Clin Chem Lab Med 2014; 52(12): 1739–1745
Chinelo P. Onyenekwu, Careen L. Hudson, Annalise E. Zemlin* and Rajiv T. Erasmus
The impact of repeat-testing of common chemistry
analytes at critical concentrations
Abstract
Background: Early notification of critical values by the
clinical laboratory to the treating physician is a require-
ment for accreditation and is essential for effective patient
management. Many laboratories automatically repeat a
critical value before reporting it to prevent possible mis-
diagnosis. Given today’s advanced instrumentation and
quality assurance practices, we questioned the validity
of this approach. We performed an audit of repeat-test-
ing in our laboratory to assess for significant differences
between initial and repeated test results, estimate the
delay caused by repeat-testing and to quantify the cost of
repeating these assays.
Methods: A retrospective audit of repeat-tests for sodium,
potassium, calcium and magnesium in the first quarter of
2013 at Tygerberg Academic Laboratory was conducted.
Data on the initial and repeat-test values and the time that
they were performed was extracted from our laboratory
information system. The Clinical Laboratory Improvement
Amendment criteria for allowable error were employed to
assess for significant difference between results.
Results: A total of 2308 repeated tests were studied.
There was no significant difference in 2291 (99.3%) of the
samples. The average delay ranged from 35 min for mag-
nesium to 42 min for sodium and calcium. At least 2.9%
of laboratory running costs for the analytes was spent on
repeating them.
Conclusions: The practice of repeating a critical test result
appears unnecessary as it yields similar results, delays
*Corresponding author: Annalise E. Zemlin, MBChB, FCPath (SA)
Chem, MMed (Chem Path), Division of Chemical Pathology, National
Health Laboratory Service (NHLS), Tygerberg Hospital, University of
Stellenbosch, PO Box 19113, Tygerberg 7505, Parow, South Africa,
Phone: +27 21 9384254, Fax: +27 21 9384640,
E-mail: azemlin@sun.ac.za
Chinelo P. Onyenekwu: National Health Laboratory Service (NHLS),
Division of Chemical Pathology, Tygerberg Hospital, University of
Stellenbosch, Cape Town, South Africa; and Department of Clinical
Pathology, Lagos University Teaching Hospital, Idi-Araba, Lagos,
Nigeria
Careen L. Hudson and Rajiv T. Erasmus: National Health Laboratory
Service (NHLS), Division of Chemical Pathology, Tygerberg Hospital,
University of Stellenbosch, Cape Town, South Africa
Brought to you by | New York University Bobst Library Technical Services
Authenticated
Download Date | 5/20/15 10:11 AM
notification to the treating clinician and increases labora-
tory running costs.
Keywords: audit; critical values; repeat-testing; turn­
around time.
DOI 10.1515/cclm-2014-0331
Received April 14, 2014; accepted May 23, 2014; previously pub-
lished online June 18, 2014
Introduction
The practice of repeating a test to ensure its accuracy
was historically necessary when testing was performed
using instruments with poor analytical precision [1]. In
many laboratories this practice has persisted for critical
values within the analytical measurement range, despite
major advances in areas, such as automation, precision
performance and quality management. Non-critical
laboratory test results are usually not repeated before
reporting [1–3].
The term ‘critical value’ was introduced by Lundberg,
over 40 years ago. It refers to a laboratory test result which
represents a pathophysiological state considered to be
life-threatening or which can result in severe morbidity
if urgent action is not taken, and for which a corrective
action could be taken [4]. Following the introduction of
this concept, various regulatory and accreditation bodies
such as the Clinical Laboratory Improvement Amend-
ments (CLIA) [5], the College of American Pathologists
(CAP) [6], the Joint Commission (JC) [7] and the South
African National Accreditation System (SANAS) [8],
require clinical laboratories to promptly notify every criti-
cal result to the clinician. These bodies, however, have
no stated requirement for a laboratory to repeat a critical
value before reporting it. There are also no standard pro-
tocols on the establishment of critical values, hence criti-
cal values may not be harmonised across laboratories [9].
In more recent years, following the Institute of Medicine’s
report ‘To err is human; building a safer health system’
[10], much emphasis has been placed on patient safety
[11]. The process of critical value reporting has gained
increasing importance and in 2004 the JC mandated
1740      Onyenekwu et al.: Impact of repeat-testing
that critical value reporting be included in its National
Patient Safety Goals [12]. However, recent literature has
placed emphasis on errors in the extra-analytical phases
(pre- and post-analytical phases). Errors or delayed
communication of critical values may be a potential for
post-analytical errors [13]. Medical decisions and patient
outcomes are highly dependent on test results [14]. It is
the responsibility of the laboratory to ensure accurate and
timely reporting [15]. In light of the attention on patient
safety and the importance of ensuring the early report-
ing of critical results, several issues arise concerning
the validity of repeating a critical value. Given that non-
critical results within an analytical run are expected to
be accurate and are therefore not repeated before report-
ing, one would hypothesise that critical results within
the same run, with values within the analytic measure-
ment range, will invariably be accurate and need not be
repeated. If these results are accurate, there will be no
significant difference in the results of the repeated tests
and the initial tests. It may then be inferred that repeat-
testing of critical values only delays the turnaround time
(TAT) of these results and the early notification which is
required to ensure patient safety. A recent CAP Q-Probes
survey reported that the practice of repeat-testing of
chemistry analytes at critical values continues in 61% of
the laboratories. The survey also suggested that the prac-
tice did not prevent analytic errors [16].
Disorders of sodium (Na), potassium (K), calcium (Ca)
and magnesium (Mg) homeostasis are common and can
be life-threatening if severe. However, they can be readily
reversed with timely and appropriate medical interven-
tion. Tygerberg Hospital is a 1384-bed facility in Cape Town.
It is one of two tertiary institutions serving the needs of
approximately 5.8 million people within the Western Cape
region of South Africa. It is the second largest hospital in
the country and serves as a referral centre for many hos-
pitals and primary healthcare clinics within the region.
The Chemical Pathology Laboratory serves both the hos-
pital and the smaller satellite hospitals and clinics. It is an
accredited facility, owned and operated by the National
Health Laboratory Service (NHLS). The laboratory handles
over 1.2 million chemistry test requests annually. Due to a
persistent problem of prolonged TAT in our laboratory we
decided to perform an audit of the repeat-testing process.
It is our policy to only change operating procedure follow-
ing documented analysis. We believed that an audit inves-
tigating the impact of repeating the critical values of these
analytes would enable us monitor and improve our labo-
ratory services. We conducted a retrospective audit of all
repeat-tests performed on these four analytes within the
first quarter of 2013. The objectives were to determine the
Brought to you by | New York University Bobst Library Technical Services
Authenticated
Download Date | 5/20/15 10:11 AM
number of repeats carried out during the 3-month study
period and the delay caused by the practice of repeating
these samples. Significant differences between the initial
and repeat values were assessed and we aimed to esti-
mate the running cost of repeating critical values of these
analytes in our laboratory. Subjectively, we decided that
if  < 5% of repeats on any of the four studied analytes dem-
onstrated a significant difference between the initial and
final result, then the practice of repeating critical values
for that analyte should be discontinued.
Materials and methods
Study background and design
This was a retrospective audit carried out on all repeat tests per-
formed on Na, K, Mg and Ca from the 1 January, 2013 to 31 March,
2013 at the Chemical Pathology Laboratory, Tygerberg Hospital, Cape
Town, South Africa. The study was approved by the Ethics Committee
of Stellenbosch University and performed according to the Declara-
tion of Helsinki. As this was a retrospective study with consent taken
from patients prior to sample collection, the Ethics Committee of Stel-
lenbosch University waived the need for individual patient consent.
Patient confidentiality was maintained throughout data analysis.
At Chemical Pathology, NHLS Tygerberg, any result exceeding
the critical limits is flagged by the machine, as either ‘H’ (high) for
values above the upper limit, or ‘L’ (low) for values less than the
lower limit of the non-critical range. Tests are automatically repeated
by the chemistry analyser at the end of the ongoing run (Na and K), or
repeated by the technologist on duty (Mg and Ca). Inclusion criteria
were all Na, K, Mg, and Ca results flagged with a ‘HR’ (high repeat)
or an ‘LR’ (low repeat) sign, in the first quarter of 2013. We excluded
all Na, K, Mg and Ca results flagged with a ‘HR’ or an ‘LR’ sign where
the initial assay was flagged with an ‘S’ for short sample. We also
excluded all test values outside the analytical measurement range
and all repeat assays where results were invalidated due to haemoly-
sis.
Methods
All repeat values for Na, K, Ca and Mg with corresponding time of
test, were extracted from our laboratory information system (LIS)
Disalab by searching for results flagged ‘HR’ and ‘LR’ within the study
period. The corresponding initial values and time of testing for these
extracted repeat values (flagged ‘H’ and ‘L’, respectively), were then
extracted. A significant difference between the initial and repeat-
tests was regarded as having occurred if the absolute bias (Na, K and
Ca) or the percentage bias (Mg) exceeded the CLIA proficiency test-
ing criteria [17]. These criteria are displayed in Table 1. Delay caused
by repeat-testing was determined by the time-difference between the
initial assay and the repeat assay. Bias was determined as the differ-
ence between the repeat value and the initial value. Percentage bias
was calculated thus: [{repeat value–initial value}/ initial value] × 100.
 Onyenekwu et al.: Impact of repeat-testing      1741

Table 1 Details of the assay methods for the four analytes.

Analyte  Method   CLIA TEa   Critical value  Reference  Analytical  Instrument flag for
high/low, mmol/L range, mmol/L range, mmol/L repeat high/low, mmol/L

Na   ISE    ± 4.0 mmol/L   155.0/120.0  135–147  100.0–200.0    ≥  155.0/  ≤  120.0

K   ISE    ± 0.5 mmol/L   6.0/3.0  3.3–5.3  1.0–10.0    ≥  6.0/  ≤  3.0
Ca   o-CPC    ± 0.25 mmol/L  2.8/1.6  0.65–1.10  0.25–3.75    ≥  2.8/  ≤  1.6
Mg   Modified xylidyl blue reaction   ± 25%   0.55  0.80–1.40  0.29–2.06    ≤  0.55

CLIA Tea, Clinical Laboratory Improvement Amendments total allowable error; ISE, ion selective electrode; o-CPC, o-cresolphthalein
complexone.

Cost assessment
An estimate of the running cost for each analyte was determined
from the proportion of repeat-tests performed in relation to the total
assays for the study period.
Assays
The analytes of interest were assayed on the Siemens ADVIA 1800®
using reagents from Siemens Medical Solutions Diagnostics, Tarry-
town, NY, USA. Table 1 is a summary of the assay methods for the
four analytes, including their critical limits, analytical measurement
range and values which the instrument flags for repeat. The Na and K
methods [18] are traceable to a flame photometric reference method
which uses reference materials from the National Institute of Stand-
ards and Technology (NIST). The Mg [19] and Ca [20] methods are
traceable to NIST atomic absorption reference methods.
Data analysis
Data was extracted into Microsoft Excel® and analysed using Ana-
lyse-It® for Excel software version 2.30 (Analyse-It Software Ltd,
Leeds, UK) and STATISTICA version 11.0 (Stat-Soft Inc., Tulsa, OK,
USA, 2012). A Bland-Altman difference plot with CLIA total allowable
error (TEa) limits, was employed to determine agreement between
initial and repeat-test values for each analyte.
Results
A total of 90,477 relevant tests were performed in the first
quarter, 31,964 (35.3%) were for Na, 34,763 (38.4%) for K,
12,392 (13.7%) for Ca and 11,358 (12.6%) for Mg. Of these
90,477 tests, 2624 (2.9%) repeat-tests were performed
during the study period. Seventy-five (2.9%) of these
repeat-tests were due to initial short sampling, while 241
(9.2%) were haemolysed. These 316 (12%) repeat-tests
were excluded from further analysis. Our sample therefore
consisted of 2308 repeats performed on 1713 samples. This
discrepancy is explained by more than one repeat occa-
sionally being performed on a single result and multiple
analytes being repeated in a single sample.
K was the most commonly repeated test, constituting
1711 (74.1%) of those repeated. This was followed by Na
with 362 (15.7%) repeats, Ca with 144 (6.2%) repeats and
finally 91 (4%) of the repeats were on Mg assays (Table 2A).
Table 2B is a summary of the delay (in minutes) in the
TAT of each analyte due to repeat-testing at critical con-
centrations. It includes the delay experienced when mul-
tiple repeats were performed on a single sample (MR). The
median and interquartile range of delays experienced on
each analyte are also shown. The longest delay observed
on any analyte was 796  min while the shortest delay
observed was 3 min. The minimum delay observed for Ca
and Mg repeats were 11 and 12 min, respectively.
There was no significant difference between the
initial and repeat-test results in 2291 (99.3%) of the tests.
Figures 1–4 show difference plots of the initial and repeat-
test values for the various analytes, with the CLIA targets
for each. In assessing for significant differences; eight
(2.2%) of repeated Na values were different from the initial
results (Figure 1), seven (4.9%) of the repeat-test values
performed on Ca were found to be different from those

Table 2A Summary of observations on the frequency of repeats for the various analytes.

Analyte  Total number  Number of single  Maximum number  Median number  Mean number
of repeats repeats (%) of multiple repeats of repeats of repeats

Na   362  275 (76.0%)  14  1.0  1.3

K   1711  1342 (78.0%)  6  1.0  1.3
Mg   144  137 (95.1%)  3  1.0  1.1
Ca   91  88 (96.7%)  2  1.0  1.0

Brought to you by | New York University Bobst Library Technical Services

Authenticated
Download Date | 5/20/15 10:11 AM
 1742      Onyenekwu et al.: Impact of repeat-testing

Table 2B Distribution of delay (in minutes) observed for the Difference plot for K
1 Identity
various analytes.
TEa (0.5)
Delay due to single repeats  Na  K  Ca  Mg
0.5

Difference, Repeat-Initial value in mmol/L

IQ1   21:57  26:06  27:30  24:41
Mean   41:44  41:19  41:47  34:33
Median   30:59  31:17  31:38  30:59 0
IQ3   36:00  38:15  40:52  37:02

IQ1, first interquartile; IQ3, third interquartile. -0.5

Difference plot for Na -1

8 Identity

TEa (4) -1.5

6
Difference, Repeat-initial value in mmol/L

4
-2
0 2 4 6 8 10 12
2 Initial value in mmol/L

Figure 3 Bland-Altman plot for K. TEa = CLIA total allowable error in

0
mmol/L (0.5 mmol/L).

-2
Difference plot for Mg
1 Identity
-4
TEa
0.8 (25%)
Difference (Repeat-Initial value in mmol/L)

-6
0.6

-8 0.4
100 120 140 160 180 200
Initial value in mmol/L 0.2

Figure 1 Bland-Altman plot for Na. Tea = CLIA total allowable error in 0

mmol/L (4 mmol/L).
-0.2

Difference plot for Ca -0.4

Identity
0.4 -0.6
TEa (0.25)
-0.8
Difference, Repeat-Initial value in mmol/L

0.2 -1
0 0.5 1 1.5 2 2.5 3 3.5
Initial value in mmol/L

E-15 Figure 4 Bland-Altman plot for Mg. TEa = CLIA total allowable error
in % bias (25%).

-0.2

of the initial assays (Figure 2), two (0.01%) of repeat-test

values on K were different from the initial values (Figure 3)
-0.4
and none of the repeat-test values on Mg were signifi-
cantly different from those of the initial assays (Figure 4).
-0.6 Table 3 shows the specifics of the 17 results whose initial
0.5 1.5 2.5 3.5 4.5 and repeat values were different.
Initial value in mmol/L
The running cost of performing repeats on the four
Figure 2 Bland-Altman plot for Ca. TEa = CLIA total allowable error in analytes was estimated to be 2.9% of the running cost
mmol/L (0.25 mmol/L). of assaying the analytes for the first quarter for 2013.

Brought to you by | New York University Bobst Library Technical Services

Authenticated
Download Date | 5/20/15 10:11 AM
 Onyenekwu et al.: Impact of repeat-testing      1743

Table 3 Details of the 17 significantly different repeat-tests.

Analyte  Initial value,  Repeat value,  Absolute difference,  Tea,  Critical limit,
mmol/L mmol/L mmol/L mmol/L mmol/L

Ca   1.04  1.47  0.43   ± 0.25    ≥  2.8 or   ≤  1.6

Ca   1.19  1.49  0.3   ± 0.25    ≥  2.8 or   ≤  1.6
Ca   3.78  3.27  –0.51   ± 0.25    ≥  2.8 or   ≤  1.6
Ca   1.52  1.82  0.3   ± 0.25    ≥  2.8 or   ≤  1.6
Ca   3.26  2.96  –0.3   ± 0.25    ≥  2.8 or   ≤  1.6
Ca   1.56  1.29  –0.27   ± 0.25    ≥  2.8 or   ≤  1.6
Ca   1.25  0.98  –0.27   ± 0.25    ≥  2.8 or   ≤  1.6
K   8.5  7.4  –1.1   ± 0.5    ≥  6.0 or   ≤  3.0
K   7.8  6.0  –1.8   ± 0.5    ≥  6.0 or   ≤  3.0
Na   161.0  168.0  7.0   ± 4.0    ≥  155 or   ≤  120
Na   169.0  176.0  7.0   ± 4.0    ≥  155 or   ≤  120
Na   163.0  169.0  6.0   ± 4.0    ≥  155 or   ≤  120
Na   162.0  167.0  5.0   ± 4.0    ≥  155 or   ≤  120
Na   165.0  170.0  5.0   ± 4.0    ≥  155 or   ≤  120
Na   160.0  165.0  5.0   ± 4.0    ≥  155 or   ≤  120
Na   158.0  163.0  5.0   ± 4.0    ≥  155 or   ≤  120
Na   177.0  171.0  –6.0   ± 4.0    ≥  155 or   ≤  120

Tea, CLIA total allowable error.

Excluding the 74 repeat-tests performed due to short sam-
pling, we estimate 2.8% of the laboratory running costs for
the analytes are spent on repeat-testing.
Discussion
The delivery of timely and accurate results is a vital com-
ponent of good laboratory practice. Clinical laboratories
are required to promptly notify a patient’s care giver of
any critical value [5–8]. Many laboratories thus repeat all
critical results before notification, in order to avoid report-
ing false positives [14, 21]. In a recent large cross-sectional
study of 599 laboratories in China, Zeng et  al. found that
more than 85% of critical values undergo repeat-testing
prior to clinician notification and also state that this may
be unnecessary [21]. This practice may lead to operational
inefficiencies by delaying the TAT of these critical results
and compromise patient safety. Here, we have examined
the process of repeat-testing of four common chemistry ana-
lytes at critical concentrations in a tertiary care laboratory.
Only 17 (0.7%) of the 2308 repeats in our study were
observed to be significantly different from the initial assay
values. None of the four analytes had a proportion of sig-
nificantly different results exceeding the 5% criterion
established at the start of the audit. This would indicate
that the practice of repeat-testing these analytes in our lab-
oratory should be terminated. Ca repeats had the largest
proportion (4.9%) of significantly different results. This,
we believe, is a reflection of this assay’s poor analytical CV
in our laboratory. This problem is currently under inves-
tigation and appears to be due to impurities in our water
supply. Most of these significantly different repeat results
(5 out of 7) observed for Ca were seen at low critical levels,
whereas the significant differences found in Na and K
repeat results were observed at high critical values.
The average delay on a single repeat for the four ana-
lytes was from 35 min (Mg) to 42 min (Na and Ca). There
was a high degree of variance with delays being as great
as 796  min or as low as 3 min. The shortest delays of 11
and 12 min observed on Ca and Mg repeat assays, respec-
tively, were longer than the shortest delay of 3 min each,
observed on K and Na assays whose repeat-testing is auto-
mated. Some of the critical results were repeated more
than once before reporting. In one instance, a Na test
was performed 14 times after the first result. This was an
unusual case and we believe the specimen may have been
stuck in the carousel, resulting in continued repeats.
Our observations show that only 2.9% repeated tests
are relevant as they are due to short sampling. Additionally,
a substantial number of repeat-tests (9.2%) are performed
on samples which ought not to have been assayed in the
first instance. These include haemolysed samples for the
K assay, which were eventually invalidated. Our findings
also imply that the average delay experienced when repeat-
testing an analyte is at least half the time required for its
notification and at times several times the required noti-
fication time. When multiple repeat-tests were performed
for an analyte there was no significant difference in the test
values, only a prolongation of the TAT and reagent wastage.

Brought to you by | New York University Bobst Library Technical Services

Authenticated
Download Date | 5/20/15 10:11 AM
1744      Onyenekwu et al.: Impact of repeat-testing
Repeating 2.9% of the total test requests for the ana-
lytes of interest in our laboratory is similar to the finding
of Deetz et al. [22], who reported repeat-testing on 3% of
the tests they examined. Our observation of 99.3% agree-
ment in the results of repeat and initial results is also
similar to the 99.5% agreement reported by this group
but slightly higher than the 97.6% agreement reported
by Chima et al. [1]. The high number (8) of significantly
different results identified in Na repeat-tests is similar
to those of Deetz et al. [22], who noted that the CLIA cri-
terion of  ± 4  mmol/L for Na assays may be analytically
significant but clinically insignificant. Among the eight
significantly different results we identified in Na repeat-
tests, the observed differences did not exceed 7.0 mmol/L.
The delays we observed in the reporting of critical values
were much longer than the 10–14 min delay experienced
by most of the participants of the Q-PROBES study [14],
or the quarter of an hour delay reported for K by Deetz
et al. [22].
Our study has certain limitations. We have exam-
ined only four chemistry analytes, therefore our find-
ings and inference cannot extend to other chemistry
analytes. Second, we have utilised data from a single
laboratory and hence the precision performance and
quality assurance practices which affect these data are
peculiar to our instrument and our laboratory. Finally,
we have not objectively assessed the impact of the
delays caused by repeat-testing on patient safety, by
examining patient case files. We would expect varying
impacts depending on the nature of illness and the
location of the patient, e.g., a patient in the intensive
care unit compared to an out-patient. Despite these
limitations, our study is a robust one which reviewed
results of common chemistry analytical procedures and
our findings can be verified in any clinical chemistry
laboratory which performs these assays. We have also
utilised the CLIA criteria for allowable error in assess-
ing for significant differences because these criteria are
universally understood.
In conclusion, our findings suggest that repeat-testing
of common chemistry analytes at critical concentrations is
irrelevant as it yields similar results, delays the notifica-
tion of actionable results, compromises patient safety and
increases laboratory running costs. We recommend that
other laboratories engaging in this practice should review
their processes with the aim of improving clinical effec-
tiveness and operational efficiency.
Acknowledgments: The authors are grateful to
Dr. M. Rensburg, S. Ratcliffe and Prof. M. Kidd for assis-
tance with the statistical analysis.
Brought to you by | New York University Bobst Library Technical Services
Authenticated
Download Date | 5/20/15 10:11 AM
Conflict of interest statement
Authors’ conflict of interest disclosure: The authors
stated that there are no conflicts of interest regarding the
publication of this article.
Research funding: None declared.
Employment or leadership: None declared.
Honorarium: None declared.
References
1. Chima HS, Ramarajan V, Bhansali D. Is it necessary to repeat
critical values in the laboratory?: today’s technology may have
the answers. Lab Med 2009;40:453–7.
2. Kost GJ. Critical limits for urgent clinician notification at US
medical centres. J Am Med Assoc 1990;263:704–7.
3. Toll AD, Liu MJ, Gulati G, Behling EM, Kocher WD. Does routine
repeat testing of critical values offer any advantage over single
testing? Arch Pathol Lab Med 2011;135:440–4.
4. Lundberg GD. When to panic over abnormal values. MLO Med
Lab Obs 1972;4:47–54.
5. Centers for Medicare & Medicaid Services, Department of Health
and Human Services. Clinical laboratory improvement amend-
ments of 1988. Fed Regist 2003;1047:68. Codified at 42 CFR
1493.1299(g).
6. College of American Pathologists. Laboratory Accreditation Pro-
gram. Laboratory general checklist [components GEN.41320,
GEN.41330, and GEN.41340]. Available from: http://www.cap.
org./apps/docs/pathology_reporting/LabGeneralChecklist_
Reporting.pdf. Accessed on 19 September, 2013.
7. The Joint Commission. National patient safety goals effec-
tive January 1, 2013. Laboratory Accreditation Program 2013.
Available from: http://www.jointcommission.org/assets/1/18/
NPSG_Chapter_Jan2013_LAB.pdf. Accessed on 19 September
2013.
8. ISO 15189:2007(E). Medical laboratories – particular require-
ments for quality and competence.
9. Hashim I A, Cuthbert JA. Establishing, harmonizing and
­analyzing critical values in a large academic health center.
Clin Chem Lab Med 2014;52:1129–35.
10. Kohn KT, Corrigan JM, Donaldson MS. To err is human: building
a safer health system. Washington, DC: National Academy
Press, 1999.
11. Plebani M. Errors in clinical laboratories or errors in laboratory
medicine. Clin Chem Lab Med 2006;44:750–9.
12. Dighe AS, Rao A, Coakley AB, Lewandroski KB. Analysis of
laboratory critical value reporting at a large academic medical
center. Am J Clin Pathol 2006;125:758–64.
13. Piva E, Plebani M. From “panic” to “critical” values: which
path toward harmonization? Clin Chem Lab Med 2013;51:
2069–71.
14. Report of the review of NHS pathology services in England.
An independent review for the Department of Health. London:
Department of Health, 2006:6.
15. Campbell CA, Horvath AR. Towards harmonisation of
­critical laboratory result management – Review of l­ iterature
 Onyenekwu et al.: Impact of repeat-testing      1745

and ­survey of Australasian practices. Clin Biochem Rev dimethylcarboxanilido)-naphthalene-1′-(2-hydroxybenzene-5-

2012;33:149–60.
16. Paxton A. Critical value repeats: redundancy, necessity? CAP
Today 2010;24:1.
17. Westgard JO. CLIA requirements/tables on CLIA proficiency test-
ing criteria for acceptable analytical performance. Available from:
www.westgard.com/clia.htm. Accessed on 14 August, 2013.
18. Eisenman G. Glass electrodes for hydrogen and other ­cations,
principles and practice. New York: Marcel Dekker Inc., 1967:2.
19. Mann CK, Yoe JH. Spectrophotometric determina-
tion of magnesium with sodium 1-azo-2-hydroxy-3-(2,4
sulfonate). Anal Chem 1956;28:202–5.
20. Gitelman HJ. An improved procedure for the d ­ etermination
of ­calcium in biochemical specimens. Ana Biochem
1967;18:521–31.
21. Zeng R, Wang W, Wang Z. National survey on critical values
notification of 599 institutions in China. Clin Chem Lab Med
2013;51:2099–107.
22. Deetz CO, Nolan DK, Scott MG. An examination of the usefulness
of repeat testing practices in a large hospital clinical chemistry
laboratory. Am J Clin Pathol 2012;137:20–5.

Brought to you by | New York University Bobst Library Technical Services

Authenticated
Download Date | 5/20/15 10:11 AM