Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Assessment of Analytical Quality through Sigma

Metrics & its Application for Selection of
Westgard Rule
Dr. Sapna Patel1 Dr. Shreyansh Sharma2 Dr. Hariom Sharma3
Department of Biochemistry, American international institute of Department of Biochemistry,
Government Medical College, medical sciences, Government Medical College,
Bhavnagar, India. Udaipur, India. Bhavnagar, India.

Abstract:- Background: In today's world, the clinical A high-quality laboratory's performance is evident in both the
laboratory is a rapidly expanding area under constant test reports it produces and the Quality Controls it conducts as
demand to give speedy and reliable results. A clinical performance checks. (2)
laboratory's performance is measured using IQC and
EQAS. However, these approaches cannot quantify the Clinical laboratories use a variety of procedures to
number of errors. A new tool called sigma metric quantifies ensure quality, including Internal Quality Control and External
the approximate amount of analytical errors, and assesses Quality Control.(3) IQC is a sample material with a matrix that
and directs the development of better quality control is identical to that of the patient's sample and a concentration
procedures. In order to minimizing error rates Six Sigma range that is available in two or three levels to cover the medical
were utilized to quantify the analytical quality of decision points. The IQC is performed according to NABL
automated clinical chemistry. guidelines and interpreted using Levy Jennings' control charts
Objective: This study was conducted to estimate Sigma and Westgard rules. IQC keeps a constant eye on the analytical
metrics and Quality Goal Index of various biochemical system to see if the results are trustworthy enough to be
analytes in order to evaluate quality control performance released.(4)(5) External quality control entails analyzing and
and execute the best quality control approach for each reporting control samples provided by a third party at a
analyte. predetermined time interval, which in clinical chemistry is once
Material and method: IQC and EQA data were examined a month. The Z score or the standard deviation indexes are used
using a chemistry auto analyzer (Architect C 8000) at the to interpret external quality control. A Z score is a calculated
Biochemistry laboratory, Sir T Hospital, Bhavnagar, from value that indicates how many standard deviations a control
January 2022 to October 2022. Mean, standard deviation, result has deviated from the expected mean value for that
coefficient of variation %, bias % and sigma metrics and material.(4)(5)While running internal and external QCs, it is
Quality Goal Index were calculated for Plasma Glucose, difficult to quantify the exact amount of errors that occur in the
serum Urea, Creatinine, alanine ALT, AST, Total protein, system and to provide a direct and integrated evaluation of the
Albumin, ALP, Total Cholesterol, Triglyceride, HDL, analytical system's performance, Sigma metrics can.(6)
LDL, Uric acid and LDH.
Results: Excellent sigma values were elicited for SGOT Six Sigma is a management approach that helps to
(Level 2), ALP, Triglycerides, HDL, and Uric acid. enhance process output quality by identifying and eliminating
Satisfactory sigma values were elicited for, Creatinine the causes of defects (errors) and limiting variability in
(both the levels) TP, LDH (Level 1), SGPT, LDL (Level 2), manufacturing and business processes.(7) Six is the number of
while Glucose, Albumin, Cholesterol, (both the levels) standard deviations from the mean, which is a statistical
SGPT, LDL (Level 1), TP (Level 2) having sigma value <3. measure of distribution. It is a data-driven and statistically
Conclusion: Sigma metrics is useful for addressing poor driven strategy to eliminating manufacturing faults. (2)
performance in assessments, improves laboratory Concept of six sigma began at Motorola in 1982 with the goal
performance and aids in the evaluation of analytical of lowering costs, enhancing manufacturing techniques,
techniques. It serves as a roadmap for developing a quality reducing variation, and promoting quality improvement.(8)
control strategy. It can be used as a self-evaluation tool for In the year 2000, laboratory medicine adopted the "Six Sigma"
clinical laboratories. technique.(2) The first study utilizing sigma metrics in the
clinical lab was published by Nevalainen et al., in the year 2000
Keywords:- IQC, EQAS, Six Sigma, Quality Goal Index and since then many similar studies have been done throughout
the world. (9) The Sigma scale, which is used to categories
I. INTRODUCTION performance, ranges from sigma level 1 to 6, with 6 being the
target for world-class quality and 3 being the least allowed
In the Healthcare Laboratory, "quality" is defined as sigma for routine performance.(11) Although achievement of
adherence to the needs and expectations of users (nurses and sigma metrics value 6 or more is not easy, but with appropriate
physicians) or customers (patients or other parties who pay the precautions to minimize the errors associated with sample
bills), as well as satisfaction of those needs and expectations.(1) processing; this goal can be approached.

IJISRT22DEC1412 www.ijisrt.com 1573

Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
II) were obtained from Randox Laboratories India (lyphochek
DMAIC was the methodology used in our Central assayed chemistry control). EQAS sample was run once in
Laboratory to deploy Lean Six Sigma. TQM and the Six Sigma every month which was obtained from Randox Laboratories,
model are comparable. In comparison with TQM’s PDCA, we UK.
can say that define corresponds to the plan step, measure to the
do step, analyze to the check step, and improve to the act step. Calculation of CV%
Control is an additional step in the Six Sigma model that is By using IQC data; Mean & S.D. were determined from
critical in modern quality management. With this step, We can which CV% was calculated for each month.(2) The test
avoid defects returning to the process with this step.(7)(10)(9) method's analytical coefficient of variation is represented by
It means "identifying" the issue that is causing our results to CV percent. It is a indicator of random error. CV has no
deviate from a set of accepted criteria. It also entails identifying dimensions and is unaffected by changes in measuring units. If
the resources that may be required to tackle the issue. (1)(4) CV is less than 5%, then the particular method used for
After that, in order to tackle the problem, its extent was determination of an analytes concentration is said to have a
"measured" by collecting relevant data and storing it in a more very good performance& precise. Precision is the degree of
presentable format, such as collecting IQC and EQC data and agreement between independent, repeatable results obtained
computing CV percent and Bias percent from it. from the same sample under certain conditions.(5)
(11)Subsequently the information was "analyzed" to work
out the basis explanation for problems. During this phase, we CV% = SD x 100 / Lab Mean
estimate the disparities between our results and also the goal
values, yet as determine their likely reasons. (12) After Calculation of Bias: Bias is the systematic discrepancy between
that, the foundation causes were eliminated by implementing the results that would be achieved using a recognised reference
certain corrective measures in accordance with Westgard sigma technique and the predicted results obtained by the laboratory's
rules to “improve” the method performance. Following the test method. Bias is used to describe the inaccuracy of the
correction of the problem, specific preventive measures were method. Lower the bias more is the accuracy.(5) The following
implemented to "control" or "check" that the formula was used to determine bias using RIQAS:
matter wouldn't recur within the future. Every step of sample
processing was meticulously checked and monitored so as to Bias (%) = (Mean of all laboratories using same instrument &
eliminate any errors or inefficient operations. To minimize method – Our Laboratory Mean) / (Mean of all laboratories
TAT and enhance sample processing quality, many remedial using same instrument & method) x 100
procedures were performed to cut back mistakes at the pre-
analytical, analytical, and post-analytical phases.(11) The total permissible error: The total permissible error (TEa)
values were derived from CLIA-88 (Clinical Laboratories
II. MATERIALS &METHODS Improvement Act) recommendations. (8)

From January to October 2022, an extensive investigation Calculation of sigma metrics: Sigma metrics (σ) were
of sample processing and quality control methods was calculated using the equation:
conducted in the Clinical Biochemistry Laboratory at Sir T
Hospital in Bhavnagar, Gujarat. The parameters which were Sigma metrics (σ) = (TEa – Bias) / CV
analyzed include Glucose, Urea, Creatinine, AST, ALT, ALP,
total protein, albumin, cholesterol, triglyceride, high Density Calculation of Quality Goal Index: The quality goal index
Lipoprotein (HDL), Low Density Lipoprotein (LDL), Uric (QGI) ratio indicates how well bias and precision meet their
acid, LDH. Both levels of IQC were done using Architect C respective quality goals. This was used to investigate the cause
8000 clinical chemistry analyzer and all data were analyzed on of the lower sigma in analytes, i.e., whether the issue is due to
each day. Only if the IQC was within the permissible range imprecision, inaccuracy, or both. (4)(11)
according to Westgard guidelines were the patients' samples
conducted and reported. The IQC material (both level I & level QGI = Bias/1.5 × CV%.

Table 1: Sigma level and ppm defects or errors per Million Opportunities (DPMO)
Sigma Level Accuracy Long-Term ppm* Defects
1 30.85% 691,462
2 69.1% 308,538
3 99.33% 66,807
4 99.38% 6,210
5 99.977% 233
6 99.99966% 3.4

Table 2: Criteria for interpreting Quality Goal Index: (4)

QGI Problem

IJISRT22DEC1412 www.ijisrt.com 1574

Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
<0.8 Imprecision
0.8-1.2 Imprecision and inaccuracy
>1.2 inaccuracy

Table 3: CV% calculated from Internal Quality Control L1 from January 2022 To October 2022
L1
Parameter Jan Feb Mar Apr May June July Aug Sep Oct AVG
Glucose 3.29 6.00 3.03 2.79 3.68 3.16 - 3.68 2.59 3.58 3.18
Urea 3.57 3.59 0.54 3.75 2.75 4.54 4.59 4.03 3.76 2.62 3.374
Creatinine 4.17 3.53 4.05 3.58 3.45 1.82 - 3.15 5.72 4.46 3.77
SGPT 6.62 4.89 5.20 5.36 6.98 5.29 5.35 6.14 5.25 6.67 5.775
SGOT 3.59 3.37 3.25 3.03 3.45 3.17 3.94 3.27 3.13 3.11 3.331
ALP 6.00 4.31 3.27 4.38 4.06 3.34 3.24 5.52 5.04 5.54 4.47
TP 2.68 2.02 2.36 3.61 2.43 2.17 1.8 2.15 2.44 1.83 2.349
ALB 2.80 2.02 2.30 3.02 3.11 3.30 2.76 0.92 3.36 2.44 2.603
Cholesterol 2.60 2.79 3.31 2.92 2.81 1.30 1.24 3.03 3.20 2.20 2.54
TG 3.16 3.80 2.6 2.03 3.56 1.91 1.93 2.89 2.39 1.67 2.594
HDL 4.56 4.18 3.81 3.56 4.48 3.0 4.86 3.43 3.57 3.05 3.85
LDL 5.26 4.94 2.96 4.03 4.27 3.70 3.16 3.98 RNS 2.71 3.89
UA 3.74 2.28 2.32 9.0 2.52 1.33 1.55 2.36 2.51 1.73 2.934
LDH 5.0 4.85 4.32 4.98 6.01 4.18 6.64 3.91 3.63 3.96 4.748
LIPASE 5.39 6.62 3.12 12.7 3.15 3.30 3.68 4.01 2.22 2.81 4.7

Table 4: CV% calculated from Internal Quality Control L2 from January 2022 to October 2022
L2
Parameter Jan Feb Mar April May June July Aug Sept Oct AVG
Glu 2.86 2.69 3.17 3.40 3.73 1.88 - 3.55 2.50 3.31 3.01
Urea 3.56 2.50 3.61 3.67 2.42 3.32 2.49 4.14 2.49 2.57 3.07
Creat 3.04 2.63 3.31 2.26 2.16 1.70 - 2.89 3.13 5.42 2.94
SGPT 3.19 2.67 4.51 2.19 3.06 3.93 3.58 4.27 2.81 3.62 3.38
SGOT 3.23 2.53 3.15 3.19 2.72 2.11 2.57 3.96 1.79 2.72 2.79
ALP 3.25 2.80 2.82 2.53 3.34 1.89 2.25 4.67 4.52 5.47 3.35
TP 3.11 2.61 2.91 2.44 3.18 3.57 3.68 2.96 2.05 2.83 2.93
ALB 3.0 2.98 3.61 3.28 2.67 5.69 2.92 1.92 2.07 2.22 3.03
CHOLE 3.02 2.84 2.48 3.13 2.74 3.36 1.59 2.91 1.43 2.78 2.62
TG 3.65 4.38 3.29 2.85 2.96 2.53 2.33 5.18 1.87 2.83 3.18
HDL 5.15 4.58 3.32 3.16 3.52 3.37 5.13 4.09 2.99 4.25 3.95
LDL 4.39 5.55 4.37 2.60 2.81 2.61 2.89 2.72 RNS 3.42 3.48
UA 3.76 2.86 2.0 2.10 2.02 1.91 1.95 2.33 1.77 1.49 2.21
LDH 2.65 2.89 2.99 4.35 4.55 3.92 3.83 3.52 3.29 2.60 3.45
LIPASE 4.19 6.64 4.20 10.85 3.09 4.36 3.15 4.18 3.50 2.42 4.65

Table5. Bias % calculated from RIQAS from January 2022 to October 2022
Paramter Jan Feb Mar April May June July Aug Sep Oct Avg
GLU 1.3 9.9 1.4 1.6 0.4 1.2 0.1 1.6 5.9 1.0 2.44
UREA 7.3 11.4 8.5 4.0 5.1 10.3 10.2 7.0 5.8 7.9 7.75
CREAT 4.9 7.2 7.4 0.5 4.0 3.7 0.1 1.8 5.3 1.6 3.65
SGPT 1.5 7.1 5.2 12.3 9.6 8.4 8.9 6.2 11.4 5.0 7.56
SGOT 2.9 0.4 1.0 8.1 5.2 6.6 4.4 5.7 0.9 5.4 4.06
ALP 8.6 16.32 0.9 3.8 4.5 3.9 2.7 7.9 13.9 12.4 7.49
TP 1.6 0.8 1.1 3.5 6.3 1.6 3.5 1.7 6.6 1.6 2.83
ALB 3.7 1.9 1.8 1.3 1.9 3.5 6.3 0.6 4.6 0.0 2.56
CHOLE 1.6 4.6 7.6 3.1 7.3 0.5 0.0 1.9 2.2 3.7 3.25
TG 0.5 6.0 0.8 5.9 3.2 2.0 3.9 0.8 4.8 3.1 3.1
HDL 4.4 2.2 0.0 0.2 1.8 0.2 9.3 0.7 2.5 0.5 2.18
LDL 1 4.7 9.5 11.6 10.9 14.3 16.8 9.6 - 4.3 9.18
UA 0.3 0.7 0.1 1.1 2.2 0.1 3.1 3.7 5.1 0.6 1.7

IJISRT22DEC1412 www.ijisrt.com 1575

Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
LDH 5.7 1.8 1.7 12.7 8.3 6.0 5.8 4.4 2.0 4.0 5.24

Table 6. Average Bias, Average CV% & sigma metrics calculated for 10 months for both levels of IQC (L1 & L2).
Paramter TAE BIAS CV 1 Sigma 1 CV2 Sigma 2
GLU 10 2.44 3.18 2.38 3.01 2.51
UREA 9 7.75 3.37 0.37 3.077 0.40
CREAT 15 3.65 3.77 3.01 2.95 3.85
SGPT 20 7.56 5.77 2.15 3.38 3.68
SGOT 20 4.06 3.33 4.78 2.80 5.70
ALP 30 7.49 4.47 5.03 3.35 6.71
TP 10 2.83 2.35 3.05 2.93 2.44
ALB 10 2.56 2.60 2.86 3.04 2.45
CHOLE 10 3.25 2.54 2.66 2.63 2.57
TG 25 3.1 2.59 8.44 3.19 6.87
HDL 30 2.18 3.85 7.23 3.96 7.03
LDL 20 9.18 3.89 2.78 3.48 3.10
UA 17 1.7 2.93 5.22 2.22 6.89
LDH 20 5.24 4.74 3.11 3.46 4.27

Table-7: Sigma metrics of various parameters

Sigma metrics L1 L2
<2 (Unacceptable) Urea Urea
2-3 (Poor) Glucose, SGPT, Albumin, Cholesterol, Glucose, TP, Albumin, Cholesterol
LDL
3-4 (Acceptable) Creatinine, TP, LDH Creatinine, SGPT, LDL
4-5 (Good) SGOT, LDH
5-6 (Excellent) Uric acid, ALP, SGOT,
>6 (world class performance) Triglyceride, HDL ALP, Triglyceride, HDL, Uric acid

Table-8: QGI
Parameter QGI (L1) Problem QGI (L2) Problem
Glucose 0.5 Imprecision 0.5 Imprecision
Urea 1.5 inaccuracy 1.7 inaccuracy
SGPT 0.9 Imprecision and inaccuracy 1.5 inaccuracy
Total protein 0.8 Imprecision and inaccuracy 0.6 Imprecision
Albumin 0.6 Imprecision 0.6 Imprecision
Cholesterol 0.8 Imprecision and inaccuracy 0.8 Imprecision and inaccuracy
LDL 1.6 inaccuracy 1.8 inaccuracy

III. DISCUSSION
>6σ –excellent tests =evaluate with 1 QC/day.(alternating
The present study was undertaken to evaluate the quality levels between days) and follow 1-3 s Westgard rule.
of the analytical performance of clinical chemistry laboratory 4 σ - 6 σ =suited for purpose –evaluate with two levels of QC
of Sir T Hospital, Bhavnagar, Gujarat, India on sigma scale. In /day, follow 1-3 s, 2-2 s, R4 s Westgard multirules.
clinical laboratories, evaluating the quality of laboratory testing 3 σ - 4 σ =poor performers-use a combination of rules with 2
is an important research topic. Six Sigma quality standards levels of qc/day, follow 1-3 s, 2-2 s, R4s, and 4-1 s Westgard’s
consider bias (system error) and CV (random error) to guide multirules
quality management in clinical laboratories while analyzing < 3 σ = max QC, 3 levels, 3 times a day. Root cause analysis
possible causes of error, identifying solutions, improving should be performed; method performance must be improved
testing quality, and optimizing the QC schedule. However, the before the method can be routinely used (10)(11)(13)
optimal TEa, bias, CV, and other indicators to calculate 6σ are
unknown, especially when the sources of bias and CV differ It can be visualized from Table1 & Table 2 that except
between laboratories. As a result, we compared two new for SGPT (CV%>5 in L1) all parameters depicted CV<5%.
methods for calculating metrics as a future reference for the This clearly indicates that our lab has achieved high level of
implementation of 6σ quality management in clinical precision in remaining 13 analytes.
laboratories. (14)
Another important calculated index in the present
Selection of westgard rules Based on the sigma values study is bias% by using the EQC data. Bias shows high degree
obtained from the QC: (16) of accuracy in our lab results. Out of all the parameters

IJISRT22DEC1412 www.ijisrt.com 1576

Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
measured in the present study Urea, SGPT, ALP, LDL, LDH [6]. Angmo D, Kant S. Six sigma implementation in
had Bias >5%. Rest all parameters depicted Bias <5%. This healthcare industry: Past, present and future. Int J Eng
indicates that our lab has achieved high level of accuracy in Res Technol. 2015;4(6):1078-82.
remaining 9 analytes. [7]. Adiga US, Preethika A, Swathi K. Sigma metrics in
clinical chemistry laboratory-A guide to quality control.
In the present study; σ>5 for Triglyceride, HDL, ALP, Al Ameen J Med Sci. 2015;8(4):281-7.
Uric acid, SGOT( Level 2) was observed (Table 4). Hence it [8]. Westgard S, Bayat H, Westgard JO. Analytical Sigma
required only 13s, 22s, and R4s to be followed. metrics. A review of Six Sigma implementation tools for
medical laboratories. Biochem Med (Zagreb).
σ>3-5 for both the level of IQC was observed for 2018;28(2):020502.
Creatinine, LDH, TP, SGOT, SGPT, LDL. It implies Westgard [9]. Nevalainen D, Berte L, Kraft C, Leigh E, Picaso L,
sigma multirole application is needed for such parameters. Morgan T. Evaluating laboratory performance on quality
indicators with the six sigma scale. Archives of
Those parameters having σ<3 requires extensive Pathology & Laboratory Medicine. 2000;124(4):516–
evaluation in terms of reducing analytical bias & imprecision. 519.
[10]. Coskun A, Inal TC, Serteser M. Six Sigma Projects and
Mahmood, Bushra, et al.have similar result for the Personal Experiences. Croatia: InTech; 2011.
Glucose, Creatinine and Urea(<3 sigma) and [11]. Smętkowska M, Mrugalska B. Using Six Sigma DMAIC
SGPT(>3).(17)(18) has similar results for Total protein and to improve the quality of the production process: a case
Albumin and Total Bilirubin(<3 sigma) and for triglyceride study. Procedia-Social and Behavioral Sciences. 2018
sigma value is >3.(19) However 20 number study does not Jan 1;238:590-6.
coincide with our study result which shows sigma value > 6 for [12]. Heher YK. A brief guide to root cause analysis. Cancer
the Glucose, Creatinine and Total Bilirubin.(20) Cytopathology. 2017 Feb;125(2):79-82.
[13]. Singh B, Goswami B, Gupta VK, Chawla R, Mallika V.
IV. CONCLUSION Application of sigma metrics for the assessment of
quality assurance in clinical biochemistry laboratory in
Sigma metrics in clinical laboratories is a crucial India: a pilot study. Indian journal of clinical
methodology to detect and rectify any lab results that deviate biochemistry. 2011 Apr;26(2):131-5.
from the established criteria. It can assist us in determining poor [14]. Westgard JO, Westgard SA. The quality of laboratory
assay performance and in evaluating the effectiveness of the testing today: an assessment of σ metrics for analytic
current laboratory procedures. The idea of sigma metrics quality using performance data from proficiency testing
allows for the elimination of time-consuming and costly surveys and the CLIA criteria for acceptable
additional stages. This will shorten the turnaround time and performance. American journal of clinical pathology.
facilitate the delivery of high-quality reports for improved 2006 Mar 1;125(3):343-54.
patient care. Additionally, sigma metrics can aid in developing [15]. Kinns H, Pitkin S, Housley D, Freedman DB. Internal
suitable plans for the prudent application of IQC & EQC in a quality control: best practice. Journal of Clinical
larger clinical laboratory.(8) The goal of six sigma is to ensure Pathology. 2013 Dec 1;66(12):1027-32.
that important specifications are met by reducing both variance [16]. Mahmood B, Rasheed MK, Khazal A, Rasheed MK.
and quality control procedures. It is also imperative to Assessment of sigma metric results of serum parameters
implement appropriate QC strategies in order to augment the of liver and kidney function tested by automated
judicious use of QC.(6) chemistry analyzer in Medical City Hospital. Iraqi
Postgrad Med J. 2018;17(3):307-14.
REFERENCES [17]. Kaftan AN, Yaseen AK, Hasan Z. Assessment of sigma
metrics results of serum glucose and lipid profile tested
[1]. Garg M, Sharma N, Das S. SIGMA METRICS by automated chemistry analyzer in medical city
ASSESSMENT AS QUALITY IMPROVEMENT hospitals in Iraq. Int J Res Med Sci. 2017;5(11):4690–6
METHODOLOGY IN A CLINICAL CHEMISTRY [18]. Iqbal S, Mustansar T. Application of sigma metrics
LABORATORY. medRxiv. 2021 Jan 1. analysis for the assessment and modification of quality
[2]. Klee GG, Westgard JO. Quality Management. In: Burtis control program in the clinical chemistry laboratory of a
AC, Ashwood ER, Brurns DE, editors. Tietz Textbook of tertiary care hospital. Indian Journal of Clinical
Clinical Chemistry and Molecular Diagnostics. Biochemistry. 2017 Mar;32(1):106-9.
Saunders: Philadelphia; 2012. p. 163–204. [19]. Hamza MM. Assessment of Quality Control in a Clinical
[3]. Koshy JS, Raza A. Sigma metrics in quality control-An Chemistry Laboratory using Six Sigma
innovative tool. Metrics (Doctoral dissertation, University of Gezira).
[4]. Westgard JO. Six Sigma quality design and control. [20]. Al-Hassan M. Assessment of Quality Control in a
Westgard QC. Inc, Madison. 2006. Clinical Chemistry Laboratory using Six Sigma Metrics;
[5]. Kumar BV, Mohan T. Sigma metrics as a tool for 2017. Available from: http://
evaluating the performance of internal quality control in repo.uofg.edu.sd/handle/123456789/2979.
a clinical chemistry laboratory. Journal of laboratory
physicians. 2018 Apr;10(02):194-9.

IJISRT22DEC1412 www.ijisrt.com 1577