Upload
others
View
18
Download
0
Embed Size (px)
Citation preview
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 31-44 www.itspoa.com/journal/em
Variable and Attribute Control Charts in
Trend Analysis of Active Pharmaceutical
Components: Process Efficiency Monitoring
and Comparative Study
Mostafa Essam Eissa1*
1Microbiology and Immunology Department, Faculty of Pharmacy, Cairo University, Cairo, Egypt
Email addresses [email protected] (Mostafa E. Eissa)
*Correspondence: [email protected]
Received: 28 May 2018; Accepted: 15 June 2018; Published: 30 June 2018
Abstract: Assessment of pharmaceutical product quality is important prerequisite to justify safe
and effective release of the medicinal dosage form to the drug market. However,
without rigorous implementation of good manufacturing practice (GMP), routine
quality control testing may be not adequate to conclude compliance with reproducible
procedures. Accordingly, the current study aimed to investigate manufacturing quality
of pharmaceutical product batches through monitoring assay results and trends
retrospectively for three components of the active ingredients using two types of
control charts and to compare the value of each in-process monitoring. This product
was manufactured in a pharmaceutical firm and subjected to the assay (expressed as
relative potency to the claimed labeled dose per tablet) in quality control laboratory.
The active components are Paracetamol (Acetaminophen) (Pa), Chlorpheniramine
Maleate (CM) and Pseudoephedrine Hydrochloride (PH). General performance and
trend of the studied batches were compared using Individual-Moving Range and
Laney U΄ chart which were constructed using statistics software. Box-and-Whisker
diagram that was constructed for the assay of the three active constituents showed that
CM relative potency was significantly higher than Pa and PH using ANOVA (p<0.05).
Capability analysis showed that Pa and PH assays have met the requirement of
analysis. In contrast to CM potency which demonstrated a failure to be maintained
within the specification window level as strong shift outside the upper border (right
drift) could be observed. Both types of control charts variable (Individual-Moving
Range) and attribute (Laney U΄) showed same control limits. But Individual-Moving
Range was more sensitive in detection of out-of-control states.
Keywords: Individual-Moving Range, Laney U΄, Capability Analysis, Upper Control Limit,
Lower Control Limit, GMP
1. Introduction
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 32-44 www.itspoa.com/journal/em
With advancement in the pharmaceutical industry, new techniques and technologies
are available to improve medicinal dosage forms quality, safety and rigorous
monitoring. However, this breakthrough is accompanied by the challenge of the ever-
increasing number of health-deficient population of patients at stake [1].
Apparently, that drug recall from the market has become an increasing trouble that
faces regulatory authorities with pharmaceutical industrials' engines, as could be seen
from Figure 1 which was derived from Gaffiney (2014) [2]. Surging in the number of
recalls of the medicinal products has just occurred in the past few years as was
discussed by Cossman (2017) [3].
Figure 1. Recall rate per year as a total, critical, major and minor [2].
Recently, Food and Drug Administration (FDA) has segregated the recalls into
what is called "three-tier system"[4]. Pharmaceutical products that violate regulations
may: be life-threatening because they may cause severe adverse side effect which may
lead to death (class I), cause reversible or temporarily adverse medical effects with
remote possibility to cause serious health effects (class II) and finally, not impose any
adverse health problems(class III) [5].
Statistical process control (SPC) tools are an effective collection of techniques and
methodologies that are used to solve problems by enhancement of process quality,
stability and capability through minimization of the variability. Shewhart control
charts are considered pivotal tool in SPC that are used to monitor the state of control
of the inspected characteristics [6].
Shewhart (process-behavior) charts offer several benefits in the pharmaceutical
industry. These advantages include improvement of the manufacturing process
economically if used correctly. Also, control charts provide a measure of the process
capability to meet the specifications and trending of data with an additional advantage
of the presence of upper control limit (UCL), lower control limit (LCL) and the
process average lines. Moreover, control charts are a graphical presentation that
facilitates visualization of the points where the process is out-of-control [7].
Several researchers have applied control charts in the monitoring of different
inspection properties (example: assay, hardness, content uniformity and disintegration)
in medicinal dosage forms such as tablets and demonstrated their value in the
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 33-44 www.itspoa.com/journal/em
pharmaceutical manufacturing industry to monitor the inspected properties and
determine the process performance accordingly [8-10].
Individual-moving range (I-MR) is a type of variable control chart that is actually
composed of two charts I and MR. I chart allows to track the process level for each
reading. While MR chart displays the variation between successive readings [11]. On
the other hand, Laney U΄ chart is a type of attribute control charts that correct for
over-dispersion or under-dispersion of data that may otherwise interfere with correct
results interpretation if conventional U chart was used [12]. Accordingly, one can
access the overall performance of the process easily visually early enough before any
true excursion beyond acceptance limits could occur.
In the light of the preceded challenges, the current study aimed to investigate and
compare process monitoring using both attribute and variable control charts
simultaneously using Laney U΄ and Individual-Moving Range (I-MR) charts,
respectively. The present work would elucidate also the manufacturing process
efficiency through monitoring the state of control of the inspected product properties.
This work is part of a broad collaborative study to monitor the level of compliance of
the pharmaceutical firm to good manufacturing practice (GMP) through SPC
monitoring tools.
2. Materials and Methods
A small pharmaceutical plant based in the industrial zone was built in South Delta,
Egypt. The firm consists of small class D production area and has launched the
manufacturing of new oral film-coated tablet (FCT) product [13]. The product is used
for the treatment of common cold signs and the active pharmaceutical ingredient (API)
is based on triple complementary components viz. Paracetamol (Acetaminophen) (Pa),
Chlorpheniramine Maleate (CM) and Pseudoephedrine Hydrochloride (PH). Other
inactive ingredients include Povidone (binder), Sodium Croscarmellose
(disintegrating agent), Microcrystalline Cellulose (filler), Colloidal Silicone Dioxide
(glidant) and Magnesium Stearate (lubricant). The coating material is composed of
coating polymer, plasticizer, opacifying agent and coloring material.
A project was established by monitoring the manufacturing process efficiency using
SPC on results collected by quality assurance team about inspection properties of the
medicinal product over the year 2016. This covered 195 batches of the product and
the property being inspected was the relative potency assay of the three APIs with
acceptance criterion of 90 - 110 % [14]. The assay results of APIs are the pooled
outcome of the whole manufacturing process [15]. A total number of 150 tablets were
sampled by a trained QC sampler for each batch and submitted to the laboratory for
analysis. The pooled bulk sample was collected after coating of the core of the tablets
in the coating machine. Each assay point (batch) is expressed as relative potency ratio,
where obtained result of the analysis is divided by the labeled theoretical value. After
that, every point is added into a cell in a column of the statistical software program
sequentially to construct the control chart and/or other statistical calculations.
Distribution fitting was assessed initially in the current study to determine the
appropriateness of data distribution to specific control chart. Determining correct
distribution is crucial for the validity of the statistical analysis and SPC performed and
any interpretation or conclusion derived from the analysis [16]. Out-of-control alarms
in control chart is interpreted as the following: Alarm "1" = One point more than 3 x
standard deviation from mean line, Alarm "2" = Nine successive points on the same
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 34-44 www.itspoa.com/journal/em
direction of the mean line, Alarm "3" = Six successive points with general trend of all
being increasing or decreasing, Alarm "4" = 14 successive points with up and down
zigzag pattern, Alarm "5" = Two out of three points that are more than 2 x standard
deviation in one direction from the mean line, Alarm "6" = Four out of five points that
are more than standard deviation in one direction from the mean line, Alarm "7" = 15
successive points that are confined in the range of one standard deviation on both
sides of the mean line and Alarm "8" = Eight successive points that fall outside one
standard deviation on both sides of the mean line. Attribute control charts can only
detect the first four types of assignable cause alarms. Derivation of control limits (CLs)
is shown in Table 1 according to Minitab® 17.1.0 manual.
Table 1.Parameters derivation for I-MR and Laney U΄ control charts using statistical software
(Minitab®
17.1.0).
Control
Limits
(CLs)
I-MR Laney U΄
Individuals chart Moving range chart
Mean Line Σxi/Σni MR(bar).d2(w) Σxi/Σni
Upper K d2 (w) + k d3 (w) ̅ + K (ui)
Lower K d2 - k d3 [If LCL <
0, LCL = 0] ̅- K (ui) [or zero
whichever is greater]
d2(w) = Unbiased constant = Standard deviation ̅ = Center line
xi = number of items in subgroup i (Laney U΄) or individual observations (I-MR)
ni = subgroup size for subgroup i (Laney U΄) or number of individual observations (I-MR)
ui = proportion of items for subgroup i = Mean line
K = the parameter that is specified for Test 1 of the tests for special causes, 1 point > K standard deviations from center line (in the current study = 3).
w = The number of observations that are used in the moving range MR(bar) = The estimate of the average moving range for the method that you use to estimate
the standard deviation
d2 = A constant used to estimate the standard deviation
d3 = A constant used to calculate control limits for ranges
Data collected for the assay were subjected to statistical analysis using One Way-
ANOVA at p<0.05 and distribution fitting by GraphPad Prism 6 for Windows and
XLSTAT Version 2014.5.03, respectively. Capability plot and histogram, in addition,
the control charts were generated by Minitab® 17.1.0. The present study would focus
on the consistency of the product manufacturing period of the 195 lots in addition to
the comparison between the application of both types of control charts I-MR and
Laney U΄ modification as an example of variable and attribute charts, respectively.
Application of these programs has been discussed previously in other works [17-19].
3. Results and Discussion
Distribution fitting study can be examined from Table 2 and it showed that the
results pattern of the assay of the three APIs demonstrates a variable degree of
normality. None of them passed Poisson or Binomial distribution tests. Accordingly,
conventional attribute charts may deem not suitable for results interpretation. In
addition, the software could identify the closest distribution that fit each result from
series of different distributions as the best-fit test.
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 35-44 www.itspoa.com/journal/em
Table 2.Distribution fitting test conducted on the three APIs of coated tablet using statistical
software program.
Distribution Fitting Pa CM PH
Gaussian Fitting Yes Yes Yes
Pass at P = 0.3747 0.1571 0.5148
Binomial Fitting no no no
P = <0.0001 <0.0001 <0.0001
Poisson Fitting no no no
P = <0.0001 <0.0001 <0.0001
Maximum Likelihood Distribution Fitting Logistic Beta 4 Weibull 3
P = 0.7974 0.5391 0.7822
Box-and-Whisker diagram that was constructed for the assay of the three active
constituents showed that CM relative potency was significantly higher than Pa and PH
using One Way-ANOVA at p<0.05 as shown in Figure 2. Interestingly, points
denoted by asterisks "*" in Pa and CM are aberrant batches that show unusual results
of the assay of both components and this requires further investigation to elucidate the
cause.
Figure 2.Box-and-Whisker Plot for the assay of three active pharmaceutical ingredients of film-
coated tablet.
On the other hand, capability analysis showed that both Pa and PH assays have met
the requirement of analysis in both long run and short-term capability study. This
contrasts with CM potency which demonstrated a failure to be maintained within the
specification window level as strong shift outside the upper border (right drift) was
evident. While Figure 3A and C histograms were nearly centered and approximate,
the bell-shaped of the normal distribution, Figure 3B was partially truncated from the
right side i.e. at the border of the upper specification limit (USL).
Figures 4-6 show the control charts for the three active components with out-of-
control batches are red marked. Two types of control charts were constructed for each
assay trend namely variable control chart (Individual-Moving Range = I-MR) and
attribute control chart (Laney U΄). Both types of charts were on the same line in terms
of CL, upper control limit (UCL), lower control limit (LCL) and some alarm points.
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 36-44 www.itspoa.com/journal/em
However, I-MR charts demonstrated additional alarm points not present in the
counterpart Laney U΄. Moreover, I-MR chart could assess process variation through
the MR chart apart from I that trends the process center.
Interestingly, out-of-control batches (marked by numbered red dots) showed
abnormal assay values in greater magnitude with Pa followed by PH then CM but
none exceeded the specification limits. This finding is currently subjected to an
extensive investigation to determine the most probable root causes. Such factors that
were not related to normal process fluctuations are critical to being identified and
corrected as they may intercept with other similar products manufacturing in the
facility.
Figure 3.Capability Histogram and plot for Paracetamol (A), Chlorpheniramine Maleate (B) and
Pseudoephedrine Hydrochloride (C).
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 37-44 www.itspoa.com/journal/em
Figure 4. I-MR and Laney U΄ chart of Paracetamol for 195 batches of film-coated tablet.
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 38-44 www.itspoa.com/journal/em
Figure 5.I-MR and Laney U΄ chart of Chlorpheniramine Maleate for 195 batches of film-coated
tablet.
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 39-44 www.itspoa.com/journal/em
Figure 6. I-MR and Laney U΄ chart of Pseudoephedrine Hydrochloridefor 195 batches of film-
coated tablet.
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 40-44 www.itspoa.com/journal/em
Preliminary data analysis before the use of suitable Shewhart chart is mandatory to
obtain valuable outcomes and conclusions. The application of conventional attribute
control charts, either P or U may suffer from over- or under-dispersion in data, in
addition, they may not follow Poisson or Binomial distributions too which is part of
important assumption in control chart construction [20-22]. In such situation, Laney
corrected control charts can bypass these barriers [23]. On the other hand, it was
recommended by other experts to apply X-mR (I-MR) charts for discrete data that
failed to follow binomial or Poisson distribution [24]. This type of variable control
charts requires that the results show a certain degree of normality [25]. Accordingly
and based on the distribution fitting results of Table 1, the application of classical
attribute charts may be deemed not appropriate with the possibility of the emergence
of false alarms which leads to a wrong interpretation of data. Thus, Laney U΄ chart
may be appropriate approach to solve this challenge. However, data distribution show
some degree of normality that may not obviate the use of I-MR charts as
recommended by some experts [26].
Visualization of data distribution and pattern could be achieved using Box Plot
diagram and any abnormally extreme points might be observed easily as could be seen
in Figure 2 [27]. In the same line, the shape of histogram provides help in the
inspection of the process behavior and detection of unusual pattern such as external
force excreted on the operation and not pertained to the process [28, 29].
Complementarily, capability plot facilitates the visualization of the ability of the
process to meet the requirements over short-term (within) and long-term (overall) [30].
Moreover, process drift or shift from the center could be easily detected [31].
The ability of the SPC software package to construct control charts either
conventional attribute or Laney modified in addition to variable control charts
facilitated the comparison and assessment of the advantages of both types [32, 33].
Minitab can detect four additional types of out-of-control states in variable control
charts over that in attribute ones including those having the ability to provide early
warning for early process shift. Interestingly, they can measure also process variation
in another chart [34-36]. Conclusively, Laney type of attribute charts is easily
implemented, time-saving and avoids assumptions needs required for conventional
control charts. They could be reserved when normality assumption is no longer works
at all for application of I-MR charts which may show an advantage over attribute
charts
Shewhart charts provide an indispensable tool to discriminate between batches that
are within normal manufacturing process variability and those ones with outlier values.
Accordingly, an immediate action(s) could be taken (when the out-of-control point(s)
emerged) before any true excursion(s) may occur if they are constructed
chronologically with manufacturing progress. Extension of SPC to other medicinal
products would help to spot the major sources of defects in the system to correct them
long before any out-of-specification (OOS) events may evolve. SPC tools could detect
the non-consistency in the production batches which requires further investigation to
determine the causative source which may be either laboratory related or production
issue. Specifically, CM potency provided a serious case that requires attention and
immediate correction, probably due to the weight problems that are significantly in
excess compared to the other two components.
4. Conclusions
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 41-44 www.itspoa.com/journal/em
When the manufacturing steps and operations are followed literally, the trended
process should show stability and controlled variation within normal range over time.
In another word, a strict implementation of GMP during operation will ensure
delivering a product with the same expected properties every time. Control charts
possess the ability to detect and isolate variability in a certain process due to an
assignable cause from that due to a common cause (normal process fluctuations). In
such instances, the aberrant result could be investigated to correct the root cause and
improve the process stability and efficiency. Thus ensure compliance with GMP.
Laney U΄ and I-MR charts have the same control limits, but the later is more sensitive
in detecting out-of-control states - including early warning of process shift. On the
other hand, Laney U΄ chart is easier to interpret and does not require special
prerequisites before implementation.
Conflicts of Interest
The author declares that there is no conflict of interest regarding the publication of
this article.
References
[1] Clontz, L. Microbial limit and bioburden tests, 2nd ed.; CRC Taylor & Francis:
Boca Raton, USA, 2009; ISBN: 97814200534941420053493.
[2] Number of Drug Recalls Surges at FDA, Led by Mid-Level Concerns RAPS.
Available online: http://www.raps.org/Regulatory-
Focus/News/2014/08/11/20005/Number-of-Drug-Recalls-Surges-at-FDA-Led-
by-Mid-Level-Concerns/# (accessed on 16 March 2017).
[3] Recalls. Available online: http://healthworldnet.com/link-directory/top-4-
more/drugs/recalls.html (accessed on 16 March 2017).
[4] Pharmaguy's Insights Into Drug Industry News. Available online:
http://www.scoop.it/t/pharmaguy-s-take-on-drug-industry-news/?tag=Drug+
Recalls (accessed on 16 March 2017).
[5] Drug recalls could hit record high in 2014. Available online:
http://munley.com/drug-recalls-hit-record-high-2014/ (accessed on 16 March
2017).
[6] Montgomery, D.C. Introduction to Statistical Quality Control, 6th ed.; John
Wiley & Sons: New York, NY, USA, 2009; ISBN 978-0-470-16992-6.
[7] Shah, S.; Shridhar, P.; Gohil, D. Control chart: A statistical process control tool
in pharmacy. Asian. J. Pharm. 2014, 4(3), 184-191, DOI:
http://dx.doi.org/10.22377/ajp.v4i3.144. Available online:
https://www.asiapharmaceutics.info/index.php/ajp/article/view/144/227
(accessed on 5 June 2018).
[8] Brochmann‐Hanssen, E.; Medina, J.C. Dosage variation in tablets. Journal of
pharmaceutical sciences, 1963, 52(7), 630-3. DOI:
https://doi.org/10.1002/jps.2600520704.
Available online: https://jpharmsci.org/article/S0022-3549(15)34022-3/pdf
(accessed on 5 June 2018).
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 42-44 www.itspoa.com/journal/em
[9] Chowdhury, M.R. Process capability analysis in pharmaceutical production.
International Journal of Pharmaceutical and Life Sciences, 2013, 2(2), 85-9.
Available online: http://www.ijlsbd.com/010505.pdf (accessed on 5 June 2018).
[10] Cleophas, T.J.; Zwinderman, A.H. Machine Learning in Medicine - a Complete
Overview, 1st ed.; Springer International Publishing: Switzerland, 2015; ISBN:
978-3-319-15195-3.
[11] I-MR Chart. Available online: http://sixsigmacharts.blogspot.com.eg/2010/02/
understand-i-mr-chart.html (accessed on 12 June 2017).
[12] Laney, D.B. Improved control charts for attributes. Quality Engineering, 2002,
14(4), 531-7, DOI: https://DOI.ORG/10.1081/QEN-120003555. Available online:
https://www.tandfonline.com/doi/abs/10.1081/QEN-120003555 (accessed on 5
June 2018).
[13] Eissa, M.E. Novel rapid method in ecological risk assessment of air-borne
bacteria in pharmaceutical facility. Mahidol Univ. J. Pharm. Sci. 2016,
43(3),115-26, DOI: https://doi.org/10.14456/mujps.2016.14. Available online:
http://www.pharmacy.mahidol.ac.th/journal/journalabstract.php?jvol=43&jpart=
3&jconnum=2 (accessed on 05 June 018).
[14] Veronin, M.A.; Nutan, M.T.; Dodla, U.K. Quantification of active
pharmaceutical ingredient and impurities in sildenafil citrate obtained from the
Internet. Therapeutic advances in drug safety, 2014, 5(5), 180-9, DOI:
https://doi.org/10.1177/2042098614543091. Available online:
http://journals.sagepub.com/doi/abs/10.1177/2042098614543091(accessed on 5
June 2018).
[15] Henson, E. Introduction to cGMPSampling: The Basics. Journal of GXP
Compliance, 2003, 7(4), 68-83.
[16] Distribution Fitting BPI Consulting. Available online:
https://www.spcforexcel.com/knowledge/basic-statistics/distribution-fitting
(accessed on 12 June 2017).
[17] Eissa, M. Shewhart Control Chart in Microbiological Quality Control of Purified
Water and its Use in Quantitative Risk Evaluation. UK Journal of
Pharmaceutical Biosciences, 2016, 4(1), 45-51, DOI: 10.20510/ukjpb/4/i1/87845.
Available online: http://www.ukjpb.com/article_details.php?id=158 (accessed on
05 June 018).
[18] Eissa, M.; Abdoh, A. Evaluation of Quality Characteristics and Process Stability
For Pharmaceutical Dosage form Using Attribute Control Charts. I.J.A.M.S. 2016,
1(1), 9-15. Available online:
http://ijams.kibanresearchpublications.com/index.php/IJAMS/article/download/1/
5 (accessed on 12 June 2017).
[19] Eissa, M.; Seif, M.; Fares, M. Assessment of purified water quality in
pharmaceutical facility using six sigma tools. Int. J. Qual. Assur. 2015, 6(2), 54-
72.
[20] Ready for Prime Time: Use P' and U' Charts to Avoid False Alarms. Available
online: http://blog.minitab.com/blog/understanding-statistics/ready-for-prime-
time:-use-p-and-u-charts-to-avoid-false-alarms (accessed on 12 June 2017)
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 43-44 www.itspoa.com/journal/em
[21] SigmaXL. How Do I Create P' Charts (Laney) in Excel Using SigmaXL?
Available online:
http://file:///E:/Local%20Disk/Projects%20for%20Mostafa/Six%20Sigma/Micro
biology%20Liquid/SigmaXL%20_%20Product%20Features%20_%20Create%2
0P'%20Charts%20(Laney)%20in%20Excel%20Using%20SigmaXL.html
(accessed on 12 March 2017).
[22] Eissa, M.E. Application of Control Charts in QualityCharacteristics Evaluation of
Microbiological Media. J. Adv. Res. Pharm. Sci. Pharmacol. Interv. 2016,
1(1&2), 1-13. Available online:
https://medical.adrpublications.in/index.php/Journal-Pharmaceutical-
Sciences/article/view/801/772 (accessed on 12 June 2017).
[23] On the Charts: A Conversation with David Laney - Minitab. Available online:
https://www.minitab.com/en-us/Published-Articles/On-the-Charts--A-
Conversation-with-David-Laney/ (accessed on 17 March 2017).
[24] Attribute Control Charts Overview BPI Consulting. Available online:
https://www.spcforexcel.com/knowledge/attribute-control-charts/attribute-
control-charts-overview (accessed on 17 March 2017)
[25] Individuals Control Charts BPI Consulting. Available online:
https://www.spcforexcel.com/knowledge/variable-control-charts/individuals-
control-charts (accessed on 17 March 2017).
[26] Crossley, M.L. The desk reference of statistical quality methods, 2nd ed.; ASQ
Quality Press: Milwaukee, Wis., USA, 2007; ISBN-13: 978-0873897259.
[27] Elseviers, M. STATISTICS CORNER: THE BOX PLOT: An alternative way to
present a distribution of observations. EDTNA-ERCA Journal, 2004, 30(2), 114-6.
DOI: https://doi.org/10.1111/j.1755-6686.2004.tb00345.x. Available online:
https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1755-6686.2004.tb00345.x
(accessed on 05 June 018).
[28] Typical Histogram Shapes and What They Mean - ASQ. Available online:
http://asq.org/learn-about-quality/data-collection-analysis-tools/overview/
histogram2.html (accessed on 17 March 2017).
[29] Tague, N. The quality toolbox, 1st ed.; ASQ Quality Press: Milwaukee, Wis.,
USA, 2005, ISBN 0-87389-639-4.
[30] A Simple Guide to Between/Within Capability Minitab. Available online:
http://blog.minitab.com/blog/applying-statistics-in-quality-projects/a-simple-
guide-to-between-within-capability (accessed on 17 March 2017).
[31] Interpret the capability plot in Capability Sixpack - Minitab. Available online:
http://support.minitab.com/en-us/minitab/17/topic-library/quality-tools/capability
-analyses/capability-graphs/interpret-the-capability-plot-in-capability-sixpack/
(accessed on 17 March 2017).
[32] Henderson, G.R, Six sigma quality improvement with minitab, 1st ed.; Wiley:
Hoboken, N.J., USA, 2011, ISBN: 978-0-470-74175-7.
[33] Newton, I. Minitab cookbook, 1st ed.; Packt Pub.: Birmingham, UK, 2014, ISBN
978-1-78217-092-1.
VOLUME 1, 2018
DOI: 10.31058/j.em.2018.11003
Submitted to Experimental Medicine, page 44-44 www.itspoa.com/journal/em
[34] Bass, I. Six sigma statistics with Excel and Minitab, 1st ed.; McGraw-Hill: USA,
2007, ISBN-13: 978-0071489690.
[35] Fuzzy Approach to Statistical Control Charts. Available online:
https://www.hindawi.com/journals/jam/2013/745153/ (accessed on 17 March
2017).
[36] Quality Control Charts. Available online:
http://www.uta.edu/faculty/sawasthi/Statistics/stquacon.html (accessed on 17
March 2017).
© 2017 by the author(s); licensee International Technology and
Science Publications (ITS), this work for open access publication is
under the Creative Commons Attribution International License (CC
BY 4.0). (http://creativecommons.org/licenses/by/4.0/)