ASTM F3172-15(2021)

Standard Guide for Design Verification Device Size and Sample Size Selection for Endovascular Devices

SIGNIFICANCE AND USE
4.1 The purpose of this guide is to provide guidance for selecting appropriate device size(s) and determining appropriate sample size(s) for design verification of endovascular devices. The device size(s) and sample size(s) for each design input requirement should be determined before testing. The device size(s) selected for verification testing should establish that the entire device matrix is able to achieve the design input requirements. If testing is not performed on all device sizes, justification should be provided.  
4.2 The sample size justification and statistical procedures used to analyze the data should be based on sound scientific principles and should be suitable for reaching a justifiable conclusion. Insufficient sample size may lead to erroneous conclusions more often than desired.  
4.3 Guidance regarding methodologies for determining device size selection and appropriate sample size is provided in Sections 5 and 6.
SCOPE
1.1 This guide provides guidance for selecting an appropriate device size(s) and determining an appropriate sample size(s) (that is, number of samples) for design verification testing of endovascular devices. A methodology is presented to determine which device size(s) should be selected for testing to verify the device design adequately for each design input requirement (that is, test characteristic). Additionally, different statistical approaches are presented and discussed to help guide the developer to determine and justify sample size(s) for the design input requirement being verified. Alternate methodologies for determining device size selection and sample size selection may be acceptable for design verification.  
1.2 This guide applies to physical design verification testing. This guide addresses in-vitro testing; in-vivo/animal studies are outside the scope of this guide. This guide does not directly address design validation; however, the methodologies presented may be applicable to in-vitro design validation testing. Guidance for sampling related to computational simulation (for example, sensitivity analysis and tolerance analysis) is not provided. Guidance for using models, such as design of experiments (DOE), for design verification testing is not provided. This guide does not address sampling across multiple manufacturing lots as this is typically done as process validation. Special considerations are to be given to certain tests such as fatigue (see Practice E739) and shelf-life testing (see Section 8).  
1.3 Regulatory guidance may exist for endovascular devices that should be considered for design verification device size and sample size selection.  
1.4 Units—The values stated in SI units are to be regarded as the standard. No other units of measurement are included in this standard.  
1.5 This standard does not purport to address all of the safety concerns, if any, associated with its use. It is the responsibility of the user of this standard to establish appropriate safety, health, and environmental practices and determine the applicability of regulatory limitations prior to use.  
1.6 This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for the Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.

General Information

• Status: Published
• Publication Date: 31-Jul-2021
• ICS: 11.040.40 - Implants for surgery, prosthetics and orthotics
• Technical Committee: F04 - Medical and Surgical Materials and Devices
• Drafting Committee: F04.30 - Cardiovascular Standards

Relations

• Refers: ASTM F2914-12(2024) - Standard Guide for Identification of Shelf-Life Test Attributes for Endovascular Devices
• Effective Date: 15-Jan-2024
• Refers: ASTM F2914-12(2018) - Standard Guide for Identification of Shelf-life Test Attributes for Endovascular Devices
• Effective Date: 01-Nov-2018
• Refers: ASTM F2914-12 - Standard Guide for Identification of Shelf-life Test Attributes for Endovascular Devices
• Effective Date: 15-Jan-2012
• Refers: ASTM E739-10 - Standard Practice for Statistical Analysis of Linear or Linearized Stress-Life (<bdit>S-N</bdit>) and Strain-Life (<bdit>&#949;-N</bdit>) Fatigue Data
• Effective Date: 01-Nov-2010
• Refers: ASTM E739-91(2004)e1 - Standard Practice for Statistical Analysis of Linear or Linearized Stress-Life (<bdit>S-N</bdit>) and Strain-Life (<bdit>&#949;-N</bdit>) Fatigue Data
• Effective Date: 01-May-2006
• Refers: ASTM E739-91(2004) - Standard Practice for Statistical Analysis of Linear or Linearized Stress-Life (<i>S-N</i>) and Strain-Life (&#949;-<i>N</i>) Fatigue Data
• Effective Date: 01-May-2004
• Refers: ASTM E739-91(1998) - Standard Practice for Statistical Analysis of Linear or Linearized Stress-Life (S-N) and Strain-Life (e-N) Fatigue Data
• Effective Date: 10-Apr-1998

Overview

ASTM F3172-15(2021) is an internationally recognized standard guide that provides best practices for selecting appropriate device sizes and determining suitable sample sizes during the design verification process for endovascular devices. Published by ASTM International, this guide focuses on in-vitro physical design verification testing, ensuring that products meet required design input specifications before market release. The guide is essential for manufacturers, quality engineers, and regulatory professionals involved with the development and approval of vascular medical devices.

Proper selection of device size(s) and sample size(s) during design verification is crucial for demonstrating that an entire device family, often called the device matrix, fulfills the specified design input requirements such as strength, durability, and performance. If not all device sizes are tested, the guide requires a robust justification, and the statistical methods chosen must be able to yield scientifically sound, defensible results.

---

Key Topics

• Device Size Selection
  The standard offers detailed methodologies for selecting which device sizes within a product family to subject to design verification testing. Key considerations include:

  • Identifying essential design outputs that influence test results
  • Defining design subgroups where critical features are constant
  • Addressing scenarios where testing each subgroup, only the worst-case scenario, or a representative subset is appropriate

• Sample Size Determination
  The guide discusses statistical approaches for establishing the minimum number of samples necessary for each test:

  • Basing sample size on scientific and risk-based principles
  • Considering confidence and reliability requirements tied to clinical risk
  • Selecting either attribute (pass/fail) or variable (measured value) data strategies

• Justification and Documentation
  Manufacturers must justify their rationale for both device size and sample size selection, provide statistical analysis, and document all choices and methodologies in compliance with regulatory standards.

• Scope and Limitations
  The guide applies strictly to physical, in-vitro verification. It does not address in-vivo, animal, computational, or multi-lot sampling. Guidance for specialized tests like fatigue (ASTM E739) or shelf-life is referenced but not central to this standard.

---

Applications

ASTM F3172-15(2021) is directly applicable to:

• Endovascular Device Development: Ensuring that angioplasty balloons, stents, catheters, and similar products are properly verified for performance across all intended configurations.
• Regulatory Submissions: Supporting robust justification of design verification test plans for pre-market approval or clearance. The standard helps demonstrate scientific rigor in sample and device size selection, meeting FDA and international requirements.
• Quality Assurance: Aligning verification strategies with quality management systems, reducing risk of erroneous conclusions, and improving confidence in product performance before clinical use.
• Test Methodology Development: Guiding the development of in-vitro testing protocols that align with the variability of the device matrix, ensuring comprehensive coverage.

By following ASTM F3172-15(2021), manufacturers help ensure patient safety, product robustness, and regulatory compliance, while potentially streamlining the path to market.

---

Related Standards

Implementing ASTM F3172-15(2021) often involves cross-referencing other key standards, such as:

• ASTM E739: Practice for Statistical Analysis of Linear or Linearized Stress-Life (S-N) and Strain-Life (ε-N) Fatigue Data - relevant for fatigue testing protocols.
• ASTM F2914: Guide for Identification of Shelf-life Test Attributes for Endovascular Devices.
• ISO 14971: Medical devices - Application of risk management to medical devices, which informs the risk-based approach to sample size selection.
• FDA Guidance Documents: Especially those for endovascular and cardiovascular devices, which may specify expectations for verification and validation plans.

For global device manufacturers, aligning with these and other standards ensures conformity with international best practices for design verification of endovascular medical devices.

---

Keywords: ASTM F3172-15(2021), endovascular device verification, sample size selection, device size selection, medical device standards, in-vitro testing, regulatory compliance, design verification, quality assurance, device matrix, risk management.

Standards Content (Sample)

This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for the
Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.
Designation:F3172 −15 (Reapproved 2021)
Standard Guide for
Design Verification Device Size and Sample Size Selection
for Endovascular Devices
This standard is issued under the fixed designation F3172; the number immediately following the designation indicates the year of
original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A
superscript epsilon (´) indicates an editorial change since the last revision or reapproval.
1. Scope responsibility of the user of this standard to establish appro-
priate safety, health, and environmental practices and deter-
1.1 This guide provides guidance for selecting an appropri-
mine the applicability of regulatory limitations prior to use.
ate device size(s) and determining an appropriate sample
1.6 This international standard was developed in accor-
size(s) (that is, number of samples) for design verification
dance with internationally recognized principles on standard-
testing of endovascular devices.Amethodology is presented to
ization established in the Decision on Principles for the
determinewhichdevicesize(s)shouldbeselectedfortestingto
Development of International Standards, Guides and Recom-
verify the device design adequately for each design input
mendations issued by the World Trade Organization Technical
requirement (that is, test characteristic).Additionally, different
Barriers to Trade (TBT) Committee.
statisticalapproachesarepresentedanddiscussedtohelpguide
the developer to determine and justify sample size(s) for the
2. Referenced Documents
design input requirement being verified. Alternate methodolo-
2.1 ASTM Standards:
gies for determining device size selection and sample size
E739 PracticeforStatisticalAnalysisofLinearorLinearized
selection may be acceptable for design verification.
Stress-Life (S-N) and Strain-Life (ε-N) Fatigue Data
1.2 This guide applies to physical design verification test-
F2914 Guide for Identification of Shelf-life Test Attributes
ing. This guide addresses in-vitro testing; in-vivo/animal stud-
for Endovascular Devices
ies are outside the scope of this guide. This guide does not
2.2 ISO Standards:
directly address design validation; however, the methodologies
ISO 14971:2012 Medical devices—Application of risk man-
presented may be applicable to in-vitro design validation
agement to medical devices
testing. Guidance for sampling related to computational simu-
lation (for example, sensitivity analysis and tolerance analysis) 3. Terminology
is not provided. Guidance for using models, such as design of
3.1 Definitions:
experiments (DOE), for design verification testing is not
3.1.1 attribute data, n—data that identify the presence or
provided.Thisguidedoesnotaddresssamplingacrossmultiple
absenceofacharacteristic(forexample,good/badorpass/fail).
manufacturing lots as this is typically done as process valida-
3.1.2 design input requirements, n—physical and perfor-
tion.Specialconsiderationsaretobegiventocertaintestssuch
mance requirements of a device that are used as a basis for
asfatigue(seePracticeE739)andshelf-lifetesting(seeSection
device design (typically defined as test characteristics such as
8).
balloon burst pressure, shaft tensile strength, and so forth).
1.3 Regulatoryguidancemayexistforendovasculardevices
3.1.3 design output, n—features of the device (that is,
that should be considered for design verification device size
dimensions, materials, and so forth) that define the design and
and sample size selection.
make it capable of achieving design input requirements.
1.4 Units—The values stated in SI units are to be regarded
3.1.4 design subgroup, n—set defined by the device sizes
as the standard. No other units of measurement are included in
within the device matrix in which the essential design outputs
this standard.
do not vary for a specified design input requirement (that is,
1.5 This standard does not purport to address all of the
device sizes that share the same design for a specified design
safety concerns, if any, associated with its use. It is the
input requirement).
1 2
This guide is under the jurisdiction of ASTM Committee F04 on Medical and For referenced ASTM standards, visit the ASTM website, www.astm.org, or
Surgical Materials and Devices and is the direct responsibility of Subcommittee contact ASTM Customer Service at [e-mail]. For Annual Book of ASTM
F04.30 on Cardiovascular Standards. Standards volume information, refer to the standard’s Document Summary page on
Current edition approved Aug. 1, 2021. Published August 2021. Originally the ASTM website.
approved in 2015. Last previous edition approved in 2015 as F3172 – 15. DOI: Available fromAmerican National Standards Institute (ANSI), 25 W. 43rd St.,
10.1520/F3172-15R21. 4th Floor, New York, NY 10036, http://www.ansi.org.
Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States
F3172−15 (2021)
3.1.5 design validation, n—establishing by objective evi- Testing the same device size(s) is typically not appropriate to
dence that the device conforms to defined user needs and verify all design input requirements. Differences in the device
intended use(s). design throughout the device matrix will drive which device
size(s) is selected for verification of each design input require-
3.1.6 design verification, n—confirmation by examination
ment.
and provision of objective evidence that the device design
5.1.1 As explained in subsequent sections, when determin-
(design output) fulfills the specified requirements (design
ing device size(s) for testing, the following should be consid-
input).
ered for each design input requirement:
3.1.7 device matrix, n—entire range of available models/
5.1.1.1 Essential design outputs,
sizes for the device family.
5.1.1.2 Design subgroups, and
3.1.8 device size, n—individual model/size (for example,
5.1.1.3 Other considerations.
6 mm diameter by 25 mm length balloon on 135 cm length
5.2 Define Essential Design Outputs (EDOs)—The design
catheter or a 6Fr 100 cm length guide catheter).
outputs of the device are the features of the device (that is,
3.1.9 endovascular device, n—device used to treat vascular
dimensions, materials, and so forth) that define the design and
conditions from within the vessel.
makeitcapableofachievingdesigninputrequirements.Notall
3.1.10 essential design output, EDO, n—designfeature(s)or design outputs are essential for each design input requirement.
characteristic(s) of the device that affects its ability to achieve
Therefore, for each design input requirement, the essential
the design input requirements (that is, design output(s) that has design outputs (EDOs) should be identified. In Table 1,
a relevant effect on the test results).
example EDOs for design input requirements of a balloon
catheter device are provided.
3.1.11 process validation, n—establishment by objective
evidence that a process consistently produces a result or device
5.3 Define Design Subgroups:
achieving its predetermined requirements.
5.3.1 The design subgroups should be defined for each
design input requirement based on the EDOs identified.
3.1.12 safety factor, n—ratio of the device performance to
5.3.2 For a specific design input requirement, the design
the specification requirement (for example, how much stronger
subgroups can be defined as one of the following:
the device is than it needs to be to meet its specification
5.3.2.1 The entire device matrix if the EDOs for the design
requirement).
input requirement are constant throughout the entire device
3.1.13 sample size, n—quantityofindividualspecimensofa
matrix,
device tested.
5.3.2.2 Subsets of the device matrix if the EDOs for the
3.1.14 variables data, n—data that measure the numerical
design input requirement vary in groups or stages throughout
magnitude of a characteristic (how good/how bad).
the device matrix, or
5.3.2.3 Each individual device size of the device matrix if
4. Significance and Use
EDOs for the design input requirement are different for each
4.1 The purpose of this guide is to provide guidance for
individual device size.
selecting appropriate device size(s) and determining appropri-
5.3.3 Fig. 1 represents the device matrix (entire range of
ate sample size(s) for design verification of endovascular
available device sizes) for a 135 cm length balloon catheter
devices. The device size(s) and sample size(s) for each design
device that has balloon diameters ranging from 3 to 7 mm and
input requirement should be determined before testing. The
balloon lengths ranging from 10 to 50 mm. Balloon catheters
device size(s) selected for verification testing should establish
are available in any combination of balloon diameter and
that the entire device matrix is able to achieve the design input
length resulting in 25 unique device sizes in the device matrix.
requirements. If testing is not performed on all device sizes,
5.3.4 Figs. 2-4 illustrate how the device matrix in Fig. 1 is
justification should be provided.
defined by different design subgroups for different design input
4.2 The sample size justification and statistical procedures
used to analyze the data should be based on sound scientific
TABLE 1 Example EDOs for Design Input Requirements for a
principles and should be suitable for reaching a justifiable
Balloon Catheter Device
conclusion. Insufficient sample size may lead to erroneous
Design Input Requirement EDOs
conclusions more often than desired.
Manifold connection/Luer lockability Luer thread dimensions
4.3 Guidance regarding methodologies for determining de-
Manifold material
vice size selection and appropriate sample size is provided in
Catheter shaft tensile strength for a Shaft material
Sections 5 and 6.
single lumen catheter Shaft cross-sectional area
(diameter and wall thickness)
5. Selection of Device Size(s)
Shaft bond design
5.1 Design input requirements are the physical and perfor-
Balloon compliance (diameter versus Balloon diameter
mance requirements of a device that are used as a basis for
pressure) Balloon material
device design. Once the device design is defined, testing is Balloon wall thickness
typicallyperformedtoverifythatthedesigninputrequirements
Balloon deflation time Balloon volume
are met. The appropriate device size(s) for verification testing
Shaft deflation lumen design
should be determined for each design input requirement.
F3172−15 (2021)
FIG. 1Device Matrix for a Balloon Catheter Device (25 unique device sizes)
FIG. 2Design Subgroup for Manifold Connection/Luer Lockability Testing (EDOs remain constant throughout the device matrix)
requirements. Fig. 2 represents a design subgroup that is subgroup that is defined by the device sizes that have shaft
defined by the entire device matrix because all device sizes design “B.” Fig. 4 represents design subgroups for balloon
share the same design for the specified design input require- compliance in which each balloon diameter defines a unique
ment (that is, the EDOs remain constant for all device sizes). design subgroup.
The design input requirement is manifold connection/luer
5.4 Design Input Requirements and Other
lockability testing, and the EDOs (luer thread dimensions and
Considerations—In addition to design subgroup definition,
manifold material) are the same for all sizes in the device
design input, device labeling, or regulatory requirements may
matrix.
make it necessary to test additional sizes.
5.3.5 Figs. 3 and 4 represent design subgroups that are
5.5 Device Size Selection Approach:
subsets of the device matrix because the EDOs for the design
5.5.1 Approach—Once the design subgroups are defined for
input requirement vary throughout the device matrix. Fig. 3
agivendesigninputrequirement,thedevicesize(s)tobetested
represents design subgroups for shaft tensile strength for a
for design verification testing can be appropriately selected by
device that contains two different shaft designs in the device
using one of the following approaches:
matrix, but the other EDOs that were identified (shaft material
5.5.1.1 Test each design subgroup,
and shaft bond design) are the same for the entire device
5.5.1.2 Test the worst-case design subgroup, or
matrix. Therefore, there is a design subgroup that is defined by
the device sizes that have shaft design “A” and a design 5.5.1.3 Test a subset of the design subgroups.
F3172−15 (2021)
FIG. 3Design Subgroups for Shaft Tensile (EDOs vary throughout the device matrix but are constant within each design subgroup)
FIG. 4Design Subgroups for Balloon Compliance (EDOs vary throughout the device matrix but are constant within each design sub-
group)
5.5.2 Test Each Design Subgroup: represents the entire device matrix, factors such as device sizes
5.5.2.1 Depending on the design subgroup definition, test-
used for other testing to minimize total test units or device size
ing each design subgroup may translate into testing one device
with the highest sales volume may be considered.
size or multiple device sizes to verify the entire device matrix.
5.5.2.3 Whenthedesignsubgroupsaredefinedbysubsetsof
5.5.2.2 When the design subgroup is defined by the entire
the device matrix, a device size should be selected from within
device matrix and the requirement is the same throughout the
each design subgroup to verify the design adequately since
device matrix, any device size may be selected for verification
EDOs vary throughout the device matrix. Fig. 6 illustrates the
testing to represent the entire device matrix. This approach is
design subgroups and example device sizes selected for veri-
appropriate since all device sizes share the same design for the
fication testing for shaft tensile strength. Note that the shaft
specified design input requirement (that is, the EDOs are the
tensile strength requirement is the same for all device sizes and
sameforalldevicesizes).Fig.5illustratesthedesignsubgroup
the other EDOs identified (shaft material and shaft bond
and example device size selection for verification testing for
design) are the same for all device sizes.
manifold connection/luer lockability. Since any device size
F3172−15 (2021)
FIG. 5Example Design Subgroup and Verification Device Size Selection for Manifold/Luer Lockability Testing
FIG. 6Example Design Subgroups and Verification Device Size Selection for Shaft Tensile Strength
FIG. 7Worst-Case Size May Be Selected Based on a Safety Factor Calculation
5.5.2.4 An alternate approach to selecting one device size to by considering how the EDOs impact performance to the
representeachdesignsubgroupwouldbetopoolmultiplesizes designinputrequirements.Ifthedesigninputrequirementlimit
within a design subgroup for testing. Refer to Section 7 for varies throughout the device matrix (for example, different
more information on data pooling. rated burst pressure (RBP) requirements for different diameter
5.5.3 Test the Worst-Case Design Subgroup: balloon catheters), a worst case could be tested for each
5.5.3.1 For certain design input requirements, testing only specification limit or one worst-case subgroup could be tested
the worst-case design subgroup adequately verifies the entire by performing a worst-case analysis that accounts for the
device matrix. The worst-case design subgroup is determined differences in the specification limits, such as a safety factor
F3172−15 (2021)
FIG. 8Example Design Subgroups to Consider for Balloon Deflation Time
FIG. 9Xs Represent Worst Case (Largest Balloon Volume) Within Each Shaft Design Subgroup (these are the device sizes selected to
verify that the entire device matrix can achieve the design input requirement)
calculation. Additionally, if the design input requirement has (1) Historical data (similar predicate device or develop-
both an upper and a lower specification limit, there may be a ment characterization of current device) or
worstcasefortheupperspecificationandadifferentworstcase (2) Engineering judgment, analysis, computational
for the lower specification. simulation, or safety factor calculation.
5.5.3.2 Testing the worst-case design subgroup is a com-
NOTE1—Whiletheengineeringorcomputationalanalysis,orboth,may
monly used verification method when EDOs vary throughout
be applied to determine the worst-case size selection, additional consid-
the device matrix and their impact to the design input perfor-
erations that could impact which device size to test may exist. For
mance is well understood/defined (for example, increasing
example, manufacturing process variations between device sizes could
result in an actual worst-case device size that is different than the
diameter has a negative impact on achieving the design input
theoretical worst case. Additionally, the assembly of a multi-component
requirement and decreasing the diameter has a positive impact
device could result in failures that would not be predicted by an
on achieving the design input requirement).
engineering analysis applied to only one component of the device. Use of
5.5.3.3 The worst-case design subgroup may be determined
historical knowledge of failures can be used to justify whether these
by one of the following methods: factors should be considered in the device size selection. The following
F3172−15 (2021)
FIG. 10A 2-by-2 Factorial May Be Selected for Evaluation When One Device Size Does Not Represent the Entire Device Matrix or a
Worst-Case Device Size is Not Known
are a couple examples of types of analysis to determine worst case:
throughout the device matrix; therefore, there are multiple
design subgroups that should be considered when selecting the
(a) Hoop stress calculation—The highest balloon hoop
devicesize(s)fortesting.Fig.8illustratesthedesignsubgroups
stress may represent the worst-case situation for balloon burst
to consider for balloon deflation time testing (two different
testingwhenitisknownthatthefinisheddevicealwaysfailsin
shaft design subgroups and 25 different balloon volume design
the balloon.
subgroups).
(i) By using a thin-walled pressure vessel assumption,
5.5.3.5 Since the relationship between deflation time and
thehoopstressofacylindricalballooncouldbecalculatedby:
balloon volume for a constant shaft design is well understood
P*D
Hoop Stress 5 (1) (that is, the larger the balloon volume, the longer the deflation
2* T
~ !
time), a worst-case approach can be used to verify each shaft
where: design subgroup. Fig. 9 illustrates that the worst-case balloon
volume device is selected within each shaft design to verify
P = pressure (rated burst pressure (RBP) design input
that the entire device matrix has acceptable deflation times.
requirement),
5.5.3.6 Note that if the design input requirement for defla-
D = diameter (EDO), and
tion time is not the same for all balloon sizes, then additional
T = wall thickness (EDO).
sizes may need to be tested to verify the worst case for each
(ii) By using the rated burst specification requirement for
specification requirement.
P in the hoop stress formula, the worst-case size (that is, the
size with the highest hoop stress at rated burst pressure) can be 5.5.3.7 Other examples of tests that may rely on the worst-
case device size rationale for selecting the device sizes for
calculated.
(b) Fatigue safety factor calculation—Appropriately vali- testing are the following: accelerated durability, particulate
generation, corrosion, and magnetic resonance imaging (MRI)
dated finite element analysis may be used on each implant
diameter or other relevant property (for example, design compatibility. The rationale and device size for each test is
different because each test evaluates a different aspect of
platform, length) to determine the fatigue safety factor as well
as the critical stress and strain values and locations. The device performance.
5.5.4 Test a Subset of the Design Subgroups—For certain
predicted stresses and strains are compared to the fatigue life
line to determine the fatigue safety factor.The implant with the design input requirements, a subset of the design subgroups
may be required for verification testing. This approach may be
lowest fatigue safety factor may be tested as the worst case in
design verification (see Fig. 7 for a stent example). used when EDOs vary throughout the device matrix and a
worst-case device size is not known. For example, a two-by-
5.5.3.4 Balloon deflation time is an example of a design
twofactorial(Fig.10)ofthelargestandsmallestdiametersand
input requirement for which a worst-case design subgroup
lengths may be an approach to device size selection to capture
approach may be acceptable to verify the entire device matrix.
the performance at the corners of the design space.
The EDOs defined for deflation time are balloon volume and
shaft deflation lumen design. For the example in Fig. 8, the
6. Statistical Approaches for Sample Size Determination
balloon volume and the shaft deflation lumen design both vary
6.1 Oncethedevicesize(s)hasbeenselectedforverification
testing per the methodology presented in Section 5, the sample
Hibbele, R. C., Mechanics of Materials, Third Edition, 1997. size needs to be defined. The sample size justification and
F3172−15 (2021)
statistical procedures used to analyze the data are to be based the 99th percentile based on a sample of 30 values. However,
on sound scientific principles and suitable for reaching a thisisonlyapointestimatesincethereisstilluncertaintyinthe
justifiable conclusion. An insufficient sample size may lead to truelocationofthe99thpercentilebecauseoftherandomdraw
erroneous conclusions more often than desired. of the samples. Therefore, an upper confidence limit on the
estimate of the 99th percentile is needed, which will take into
6.2 This section provides an overview of determining sta-
account the sample size and provide a margin of error on the
tistically based sample sizes for the following design verifica-
percentile estimate in the direction of the specification limit(s).
tion test methodologies:
Placing a confidence limit on this percentile estimate creates a
6.2.1 Attribute testing to a predefined specification (pass/
statistical tolerance limit.
fail), and
6.3.1.3 Finally, select a sampling plan that meets the toler-
6.2.2 Sampling by variables for proportion nonconforming.
ance limit requirement. There are usually many different
6.3 If a predefined specification is not initially available, a
sampling plans that can satisfy the same tolerance limit
comparison to a predicate or similar device may be performed
requirement.Assumingthetruepopulationcharacteristicmeets
to justify acceptability. It is recommended that before design
thepercenttolerancelimit,theselectionofthesamplingplanis
verification, the predicate or similar device should be charac-
a trade-off between efficiency (lower sample size) and likeli-
terizedandaspecificationlimitshouldbedefinedbasedonthat
hood of passing (higher sample size). Therefore, the best
characterization. Once the specification limit is defined, either
practice is to select a sampling plan that efficiently provides
the attribute or variable testing approaches can be used to
evidence that the product meets tolerance requirements and is
verify the design. In Table 2, some of the considerations/
unlikely to give false conclusions.
limitations of the approaches discussed in this guide are
6.4 Attribute Testing:
summarized.
6.4.1 Description—When data are assessed as attribute,
6.3.1 Risk and Sample Size—Selection of a sample size is a
eachunitreturnsaresultofpassorfail.Thedatacollectedmay
three-step process:
be binary data (for example, successful or not successful) or
6.3.1.1 First,determinetherisklevelbasedontheperceived
variable data individually assessed against the criteria (for
risktothepatientasaresultofthefailureofthespecificdesign
example, is the measured value greater than the specified
input requirement. Risk level may be defined as the combina-
limit). This section provides guidance on how to determine a
tion of the severity outcome of the device failure and the
sample size for attribute data to establish a desired confidence
likelihood of that failure happening (occurrence). For addi-
and reliability level.
tional detail regarding how to determine the risk level, refer to
6.4.2 Sam
...