibrator
traceability
Precision /
Imprecision
(Type A uncertainty)

Tyroid stimulating hormone (TSH)

TSH concentration in human serum.
UI/mL
0,4 6,0 UI/mL
The TSH ELISA test is based on the principle of a solid phase enzyme-linked immunosorbent assay. The assay system utilizes a unique mouse monoclonal antibody directed against a distinct antigenic determinant on the intact TSH molecule.
6 calibrators (included in kit): 0; 0,5; 2; 5; 10; 25 UI/ml
is not traceable to CRM
internal QC data on multiparameter control serum for: 1.01.2009 31.03.2009
Number of measurements
QC Mean
SD
CV (%)
48
7.94%
4.03 UI/ml
0.32 UI/ml
Standard deviation of Bias from external quality control (intercomparisons schemes)
Bias = measured result intercomparison average
The Bias values from the same period for Umeasurement procedure have a rectangular distribution, so

SD Bias =
Type B uncertainty

(Figure 6)

BIAS1 (Cycle 6/Sample 8) = 2.98 3.02 = - 0,04

BIAS2 (Cycle 6/ Sample 10) = 3.64 3.81 = - 0,17
BIAS3 (Cycle 6/ Sample 11) = 4,26 4,20 = +0,06

SD Bias =

BIAS max .

0,17

= 0,09 UI/ml

UMc = UMcombined

2
2
UM c = U A + U B
UMc

2
2
UMc = SD (precision from internal control) + SD Bias
=
(external control)
=

Reporting UM

2
2
UI/ml
0.32 + 0,09 = 0,33

Note: The medical laboratory is responsible for taking into account other UB types from apriori
data beside UMBias (from external control)
U = UMexpanded = UMc  k = 0,33  2 = 0,66 UI/ml
k =2, for level of confidence of 95%
Y U, where
Y = measured result
U = expanded uncertainty
Mean value of TSH 4.03 0,66 UI/ml

7. Assessing and reporting measurement uncertainty will help reduce the differences
between laboratories results, which translates
into prompt clinical decision, low costs, higher
efficiency and highest confidence of clinicians
and patients in testing results.

Acknowledgements
This study was supported by Medical
Investigations Praxis S.R.L., Iasi, Romania. and
the National University Research Council
CNCSIS, grant PN_2_ID_PCE_2670/2008.

Revista Romn de Medicin de Laborator Vol. 18, Nr. 1/4, Martie 2010

Abbreviations list
CRM
CV
EA
GUM
OD
SD
UM
UMc
UMexpanded, U

certified reference materials

coefficient of variation
European Accreditation
Guide to the expression of
uncertainty in measurement
optical density
standard deviation
uncertainty of measurement
combined standard uncertainty
expanded standard uncertainty

References
1. International Organization of Standardization, ISO/TC
212. ISO 15189:2007, Medical laboratories - Particular requirements for quality and competence, 2nd Edition.
Geneva, 2007: p37
2. ISO/IEC Guide 98-3:2008. Uncertainty of measurement
- Part 3: Guide to the expression of uncertainty in measurement (GUM:1995). International Organization for Standardization (ISO)/ International Electrotechnical Commission
(IEC): Geneva, 2008: p 3-58
3. BIPM, IEC, IFCC, ILAC, ISO, IUPAC, IUPAP, OIML,
The International Vocabulary of Metrology, Basic and General Concepts and Associated Terms - VIM, Joint Committee
for Guides in Metrology- JCGM 200:2008 [ISO/IEC Guide

77

99]: p 22-28
4. Eurachem / Citac Guide CG4, Quantifying Uncertainty
in Analytical Measurement, Second Edition, 2001: p16-30
5. Technical Report EA- 4/16 (rev00). Guidelines on the
expression of uncertainty in quantitative testing. European
Co-operation for Accreditation (EA), 2003: p 6-16
6. SR ENV 13005:2005. Romanian Standard for estimation of measurement uncertainty, Third Edition; 2005: p 8-18
7. White GH, Farrance I. Uncertainty of measurement in
quantitative medical testing: a laboratory implementation
guide. Clin Biochem Rev. 2004;25(4):S1-S24
8. Hibbert B. Quality assurance for the analytical chemistry laboratory. Oxford University Press US, 2007: p 255
9. Taylor BN, Kuyatt KE. Guidelines for Evaluating and
Expressing the Uncertainty of NIST Measurement Result National Institute of Standards and Technology Technical
Note 1297. U.S. Government Printing Office, Washington
DC, 1994: p 2-6
10. Are there similar performance characteristics for QC
and diagnostic tests? "Westgard Rules" and Multirules.
Westgard QC (n.d.) Retrieved November 01, 2009 from
http://www.westgard.com/mltirule.htm#performance
11. European Diagnostic Manufactures Association
(EDMA). Estimation of uncertainty of measurement in medical laboratories, 2006: p 2-5
12. The Romanian Accreditation Association (RENAR).
Instruction for validation of methods used in medical laboratories, RENAR Code: I-12, First Edition, 2007: p 1-9
13. White GH. Basics of estimating measurement uncertainty. Clin Biochem Rev. 2008 Aug;29 Suppl 1:S53-60