ARTICLES

Planning Risk-Based Statistical Quality Control

Strategies: Graphical Tools to Support the New
Clinical and Laboratory Standards Institute
C24-Ed4 Guidance

Hassan Bayat,1 Sten A. Westgard,2 and James O. Westgard2,3*

Background: Clinical and Laboratory Standards Institute (CLSI)'s new guideline for statistical quality control
(SQC; C24-Ed4) (CLSI C24-Ed4, 2016; Parvin CA, 2017) recommends the implementation of risk-based SQC strate-
gies. Important changes from earlier editions include alignment of principles and concepts with the general patient
risk model in CLSI EP23A (CLSI EP23A, 2011) and a recommendation for optimizing the frequency of SQC (number
of patients included in a run, or run size) on the basis of the expected number of unreliable final patient results. The
guideline outlines a planning process for risk-based SQC strategies and describes 2 applications for examination
procedures that provide 9-σ and 4-σ quality. A serious limitation is that there are no practical tools to help
laboratories verify the results of these examples or perform their own applications.
Methods: Power curves that characterize the rejection characteristics of SQC procedures were used to predict
the risk of erroneous patient results based on Parvin's MaxE(Nuf) parameter (Clin Chem 2008;54:2049 –54). Run
size was calculated from MaxE(Nuf) and related to the probability of error detection for the critical systematic error
(Pedc).
Results: A plot of run size vs Pedc was prepared to provide a simple nomogram for estimating run size for
common single-rule and multirule SQC procedures with Ns of 2 and 4.
Conclusions: The “traditional” SQC selection process that uses power function graphs to select control rules and
the number of control measurements can be extended to determine SQC frequency by use of a run size nomo-
gram. Such practical tools are needed for planning risk-based SQC strategies.

IMPACT STATEMENT
The new fourth edition of the Clinical and Laboratory Standards Institute (CLSI) guideline for statistical
quality control (SQC) focuses on the application of risk-based SQC strategies but lacks practical planning or
design tools. The methodology outlined here will help medical laboratories select SQC strategies that limit
the risk of erroneous patient test results. Simple graphical tools—power function graphs and run size
nomograms—make it practical for laboratories to select appropriate control rules, the total number of control
measurements/event, and the number of patient samples between quality control events (or SQC frequency).

1
Sina Medical Laboratory, Qaem Shahr, Iran; 2Westgard QC, Inc., Madison, WI; 3University of Wisconsin School of Public Health, Madison, WI.
*Address correspondence to this author at: Westgard QC, 7614 Gray Fox Trail Madison, WI 53717. Fax 608-833-0640;
e-mail james@westgard.com.

..............................................................................................................

September 2017 | 02:02 | 211–221 | JALM 211

ARTICLES Planning Risk-Based SQC Strategies

The fourth edition of the Clinical and Laboratory CLSI EP23. C24 does not recommend a specific QC
Standards Institute (CLSI)4 guideline on statistical strategy for any individual device or technology.
quality control (SQC)was published in 2016 (CLSI Likewise, while a number of the QC performance
C24-Ed4) (1, 2) and was recently discussed in this metrics discussed in the document require com-
journal (2). Earlier editions were published in 1991, puter software to compute, the guideline neither
1999, and 2006 and have a long history of use in makes recommendations nor gives examples of
medical laboratories. The new edition introduces the use of any specific software”.
several important changes: Our purpose here is to consider the practical
• “Alignment of principles and definitions to be details of implementing the roadmap for selecting
consistent with and to supplement the general control rules, number of control measurements,
patient risk model described in CLSI document and number of patient samples between QC
EP23 (3); events (run size, SQC frequency) for an SQC strat-
egy. Our approach builds on the C24-Ed4 guidance
• Introduction of additional performance mea- that recommends the use of Sigma-metrics for
sures useful for evaluating the performance
characterizing the quality of an examination proce-
characteristics of a quality control (QC) strategy;
dure (4), power functions for characterizing the
• A greater focus on QC frequency and QC sched- performance of SQC control rules and number of
ules as a critical part of a QC strategy; control measurements (5), and Parvin's MaxE(Nuf)
• Expanded guidance on setting target values parameter for optimizing SQC frequency (6). The
and SDs for QC materials; recommended planning process extends an ear-
lier approach using a Sigma-metric SQC Selection
• A substantial chapter on recovering from an Tool (7) based on calculation of a process capability
out-of-control condition.” index for the critical systematic error that must be
The first 3 changes are concerned with the plan- detected to maintain a defined quality require-
ning of a risk-based SQC strategy, which is defined ment (8). This “traditional” planning process has
as “the number of QC materials to measure, the been adapted to include run size by considering
number of QC results and the QC rule to use at the relationship between the probability of detect-
each QC event, and the frequency of QC events,” ing the critical systematic error (Pedc) and Parvin's
more commonly called an SQC procedure. The last MaxE(Nuf) patient risk parameter, as described by
2 relate to the proper application and implemen- Yago and Alcover (9) for single-rule SQC proce-
tation of SQC procedures in medical laboratories. dures. Bayat (10) has evaluated those relationships
According to Parvin (2), the intention of the guid- for multirule SQC procedures to support more
ance is to provide principles and definitions rather widespread applications for planning risk-based
than a specific approach, performance metrics, or SQC procedures. Although Yago and Alcover and
software tools: “[T]he objective…was to provide a Bayat provide nomograms that can be used to se-
helpful roadmap for designing, assessing, and im- lect SQC procedures based on the relationship be-
plementing a statistical QC strategy that is consis- tween the observed Sigma-metric vs MaxE(Nuf),
tent with the patient risk concepts introduced in those nomograms do not consider the probability

DOI: 10.1373/jalm.2017.023192
© 2017 American Association for Clinical Chemistry
4
Nonstandard abbreviations: CLSI, Clinical and Laboratory Standards Institute; SQC, statistical quality control; QC, quality control; Pedc, proba-
bility of error detection for critical systematic error; MaxE(Nuf), maximum number of unreliable final patient results before an out-of-control error
condition is detected; ΔSEcrit, critical systematic error that needs to be detected to maintain a defined ATE quality requirement; ATE or TEa, allowable
total error; Pfr, probability of false rejection.

...............................................................................................................

212 JALM | 211–221 | 02:02 | September 2017

Planning Risk-Based SQC Strategies ARTICLES

Fig. 1. Power function graph for 2 levels of controls.

Probability of rejection is plotted on the y axis vs the size of a medically important systematic error (ΔSEcrit) on the lower x axis
and vs the Sigma-metric on the upper x axis. Power curves (top to bottom) correspond to the control rules and total number
of control measurements/event (top to bottom) in the key at the right. Pfr in the key corresponds to the y intercept of the
power curve. Ped corresponds to the probability of error detection that is assigned once additional planning parameters have
been specified, e.g., the specification for the ATE and the bias and imprecision observed for the measurement procedure. N
corresponds to the total number of control measurements/event made on the same or different levels of controls. R
corresponds to the number of runs in which the control rules are applied, which is 1 when all the rules can be applied in an
individual run, but it may be higher if rules require more control measurements than available in a single run.

for false rejection, which is an important design individual test result will exceed the defined ATE.
parameter incorporated in the traditional design The x axis of a power function graph can also be
approach that uses power function graphs. scaled directly in terms of a Sigma-metric, which is
calculated as (ATE − |bias|)/SD when concentra-
METHODS AND MATERIALS tion units are used or (%ATE − |%bias|)/%CV for
percentage units; therefore, the Sigma-metric cor-
Performance characteristics of SQC responds to ΔSEcrit + 1.65 (4).
procedures Fig. 1 shows a set of power curves that are com-
SQC procedures have been traditionally se- monly of interest when 2 levels of controls are an-
lected through the evaluation of power function alyzed. The curves in these figures correspond
graphs (8, 11) that describe the probability of rejec- from top to bottom with the list of SQC rules and
tion on the y axis as a function of the size of the number of control measurements (N) shown (top
error (ΔSE for systematic error) on the x axis. The to bottom) in the key at the right side of the graph.
size of the systematic error that is critical for detec- Typically, the critical systematic error (ΔSEcrit) is cal-
tion, ΔSEcrit, can be calculated as [(ATE − bias)/SD] − culated and a goal of 0.90 is set for the probability
1.65, in which 1.65 corresponds to a 5% risk that an of error detection (Ped) and a goal of ≤0.05 (as low

..............................................................................................................

September 2017 | 02:02 | 211–221 | JALM 213

ARTICLES Planning Risk-Based SQC Strategies

as possible) for the probability of false rejection calculated from Excel spreadsheets, and those re-
(Pfr) to assess the suitability of different control sults were used to prepare the nomogram. Run
rules and different numbers of control measure- size was calculated as 100/MaxE(Nuf) in accor-
ments. Yago and Alcover (9) use the term Pedc to dance with Parvin's model, in which QC events
represent the probability of detecting the critical bracket 100 patient samples (i.e., M = 100). Such
systematic error. nomograms can be developed whenever com-
plete power curves are available for the candidate
Risk prediction for SQC procedures SQC procedures of interest.

The risk of patient harm is related to the errone-

ous test results that are reported when analyzing Approach for planning risk-based SQC
Procedures
patient samples. C24-Ed4 describes this as “the
expected number of unreliable final patient re- 1. Define the quality required for intended use in
sults”. This expression is based on Parvin's defini- the form of an ATE.
tion of a parameter MaxE(Nuf) that represents the 2. Determine the precision (SD, CV) and trueness
“maximum expected increase in the number of un- (bias) of the examination procedure from experi-
acceptable patient results reported during the ex- mental data.
istence of an undetected out-of-control error 3. Calculate the Sigma-metric as (ATE − bias)/SD for
condition” (6). In short, these are increased defec- concentration units or (%ATE − %bias)/%CV for
tive results that may be reported, although con- percentage units.
trols are being analyzed. The number of defects 4. Assess the probability for false rejection (Pfr)
depends on the quality required for intended use from the y intercept of the power curves and the
(ATE, TEa), precision and bias of the examination probability of error detection (Ped) from the inter-
procedure, control rules and number of control section of the power curves and the observed
measurements being used for SQC, and the num- σ-metric or critical-sized systematic error.
ber of patient samples in the analytical run (run 5. Select control rules and the total number of con-
size, frequency of SQC). The calculation of the trol measurements to achieve a probability of er-
MaxE(Nuf) risk parameter is complicated and re- ror detection (Pedc) of ≥0.90 and a probability of
quires informatics support (12). However, a graphical false rejection (Pfr) as low as possible.
approach may be used to approximate MaxE(Nuf) 6. Convert the observed Pedc to the maximum
based on the relationship between MaxE(Nuf) and run size (or frequency of SQC) using a run size
Pedc as described for single-rule SQC procedures nomogram.
by Yago and Alcover (9) and for multirule SQC pro- 7. Consider other practical factors that will affect
cedures by Bayat (10). Then it is a simple matter to the frequency of SQC, and shorten the run size, as
calculate run size and prepare a run size nomo- necessary, to provide both an effective and effi-
gram for different SQC procedures. cient quality management process.

Run size nomogram

RESULTS
To develop the nomogram, the example condi-
tions represented an HbA1c examination proce- Fig. 2 provides a run size nomogram for single-
dure in which ATE was 6.0%, CV was 1.0%, and bias and multirule SQC procedures with 2 and 4 control
varied from 0.0% to 3.5% to change the Sigma- measurements per event (N, total number across
metric from 6.0 to 2.5. Pedc and MaxE(Nuf) were levels of controls). The x axis describes the probability

...............................................................................................................

214 JALM | 211–221 | 02:02 | September 2017

Planning Risk-Based SQC Strategies ARTICLES

Fig. 2. Run size on y axis is plotted vs probability of error detection for the critical systematic error (Pedc)
on the x axis.
Run size is calculated as 100/MaxE(Nuf) and represents the number of patient samples between 2 QC events. Lines, top to
bottom, are identified in the key at the right. SR2 refers to 13s with n = 2; MR2 is 13s/22s/R4s with n = 2; SR4 is 13s with n = 4; MR4
is 13s/22s/R4s/41s with n = 4.

of detection for the critical-sized systematic error Note that for a MaxE(Nuf) goal of 1.00, which cor-
(Pedc), which would be determined for the selected responds to a run size of 100 under the conditions
SQC rules and N/event from a power function graph. of the risk model, these common single- and mul-
The y axis shows run size, which is the number of tirule procedures provide a Pedc of 0.76 – 0.86;
patient samples between QC events, calculated as thus, any SQC procedure that is designed to
run size = 100/MaxE(Nuf). The logarithmic scale achieve a Pedc of ≥0.90 will achieve Parvin's goal for
spreads the data around Parvin's suggested Max- low patient risk. A higher Pedc will allow laboratories
E(Nuf) goal of 1.0, which corresponds to a run size of to increase the run size, e.g., a Pedc of 0.90 would
100. The different lines correspond to different SQC correspond to maximum run sizes from approxi-
procedures, as shown in the key at the right side of mately 150 –300 patient samples, depending on
the figure. The highest line is for 13s with n = 2 (SR2), the particular SQC procedure selected. In contrast,
the next lower line for 13s/22s/R4s with n = 2 (MR2), a lower Pedc will lead to smaller run sizes, e.g., a Pedc
then 13s with n = 4 (SR4), and finally (lowest line) 13s/ of 0.60 leads to run sizes of about ≤40 and a Pedc
22s/R4s/41s with n = 4 (MR4). <0.50 leads to very short run sizes of about ≤25.

..............................................................................................................

September 2017 | 02:02 | 211–221 | JALM 215

ARTICLES Planning Risk-Based SQC Strategies

DISCUSSION dures. Another alternative, shown here, is to

extend the traditional SQC design approach and
The frequency of SQC became a major issue determine the frequency of SQC using a run size
in laboratory practice when the Centers for nomogram that relates the number of patient
Medicare and Medicaid Services published the samples between QC events to Pedc, the probabil-
Final CLIA Rule in 2003 (13) and introduced equiv- ity of error detection for the critical systematic er-
alent quality control as an option for compliance in ror that needs to be detected by an SQC
the Interpretive Guidelines found in the Centers procedure. Pedc is an SQC planning parameter that
for Medicare and Medicaid Services' State Opera- has been used for many years in the traditional
tions Manual. Equivalent quality control (or EQC, as approach that uses power function graphs for se-
it was known) allowed a laboratory to perform cer- lecting control rules and the total number of con-
tain validation experiments and then reduce the trol measurements/event (15).
frequency of SQC to once per week or even once An application is illustrated in Fig. 3 for an exam-
per month. However, those experiments were not ination procedure having 4-σ quality (or 2.35s crit-
scientifically valid; for example, a stability study ical systematic error). The traditional goal has been
over 10 days was used to justify reducing SQC fre- to achieve a Pedc of 0.90 with a Pfr as low as possi-
quency to once every 30 days. That shortcoming ble. Pfr is evaluated from the y intercepts of the
(along with others) eventually led the Centers for power curves, and the estimates are shown in the
Medicare and Medicaid Services to replace equiv- first column of the key at the right of the figure. Pedc
alent quality control in January 2016 with risk- is evaluated at the intersection of the perpendicu-
based individualized QC plans (now known as lar line and the power curves for the different SQC
IQCP). That change in regulations makes the guid- procedures, and estimates are shown in the second
ance from the new edition of the CLSI C24-Ed4 column of the figure key. A 13s/22s/R4s multirule SQC
document critical for SQC practices today. procedure having 2 controls/event would provide
Key guidance in C24-Ed4 is to define the frequency a Pedc of 0.59, whereas a Pedc of 0.91 could be
of SQC on the basis of the number of patient sam- achieved with a 13s/22s/R4s/41s multirule proce-
ples analyzed between 2 QC events, i.e., bracketed dure with 4 controls/event. Fig. 4 shows how to use
QC. A QC event is the term used for “the occurrence of the run size nomogram to determine the appropri-
one or more QC measurements and a QC rule eval- ate number of patient samples between QC
uation using the QC results”. For bracketed QC oper- events. The maximum run size is determined to be
ation, the number of patient samples, or frequency approximately 40 patient samples for the n = 2
of SQC, is supposed to be determined by the risk of multirule procedure and approximately 170 pa-
harm to patients if erroneous results are reported, tient samples for the n = 4 multirule procedure.
which can be estimated by calculating Parvin's Max- This application corresponds to the 4-σ example
E(Nuf) parameter (6). The practical problem for labo- that appears in the C24-Ed4 document, in which
ratories is that this calculation is complicated and the recommendation is “a candidate strategy us-
requires specialized informatics support (14). ing 13s, 22s, 41s, and R4s together with two QC
As an alternative, Yago and Alcover (9) provided concentrations at every QC event” (1, p. 44). It is
a rule selection nomogram that relates MaxE(Nuf) not clear whether N should be 2 or 4, although
to the observed Sigma-metric of an examination inclusion of the 41s rule suggests that N be 4. A run
procedure and the performance of different size of 125 is recommended but does not corre-
single-rule SQC procedures. Bayat (10) has ex- spond to either of the estimates above. The run
tended that approach for multirule SQC proce- size of 125 seems to come from the earlier

...............................................................................................................

216 JALM | 211–221 | 02:02 | September 2017

Planning Risk-Based SQC Strategies ARTICLES

Fig. 3. Determination of Pedc for an example examination procedure having a Sigma-metric of 4.0 or a
critical SE of 2.35*SD.
Power curves (top to bottom) correspond to the control rules and total number of control measurements/event (top to
bottom) in the key at the right. Pfr in the key that corresponds to the y intercept of the power curve; Ped corresponds to the
intersection of the vertical line with the power curve. The multirule SQC procedure with n = 2 shows Pedc of 0.59, and the
multirule SQC procedure with n = 4 shows Pedc of 0.91.

proposed edition of the C24 document that rec- Examination procedures with higher Sigma-met-
ommended a 13s/2 of 32s/R4s/31s multirule with 3 rics (better quality) would permit simpler SQC proce-
levels of controls. Thus, the control rules and N dures (single rules, lower N) and allow larger
have changed in the final document, but the run maximum run sizes, e.g., Pedc could approach 1.00
size remains the same. Some clarification of this for a 6-σ process for a 13s n = 2 SQC procedure, or a
example was provided in December 2016 when 13s/22s/R4s SQC procedure with n = 2, both of which
CLSI issued an editorial omission that corrected would allow maximum run sizes of at least 250 pa-
the SQC recommendation to read “A candidate tient samples. This assessment is consistent with the
strategy is using 13s, 22s, and R4s rules together 9-σ example shown in C24-Ed4, in which a run size of
with two QC concentrations at every QC event” 200 is recommended. Other practical factors may, of
(16). That correction not only removes the ambigu- course, impose smaller run sizes and must be care-
ity about the control rules and number of control fully considered when defining the final SQC strategy.
measurements/event that are recommended (n = It is also apparent that examination procedures
2), but also makes it clear that the recommended with low Sigma quality cannot be adequately con-
run length of 125 is wrong or, at best, arbitrary trolled by SQC procedures to minimize patient risk.
rather than objective. Use of a multirule SQC Industrial guidelines suggest that processes with
procedure with 4 control measurements/event lower than 3-σ quality are not suitable for routine
would be more appropriate and would justify a service because they cannot be adequately con-
run size of 125 patient samples. trolled. An essential part of risk management

..............................................................................................................

September 2017 | 02:02 | 211–221 | JALM 217

ARTICLES Planning Risk-Based SQC Strategies

Fig. 4. SQC planning application for an examination procedure having 4-σ quality.
For a 13s/22s/R4s n = 2 multirule SQC procedure (MR2, second from top line), Pedc is 0.59 and the corresponding run size is
approximately 40 patient samples. For a 13s/22s/R4s/41s n = 4 multirule procedure (MR4, bottom line), Pedc is 0.91 and the run
size may be as large as approximately 170 patient samples.

should be the validation of safety characteristics, SQC rules. SQC rules are essentially statistical tests
such as precision and bias, to ensure they satisfy of significance whose performance characteristics
the requirements for intended use (e.g., ATE); thus, can be determined on the basis of the probability
the selection and validation of methods are critical theory for simple single-rule procedures. For com-
in a medical laboratory, and the prerequisite to binations of rules, computer simulations have
implementation should be of a quality better than been performed to describe the probabilities for
3-σ. Methods having low Sigma quality will require rejection under different error conditions (5) and
well-trained operators, rigorous adherence to a have been available in the literature for decades
manufacturer's directions for use and preventive (17). C24-Ed4 recommends some new SQC rules—
measures, frequent control, short run lengths, and 81s, 61s, 101s—whose power curves have not been
thorough corrective actions with an emphasis on explicitly documented. These rules are described
elimination of error sources and failure modes. as having been empirically evaluated (17) based on
A prerequisite for these graphical tools, as well a set of glucose data that is included as Fig. 2A in
as the calculation of MaxE(Nuf), is the availability of C24-Ed4. According to Miller and Nichols (18), such
information about the rejection characteristics of empirical validation is the basis for a recommendation

...............................................................................................................

218 JALM | 211–221 | 02:02 | September 2017

Planning Risk-Based SQC Strategies ARTICLES

Fig. 5. Sigma-metric run size nomogram, in which run size is plotted on the y axis vs the observed
Sigma-metric on the x axis.
Maximum false rejection probability is 0.03 or 3% for the included SQC procedures. The different lines, from left to right,
correspond to 13s/22s/R4s/41s with n = 4 (MR4, Pfr = 0.03), 13s with n = 4 (SR4, Pfr = 0.01), 13s/22s/R4s with n = 2 (MR2, Pfr = 0.01),
and a 13s with n = 2 (SR2, Pfr = 0.00).

of a 13s/22.5s/R4s/81.5s multirule procedure “based as shown by its power curve in Fig. 1. That obser-
on the clinical requirements for patient care, the vation suggests that the introduction of new con-
observed long-term method performance, and the trol rules should require proper performance
need to identify potential method issues with a characterization of their power curves, both as sin-
false alert rate ≤1%”. Note the introduction of 2 gle rules and as particular combinations recom-
additional new control rules—22.5s and 81.5s. Inter- mended for multirule procedures.
estingly, the set of glucose data shows an SD of It is also possible to develop nomograms that
about 4 mg/dL at a concentration of 273 mg/dL, or relate run size directly to the observed Sigma-
a CV of about 1.5%. Given the CLIA criterion of 10% metric by substituting run size for MaxE(Nuf) in the
for acceptable performance in proficiency testing graphical relationships shown by Yago and Alcover
and assuming no bias, the Sigma-metric would be (9) and Bayat (10). Such nomograms are simpler in
10%/1.5% or 6.7. Such a method could be ade- principle because they eliminate the need for de-
quately controlled by a 13s control rule with n = 2, termining Pedc; however, they risk overlooking the

..............................................................................................................

September 2017 | 02:02 | 211–221 | JALM 219

ARTICLES Planning Risk-Based SQC Strategies

false rejection characteristic of the SQC proce- International standards, such as ISO 15189 (19),
dure. The prerequisite for such run length vs typically provide high-level guidance that sets re-
Sigma-metric nomograms should be the elimina- quirements for what to achieve without describing
tion of high Pfr procedures, as per the example how to do it. For example, ISO 15189 requires that
shown in Fig. 5, where all the SQC procedures have “the laboratory shall design quality control proce-
Pfr values of ≤0.03 or 3%. For a 4-σ process, similar
dures that verify the attainment of the intended qual-
run lengths can be determined, e.g., approxi-
ity of results,” but does not describe how to
mately 40 for the n = 2 multirule procedure and
accomplish this. CLSI documents typically fill the gap
approximately 180 for the n = 4 multirule proce-
dure. Ultimately, such Sigma-metric run size no- between the “what to achieve” and “how to do it,” but
mograms are simpler and may be preferred, but that is not the case for C24-Ed4. The recommended
analysts should appreciate the underlying require- SQC planning approach cannot be implemented
ment for power curves for the SQC procedures. solely on the basis of the guidance provided in the
In conclusion, there are serious limitations with the document. Simple graphical tools, such as Yago and
guidance from the new CLSI C24-Ed4 document. The Alcover's MaxE(Nuf) nomogram for single-rule SQC
lack of practical tools to support the planning of risk- procedures (9) and Bayat's MaxE(Nuf) nomogram for
based SQC procedures, as well as the lack of perfor- multirule SQC procedures (10), offer alternatives, but
mance characteristics for some new control rules
they may be limited because the probability of false
that are proposed, is a problem. Other issues, such
rejection is not included as a planning parameter.
as the practicality of bracketed QC for continuous
That limitation can be overcome by extending the
operation and result reporting, may also be prob-
lematic for laboratory practice. Even if the guidance is traditional SQC planning approach based on power
intended to only emphasize principles and provide a curves and adding run size nomograms to deter-
roadmap for risk-based SQC strategies, it leaves lab- mine SQC frequency, as described here. We hope
oratories without sufficient direction to implement and expect others will also develop new tools to sup-
the recommended practices. port the planning of risk-based SQC strategies.

Additional Content on this Topic

Selecting Statistical Quality Control Procedures for Limiting the Impact of Increases in
Analytical Random Error on Patient Safety
Martín Yago. Clin Chem 2017;63:1022–30
What's New in Laboratory Statistical Quality Control Guidance? The 4th Edition of CLSI C24,
Statistical Quality Control for Quantitative Measurement Procedures: Principles and Definitions
Curtis A. Parvin. J Appl Lab Med 2017;1:581– 4

...............................................................................................................

220 JALM | 211–221 | 02:02 | September 2017

Planning Risk-Based SQC Strategies ARTICLES

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following
4 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b)
drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for
all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately
investigated and resolved.

Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure
form. Employment or Leadership: S.A. Westgard and J.O. Westgard, Westgard QC, Inc. Consultant or Advisory Role: S.A.
Westgard, Abbott Diagnostics. Stock Ownership: J.O. Westgard, Westgard QC, Inc. Honoraria: S.A. Westgard, ThermoFisher
Diagnostics. Research Funding: None declared. Expert Testimony: None declared. Patents: None declared.

Role of Sponsor: No sponsor was declared.

REFERENCES
1. CLSI C24-Ed4. Statistical quality control for quantitative based on patient risk. Clin Chem Lab Med. [Epub ahead
measurement procedures: principles and definitions. 4th of print] Clin Chem Lab Med February 25, 2017 as doi:
Ed. Wayne (PA): Clinical and Laboratory Standards 10.1515/cclm-2016-1077.
Institute; 2016. 11. Westgard JO. Internal quality control: planning and
2. Parvin CA. What's new in laboratory statistical quality implementation strategies. Ann Clin Biochem 2003;40:
control guidance? The 4th edition of CLSI C24, statistical 593– 611.
quality control for quantitative measurement 12. Yundt-Pacheco J, Parvin CA. Validating the performance
procedures: principles and definitions. J Appl Lab Med of QC procedures. Clin Lab Med 2013;33:75– 88.
2017;1:581– 4. 13. Centers for Disease Control and Prevention (CDC) (2)
3. CLSI C23-A. Laboratory quality control based on risk Centers for Medicare & Medicaid Services (CMS), HHS.
management. Wayne (PA): Clinical and Laboratory Medicare, Medicaid, and CLIA programs: laboratory
Standards Institute; 2011. requirements relating to quality systems and certain
4. Westgard JO, Westgard SA. Quality control review: personnel qualifications. Final rule. Fed Regist 2003;16:
implementing a scientifically based quality control 3639 –714.
system. Ann Clin Biochem 2016;53:32–50. 14. Parvin CA, Yundt-Pacheco J. System and method for
5. Westgard JO, Groth T. Power functions for statistical analyzing a QC strategy for releasing results. US patent
control rules. Clin Chem 1979;25:863–9. 8,938,409 B2, 2015 Jan 20.
6. Parvin CA. Assessing the impact of the frequency of 15. Westgard JO. Six sigma quality design and control. 2nd
quality control testing on the quality of reported patient Ed. Madison (WI): Westgard QC, Inc.; 2006.
results. Clin Chem 2008;54:2049 –54. 16. Adams JK. Editorial omission in subchapters 8.2 and 8.4.
7. CLSI C24A3. Statistical quality control for quantitative CLSI letter of December 2016 addressed to recipients of
measurement procedures: principles and definitions. C24, 4th Ed.
3rd Ed. Wayne (PA): Clinical and Laboratory Standards 17. Westgard JO, Barry PL. Cost-effective quality control:
Institute; 2006. managing the quality and productivity of analytical
8. Koch DD, Oryall JJ, Quam EF, Feldbruegge DH, Dowd DE, processes. Washington (DC): AACC Press; 1986. p. 195–
Barry PL, Westgard JO. Selection of medically useful 217.
quality-control procedures for individual tests done in a 18. Miller WG, Nichols JH. Quality control. In: Clark W, editor.
multi-test analytical system. Clin Chem 1990;36:230 –3. Contemporary practice in clinical chemistry. Washington
9. Yago M, Alcover S. Selecting statistical procedures for (DC): AACC Press; 2016. p. 47– 62.
quality control planning based on risk management. Clin 19. ISO 15189. Medical laboratories—requirements for
Chem 2016;62:959 – 65. quality and competence. Geneva, Switzerland: ISO;
10. Bayat H. Selecting multi-rule quality control procedures 2012.

..............................................................................................................

September 2017 | 02:02 | 211–221 | JALM 221