Clinica Chimica Acta 450 (2015) 3138

Contents lists available at ScienceDirect

Clinica Chimica Acta

journal homepage: www.elsevier.com/locate/clinchim

Developing optimized automated rule sets for reporting hemolysis,

icterus and lipemia based on a priori outcomes analysis
Jessica M. Boyd a, Richard Krause a, Gayle Waite b, William Hui b, Elham Yazdi b,
David Wilmink b, Isolde Seiden-Long a,
a
b

Department of Pathology and Laboratory Medicine, University of Calgary and Calgary Laboratory Services, Canada
Gamma-Dynacare Medical Laboratories, Canada

a r t i c l e

i n f o

Article history:
Received 2 June 2015
Received in revised form 3 July 2015
Accepted 6 July 2015
Available online 8 July 2015
Keywords:
Hemolysis
Icterus
Lipemia
Automated indices

a b s t r a c t
Background: There is limited information about the effects of instituting CLSI Document C56A recommended
workows for the automated detection of hemolysis, lipemia and icterus (HIL) in different clinical laboratories
and patient populations. We describe a process to develop and tailor automated reporting rules that are appropriate for the local laboratory population.
Methods: Automated decision algorithms were generated and applied to 2 high volume labs serving community
and hospital populations. Proposed rules were applied to the datasets ofine to predict the outcomes, and then
were further optimized prior to implementation.
Results: Introduction of automated serum indices decreased HIL agging compared to manual agging. Hemolysis agging was the greatest in all 3 patient populations, and was successfully reduced for LD, CK and AST by optimized rules that incorporated both the H-index result and the analyte result. Changes in agging rates were also
patient population specic, particularly for icterus which was a problem in hospitalized populations but not in
the community. Overall, concordance between manual and automated agging methods was very low in both
laboratories.
Conclusions: We demonstrate that agging algorithms may not be universally transferable due to lab specic and
population specic factors and demonstrate the benets of local, a priori testing of algorithms prior to
implementation.
2015 Elsevier B.V. All rights reserved.

1. Introduction
Hemolysis, icterus and lipemia (HIL) interferences are common preanalytical sources of error in the clinical laboratory [14]. Determination
of HIL interferences has traditionally been done by manual visual grading by technologists; however, this process is highly subjective and variable [57]. Serum indices are a semi-quantitative measurement of HIL
interference, using spectrophotometric measurement and mathematical correction to determine the level of interference. Serum indices are
less subjective than manual grading, can be automated, and negligibly
impact turnaround times [8].
CLSI document C56A provides guidance on the use of serum indices
for measurement of HIL interference [1]. It recommends selection of
assay specic HIL cut-offs, above which HIL interferences will affect results, and development of algorithms to deal with samples exceeding

Abbreviations: HIL, Hemolysis, icterus, lipemia; CLSI, Clinical Laboratory Standards

Institute; LIS, Laboratory information system.
Corresponding author at: Foothills Medical Centre, 1403-29 St. NW, Calgary, Alberta
T2N 2T9, Canada.
E-mail address: Isolde.seidenlong@cls.ab.ca (I. Seiden-Long).

http://dx.doi.org/10.1016/j.cca.2015.07.006
0009-8981/ 2015 Elsevier B.V. All rights reserved.

the HIL cut-offs (e.g., cancelation, report with comment). HIL cut-off selection and algorithm design is left to the discretion of each laboratory.
Although serum indices have been available for several years, there
is limited information about their implementation in clinical laboratories, either for selection of assay specic HIL cut-offs or development
of algorithms. One study has looked at retrospective development and
implementation of algorithms for processing HIL interferences [6]. The
authors investigated the changes to HIL agging after switching from
visual inspection to automated indices and found large increases in agging for all 3 interferences. Test specic agging rates were not included
so it is uncertain if the agging increases were seen across many tests or
dominated by a few interference prone assays. It is also unknown if increased agging would be seen in other patient populations (community vs. hospital). The large increase in agging suggests that agging
analysis prior to implementation may be necessary in order to minimize
the impact of introducing automated serum indices on lab workow
and provide adequate notication to physicians of the anticipated
changes to reporting.
We outline a process to develop and test an HIL algorithm with the
overall goal of producing a process that can identify specimens with interference levels relevant to the tests ordered; alert technologists to the

32

J.M. Boyd et al. / Clinica Chimica Acta 450 (2015) 3138

interference if intervention is required; provide interpretive information to physicians about specimens with interference; cancel the least
number of tests; and avoid adverse effects caused by failure to report results with HIL interference.
2. Materials and methods
2.1. Ethics
This project is classied as quality assurance in nature by our institutional Co-joint Health Research Ethics Board and is granted exempt
status for research ethics review.
2.2. Demographics
Data from 2 high-volume community labs and one group of hospital
labs were analyzed in this study (Table 1). Lab A is a high volume lab in
Ontario, Canada with 3 locations serving western, central and eastern
Ontario. Labs B1 and B2 serve the same city in Alberta, Canada. Lab B1
is a centralized, high volume laboratory serving the community population while Lab B2 is a combination of 5 urban hospital labs serving different sub-populations. Labs B1 and B2 were analyzed separately
because of the differences in population served, instrumentation and
test menus.
2.3. Lab inclusion rationale
2.3.1. Patient population
No data is currently available comparing the application of identical
HIL algorithms to different patient populations (e.g. hospital vs. community, community vs. community). To address this, 2 community labs
(Labs A and B1) and one hospital lab (Lab B2) were included in the present study.
2.3.2. Effect on existing methods to detect HIL interference
Only one study investigating one lab has been published describing
the effects of switching from manual grading to automated serum indices. To determine if the same effects are seen in other labs making this
transition, data from 2 labs were included (Labs A and B2).
No data is available to determine if the index cut-offs and algorithms
need to be periodically evaluated and revised or if they can be left alone
indenitely. In this study, Lab B1 was included to evaluate this question,
as it was re-evaluating its automated serum index rules as it changed its
analyzers from Roche Modular P/E to Roche Cobas c701/e602. Cut-off
changes were observed for some tests because of methodology changes
(e.g. calcium from o-cresolphenol to NM-BAPTA) or because Roche stated different HIL cut-offs for the same test run on different analyzers (e.g.
TBIL c701 I index = 90, c501 I-index = 1000). In addition, there was interest in determining if the algorithm could be modied to cancel fewer
tests, as the lab had experienced some calls about too many tests being
canceled.

2.4. Measurement of serum indices

All instruments used in this study were from Roche Diagnostics
(Laval, QC) (Table 1) and used the same serum index measurement
principle [9]. To conrm that serum index values were consistent
among analyzers within each lab, several specimens representing a
range of HIL values were split and analyzed across all analyzers within
each lab. All results were within 10%.
2.5. Tests included in the analysis
Serum indices were collected on a combined 80 tests between the 3
labs (see Supplemental Table S1 for test abbreviations and methodologies). HIL agging rates were not available if the lab did not perform
the test or if it was performed on an instrument that was not collecting
serum indices.
2.6. Selection of HIL cut-offs and decision algorithm development
2.6.1. Desired outcomes and assumptions
Five objectives were used to guide cut-off selection and algorithm
development: identify specimens with interference levels relevant to
the tests ordered; alert technologists to the presence of interference
when intervention is required; aid physicians with interpretation of
test results from samples with interference; cancel the least number
of tests; and avoid adverse effects caused by failure to report results
with HIL interference.
The algorithm was also designed to reduce manual technologist interventions as much as possible. The effect of introducing the algorithm
was determined by assessing sample agging and cancelation rates in 2
different laboratories, serving 3 different populations. This allowed for
assessment of lab specic and population specic factors requiring renement of the algorithms with the ultimate goal of reporting more results with better interpretive information to physicians. By performing
this analysis a priori, the impact on laboratory workows could be determined and the process rened prior to implementation.
2.6.2. HIL cut-off selection
Selection of test-specic HIL cut-offs is discussed in Section 3.1.
2.6.3. Rule sets
The 3 sets of comment and cancelation rules assayed in this study
are baseline rules (the existing HIL agging process in place in each
lab), proposed rules (automated HIL indices with test specic comment
and cancelation cut-offs chosen using manufacturers' package inserts,
information from literature [10], and CLSI guideline C56A [1]); and optimized rules (use of alternative cut-offs or interference clearing strategies to increase the number of reportable results). Five or more ags
under proposed rules were used as the trigger for consideration of an
optimized rule for any test.
The general process used to develop and rene the serum index rule
sets is illustrated in Supplemental Fig. S1. Specic details about the 3

Table 1
Key demographics of Labs A, B1 and B2.
Patient population

Average # of tests per day during data collection period

# of chemistry tests on menu
# of days data collected
Total number of tests performed during data collection period
Total number of tubes analyzed during data collection period
Instrumentation
Location where rule sets are applied
LIS

Lab A

Lab B1

Lab B2

Community

Community

Hospital

77,319
55
7 (Nov 28Dec 4, 2011)
541,236
66,668
Roche Modular P and E170
In house developed LIS IBM AS400
In house developed LIS IBM AS400

22,713
45
14 (Oct 1630, 2013)
317,986
48,245
Roche c701 and e602
Roche Cobas IT Middleware
Cerner Millennium

15,381
47
21 (Jan 13Feb 3, 2014)
323,018
48,460
Roche c501 and e601
Roche Cobas IT Middleware
Cerner Millennium

J.M. Boyd et al. / Clinica Chimica Acta 450 (2015) 3138

33

rule sets for each interference in each lab are given below and in Supplemental Table S2.

2.6.3.3.3. Optimized rules. Lipemia optimized rules were the same as

the proposed rules.

2.6.3.1. Hemolysis (H)

2.6.3.1.1. Baseline policies. Lab A and Lab B2 both used manual visual
inspection with comparison to a hemolysis grading chart. In Lab A, hemolyzed tubes had a hemolysis letter value ag applied to all the tests
on the tube but no chartable comment. In Lab B2, a ag and chartable
comment indicating slight, moderate or gross hemolysis was applied
to one test on the hemolyzed tube, chosen by the technologist.
Lab B1 had previously instituted automated serum indices and decision rules. All tests were canceled with comment with H-index N 500,
unless a lower cut-off was indicated by the package insert (e.g., potassium cancelation cut-off was 300).
2.6.3.1.2. Proposed rules. Identical proposed rules were applied at all 3
labs. Tests were canceled at H-index 600 (unless a higher cut-off was
indicated by the package insert) and comment code HEMC applied
(Supplemental Table S3). For tests where an H-index cut-off b 600
was indicated by the manufacturer, results were reported with a comment indicating the direction of interference if known (comment
codes HEM1, HEMU, and HEMD).
2.6.3.1.3. Optimized rules. Hemolysis optimized rules used reference
intervals in addition to the H-index to determine clinically signicant
interference. The criteria for hemolysis optimized rule generation
were N5 comment ags under proposed rules (chosen by the authors
as a signicant agging rate in the absence of any recommended quality
indicators for acceptable incidence of hemolysis); unidirectional interference only, and critical values must be only at one end of the reference
interval (high or low, but not both).
Under optimized rules, samples with H-index above the agging
cut-off but b 600, would be given a hemolysis ag and comment code
only if the test result was outside reference range. In this study, optimized rules were generated for AST, CK and LD.

2.7. Data collection and analysis

2.6.3.2. Icterus (I)

2.6.3.2.1. Baseline policies. Lab A used manual visual inspection by
comparison to a visual grading chart with results reported with an icterus ag but no comment. Labs B1 and B2 did not have any policies regarding icterus prior to this study; all test results were reported
without review or comment.
2.6.3.2.2. Proposed rules. Identical proposed rules were applied at all 3
labs. Samples exceeding I-index cut-offs were reported with a comment
indicating the direction of interference if known (comment codes ICT1,
ICTU, and ICTD).
2.6.3.2.3. Optimized rules. Optimized rules required a dilution of specimens with a high degree of icterus. Optimized rules were considered
for tests with I-index interference levels b15 (GGT, creatinine) in hospital specimens.
2.6.3.3. Lipemia (L)
2.6.3.3.1. Baseline policies. Labs A and B2 visually inspected tubes for
lipemia. In Lab A, moderately lipemic tubes were reported with a
lipemia letter value ag without comment. Grossly lipemic tubes had
all tests canceled with comment. In Lab B2, lipemic samples were
ultracentrifuged, rerun and reported with a comment. In Lab B1, the
middleware alerted the technologist if the instrument L-index was
exceeded for a particular test. The sample would be ultracentrifuged,
rerun with indices and reported with a comment if the ultracentrifuged
L-index b lipemia cut-off. In Labs B1 and B2, ultracentrifugation was not
performed for triglycerides, HDL, cholesterol, or ammonia tests. Ammonia samples were diluted to reduce handling time.
2.6.3.3.2. Proposed rules. Identical proposed rules were applied at all 3
labs. Samples N L-index cut-off were ultracentrifuged. The test and indices were there rerun and the result reported with a comment (LIP1) if
the ultracentrifuged L-index b L cut-off. Ultracentrifugation was not performed for triglycerides, HDL, cholesterol, or ammonia tests.

Serum index data were collected from all 3 labs for the time periods
specied in Table 1. In Labs A and B2, automated indices were turned on
for data collection purposes only, which prevented impact to patient result reporting. In Lab B1, where automated indices were already in use,
data were collected without affecting the existing agging processes.
A different process was required in each lab to obtain baseline agging rates. In Lab A, manual comment ags were obtained from the
LIS. In Lab B2, manual comment and cancelation ags were obtained
from the LIS. In Lab B1, baseline agging rates were modeled using the
instrument data les in order to capture the canceled tests that were
not transferred to the LIS. Flagging rates were modeled using Microsoft
Excel 2007.
2.8. Study limitations
Several limitations were identied with this study. Indices were only
employed on the high volume analyzers at all 3 labs. Although more
tests are run at these locations, other analyzers were not included or
not able to measure serum indices. Data from Lab A do not include
any samples canceled under baseline policies as this information could
not be retrieved from the LIS. Therefore, cancelation rates for Lab A
may be underestimated in this analysis. In Labs A and B1, baseline policies did not collect index data on tubes that only had immunoassay tests
ordered. Indices were performed if the specimen had both immunoassay and chemistries ordered. Therefore, agging rates for immunoassay
module tests may be underestimated. In Lab B2, baseline policies indicated adding comment ags to one result on the tube, chosen by the
technologist. For the agging analysis, we have only counted the test
to which the ag was applied, not the other tests on the tube. For canceled specimens, all tests on the tube were canceled and were all included in the analysis. Physicians were not surveyed to determine if the
interpretive comments were of use to them.
3. Results and discussion
A process for developing HIL cut-off and decision algorithms for use
with automated serum indices was developed. The implementation of
HIL cut-offs and decision algorithms was studied in 3 high volume laboratories that served distinct patient populations (community or hospital) and had different pre-existing policies for detecting HIL interference
(manual inspection or automated serum indices) (Tables 24).
3.1. Proposed rule set development
3.1.1. Selection of test specic cut-offs
Cut-offs for HIL interference were mainly drawn from the manufacturers' package inserts (objective: identify specimens with interference
levels relevant to the tests ordered). Most manufacturers provide this information; however, it is important to understand how they report
serum indices (e.g., numeric value, bin number) and the criteria used
to determine signicant interference with the assay (e.g., biological variation, consensus of an advisory panel [1], 10% recovery from the initial concentration [1012]) as this can cause variation in HIL cut-offs
between different platforms and vendors. Roche (whose instruments
were used in this study) reports serum indices as a value (e.g., Hindex of 101) and uses 10% recovery from the initial concentration
as their criteria for signicant interference with an assay. Roche HIL
cut-offs were used to generate the proposed rule set, with the exception
of hemolysis cut-off for potassium. The Roche cut-off of H-index = 300,
was considered to be too high to be used as a warning level [10], and the
lower cut-off of H-index = 100 was chosen instead.

34

J.M. Boyd et al. / Clinica Chimica Acta 450 (2015) 3138

Table 2
Hemolysis index agging counts for selected tests.
Test

ALT
ALP
AST
CO2
DBIL
NBIL
TBIL
CL
CHOL
CK
CREA
FERR
GGT
GLU
HDL
LD
MG
PHOS
K
NA
TSH
TRIG
UIBC

Lab A

Lab B1

Lab B2

# of tests

Baselinea

Proposed/Optimizedb

# of tests

Baseline

Proposed/Optimizedb

# of tests

Baselinea

Proposed/Optimizedb

36,202
22,038
11,901
1627
168
1
14,796
24,858
30,214
16,369
43,144
22,559
5077
32,673
29,883
1323
1647
2511
31,546
30,115
29,854
30,108
1742

563
325
187
8
4
1
146
310
218
230
303
206
117
292
213
20
26
24
470
404
134
266
10

0
0
86/29
0
13
1
0
0
0
33/5
0
1
5
0
0
257/38
0
0
42
0
0
0
5

22,846
12,616
2377
20,757
3518
NP
7935
20,950
20,857
6246
26,099
16,888
10,495
22,625
19,591
2598
3366
3663
21,706
21,471
23,461
19,863
8245

34
0
12
0
9
NP
0
0
0
28
0
0
4
0
0
25
0
2
32
0
0
0
29

3
7
23/7
0
45
NP
0
0
0
12/1
0
0
5
0
0
98/25
0
2
18
0
0
0
52

12,642
10,732
1888
35,809
2684
1099
11,399
35,846
NP
2423
34,898
NP
8388
17,584
NP
8276
10,792
8440
36,202
36,221
NP
NP
NP

9
6
2
37
6
32
32
37
NP
4
51
NP
0
9
NP
7
12
9
567
50
NP
NP
NP

32
33
76/50
5
333
54
0
3
NP
21/13
2
NP
23
1
NP
1175/531
1
14
288
3
NP
NP
NP

NP = test was not performed by this lab.

a
Baseline agging for Labs A and B2 was performed manually.
b
Optimized rules counts are only shown for the tests for which they were developed (AST, LD and CK for H-index).

alert technologists to the presence of interference when intervention is required). Following ultracentrifugation or dilution, the sample was
rerun and reported with a comment if the I or L index was below the
cut-off (objectives: cancel the least number of tests, avoid adverse effects
caused by failure to report results with HIL interference). For hemolysis,
a different approach was used by setting separate alert and cancelation
cut-offs. The Roche H-index cut-off from the package insert was used as
an alert cut-off, above which the result was reported with a comment. A
cancelation cut-off of H-index = 600 was chosen as representative of
gross hemolysis and that the tube was not acceptable for testing (unless
otherwise indicated by the package insert). This was done in order to increase the number of reportable results by only canceling grossly hemolyzed specimens (objective: cancel the least number of tests). The only
exception to the cancelation rule at H-index = 600 was potassium,
where the alert cut-off was 100 (as mentioned above) and the cancelation cut-off was 300.

3.1.2. Decision algorithm development

An algorithm was developed to determine the processing of samples
exceeding HIL cut-offs. Decision algorithms can have varying degrees of
complexity ranging from sample cancelation to implementation of
strategies aimed at reducing or compensating for interference in order
to produce a reportable result (e.g., agging of only clinically signicant
results [1,11], mathematical correction [1315], reduction of interference (e.g., dilution, ultracentrifugation) [13,16,17], or analysis using a
different method).
In the developed algorithm, an action (e.g. appending a comment,
manual reduction of interference, and cancelation) was required any
time a sample exceeded the HIL cut-off for a particular test. For lipemia
and icterus, this involved ultracentrifugation (lipemia) or dilution (icterus) to reduce or clear the interference. The technologist was alerted
by the middleware (in Labs B1 and B2) or the laboratory information
system (in Lab A) that manual intervention was required (objective:

Table 3
Icteric index agging counts for selected tests.
Test

ALT
ALB
ALP
AST
CHOL
CK
CREA
GGT
GLU
HDL
PHOS
TRIG
NTRIG
URATE

Lab A

Lab B1

Lab B2

# of tests

Baselinea

Proposed

# of tests

Baseline

Proposed

# of tests

Baselinea

Proposed/Optimized

36,202
9093
22,038
11,901
30,214
16,369
43,144
5077
32,673
29,883
2511
30,108
NP
15,432

43
17
39
20
19
17
41
12
27
19
2
19
NP
17

0
0
0
0
0
0
0
0
0
0
0
0
NP
0

22,846
9742
12,616
2377
20,857
6246
26,099
10,495
22,625
19,591
3663
19,863
NP
4085

0
0
0
0
0
0
0
0
0
0
0
0
NP
0

0
0
0
0
8
0
13
8
0
0
0
11
NP
0

12,642
6566
10,732
1888
NP
2423
34,898
8388
17,584
NP
8440
NP
152
998

0
0
0
0
NP
0
0
0
0
NP
0
NP
0
0

0
0
0
0
NP
0
125/28
53/5
0
NP
6
NP
9
0

NP = test was not performed by this lab.

a
Baseline agging for Labs A and B2 was performed manually.

J.M. Boyd et al. / Clinica Chimica Acta 450 (2015) 3138

35

Table 4
Lipemic index agging counts for selected tests.
Test

ALT
ALB
ALP
AST
TBIL
CL
CHOL
CK
CREA
FERR
FRUC
GGT
GLU
HDL
K
TP
NA
TSH
URATE
UREA

Lab A

Lab B1

Lab B2

# of tests

Baselinea

Proposed

# of tests

Baseline

Proposed

# of tests

Baselinea

Proposed

36,202
9093
22,038
11,901
14,796
24,858
30,214
16,369
43,144
22,559
555
5077
32,673
29,883
31,546
2765
30,115
29,854
15,432
4466

369
76
117
215
149
176
59
44
185
54
95
125
115
57
244
49
226
51
49
118

3
1
0
1
3
3
1
3
0
0
0
0
2
0
0
4
4
0
1
4

22,846
9742
12,616
2377
7935
20,950
20,857
6246
26,099
16,888
NP
10,495
22,625
19,591
21,706
2974
21,471
23,461
4085
4619

47
0
0
9
0
0
2
0
0
0
NP
0
0
2
0
0
0
0
0
0

47
0
0
9
125
2
0
0
0
0
NP
0
0
2
2
0
2
0
0
0

12,642
6566
10,732
1888
11,399
35,846
NP
2423
34,898
NP
NP
8388
17,584
NP
36,202
2352
36,221
NP
998
11,314

1
2
2
0
2
0
NP
0
1
NP
NP
0
0
NP
25
0
0
NP
0
2

18
1
0
5
1
0
NP
0
0
NP
NP
0
0
NP
0
0
0
NP
0
0

NP = test was not performed by this lab.

a
Baseline agging for Labs A and B2 was performed manually.

All tests where HIL cut-offs were exceeded had an interpretive comment appended, providing additional information to the physician as to
the interference detected and effect on the result (Supplemental
Table S3) (objective: aid physicians with interpretation of test results
from samples with interference).
The developed proposed rules were modeled on data from all 3 labs
to assess changes in agging rates and to determine if further rule renement was required (optimized rules) (objective: avoid adverse effects caused by failure to report results with HIL interference).
3.2. Flagging analysis
3.2.1. Hemolysis
Introduction of automated serum indices at Lab A resulted in a 91%
decrease in the total number of samples agged for hemolysis (comment and cancel) compared to manual agging (Baseline N = 5076
[0.9% of total]; Proposed N = 450 [0.1% of total]). This suggests that
manual inspection at Lab A resulted in application of hemolysis ags
to tests that were not affected at that level of hemolysis. This is illustrated in the agging analysis from several tests from Lab A including ALT,
ALP, CL, CHOL, CREA, GLU, HDL, TRIG, and NA (Table 2). A signicant
part of this decrease may be attributed to a change in agging policy
where only affected tests on a given specimen received ags under optimized rules, whereas all tests on an identied specimen would receive
ags under baseline rules. A major exception to the overall agging decrease was LD, where index agging increased by 1100% (Baseline N =
20 ags [1.5% of LD results]; Proposed N = 257 [19.4% of LD results]),
which is due to the low H-index LD cut-off value of 15.
In Lab B1, implementation of proposed rules caused only a slight increase in the total hemolysis agging rates (comment and cancel)
(Baseline N = 277 [0.05% of total]; Proposed N = 283 [0.05% of total];
2% increase). The changes are due to different H-index cut-offs for certain tests between the Modular P (baseline rules) and the c702 and
e601 (proposed rules). Although expected to be similar, there are sometimes signicant differences in the HIL cut-offs provided by the manufacturer for the same test on different platforms. A good example of
this from our study is DBIL, where despite having similar chemistries
on the 2 platforms, the H index cut-offs were very different (60 for the
Modular P; 25 for the c702). This cut-off change resulted in an increase
in DBIL agging from 9 to 45, and demonstrates how existing index
rules need to be re-evaluated after reagent and vendor changes. Other

examples included increased LD agging (Baseline: 25, Proposed: 98)

due to the change in H-index cut-off of 60 to 15. Conversely, ALT agging decreased from 34 ags under baseline to 3 ags under proposed
rules due to H-index cut-off changes of 60 to 200.
In Lab B2, proposed rules resulted in a 110% increase in agging
(Baseline N = 1047 [0.2% of total]; Proposed N = 2204 [0.40% of
total]). This was largely due to increased agging of LD and DBIL (LD
Baseline N = 7 [0.1% of LD results], LD Proposed N = 1175 [14.2% of
LD results]; DBIL Baseline N = 6 [0.2% of DBIL results], DBIL Proposed
N = 333 [12.4% of DBIL results]). For LD, this increase (which was also
observed to a lower extent in Lab A), is due to the low H-index cut-off
for this test of 15, which can be difcult to detect visually. Further analysis of the DBIL results showed that the majority of the H-index ags for
this test were on samples from babies, which is likely due to the collection method (heel poke) and the collectors (nurses instead of lab staff).
Hemolysis falsely lowers DBIL results, which could make an elevated
DBIL appear to be in the normal range. Further investigation of DBIL orders on children b 1 year of age revealed that most samples with Hindex 25500 did not have a hemolysis comment appended to the result and had DBIL values within the reference range. Of samples with
H-index N 500, 4 of 13 did not have an associated hemolysis comment,
demonstrating that even grossly hemolyzed samples were not always
agged under baseline policies; however all were identied using developed decision algorithm. Conversely, proposed rules decreased potassium agging (Baseline N = 567 [1.6% of potassium results],
Proposed N = 288 [0.8% of potassium results]; 49% decrease overall).
Under Lab B2 baseline policies, the hemolysis comments used potassium as an example of a test affected by hemolysis. This may have resulted in hemolysis ags being consistently attached to the potassium result
by technologists, even though the Roche c6000 H-index cut-offs are
lower for DBIL (25) and LD (15) than for potassium (100). Compared
to Lab A, which also used manual agging but saw a decrease in overall
agging following implementation of automated indices, the results
from Lab B2 further demonstrate the variability seen with manual agging and that agging can increase or decrease depending on the lab
when switching to automated indices.
3.2.2. Optimized rules
Proposed rules analysis greatly increased hemolysis agging for
some tests. In order to rene hemolysis agging, optimized rules were
generated that only agged a result if clinically signicant changes

36

J.M. Boyd et al. / Clinica Chimica Acta 450 (2015) 3138

were observed (Section 2.6.3.1.3) (objectives: identify specimens with interference levels relevant to the tests ordered, avoid adverse effects caused
by failure to report results with HIL interference).
Based on these criteria, only AST, CK and LD qualied for hemolysis
optimized rule generation. Since hemolysis falsely elevates results for
all 3 theses, the optimized rules were designed so that a hemolysis
ag was applied to these tests only if the H-index was above the test
specic cut-off and the test result was above the reference range.
When the new rules were modeled using the data, hemolysis agging
rates were reduced for all 3 tests by N60% in Labs A and B1 and N 30%
in Lab B2 (Table 2).
3.3. Icterus
3.3.1. Proposed rules
As with hemolysis, use of automated I-index reduced total icteric
agging in Lab A (Baseline 332 [0.06% of total]; Proposed 0; 100% reduction) (Table 3). This also shows that Lab A had relatively low rates of
icteric specimens which is expected in a community setting. One potential limitation of this data analysis is that Lab A may not have captured
sufcient icteric specimens in the time frame being examined due to
the low frequency of icterus in the studied population.
Labs B1 and B2 did not have a baseline process for investigating icteric interference. In Lab B1, application of proposed rules determined
that 13 creatinine, 11 triglycerides, 8 cholesterol and 8 GGT results
would have been agged (total of 40) over the 3 week data collection
period. This indicated that icteric interference was not a large concern

in this community population, which is similar to the observation in

Lab A. However, in the absence of icterus reporting rules, even the relatively rare icteric community samples could have been missed by the
clinician (objective: avoid adverse effects caused by failure to report results with HIL interference). Upon audit, it was noted that 4/40 of these
icteric samples did not actually have a bilirubin on the order, thus the
ag adds clinical value to the reporting.
Analysis of the hospital population from Lab B2 showed that icteric
interference was an issue (200 icteric ags generated) that had been
largely overlooked until the implementation of these rule sets. The
most highly agged test was creatinine (N = 125 [0.4% of creatinine results]) followed by GGT (N = 53 [0.6% of GGT results]). This is a reection of the lower icteric cut-offs for these tests of 15 and 20 respectively,
on the Roche c501 platform. None of these samples were agged under
baseline policies, as no rules for monitoring icterus were in place. In addition, when compared to the 2 community populations from Labs A
and B1, these results clearly demonstrate how patient population can
affect agging rates for different interferences.
3.3.2. Optimized rules
Optimized rules for icterus involved diluting the sample to below the
cut-off I-index and then rerunning it. They were considered for any test
with N 5 icteric ags under proposed rules. No icteric optimized rules
were developed for Labs A and B1 because of the low prevalence of icteric specimens. However, optimized rules were developed for CREA
and GGT in Lab B2. This could not be evaluated on our test data set as
the samples had already been analyzed and no existing icterus

Fig. 1. Optimized rule workow.

J.M. Boyd et al. / Clinica Chimica Acta 450 (2015) 3138

37

Table 5
Tube agging concordance between manual agging and automated serum indices in Labs A and B2.
Tubes agged
Hemolysis

Lab A
Lab B2

Lipemia

Icterus

Baseline

Proposed

Both

Baseline

Proposed

Both

Baseline

Proposed

Both

480
754

411
1750

72
364

369
37

4
19

2
4

54
0

0
139

0
0

intervention step was in place. Testing of the CREA and GGT optimized
rules was performed, and indicated that the CREA rule needed further
optimization to make the workload manageable for technologists (293
samples over 6 months). Investigation of the I-index cut-off for CREA indicated that results did not signicantly change until I-index of 30 (Supplemental Fig. S2). Further investigation determined that the agging
rate could be further reduced by only performing dilution on samples
with a CREA N 100. The nal optimized rule (dilution only if I
index 30 and CREA N 100) reduced sample dilution by 77%. The
same dilution study was conducted for GGT and there was no signicant
interference from Icterus until I index 55, which reduces GGT dilutions
from 53 to 5 (91%). This further highlights the need to assess the effects
of automated indices on each patient population and the value of testing
rule sets prior to implementation.

3.4. Lipemia
3.4.1. Proposed rules
Proposed rules in Lab A resulted in a decrease in agging for all tests
(Baseline N = 2814 [0.5% of total]; Proposed N = 37 [0.007% of total];
99% decrease) (Table 4).
In Lab B1, total lipemia agging rates increased by 160% (Baseline
74 [0.01% of total], Proposed 193 [0.04% of total]). This was due to an
increase in TBIL agging (Baseline N = 0 [0.0% of TBIL results]; Proposed N = 125 [1.6% of TBIL results]), as a result of a lower lipemia
cut-off (L-index of 90) indicated in the package insert for the c701
reagent compared to the Modular P reagent. This appeared unusual
and was further investigated. When TBIL interference was veried
in-house using an intralipid reagent, no signicant lipemic interference was found to an index level of 600. Changing the lipemia agging settings to this level reduced the agging rates by 98%. This
example highlights that it is sometimes insufcient to implement
the manufacturer's recommendations without at least some inhouse interference validation studies.
In Lab B2, proposed rules reduced total agging by 33% (Baseline
N = 39 [0.007% of total]; Proposed N = 26 [0.005% of total]). This
encompassed a decrease in agging for potassium (Baseline N = 25
[0.1% of potassium results]; Proposed N = 0) and increases in agging
for ALT (Baseline N = 1 [0.0% of ALT results]; Proposed N = 18 [0.1%
of ALT results]) and AST (Baseline N = 0 [0.0% of AST results]; Proposed
N = 5 [0.3% of AST results]).
No optimized rules were developed as ultracentrifugation procedures were already included in the proposed rules. Ultracentrifugation was successful in clearing lipemic interference using data from
Labs B1 and B2 which already had ultracentrifugation procedures
in place under baseline policies. Ultracentrifugation has been
recommended as the preferred method to reduce lipemia interferences [3].

3.6. Flagging concordance

Tube agging concordance between baseline (manual) and
proposed (automated indices) rules sets were performed for Labs A
and B2 (Table 5). Poor concordance was observed between manual
and automated HIL agging at both labs, which has been observed
previously [5,7].
4. Conclusions
A process is described to select test specic cut-offs and develop decision algorithms for specimens with HIL interference. Interference was
identied using automated serum indices with test specic HIL cut-offs.
The instrument middleware or LIS alerted technologists to the presence
of lipemia or icterus that needed ultracentrifugation or dilution. Tests
with interference above the cut-off had an interpretive comment
appended to the result. The number of test cancelations was limited
by agging tests with lower levels of hemolysis interference instead of
canceling them and by using interference reduction strategies for
lipemia and icterus. Flagging analysis allowed for the development of
optimized rules tailored for certain tests and patient populations to ensure that clinically signicant interference was reported. The development of optimized rules also ensured that more test results were
reported (with a comment) in order to avoid any adverse events by failing to report results. Analysis of the HIL agging rates clearly demonstrated the effect of lab specic and population specic factors.
Hemolysis agging under proposed rules decreased in Lab A, but increased in Lab B2 demonstrating how lab specic processes for manual
agging play a role in agging analysis. Lab B1 showed that agging
rules need to be re-evaluated when changing instrument platforms,
even from the same manufacturer. The effect of patient population
was demonstrated by the differences seen in icteric agging for creatinine and GGT between community and hospital populations. Taken together, this indicates that serum index algorithms are not universally
transferrable between labs or patient populations. An increase or decrease in overall HIL agging will depend on the lab in question, their
existing process and the population they serve. The authors present
this workow for developing site-appropriate rule sets as a resource
for labs intending to implement automated index reporting in their laboratories. It also applies to labs seeking strategies to further optimize
their interference reporting processes. The lack of publicly available
rule sets in this area of lab medicine leaves every lab trying to design
the whole system from rst principles. The current work clearly
demonstrates the benets of determining agging rates prior to implementation of automated serum index algorithms in order to better understand the effect on the laboratory workow.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.cca.2015.07.006.
Acknowledgements

3.5. Final algorithm

The nal version of the algorithm is shown in Fig. 1, including Lab B2
specic rules for icteric interference with creatinine and GGT. Examples
of the optimized rules are shown in Supplemental Table S4.

The authors thank Tihomir Curic and Dan Henne for their assistance
in obtaining data from the laboratory information system in Lab B2, and
Jenette Bourbonnie and Bernice Frandle for their assistance in comparing index measurements between analyzers in Labs B1 and B2.

38

J.M. Boyd et al. / Clinica Chimica Acta 450 (2015) 3138

References
[1] CLSI, C56A: Hemolysis, Icterus and Lipemia/Turbidity Indices as Indicators of Interference in Clinical Laboratory Analysis; Approved Guideline. CLSI Document C56A, Clinical and Laboratory Standards Institute, Wayne, PA, 2012.
[2] G. Lippi, N. Blanckaert, P. Bonini, et al., Haemolysis: an overview of the leading cause
of unsuitable specimens in clinical laboratories, Clin. Chem. Lab. Med. 46 (2008)
764772.
[3] P.L. Walker, M.A. Crook, Lipaemia: causes, consequences and solutions, Clin. Chim.
Acta 418 (2013) 3032.
[4] K.W. Ryder, M.R. Glick, S.J. Glick, Incidence and amount of turbidity, hemolysis, and
icterus in serum from outpatients, Lab. Med. 22 (1991) 415418.
[5] M.R. Glick, K.W. Ryder, S.J. Glick, J.R. Woods, Unreliable visual estimation of the incidence and amount of turbidity, hemolysis, and icterus in serum from hospitalized
patients, Clin. Chem. 35 (1989) 837839.
[6] H.J. Vermeer, E. Thomassen, N. de Jonge, Automated processing of serum indices
used for interference detection by the laboratory information system, Clin. Chem.
51 (2005) 244247.
[7] R. Hawkins, Discrepancy between visual and spectrophotometric assessment of
sample haemolysis, Ann. Clin. Biochem. 39 (2002) 521522.
[8] G. Lippi, P. Avanzini, D. Campioli, et al., Systematical assessment of serum indices
does not impair efciency of clinical chemistry testing: a multicenter study, Clin.
Biochem. 46 (2013) 12811284.
[9] Roche, Serum indices: reduction of clinical errors in laboratory Medicine, 2007.

[10] G. Lippi, G.L. Salvagno, M. Montagnana, G. Brocco, G.C. Guidi, Inuence of hemolysis
on routine clinical chemistry testing, Clin. Chem. Lab. Med. 44 (2006) 311316.
[11] G. Steen, H.J. Vermeer, A.J. Naus, B. Goevaerts, P.T. Agricola, C.H. Schoenmakers, Multicenter evaluation of the interference of hemoglobin, bilirubin and lipids on
Synchron LX-20 assays, Clin. Chem. Lab. Med. 44 (2006) 413419.
[12] J.Z. Ji, Q.H. Meng, Evaluation of the interference of hemoglobin, bilirubin, and lipids
on Roche Cobas 6000 assays, Clin. Chim. Acta 412 (2011) 15501553.
[13] H.J. Vermeer, G. Steen, A.J.M. Naus, B. Goevaerts, P.T. Agricola, C.H.H. Schoenmakers,
Correction of patient results for Beckman Coulter LX-20 assays affected by interference due to hemoglobin, bilirubin or lipids: a practical approach, Clin. Chem. Lab.
Med. 45 (2007) 114119.
[14] R.C. Hawkins, Correction and reporting of potassium results in haemolysed samples,
Ann. Clin. Biochem. 43 (2006) 8889.
[15] M.M. Mansour, H.M. Azzazy, S.C. Kazmierczak, Correction factors for estimating potassium concentrations in samples with in vitro hemolysis: a detriment to patient
safety, Arch. Pathol. Lab. Med. 133 (2009) 960966.
[16] H.J. Vermeer, G. Steen, A.J. Naus, B. Goevaerts, P.T. Agricola, C.H. Schoenmakers, Correction of patient results for Beckman Coulter LX-20 assays affected by interference
due to hemoglobin, bilirubin or lipids: a practical approach, Clin. Chem. Lab. Med. 45
(2007) 114119.
[17] A. Saracevic, N. Nikolac, A.M. Simundic, The evaluation and comparison of consecutive high speed centrifugation and LipoClear(R) reagent for lipemia removal, Clin.
Biochem. 47 (2014) 309314.