Development and Evaluation of a Logical Delta Check
for Identifying Erroneous Blood Count Results
in a Tertiary Care Hospital
Ira Miller, MD, PhD
Context.—Delta checks have been suggested to increase
patient safety by identifying preanalytic and analytic
errors, including wrong name mislabeling on the sample
tube.
Objective.—To implement an effective and practical
complete blood cell count (CBC) delta check by optimizing
specificity and sensitivity using weighted deltas of multiple
parameters.
Design.—The mean red blood cell volume (MCV) delta
(.3.0 fL) check was retrospectively assessed. The composite CBC delta (CCD) test was formulated using serial
same-patient CBC data and random interpatient CBCs. The
logical delta check (LDC) ignores CCD failures due to
platelet change only. The effect of LDC implementation
was evaluated.
Results.—The MCV delta check test recognized only 3 of
6 confessed mislabeled specimens in the initial review
period, whereas all were identified using the CCD. The
LDC flagged 2% (205 of 13 234) of eligible results, one-
third as many as the MCV delta check. The CCD and LDC
checks revealed 20 presumed or confirmed mislabeling
events, only half of which were caught by the MCV delta
check. Thirty-four percent of LDC failures not due to
transfusion reflected problematic results, including presumed or confirmed wrongly labeled patient samples (36%
of flags for real problems). Implementation of the LDC,
requiring immediate verbal feedback to the caregivers, was
associated with more retracted erroneous results in
patients’ medical records.
Conclusions.—The MCV delta check test was found not
to have led to correction of errors in our laboratory due to
impractically low specificity and sensitivity. The LDC is a
useful tool for identifying preanalytic and analytic specimen problems, including wrong name mislabeling on the
sample tube.
(Arch Pathol Lab Med. 2015;139:1042–1047; doi:
10.5858/arpa.2014-0494-OA)
M
quality management plan that covers all areas of the
laboratory and includes benchmarking key measures of
laboratory performance.’’2(GEN.13806, .20316) Two programs are
offered in the Q-Tracks Continuous Quality Monitoring
Program, namely, Patient Identification Accuracy and
Corrected Results. In the former instance, phlebotomists
record problems with patient wristbands; in the latter, the
fraction of corrected laboratory results (a portion of which
may have been due to mislabeled specimens) is recorded.
The Clinical Laboratory Improvement Act also requires
laboratories to monitor, assess, and correct problems
identified in preanalytic, analytic, and postanalytic systems.2
The greatest emphasis on preventing labeling errors has
been in the area of blood banking, where transfusion of the
wrong blood product may result in a rapid and fatal
reaction. In a multi-institutional study, Grimm et al3
estimated a frequency of at least 1 per 2500 samples for
misidentified specimens sent for blood typing, and incidence was positively correlated with phlebotomy by nonlaboratory personnel. Their estimate included errors
unveiled through a mismatch of the sample blood type
with the historical patient blood type. The safety standard of
The Joint Commission is for 2 qualified people to verify
blood samples sent to the blood bank. However, any
mislabeled blood sample has the potential for adverse
repercussions.
islabeled laboratory specimens have the potential to
lead to adverse patient outcomes through misdiagnosis, delayed treatment, and inappropriate treatment. For
this reason, The Joint Commission, formerly The Joint
Commission on Accreditation of Healthcare Organizations,
has listed ‘‘Improve the Accuracy of Patient Identification’’
as the first goal of the 2014 National Patient Safety Goals.1
The standards instituted are to use at least 2 patient
identifiers for any specimen and to label the specimens in
the presence of the patient. The Joint Commission requires
accredited hospitals to regularly collect and analyze performance data. Proof of compliance with these standards will
undoubtedly be incorporated into upcoming accountability
measures used in accreditation and determining top
performers.
According to the general checklist of the College of
American Pathologists, accredited laboratories must have ‘‘a
Accepted for publication December 2, 2014.
From the Department of Pathology, Rush University Medical
Center, Chicago, Illinois.
The author has no relevant financial interest in the products or
companies described in this article.
Reprints: Ira Miller, MD, PhD, Department of Pathology, Rush
University Medical Center, 1653 W Congress Pkwy, Chicago, IL
60612-3833 (e-mail: Ira_Miller@rush.edu).
1042 Arch Pathol Lab Med—Vol 139, August 2015
CBC Logical Delta Check—Miller
Wagar and colleagues tracked mislabeling events at their
institution before and after implementation of performance
improvement strategies, which included an electronic event
reporting system, and concluded that ‘‘strategies that focus
longitudinally on specimen labeling errors can significantly
reduce errors, therefore improving patient safety.’’4(p1662)
They lowered the rate of recognized mislabeled specimens
from an average of more than 20 per month in the first 7
months to less than 1 per month in the last 10 months in
their study period. However, they did not state the details of
how mislabeled specimens were recognized, and they
suggested that ‘‘the number of mislabeled specimens
observed in this study underestimates the actual frequency
of mislabeled specimens.’’4(p1667)
A misidentified specimen test result may be obvious to a
caregiver if a result is reported on a test that was not ordered
or if the result differs substantially from the expected result.
It may be identified by laboratory personnel to reconcile an
entry on the preceding day’s log of rejected specimens,
which includes those for which no test has been ordered.
Also, it might be suspected by the caregiver or laboratory
when there is significant variation from a recent prior result
from that test on the same patient, the so-called delta
checks. The practice of delta checking in hematology was
first reported by Nosanchuk and Gottman5 in 1974 as a
broad quality control for preanalytic and analytic errors. In
an analysis of simulated delta check rule performance,
Strathmann and colleagues6 found that the mean red blood
cell volume (MCV) had the highest positive predictive value
and the fewest false positives. In 2005, the International
Consensus Group for Hematology Review7 published
suggested criteria for action following automated complete
blood cell count (CBC) and white blood cell differential
analysis. This list of actionable items included delta rules for
changes in white blood cells, MCV, and platelets. The action
suggested for a failure of the MCV delta was to ‘‘verify
sample integrity/identity.’’ The authors did not suggest a
specific delta value to trigger action.
Our laboratory had long used a delta check of the MCV
value in patients’ CBCs to suggest the possibility of a
specimen error, including misidentification. We followed a
protocol whereby serial MCVs within 3 days on the same
patient that differ by more than 3.0 fL prevented the
automatic release, or autoverification, of the entire CBC
panel and prompted the instrument operator to release the
results with a footnote in the medical record indicating
either that the change correlates with a charted interim
transfusion or that the caregiver should confirm the
specimen identity or redraw the sample. This MCV delta
check procedure was instituted at a time before autoverification was in place, when test volume was lower, staffing
was higher, and test turnaround time was longer. In their
original study, before introduction of computers into the
laboratory, Nosanchuk and Gottmann5 reported that the
results of 10% of tests failed 1 or more delta checks, but they
believed that improvement in the ‘‘standard of reliability
and confidence’’ justified the expenditure. The present study
began as an evaluation of the effectiveness of the MCV delta
check as currently used in our institution for unveiling
labeling errors. That analysis led to the proposal and testing
of a composite CBC delta (CCD) check and subsequently to
the logical delta check (LDC), with improved specificities
and sensitivities. This streamlined tool is suitable for
implementation in the institutional patient safety insurance
program.
Arch Pathol Lab Med—Vol 139, August 2015
MATERIALS AND METHODS
Initial Survey of the MCV Delta Check Failures
On 110 days between November 2010 and March 2011, CBCs
were performed on 1 or more of 3 automated blood analyzers
(2100XE; Sysmex Corporation, Kobe, Japan). Instrument printouts
were collected for patient samples failing the MCV delta check,
defined as a change of 3.0 fL or more compared with the most
recent test within the preceding 3 days. The electronic medical
record of the patient was interrogated to determine the following:
(1) whether the MCV delta flag was triggered accurately, (2)
whether comparison with subsequent and prior counts suggested
that a mislabeling error had occurred, (3) whether the patient had
been transfused in the interim, and (4) whether other conditions or
therapies might account for the change in MCV. Test results were
presumed to have come from mislabeled specimens if multiple
CBC parameters changed, with cell counts moving in opposite
directions, and subsequent results, which were almost always
available, returning to the previous values. Cases were not
considered mislabeled if changes could be attributed to fluid
dilution, poor mixing (no change in red blood cell distribution
width [RDW] or the mean cell hemoglobin [MCH]), or clinical
situations such as postoperative samples without transfusion or
hydration of patients.
Optimization of the CCD
Forty-nine random patients with MCV delta check failures were
chosen from the initial survey group, each with more than 30
sequential CBC tests (up to 280 tests per patient, with an average of
approximately 50) performed on Sysmex 2100XE instruments. For
some patients, data from more than 1 hospitalization were
analyzed. The CBC results were transferred into a spreadsheet
(Excel; Microsoft Corporation, Redmond, Washington). For each of
the independent parameters of the CBC, comparison was made
between all serial CBCs performed within 5 days of each other for
each patient’s data set (intraindividual deltas), and a separate
comparison was made between the most recent CBCs of each
patient (interindividual deltas, up to 5 million comparisons). For
intraindividual deltas, serial cases with increases in hemoglobin of
more than 1.7 g/dL were excluded as indicative of interim
transfusion. Using the spreadsheet frequency function with bins
scaled to each CBC parameter, the frequency distribution of
intraindividual and interindividual deltas for each variable was
plotted. These plots were assessed visually to determine the
usefulness of each parameter in discriminating interindividual
comparisons from serial intraindividual comparisons. An optimized
CCD was then derived by serially repeating all comparisons using
the square root of the sum of the squares of the discriminating CBC
parameters, each with a scaling or weighting factor derived by trial
and error. The following computation for CCD was chosen: CCD ¼
SQRT ([50 3 DHb]2 þ [100 3 DMCH]2 þ [100 3 DRDW]2 þ [1.5 3
DPLT]2), where SQRT is square root, the hemoglobin concentration
(Hb) is in grams per deciliter, MCH is in picograms, RDW is a
percentage, and PLT (platelet) is in thousands per microliter. An
initial cutoff value of 250 for flagging was suggested based on the
initial training data set. (To simplify the computer programming,
the same result is achieved by comparing the term within the
square root sign with the cutoff value of 62 500.)
Testing the CCD and LDC
The CCD was programmed into the Workspace Application
Module software by programmers at Sysmex. For each CBC, both
MCV delta checks and CCD checks were performed on Sysmex
9000XN instruments. Data from the course of a 6-month period
were reviewed to determine the discriminatory power of the MCV
delta and CCD checks for specimens subsequently discovered to
have been mislabeled. In addition, Workspace Application Module
manager reports of failures were generated. All data for a period of
2 weeks (CCD) or 5 weeks (LDC) were assessed to determine the
number of specimens flagged by each delta test, which changes
CBC Logical Delta Check—Miller 1043
were attributable to transfusion, and any clinical or laboratory
associations with a delta check failure.
RESULTS
Initial Survey of the MCV Delta Check Failures
In the initial evaluation period, Sysmex operators were
requested to submit instrument printouts for tests failing the
MCV delta check. Five hundred ninety-one CBC results
were assessed. This represented an average of 5.4 per day
out of an average of 440 CBC tests per day (1%) and serves
as a lower limit estimate of the MCV delta failure rate
because there was no monitor of operator compliance.
Review of the prior CBC results for each patient revealed
that 89 tests were false calls of an MCV failure due to lack of
finalization of all components of the most recent test results,
causing comparison with earlier results, or due to prior
retracted results read as zero values. Of the 502 true MCV
delta failures, review of the medical record indicated that
288 were associated with transfusion in the interim between
tests. Of the 205 MCV delta failures without interim
transfusion, review of the results of subsequent CBCs,
which were available in almost every case, showed similar
values, indicating that the CBC test with the MCV delta
failure was accurate. Analytic problems due to interinstrument calibration differences or random error contributed to
MCV delta failures. For 46 tests of the MCV delta failures,
the sample was retested on another similar instrument, with
a mean (SD) difference of 1.0 (0.8) fL. In 28% of those cases,
the MCV change due to interinstrument variability was 1.5
fL or more, or half of the change required to trigger a MCV
delta check failure. Possible physiological reasons suggested
from review of the patient records included red blood cell
volume changes due to alterations in plasma osmolarity in
patients on hemodialysis or with aggressive fluid hydration
associated with diarrhea or emesis, as well as patients who
had undergone surgery with no blood products given.
Six mislabeled specimens were identified. The MCV delta
check failure flag occurred on the erroneous test and on the
subsequent test in 3 cases. In 2 cases, the flag occurred only
on the subsequent test. Failure to flag the truly mislabeled
specimen probably occurred because other results were still
pending when the next CBC test was reported. In 1 case,
there was no follow-up specimen, so it is uncertain which
was mislabeled. There was no evidence of any result
corrections initiated from charted MCV delta check failures
because erroneous results remained in the medical record.
Formulating an Optimized CCD Check
An index population of 49 patients with many sequential
CBC tests was used to derive a CCD check based on the
hypothesis that the individual CBC parameters varied
independently in a population and that sensitivity and
specificity of the delta test could be improved by combining
the most informative parameters. In this index patient
group, more than 98% of the sequential samples were
acquired less than 2 days apart, and almost all of these were
next-day comparisons. Most but not all sequential samples
with interim transfusion were eliminated from analysis by
excluding serial samples with a rise of 1.7 g/dL or more.
Comparison between random interpatient deltas and intrapatient serial CBC deltas for each of the CBC parameters
suggested that hemoglobin concentration (or hematocrit),
MCV, MCH, RDW, and platelet count would be useful delta
parameters (representative curves are shown in Figure 1).
1044 Arch Pathol Lab Med—Vol 139, August 2015
By trial and error, the CCD check was optimized by
applying scaling and weighting factors to give the greatest
separation between the curves (Figure 2, A). The MCV and
MCH gave similar discriminatory power and were not
independently useful. The MCH was incorporated rather
than the MCV because of the observation that MCV changes
occurred with hydration or dialysis, which should not affect
the MCH. Adding the white blood cell delta did not provide
additional discriminatory power. Based on the initial sample
comparisons, the computation for CCD and the cutoff value
stated in the ‘‘Materials and Methods’’ section were derived.
The receiver operating plot for this function is shown in
Figure 2, B. This test gave a specificity of 97.6% and a
sensitivity of 92.5% using the training data set from the 49
index patients. That is to say, 2.4% of the more than 2000
serial intrapatient CBC tests gave CCD values of greater
than 250, and 7.5% of the 5 million interpatient comparisons gave CCD values of 250 or less, flagging one-third as
many as the MCV delta check. In the same data set, using
the single parameter MCV delta with a cutoff of 3.0 fL had a
specificity of 94.4% and a sensitivity of only 77.4%.
Testing the CCD Parameter
The CCD check and the MCV delta check were
compared in the subsequent 6-month period for sensitivity
in detecting actually mislabeled samples discovered
through other means (error forms submitted by nurses).
Of 6 such specimens with prior results for comparison, all
failed the CCD check with a cutoff of 250, while only 3
failed the MCV delta check (Table 1). In further assessment, the number and characteristics of all 110 specimens
flagged by the CCD check out of a possible 5792 specimens
(with a 3-day look-back to prior window) during a
randomly chosen 2-week period were also evaluated
(Table 2). The CCD check with a cutoff of 250 flagged less
than half of the number of specimens flagged by the MCV
delta check. The difference is significant, with a 2-tailed P
, .001 (v21 ¼ 66.704). Approximately half were readily
explained by interim transfusion. Review of prior and
subsequent values showed that in approximately 13% of
cases a large change in platelets contributed disproportionately to the high CCD value, usually in patients with
platelet counts that were elevated. In another 36% of cases,
review of subsequent values confirmed the shifted CBC
results, and review of the medical record usually provided
an explanation such as hemorrhage, surgery, delivery, or
sickle cell disease, with the latter associated with a large
change in RDW. In 17 cases (0.3% of all samples with
recent priors for comparison), the large change in CCD
value reflected an error. Dilution from intravenous fluid
was apparent in 3 cases by a proportional change in
concentrations of all components, usually accompanied by
a rise in the MCV in the diluted sample. Erroneous red
blood cell parameters due to cold agglutinins were evident
as increases in the MCH in 3 cases in which the operator
deviated from protocol, and improper mixing of tubes run
in manual mode was the likely explanation for 2 errors.
Cases of CCD failure reflecting clotted samples were
excluded because these were always independently recognized and flagged by other sample or result characteristics.
Finally, 5 mislabeled specimen events triggering 9 CCD
failures were discovered in the 2-week period. Four
mislabeled specimen events each triggered 2 high CCD
values, one for the mislabeled specimen compared with the
results of the correctly labeled prior and another for the
CBC Logical Delta Check—Miller
Figure 1. Delta curves comparing frequency distributions of representative single parameters (MCV, Hb, MCH, and RDW in A, B, C, and D,
respectively) from the index population, frequency plotted on the y-axis, show random interindividual comparisons (red) and intraindividual
comparisons (blue). Abbreviations: MCV, mean red blood cell volume; Hb, hemoglobin concentration; MCH, mean cell hemoglobin; and RDW, red
blood cell distribution width.
subsequent correctly labeled specimen compared with the
mislabeled specimen results. In only 1 of these instances
was the MCV delta greater than 3.0 fL. In 1 additional case,
the actually mislabeled specimen was reported before the
beginning of the 2-week assessment period. Several clotted
specimens would have been detected but were never
reported because they were already flagged based on
abnormal platelet parameters.
Assessment of the 14 cases with high CCD values due to
a disproportionately large platelet delta suggested the
possibility of further refinement of the CCD test. For all 14
of these false-positive cases, eliminating platelets from the
CCD calculation—using the square root of the sum of
squares of deltas of hemoglobin, MCH, and RDW (termed
HMR calculation)—gave values of less 116, but all of the
true-positive problem cases had HMR values of greater
than 116. The LDC, requiring violation of both criteria
(CCD .250 and HMR .116), was enabled in the
laboratory, and at this point we changed our delta check
protocol to only require follow-up of LDC failures and not
Arch Pathol Lab Med—Vol 139, August 2015
of MCV delta failures. Technicians were required to
determine if flagged values were associated with interim
transfusion, to consider the possibilities of inadequate
mixing of samples run in manual mode, and to be alert to
the possibility of inadequate warming of samples with cold
agglutinins in cases with increased MCH. All unexplained
changes required a phone call that suggested the need for a
redraw if the recent change in values could not be
explained clinically. Technicians were requested to save
the histogram and CBC result printouts for LDC failures
and to indicate the cause and the resolution of the
discrepancy.
For a period of 5 weeks, characteristics of samples failing
the LDC test (with a 5-day look-back window) were
assessed (Table 2). Retrospective review of the patient
records and laboratory information indicated that the LDC
test caught 15 misidentified specimens that caused 16 delta
check failures. In all but 3 cases, the erroneous results had
been deleted from the record. In 2 cases, the test results had
not been reported because a duplicate specimen was
CBC Logical Delta Check—Miller 1045
Table 1. Comparison of CCD and MCV Delta Values
for Independently Established Mislabeled Blood
Specimens Sent for CBC Testing
During a 6-Month Period
Mislabeled
Specimen No.
Composite
CBC Delta
MCV Delta, fL
1
2
3
4
5
6
Mean
414.6
655.8
782.8
554.3
300.6
398.3
518.0
2.4
18.5
0.5
15.0
2.6
7.1
7.7
Abbreviations: CBC, complete blood cell count; CCD, composite CBC
delta; MCV, mean red blood cell volume.
Figure 2. Shown are optimized composite complete blood cell count
(CBC) delta (CCD) frequency distribution curves (A) of random
interpatient comparisons (red) and serial intrapatient comparisons
(blue) using the index population CBC test results and the receiver
operating curve (B) for these data. The CCD calculation uses the
weighted combination of deltas of hemoglobin concentration, red
blood cell distribution width, mean cell hemoglobin, and platelet count
as indicated in the ‘‘Materials and Methods’’ section.
recognized before release of the erroneous results. In 4
cases, error reports were submitted by the nurse responsible. One error was committed by a phlebotomist, who
received a reprimand. In the other 5 cases, the results were
deleted from the medical record, with a note that the
specimen had ‘‘questionable integrity.’’ Specimens with
problematic results for reasons other than mislabeling were
also detected (Table 2). Shortening the period of look-back
to prior CBCs from 5 days to 3 days would have excluded 22
of 85 false-positive specimens unrelated to transfusion,
without reducing sensitivity. Since that observation was
made, we have shortened our look-back to 3 days.
1046 Arch Pathol Lab Med—Vol 139, August 2015
COMMENT
This study was initiated in an attempt to evaluate the
effectiveness of the MCV delta test for identifying
mislabeled blood specimens in our hospital, which is a
tertiary care center with a varied inpatient population with
medical, surgical, obstetric, psychiatric, and pediatric
illness. Review of the MCV delta test indicated that, as it
had been used in our hospital, it was ineffective and a
nuisance. The MCV changes were often associated with
interim transfusion or occurred in patients with probable
osmotic imbalance such as those on dialysis. Also,
calibration differences between instruments could significantly contribute to the delta value if serial samples were
run on different instruments. Approximately 1% of the 860
investigated MCV delta check–flagged tests were mislabeled specimens. However, these values remained in the
record, indicating either that they were not recognized by
the caregiver to be erroneous or that there were barriers to
documenting correction of the results in the medical
record. Approximately half of the MCV delta failures were
due to interim transfusion. Robust implementation of the
MCV delta test by laboratory computer software was
hampered by consideration only of prior samples for which
all reportable tests had been released, preventing consideration of many prior samples for which there were
pending unreported tests. To improve upon the MCV
delta check, I used random sample results from our
hospital population to initially determine which of the
CBC values were most informative to distinguish interpatient samples from serial intrapatient samples, followed
by the best weighting coefficients for the CCD calculation.
Retrospective review of the CCD performance proved that
it had higher sensitivity and specificity than the MCV delta
check. Review of false-positive and false-negative cases
suggested further improvement, with the LDC calculation
adjusting for rapid platelet changes. Review of the patients’
medical records after implementation of the LDC showed
that only 3 of the 15 erroneous result sets were not
removed, suggesting that action by the technician led in
most cases to real-time recognition of the error. Finally,
shortening the look-back period to 3 days will further
improve specificity. Based on the initial 5-week period of
implementation, the LDC is predicted to flag less than 1%
of total CBC results. Posttransfusion results are released
after consulting the record, and other results are released
after notification of the caregiver or withheld (along with
other concurrent test results) at his or her request, pending
CBC Logical Delta Check—Miller
Table 2.
Characteristics of Failures for CCD and LDC Using Cutoff Values Indicated in the Text
Variable
CCD
Length of evaluation period, d
Total CBCs performed during the evaluation period, No.
Total delta checks, No. (% of CBCs with recent priors)
Failed delta checks
Interim transfusion, No. (%)
No interim transfusion, valid based on medical record review, No. (%)
Failure due to platelet change only, No. (%)
Presumed or confirmed mislabeled, No. (%)
Failure due to other problem, No. (%)
Specimen dilution from intravenous fluid in line, No.
Analytic (agglutinin or unmixed specimen), No.
Failed tests
Both MCV delta (.3.0 fL) and CCD, No.
Both MCV delta (.3.0 fL) and CCD, presumed or confirmed mislabeled,
No./total No.
LDC
14
11 193
5792 (52)
(n ¼ 110)
54 (49)
39 (36)
14 (13)
35
26 566
13 234 (50)
(n ¼ 205)
76 (37)
85 (42)
0
9 (8), comprising 5 events
8 (7)
3
5
16 (8), comprising 15 events
28 (14)
11
17
38a
Not determined
2/38, comprising 1 event
8/13 assessable LDC failures
Abbreviations: CBC, complete blood cell count; CCD, composite CBC delta; LDC, logical delta check; MCV, mean red blood cell volume.
a
There were 269 MCV delta check failures (5% of values with recent priors) in this period.
a redraw. The clinical effect of this innovation is 3-fold.
First, there is increased turnaround time for the approximately 2% of specimens that would have been investigated for possible error using the MCV delta check but not
with the more specific LDC. Second, early identification of
mislabeled or problem specimens—with retraction of
erroneous results for not only the CBC but also other tests
ordered on blood tubes drawn at the same time—allows
for rapid redraw and may prevent unnecessary or
inappropriate clinical action. At the current rate, approximately 40% of LDC flags of nontransfused patients’ results
will correspond to problem specimens or to their follow-up
specimens. Third, identifying more mislabeled specimens
provides a tool for laboratory personnel, nurses, and others
to identify patterns of error and remediate them. Monitoring these data may be useful for ongoing quality
improvement efforts.
In consideration of possible limitations of this study, I
note that the LDC calculation was derived and tested on a
set of specimens from a single medical center with a diverse
patient population. As such, it is likely to be useful for most
high-volume laboratories that serve large medical centers.
Additional modification may be useful to laboratories that
encounter false-positive LDC failures resulting from unique
clinical situations. For example, if the laboratory receives
specimens from a medical center where blood loss from
Arch Pathol Lab Med—Vol 139, August 2015
surgery is frequently not corrected by transfusion due to
practice variation or patient values, then an additional
logical step to ignore LDC failures due to hemoglobin
change only may be desirable.
I thank Kristin Epperson, MT, and Kathryn Williams, MT, for
assistance in data collection, as well as the programmers at Sysmex
Corporation.
References
1. Joint Commission. National Patient Safety Goals. http://www.
jointcommission.org/standards_information/npsgs.aspx. Accessed October 31,
2014.
2. College of American Pathologists. Quality Management Tools. Footnote to
table, page 3. http://www.cap.org/apps/docs/proficiency_testing/2014_qmt_
catalog.pdf. Accessed October 31, 2014.
3. Grimm E, Friedberg RC, Wilkinson DS, AuBuchon JP, Souers RJ, Lehman
CM. Blood bank safety practices: mislabeled samples and wrong blood in tube: a
Q-PROBES analysis of 122 clinical laboratories. Arch Pathol Lab Med. 2010;
134(8):1108–1115.
4. Wagar EA, Tamashiro L, Yasin B, Hilborne L, Bruckner DA. Patient safety in
the clinical laboratory: a longitudinal analysis of specimen identification errors.
Arch Pathol Lab Med. 2006;130(11):1662–1668.
5. Nosanchuk JS, Gottmann AW. CUMS and delta checks: a systematic
approach to quality control. Am J Clin Pathol. 1974;62(5):707–712.
6. Strathmann FG, Baird GS, Hoffman NG. Simulations of delta check rule
performance to detect specimen mislabeling using historical laboratory data. Clin
Chim Acta. 2011;412(21–22):1973–1977.
7. Barnes PW, McFadden SL, Machin SJ, Simson E; International Consensus
Group for Hematology. The International Consensus Group for Hematology
Review: suggested criteria for action following automated CBC and WBC
differential analysis. Lab Hematol. 2005;11(2):83–90.
CBC Logical Delta Check—Miller 1047