You are on page 1of 13

Journal of Medical Engineering & Technology

ISSN: 0309-1902 (Print) 1464-522X (Online) Journal homepage: https://www.tandfonline.com/loi/ijmt20

OQ & PQ protocols development for medical


device two-level process validation

Yeong-Lin Chen

To cite this article: Yeong-Lin Chen (2019) OQ & PQ protocols development for medical device
two-level process validation, Journal of Medical Engineering & Technology, 43:7, 431-442, DOI:
10.1080/03091902.2019.1692939

To link to this article: https://doi.org/10.1080/03091902.2019.1692939

Published online: 28 Nov 2019.

Submit your article to this journal

Article views: 11

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=ijmt20
JOURNAL OF MEDICAL ENGINEERING & TECHNOLOGY
2019, VOL. 43, NO. 7, 431–442
https://doi.org/10.1080/03091902.2019.1692939

RESEARCH ARTICLE

OQ & PQ protocols development for medical device two-level


process validation
Yeong-Lin Chen
Validation Department, SHL Group, Taoyuan, Taiwan

ABSTRACT ARTICLE HISTORY


Medical device industry encompasses a wide range of technologies and applications, which Received 1 August 2019
makes the process validation approach (IQ/OQ/PQ framework) and OQ/PQ implementation Revised 7 November 2019
method difficult to be defined in the industry’s regulation and process validation guidance Accepted 8 November 2019
documents. Based on two-level (component production level and device assembly level) process
KEYWORDS
validation approach, this article proposes 1) a typical section structure & contents for production Medical device process
equipment/system/process validation OQ/PQ protocols, 2) a typical subsection structure & con- validation; two-level process
tents for the protocols’ most critical section - Test Plan, and 3) the validation approach and OQ validation; Equipment
Detailed Test Plan for both common equipment and test equipment. Validation IQ/OQ/PQ;
Process Validation PQ;
detailed test plan

1. Introduction level process validation and final device assembly level


process validation. For each level’s process validation,
In medical device industry, process validation is critical
it should have Installation Qualification (IQ),
to the regulatory compliance and product quality
Operational Qualification (OQ), and Performance
assurance. A medical device can be an instrument,
apparatus, appliance, software, implant, reagent, Qualification (PQ) as the validation framework, which
material or other articles, and from syringes and can be further divided into Equipment Validation IQ,
wheelchairs to cardiac pacemakers and medical imag- OQ, PQ and Process Validation PQ. Equipment valid-
ing technologies (e.g., Magnetic Resonance Imaging ation PQ is an option depending on equipment qual-
(MRI), Computed Tomography (CT) and X-ray ity risk, validation financial risk, and machine design
machines). Due to a wide range of technologies and and development activities. Process Validation PQ is
applications, medical device manufacturing could mandatory, which is aimed at the performance test of
involve multiple technical fields, such as chemical, the manufacturing process as a whole under the regu-
mechanical, electronics, software and their combina- lar production environment, and is equivalent to stage
tions. Except the devices with chemical and software 2 - Process Performance Qualification (PPQ) shown in
in nature, device mainstream products (e.g., mechan- the FDA process validation guidance [2]. This two-level
ical-based, electromechanical-based) are manufactured process validation approach is the base for the subject
in two levels (i.e., component production level, device article to develop OQ, PQ protocols.
assembly level) based on their manufacturing order Process validation for medical device manufacturing
practice. According to the published paper “Two-level is considered a challenge task since 1) regulatorily, the
Process Validation Approach for Medical Devices” [1], medical device Global Harmonisation Task Force
it mentioned that the process validation (or stage 2 - (GHTF) process validation guidance (GHTF/SG3/N99-10,
Process Qualification shown in Food and Drug Quality Management Systems - Process Validation
Administration (FDA) Guidance for Industry, Process Guidance [3]) only provided general suggestions on
Validation: General Principles and Practices [2], or trad- how to execute process validation for an individual
itional process validation) approach should align with production equipment/process, and it did not illustrate
the manufacturing order practice and be divided into how to execute process validation for a finished
two levels including device component production device product manufacturing which involves an

CONTACT Yeong-Lin Chen ylchenshl283@gmail.com Validation Department, SHL Group, No. 38, Lane 51, Jianguo Road, Lungtan, Taoyuan
325, Taiwan
ß 2019 Informa UK Limited, trading as Taylor & Francis Group
432 Y.-L. CHEN

integrated production system. A device integrated  Title & Signature Page


production system could include several individual Show the document title and approval record with
production systems, such as plastic component manu- role (e.g., author, reviewer, approver), title, name,
facturing system, metal component manufacturing sys- signature, and date.
tem, printing system, and device assembly system, 2)  Table of Contents
technically, it involves multiple components produc- Show the protocol section title and page number.
tion and device assembly performance verifications.  Purpose
To remedy the insufficiency of GHTF process valid- – Equipment (including component production level
ation guidance, the paper “Two-level Process and device assembly level) Validation OQ & PQ
Validation Approach for Medical Devices” provides a Describe what validation, what equipment, and
clear validation approach on how to establish the what medical device are aimed at, and also the
process validation IQ, OQ, PQ framework for a med- prerequisite validation (e.g., Equipment Validation
ical device’s integrated production system. As for the IQ) should be completed before starting the
medical device process validation technical difficulty protocol execution. In addition, Equipment
involving how to implement the IQ, OQ, PQ exercises Validation OQ purpose (e.g., documented verifica-
for both device component production level and tion that each individual component/feature of
device assembly level, the most difficult thing is to the equipment and the equipment under the
develop suitable and justifiable process validation IQ, specified operating parameter range (or process
OQ, PQ protocols. For any type of equipment (e.g., window) can operate according to their predeter-
test equipment, production equipment, facilities, util- mined requirements and quality characteristics),
ities), IQ verification items are very similar, and thus and Equipment Validation PQ purpose (e.g.,
IQ protocol development generally is not considered documented verification that the equipment
a challenge for process validation practitioners. under the normal (or nominal) operating condi-
However, for Equipment Validation OQ, PQ and tion can consistently produce a product meeting
Process Validation PQ, the validation verification the predetermined requirements and quality
items/methods vary widely, also their validation exe- characteristics) should be stated.
cutions are much more complex and critical. For the
Note: Another important OQ purpose is to iden-
reasons, the subject article is aimed at how to
tify critical process parameters and optimise their
develop appropriate OQ, PQ protocols by providing a
operating parameter ranges for worst case test-
typical section structure & contents for the protocols.
ing. Generally, the parameter screening and opti-
Detailed Test Plans, describing the test procedures
misation are conducted via Design of Experiment
for the performance verification of device production
(DOE) study, which does not have predetermined
equipment/system/process, are deemed an essential
acceptance criteria. Since predetermined accept-
part of the OQ, PQ protocols. Therefore, a typical
ance criteria are indispensable for process valid-
subsection structure & contents for the Detailed Test
ation, DOE study should not be categorised as an
Plan are described and analysed. In addition, the val-
event of process validation. Therefore, it is pro-
idation approach and OQ Detailed Test Plan for com-
posed that DOE study had better be separated
mon equipment and test equipment are illustrated
from Equipment Validation OQ documents.
as well.
– Process Validation PQ for Component
Production Level
2. Typical section structure and contents for Describe what validation, what component pro-
OQ and PQ protocols duction process, and what medical device are
aimed at, and also the prerequisite equipment
In order to simplify the process validation protocol validation (e.g., Equipment Validation IQ/OQ)
preparation, the main structures (or main section should be completed before starting the protocol
titles) of OQ and PQ protocols can be kept the same. execution. In addition, Process Validation PQ pur-
A typical section structure and contents for OQ and pose (e.g., documented verification that the com-
PQ protocols, based on the two-level process valid- ponent production process under the normal
ation approach (including Equipment Validation OQ & operating condition and regular production envir-
PQ for both levels, Component Production Level onment can consistently produce a component
Process Validation PQ, and Device Assembly Level product meeting the predetermined requirements
Process Validation PQ), are proposed below. and quality characteristics) should be stated.
JOURNAL OF MEDICAL ENGINEERING & TECHNOLOGY 433

– Process Validation PQ for Device Assembly Level  System Description


Describe what validation, what device assembly – Equipment Validation OQ & PQ
process (e.g., device assembly lines, plastic Describe the equipment (e.g., equipment name,
moulding toolsets), and what medical device are equipment number, equipment model number,
aimed at. In addition, the prerequisite component equipment serial number, equipment type (e.g.,
production level validation (e.g., plastic compo- production, testing), supplier, equipment dia-
nents, metal components), device assembly level grams/pictures) and equipment operating pro-
equipment validation, and device test equipment cess (e.g., operating process flow chart, flow
validation should be completed before starting chart description, product illustrate).
the protocol execution. Further, Process – Process Validation PQ for Component
Validation PQ purpose (e.g., documented verifica- Production Level
tion that the device assembly process under the Describe the component production process
normal operating condition and regular produc- used equipment (e.g., equipment name, equip-
tion environment can consistently produce a ment number) and component production pro-
device meeting the predetermined requirements cess (e.g., production process flow chart, flow
and quality characteristics) should be stated. chart description, product illustrate).
 Scope – Process Validation PQ for Device Assembly Level
– Equipment Validation OQ & PQ Describe the device assembly used components
Describe the validation activity applies to what (e.g., part name, part number, moulding tool
equipment, and follows what standard operating number, part drawing number/revision), device
procedure (SOP), validation master plan, valid- assembly line number and equipment (e.g.,
ation plan, etc. In addition, Equipment Validation equipment name, equipment number), and
OQ test items, and Equipment Validation PQ test device assembly process (e.g., assembly process
items & how many test batches (e.g., three con- flow chart, flow chart description, prod-
secutive batches) should be stated. uct illustrate).
– Process Validation PQ for Component  References
Production Level – Equipment Validation OQ & PQ
Describe the validation activity applies to what List the validation related documents (e.g., pre-
component production process, and follows what requisite validation protocol and report num-
SOP, validation master plan, validation plan, etc. bers, URS, FS, DS, FMEA, assembly drawing,
In addition, Process Validation PQ test items & equipment operating instruction, QC inspec-
how many test batches should be stated. tion/measurement/test instructions).
– Process Validation PQ for Device Assembly – Process Validation PQ for Component
Level Production Level
Describe the validation activity applies to what List the validation related documents (e.g., pre-
device assembly process, and follows what SOP, requisite component production equipment &
validation master plan, etc. In addition, Process test equipment validation protocol and report
Validation PQ test items & how many test numbers, production equipment operating
batches should be stated. instructions, QC inspection/measurement/test
 Responsibilities instructions).
Describe what documents (e.g., validation protocol/ – Process Validation PQ for Device Assembly
report, User Requirement Specification (URS), Level
Functional Specification (FS), Design Specification (DS), List the validation related documents (e.g., pre-
Failure Mode and Effect Analysis (FMEA), product requisite component production process valid-
drawing, Quality Control (QC) inspection instruction, ation protocol and report numbers, assembly
equipment operating instruction, instrument calibra- equipment and test equipment validation
tion, maintenance instruction) should be prepared, protocol and report numbers, assembly equip-
and what activities for the protocol (e.g., protocol ment operating instructions, QC inspection/
review and approval, protocol implementation) should measurement/test instructions).
be performed, and by what roles (e.g., validation  Test Plan
leader, Quality Assurance (QA) leader, engineering For OQ and PQ validation protocols, test plan (or test
leader, calibration leader) in an organisation. procedures) section is much more important than any
434 Y.-L. CHEN

other section in the protocols, and it describes how Standard Operating Procedure (SOP)/validation
to validate (or qualify) the equipment/system/process. protocol) and their handling during the validation
Depending on the test plan’s complexity and quality should be summarised in the validation report, and
risk, generally two formats (simple & detailed) of test it should include deviation description, investiga-
plan are used, which are detailed below. tion & root cause analysis, corrective action, re-test
results, and conclusion. While necessary, a formal
– Simple Test Plan and separate deviation report should be prepared
Simple Test Plan is in a checklist table form used to facilitate the deviation handling. For abnormal
for simple verification tests. The table form test results (e.g., outlier data, unexpected data trend)
plan includes multiple columns, such as verifica- but still within the specification, the investigation
tion item Identification (ID), verification item extent should be proportional to the incident risk
description, test method description, equipment level, and the investigation results need to be
expected response, whether test results are as reported in the validation report.
expected, reference to URS item number/FS item  Revision History
number/Process-FMEA (P-FMEA) item number, Describe the protocol revision number, effective
verified by/date. This Simple Test Plan is gener- date, prepared by, change control number, and
ally used for equipment’s individual component/ change description.
feature functional verification tests (e.g., safety,
alarm, error proof, interlock feature, parameter
3. Typical subsection structure & contents for
control, computer system, HMI navigation, power
detailed test plan
failure test), and therefore it is more suitable to
be used in Equipment Validation OQ protocol. A Detailed Test Plan is needed while a test item’s test
Some of the activities in the Simple Test Plan procedures are complex and critical, and generally the
can be moved to Equipment Validation IQ proto- test procedures information (e.g., how to conduct the
col while appropriate. test, what are the rationales for setting the test batch
– Detailed Test Plan size and sampling plan, how to express the test
Detailed Test Plan is used for complex and inte- results, how to judge pass/fail for the test results)
gration verification tests (e.g., Equipment needs to be described in details. For some test items,
Validation OQ worst case test, Equipment each of them needs to utilise a separate Detailed Test
Validation PQ normal operating condition test, Plan, so multiple Detailed Test Plans could be needed
Process Validation PQ test for Component in one validation protocol. For other test items, all of
Production Leve, Process Validation PQ test for them could be covered in a single Detailed Test Plan
Device Assembly Level) in validation OQ & PQ if the test information, including raw material, operat-
protocols, and is in a form comprising multiple ing parameter setting, and test batch setting, is the
subsections shown below. same. To simplify a validation protocol, minimising the
 Raw Material & Instrument Detailed Test Plan number could be better. Detailed
 Operating Parameter Setting Test Plan is an essential part of a validation protocol,
 Test Batch Setting and therefore it should be carefully prepared,
 Test Item reviewed, and approved by the validation team mem-
 Inspection & Test Method bers (including QA). A typical subsection structure &
 Sampling Plan and Sample Number contents for Detailed Test Plan is proposed and ana-
Designation lysed below for production equipment/system/process
 Acceptance Criteria validation. As for common equipment & test equip-
 Test results and Statistics ment validation, their validation approach and OQ
Detailed Test Plan is complex and critical, and its con- Detailed Test Plans will be separately described and
tents can vary widely depending on different valid- analysed in the following Section 4.
ation cases, Therefore, it will be described and
analysed separately in the following Section 3.
3.1. Raw material & instrument
 Deviation & Abnormal Results Handling  Subsection Purpose
Any deviations (e.g., non-conformance of test To describe what raw materials are used for the
results, validation execution non-compliant with validation test and what instruments are needed
JOURNAL OF MEDICAL ENGINEERING & TECHNOLOGY 435

for the validation sample inspection/measure- parameter targets optimisation and CPP parameter
ment/test. process windows finding.
 Raw Material  Major Test Cases for Operating Parameter Setting
For instruction purpose, all raw materials (e.g., plas- For Equipment Validation OQ worst case test, both
tic resin, metal wire coil/tube, component, subas- CPP parameter upper limits and lower limits, NCPP
sembly) to be used in OQ/PQ validation exercises parameter settings, and NQPP parameter settings
should be described with raw material name, part (if existing) should be listed. Normal parameter
number (or drawing number/revision), etc. in the (generally the mid-points between the upper and
validation protocols. For traceability and investiga- lower limits of the CPPs) case test is better to be
tion purposes, all raw materials used for the valid- included in the OQ worst case test, but not manda-
ation exercises should be recorded with raw tory. For Equipment Validation PQ and Process
material name, batch number, etc. in the valid- Validation PQ tests, generally only normal parame-
ation reports. ters are listed. Further, during the batch running
 Instrument (or sample build), CPP and NCPP parameters must
For instruction purpose, all instruments (e.g., 3D follow the protocol settings (cannot be changed),
vision measuring system, functional test machine, however, NQPP parameters are allowed to be
Instron test machine, stand-alone balance) to be adjusted while necessary.
used for OQ/PQ validation sample inspection/meas-  Other Test Cases for Operating Parameter Setting
urement/test should be described with instrument Except for Equipment Validation OQ worst case
name, model number, etc. in the validation proto- test, Equipment Validation PQ test, and Process
cols. For traceability and investigation purposes, all Validation PQ test cases, there may be some other
instruments used for the validation exercises test cases that also need to describe operating par-
should be recorded with instrument name, equip- ameter settings. For example, assembly equip-
ment’s inspection sensors for checking correct
ment number, etc. in the validation reports.
assembly positions should be verified for their
functionality/capability (e.g., challenge test, Gage
3.2. Operating parameter setting Repeatability and Reproducibility (R&R) test) with a
Detailed Test Plan in the Equipment Validation OQ
protocol, and then the inspection sensor parameter
 Subsection Purpose settings should be listed in the subsection.
To describe what equipment operating parameters
and their set values should be used for the valid-
ation test. 3.3. Test batch setting
 Operating Parameter (or Controllable Process  Subsection Purpose
Parameter) Classification (as an example) To describe how many test batches (or test runs)
– Critical Process Parameter (CPP): A process par- (e.g., 3 consecutive batches for PQ test, parameter
ameter whose variability has a significant upper limit condition & lower limit condition
impact on a critical quality attribute. batches for OQ worst case test) should be con-
– Non-Critical Process Parameter (NCPP): A pro- ducted, what is the test batch size (e.g., producing
cess parameter whose variability has a minor 1200 pcs/batch for sample build, 4 h running/batch
impact on a critical quality attribute. for sample build), and other important information.
– Non-Quality Process Parameter (NQPP): A pro-  How many PQ Consecutive Test Batches
cess parameter whose variability has no effect Consideration
on critical quality attributes, but has a signifi- According to GHTF process validation guidance,
cant impact on process performance (e.g., yield, the PQ test should be repeated enough times to
cycle time). assure that the results are meaningful and consist-
 Operating Parameter and Range Setting Study ent. Based on FDA 2011 Process Validation
Before setting equipment operating parameters for Guidance: Process Validation Revisited [4], it men-
CPP and NCPP, their values need to be obtained tioned that “despite the pervasive practice of
through a DOE study. This study encompasses three-batch validation, FDA’s position remains that
screening experiments for CPP parameters identifi- there never was a three-run requirement, and the
cation, and response surface studies for both CPP protocol should specify a sufficient number of
436 Y.-L. CHEN

replicate process runs to demonstrate reproducibil- repeatability and long-term stability. The test batch
ity and provide an accurate measure of variability setting should demonstrate the PQ purpose can
among successive runs. Accordingly, the FDA 2011 be achieved.
guidance recommends an approach to process val-  Actions to differentiate Batches
idation that is tailored to and based upon up-front For the OQ worst case test, since the parameter
learning and knowledge about the product and settings will be changed between batches, there is
process rather than simply getting to the goal of no need to describe any actions between OQ worst
three acceptable batches.” In addition, from case test batches (e.g., upper limit, lower limit con-
Pharmaceutical Inspection Co-operation Scheme ditions). For the PQ tests, since the process param-
(PIC/S) process validation guidance eter settings are the same (at normal values)
(Recommendation on validation master plan, instal- between test batches, what actions (e.g., shut
lation and operational qualification, non-sterile pro- down the machines, change the machine opera-
cess validation, cleaning validation [5]), it described tors, use raw material of different batches or differ-
that “it is generally considered acceptable that ent suppliers, implement line clearance) should be
three consecutive batches/runs within the finally taken between batches to differentiate them
agreed parameters, giving product of the desired should be described.
quality would constitute a proper validation of the  Device Assembly Level Process Validation PQ
process.” Therefore, there is no rule of how many Compared with other validation tests, device
PQ batches should be run, however, if there are no assembly level process validation PQ is the most
special reasons that can be justified, at least 3 con- important validation activity since it combines the
secutive batches should be conducted for PQ tests. actual facility, utilities, equipment, and the trained
 Process Validation PQ Test Batch Size Consideration personnel with the commercial manufacturing pro-
For medical devices manufacturing, generally their cess, control procedures, and all components to
productions are not batch-wise operation, and produce simulated commercial batches for final
therefore there is no requirement that “the Process devices, and that will demonstrate the commercial
Validation PQ should be maintained as the same device manufacturing process performs as
batch size as the commercial production” described expected. The PQ test is generally considered
in pharmaceutical process validation guidance costly because it allocates most resources (includ-
documents. For medical device Process Validation ing operators, engineers, all components, assembly
PQ, challenges to the process should include the lines, test instruments/machines) and is deemed
conditions that will be encountered during regular the last validation activity for a medical device pro-
commercial production, which includes human fac- ject. Because the test is for production system as a
tor challenge (e.g., stress, shift change) and envir- whole, except how many PQ batches should be
onment factor challenge (e.g., temperature, conducted, what is the PQ batch size, and how to
humidity). In addition, the batch size setting should differentiate PQ batches, there are other important
consider the following factors: risk assessment of items should be described in the subsection as
the equipment/process/product, how long each well, such as how many assembly pieces should go
machine should operate to test tear and wear of through each assembly line/machine, how to make
equipment, the PQ batch size had better be two- the components/sub-assemblies from different
time larger than the inspection sample size so that moulding toolsets evenly go through the assembly
the spare samples can be used for investigation lines/machines, how many production batches of
purpose while deviation occurs, customer’s require- each component should be used, whether different
ments, etc. suppliers’ components should be used, etc.
 Test Batch Setting Rationale  Training
The rationale for OQ test batch setting may not be Production, QC personnel should be trained for the
mandatory in the protocol, however, for PQ tests related production/test equipment operation, SOPs,
(either Equipment Validation PQ or Process and the test plans of validation protocols.
Validation PQ), it should be mandatory to rational-  Product Disposition in Device Assembly Level
ise the test batch setting (including how many test Process Validation PQ
batches and test batch size), because one of the Only the product (or finished device) from Device
PQ purposes is to test the operation/process Assembly Level Process Validation PQ batches can
JOURNAL OF MEDICAL ENGINEERING & TECHNOLOGY 437

be for commercial use, and therefore in the subsec- inspection, measuring, and test equipment is suit-
tion, it should mention that the disposition of the able for its intended purposes and is capable of
product should be documented in the PQ report. producing valid results.” For medical devices, gen-
erally, the inspection/measurement/test are for
physical property (e.g., dimension, force, time,
3.4. Test item
weight) instead of chemical property (e.g., impurity,
 Subsection Purpose assay). The qualification of physical property meas-
To describe what items (e.g., appearance, dimen- urement is focussed on accuracy and precision.
sion, force, ink permanency, function, assembly Accuracy mainly relies on instrument calibration,
condition) should be inspected, measured and and precision is secured by Gauge R&R study. Both
tested. While there is no separate validation trace- of which should be executed to prove the method/
ability matrix, the test items sourced requirements instrument is suitable for its intended use. The
should be referenced for traceability. term of “test method validation” is more used for
 Test Item Source chemical property measurement (or analytical
Test items are generally sourced from URS, p-FMEA, method) qualification, which includes more test
QC inspection instruction, equipment work instruc- items, such as accuracy, precision, specificity, linear-
tion, process validation SOPs, etc. ity, range, detection limit, etc.
 P-FMEA Source
In the p-FMEA score sheet, the critical risk items
3.6. Sampling plan and sample number
with Severity (SEV)  xx and/or RPN (Risk Priority
designation
Number)  yy should be addressed by the valid-
ation tests.  Subsection Purpose
 Plastic Injection moulding Process Test To describe what acceptance sampling plans and
Generally, it could include the test items of part their sample number designations should be used
dimension, appearance, function, etc. for the validation test.
 Device Assembly Level Equipment Validation OQ  Acceptance Sampling Plan
worst Case Test and PQ Test According to GHTF process validation guidance,
Generally, it could include the test items of sub- Section A1, it mentioned that “Acceptance sam-
assembly dimension, appearance, function, assem- pling plans can be useful in optimising the number
bly condition, etc. of samples to be tested and to demonstrate con-
 Device Assembly Level Process Validation PQ Test formance to specification.” Section A3 described
Generally, it could include the test items of final that “an acceptance sampling plan takes a sample
assembly (or final device) dimension, appearance, of product and uses this sample to make an accept
function, etc. or reject decision.”
 High Degree of Assurance
According to 21 CFR Part 820 - Section 820.75, (a),
3.5. Inspection & test method
“where the results of a process cannot be fully veri-
 Subsection Purpose fied by subsequent inspection and test, the process
To describe what methods (or instructions) should shall be validated with a high degree of assurance.”
be used for each test item’s sample inspection/ Based on FDA Process Validation Guidance: General
measurement/test. Principles and Practices [2], Section IV, C, 3, PPQ
 Method Validation Requirement Protocol, it described that “The sampling plan
Except for simple inspection (e.g., visual), test should include sampling points, number of sam-
methods/instruments should be qualified/validated ples, and the frequency of sampling for each unit
to secure the valid results before being used for operation and attribute. The number of samples
OQ/PQ validation sample inspection/measure- should be adequate to provide sufficient statistical
ment/test. confidence of quality both within a batch and
 Method Validation Approach between batches. The confidence level selected
According to 21 CFR Part 820, Medical Devices - can be based on risk analysis as it relates to the
Quality System Regulation QSR [6], Section 820.72, particular attribute under examination. Sampling
it mentioned that “Inspection, measuring, and test during this stage should be more extensive than is
equipment: Each manufacturer shall ensure that all typical during routine production.” Therefore, a
438 Y.-L. CHEN

high degree of assurance means a relative compari- procedures to ensure that sampling methods are
son. For example, for the same test item, validation adequate for their intended use.” For Equipment
test’s acceptance sampling plan should be stricter Validation OQ/PQ, it could be better if sampling
and/or with more extensive sampling than that of plan rationale is provided, however, it is not man-
routine production QC lot release test to demon- datory. For Process Validation PQ, since it is a simu-
strate a high degree of assurance that the process lation of commercial production, a valid statistical
capability of conformance to the test item specifi- rationale for sampling plan should be documented
cation can be reached in validation. to provide evidence that it will have a higher level
 Sampling Plan Procedure of sampling and/or greater scrutiny of process per-
Each device company should establish sampling formance than would be typical of routine com-
plan SOPs that describe how to the select statistic- mercial production.
ally valid acceptance sampling plans for inspection  Sample Number Designation
by attribute (pass/fail, go/no go results) and vari- Generally, sample number should be designated
able (numerical results) data types for both routine and marked on the samples to prevent test data
production and process validation. In addition, mix-up, secure the correct test data trend, and
acceptance sampling plan selection should be facilitate investigation once non-conformity is
based on defect risk severity level (e.g., low, moder- detected. Sample number should indicate which
ate, critical, catastrophic) which could bring out dif- validation batch the sample is used for, what test
ferent confidence levels, probability content levels, item the sample is used for, the sampling point,
AQLs, RQLs, etc., and then lead to different sam- sampling time sequence along the validation batch
pling plans. running, etc.
 Sampling Plan Standard
Where appropriate, sampling plan standard ISO
3.7. Acceptance criteria
2859-1 [7] (or ANSI/ASQ Z1.4 [8]) for attribute test
and ISO 3951-2 [9] (or ANSI/ASQ Z1.9 [10]) for vari-  Subsection Purpose
able test can be used. At the same Acceptance To describe what specifications and their accept-
Quality Limit (or Acceptable Quality Level) (AQL), ance criteria for each test item should be used for
process validation is recommended to adopt the validation test.
General Inspection Level II and single sampling  Specification
plan for tightened inspection while the routine pro- For each test item, the test specifications, including
duction uses the single sampling plan for nor- both individual sample test specification (e.g.,
mal inspection. appearance, dimension) and statistical specification
 Defect Risk Severity Level Analysis (e.g., Cpk, Ppk, k value, data normal distribution),
P-FMEA is a source of risk severity level analysis. should be clearly described so the test result
However, if the test item is not covered by p- “conformed/non-conformed” judgement can be
FMEA, then the risk severity level can be assessed correctly made.
by several aspects, such as device end user safety  Acceptance Criteria
(e.g., life threatening injury, non-life threatening For each test item, both acceptance criteria (e.g.,
injury, temporary discomfort), device function accept: 0/reject: 1, accept: 2/reject: 3) for individual
impact (e.g., loss of device function, cosmetic sample test specification and the acceptance crite-
defect), and production impact (e.g., major disrup- ria (e.g., conformed to the specification) for statistic
tion of production, minor disruption of production). specification should be clearly described so the test
 Sampling Plan Rationale item results pass/fail judgement can be correctly
According to FDA 21 CFR Part 820, Section O – made. Except for cosmetic defect, the acceptance
Statistical Techniques, “(a) Where appropriate, each criteria for individual sample test specification is
manufacturer shall establish and maintain proce- better to set at accept: 0, reject: 1. Otherwise, while
dures for identifying valid statistical techniques some failure samples are detected, the test item
required for establishing, controlling, and verifying result still could be deemed passed. Under the cir-
the acceptability of process capability and product cumstance, the defect samples generally still need
characteristics. (b) Sampling plans, when used, shall to be investigated for their failures, and it will cost
be written and based on a valid statistical rationale. time. In addition, choosing accept: 0, reject: 1 sam-
Each manufacturer shall establish and maintain pling plan (if allowed) means a smaller sample size
JOURNAL OF MEDICAL ENGINEERING & TECHNOLOGY 439

is used for the test, and that is considered judgement, and results expression forms (e.g.,
more efficient. table, graph) that should be presented in the valid-
 Acceptance Criteria for statistical Specification ation report.
For the test data in variable, it is preferred to use  Device Assembly Level Process Validation PQ Test
the variable test sampling plan (e.g., ISO 3951-2, Results
ANSI/ASQ Z1.9) with statistical specification (e.g., k Device Assembly Level Process Validation PQ Test
value) to verify the process capability. And the rea- (equivalent to the Process Performance
son is: under the same level of quality assurance Qualification (PPQ) shown in FDA process valid-
(e.g., same AQL) the test sample size of variable ation guidance [2]) is most important and last val-
test sampling plan can be significantly reduced idation activity in a device project, and it integrates
while compared with the attribute test sampling all the component production performance and
plan (e.g., ISO 2859-1, ANSI/ASQ Z1.4). However, assembly equipment operation to demonstrate that
there are some concerns over the acceptance crite- the final medical device can be consistently pro-
ria for variable test sampling plan, such as the test duced that meets its predetermined specifications
data resolution should be sufficient (e.g., It is better and quality attributes (e.g., appearance, dimension,
to use at least two more decimal places of the function). For variable test results, except for the
measured data while compared with the specifica- statistical result table for descriptive statistics and
tion ones.). Otherwise, it will affect the Ppk, k value, process capability indices, process stability should
and normality test (e.g., Anderson Darling p value) be checked by using a scattered data graph or
results, and consequently impair the meaning of Statistical Process Control (SPC) chart for sample
the statistical results (e.g., 99.7% of all data points data trend. If any abnormal data trend (e.g., con-
will be within ±3standard deviation). In addition, tinuous increase/decrease along the validation
variable data with non-normal distribution could batch running, data points jump/fall to another
be a problem for meaningful statistical results. To steady level, outlier data occurrence) is identified
overcome the concerns, sometimes attribute test and considered risky to the product quality, the
sampling plan is preferred to variable test sam- incident should be investigated for root cause find-
pling plan. ing (e.g., component batch number change, assem-
 Commonly Used Process Capability Acceptance bly equipment status change, test equipment
Criteria status change), and corrective actions should be
If there is no variable statistical specification taken to prevent the recurrence.
requirement, a simple way is to use process cap-  Statistical Results Pass/Fail Judgment
ability index Ppk (or Cpk), which had better be set The statistical techniques of process capability
at ⭌1.33. Ppk set at ⭌1.0 could be acceptable index, process stability, and data distribution nor-
while the risk severity level is not high or there is a mality are rough rules of thumb, and they should
proper justification. For example, during plastic be carefully used while making important decision
injection moulding process validation, the Ppk of on the validation results pass/fail judgment. In
critical dimension for each cavity could be set at other words, if any negative statistical result (e.g.,
⭌1.33, however, the Ppk of critical dimension for capability indices do not meet the specifications,
cross-cavity (or total population) set at ⭌1.0 is process is unstable, test data are not normally dis-
deemed reasonable/acceptable because cross-cav- tributed) occurs, unless the failure is obvious and
ity Ppk value is generally lower than that of single certain, a more detailed and formal statistical ana-
cavity. In addition, the test sample size should be lysis should be performed, or statisticians should
large enough (e.g., better ⭌30 pieces) to make the be consulted to further clarify the issue and make
Ppk result more meaningful. the right decision.

3.8. Test results and statistics 4. Validation approach and OQ detailed test
plan for common equipment and
 Subsection Purpose
test equipment
To describe how test results should be analysed, as
well as what test results, defect amounts, statistical Since the production equipment/process are the focus
data (e.g., mean, maximum, minimum, standard for GHTF process validation guidance, their validation
deviation, k value, Ppk), test result pass/fail approach is more described in the guidance. For
440 Y.-L. CHEN

common equipment and test equipment, they are validation is the common equipment “Instron test
important as well, however, not much validation machine” needs to conduct Equipment Validation
approach is illustrated in the guidance. For the reason, IQ/OQ, and 2nd step is the project specific test fix-
the following provides the validation approach and tures needs to conduct Equipment Validation IQ,
OQ Detailed Test Plan for these two types OQ, and Process Validation PQ (where applicable).
of equipment.
The advantages of the two-step validation
approach include: 1) Simplified re-validation: Once
4.1. Validation approach and OQ detailed test
common equipment is relocated, the moved common
plan for common equipment validation
equipment only needs to be re-validated for its IQ, OQ
Common equipment means non-project specific with using a reduced-scale validation approach (e.g.,
equipment which needs to install additional project re-validate those items, shown in the initial common
specific tools or fixtures to perform their functions. For equipment validation protocols, impacted by the
example, injection moulding machine needs to install relocation), and there is no connection to the project
a project specific moulding tool to perform plastic specific tools or fixtures; 2) Efficient documentation:
injection moulding production, ink printing machine The separation of the common equipment’s validation
needs to install a project specific printing plate to per- documents from the project specific tools or fixtures’
form printing production, and Instron test machine validation documents is beneficial to the documents
needs to install project specific test fixtures to perform archiving and retrieval.
testing function. The OQ test items of a common equipment should
For common equipment, it is better to utilise two- be aimed at the performance by itself, and there
step validation approach, which includes 1st step of should be no linkage to the project specific tools or
common equipment’s Equipment Validation IQ, OQ to fixtures of the common equipment. For example,
verify the common equipment (without installing pro- Instron test machine OQ test could design a test pro-
ject specific tools or fixtures) performance, and 2nd gramme which does not need to use project specific
step of project specific tools or fixtures’ Equipment test fixtures, and the test programme could include
Validation IQ, OQ and Process Validation PQ (where four testing parameter items (displacement, speed,
applicable) to verify the combined equipment (com- force and time) with different settings for each item to
mon equipment þ project specific tools or fixtures) verify the Instron’s key functional performance by
performance. For clarification, three examples are itself. The common equipment’s Equipment Validation
given below. OQ test plan can use the same subsection structure of
Detailed Test Plan shown in the Section 3 above.
 Example 1
For the combined equipment of “injection mould-
4.2. Validation approach and OQ detailed test
ing machine þ project specific moulding tool”, 1st
plan for test equipment validation
step validation is the common equipment
“injection moulding machine” needs to conduct According to EU Guidelines for Good Manufacturing
Equipment Validation IQ/OQ, and 2nd step valid- Practice for Medicinal Products for Human and
ation is the project specific moulding tool needs to Veterinary Use, Annexe 15: Qualification and
conduct Equipment Validation IQ, OQ, and Process Validation [11], Section 5.9 described that “equipment,
Validation PQ. facilities, utilities and systems used for process valid-
 Example 2 ation should be qualified, and test methods should be
For the combined equipment of “ink printing validated for their intended use.” Section 9.1 men-
machine þ project specific printing plate”, 1st step tioned that “all analytical test methods used in qualifi-
validation is the common equipment “ink printing cation, validation or cleaning exercises should be
machine” needs to conduct Equipment Validation validated with an appropriate detection and quantifi-
IQ/OQ, and 2nd step is the project specific printing cation limit where necessary.” In addition, based on
plate needs to conduct Equipment Validation IQ, GHTF process validation guidance, Section A.3, it
OQ, and Process Validation PQ. described “Gauge R&R Study – study for evaluating
 Example 3 the precision (repeatability & reproducibility) and
For the combined equipment of “Instron test accuracy of a measurement device should be con-
machine þ project specific test fixtures”, 1st step ducted.” As mentioned earlier, the term of test
JOURNAL OF MEDICAL ENGINEERING & TECHNOLOGY 441

method validation is more used for chemical property cell), instrument calibration should be sufficient to
measurement qualification especially in pharmaceut- prove the measurement accuracy. However, while
ical industry. For physical property measurement quali- the measurement is indirect or the accuracy could
fication in medical device industry, only accuracy test be impacted by the test programme, the accuracy
and precision test (Gauge R&R test) from the test test (e.g., compare the test data obtained in the
method validation are required. Based on the above, indirect measurement with those obtained in the
for medical device test equipment, except for simple direct measurement) should be performed.
measurement instruments (e.g., balance, calliper, force  Good and Bad Samples Test
gauge), all the other test equipment (e.g., Instron test To check whether a project specific testing can dif-
machine/test fixture, bespoke test machine), involving ferentiate good samples from bad ones. Generally,
custom-made test fixtures and test instructions, need the bad samples (or failure samples) should be pur-
test equipment validation. For which, it includes 1st posely made. The good and bad samples test
step of common test equipment’s Equipment should be conducted especially while the test fix-
Validation IQ, OQ, and 2nd step of project specific test tures, test programme are custom-made and the
fixtures’ Equipment Validation IQ, OQ. To simplify val- test system (i.e., common test equipment þ test fix-
idation practice and documentation, the accuracy test tures þ test programme) is equipped with auto-
and precision test from the test method validation can matic pass/fail judgment.
be performed in conjunction with the test fixtures’  Repeatability Test & Reproducibility Test
Equipment Validation OQ. This is commonly known as Gauge R&R test. For
For any type of equipment (e.g., test equipment, non-destructive test (e.g., dimension measurement),
production equipment, facilities, utilities), the IQ check Gauge R&R test is straightforward. However, for
items are very similar, which generally include equip- destructive test, Gauge R&R test is more difficult
ment design features, installation conditions, environ- since the test samples cannot be re-tested, and
mental conditions, safety features, supplier documents, controlled sample groups used for the test could
calibration, preventive maintenance, spare parts, and be difficult to be prepared.
computer system (including software) verification, etc.  Comparison Test
However, there is almost no description of how to con- To compare the test results from a new and differ-
duct OQ/PQ for test equipment in medical device GHTF ent type test system with the existing one to
process validation guidance. Hence, the test items ensure the test results between the new and old
below are proposed for the Detailed Test Plans of test test systems are equivalent.
fixtures’ Equipment Validation OQ protocol.
Generally, there is no need to develop a dedicated
 Process Parameter Window Test PQ protocol for test equipment, because the long-
Generally, all the equipment test parameters are set term stability and repeatability performance require-
at the appropriate values before the OQ test, and ment of the test equipment can be verified in the
there is no allowable parameter range that can be Process Validation PQ (either in the Component
adjusted during sample testing. Under the general Production Level Process Validation PQ or Device
case, there is no need to conduct the parameter win- Assembly Level Process Validation PQ depending on
dow test. However, for a special case while the where the test equipment is used). The test fixtures’
equipment test parameter settings (e.g., position Equipment Validation OQ test plan for all the test
range parameter settings for peak force searching in items above can use the same subsection structure of
Instron testing) may need to be changed during sam- Detailed Test Plan shown in the Section 3 above.
ple testing, the window test should be conducted.
 Test Programme Verification Test
5. Conclusion
While a test programme is needed for a project
specific testing (e.g., Instron þ test fixtures), the test Based on the description and analysis above, the fol-
graphs should be checked to verify whether the lowing conclusions can be drawn:
test data captured is correct by following the test
description.  The terms of qualification, validation, and process
 Accuracy Test validation are a little bit tricky, and should be
To check a project specific testing accuracy. While clearly defined in the medical device company’s
the measurement is direct (e.g., force test by load SOPs, validation master plans, etc. In the subject
442 Y.-L. CHEN

article, the term of validation is used instead of production process’ Process Validation PQ, valid-
qualification even though these two terms should ation approach and OQ Detailed Test Plan used for
be interchangeable in medical device industry. For common equipment & test equipment are pro-
example, Equipment Validation IQ/OQ/PQ is used vided and analysed as well, and that should make
instead of Equipment Qualification IQ/OQ/PQ. The the validation approach and Detailed Test Plan
term of process validation (or traditional process more clear and complete for the whole picture of
validation) includes individual equipment’s medical device process validation.
Equipment Validation IQ, OQ, PQ and integrated
process/system’s Process Validation PQ.
 For process validation practitioners, one of the most Disclosure statement
challenge tasks is to develop process validation pro- No potential conflict of interest was reported by the author.
tocols, especially for OQ and PQ. A typical section
structure (including Title & Signature Page, Table of
References
Contents, Purpose, Scope, Responsibilities, System
Description, References, Test Plan, Deviation & [1] Chen Y-L. Two-level process validation approach for
Abnormal Results Handling, and Revision History) medical devices. J Med Eng Technol. 2019;43(2):139–149.
[2] FDA guidance for industry, process validation: general
and contents for OQ and PQ validation protocols
principles and practices; January 2011. Available from:
are provided based on the medical device two-level https://www.fda.gov/media/71021/download
process validation approach, and that should facili- [3] GHTF/SG3/N99-10, quality management systems - pro-
tate the OQ/PQ validation protocols development. cess validation guidance; January 2004. Available from:
 Test plan describing how to validate the equip- http://www.imdrf.org/docs/ghtf/final/sg3/technical-docs/
ment/process/system is an essential part of OQ/PQ ghtf-sg3-n99-10-2004-qms-process-guidance-04010.pdf
validation protocols, and two formats (simple & [4] Katz P, Campbell C. FDA 2011 process validation
detailed) of test plan are proposed. Simple Test guidance: process validation revisited. J GXP
Compliance. 2012;16(4):18–29.
Plan with a simple table form is mainly used for
[5] PIC/S. Recommendation on validation master plan,
the equipment’s individual and simple component/ installation and operational qualification, non-sterile
feature functional verification test and that is gen- process validation, cleaning validation; September
erally with characteristics of lower risk and less 2007. Available from: https://www.picscheme.org/lay-
complex. Detailed Test Plan with multiple subsec- out/document.php?id=152
tions in the OQ/PQ protocols is more suitable to be [6] 21 CFR Part 820, medical devices - quality system
used for both individual and complex component/ regulation QSR; April 2019. Available from: https://
www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/
feature functional verification test and integrated
CFRSearch.cfm?CFRPart=820
equipment/process/system functional verification [7] ISO 2859-1, Sampling procedures for inspection by
test that are generally with characteristics of higher attributes - Part 1: sampling schemes indexed by
risk and more complex. Detailed Test Plan is more acceptance quality limit (AQL) for lot-by-lot inspec-
critical than Simple Test Plan. tion; 1999. Available from: https://www.iso.org/obp/
 Detailed Test Plan’s subsections, including Raw ui/#iso:std:iso:2859:-1:en
Material & Instrument, Operating Parameter Setting, [8] ANSI/ASQ Z1.4, Sampling procedures and tables for
Test Batch Setting, Test Item, Inspection & Test inspection by attributes, 2003. Available from: https://
asq.org/quality-resources/z14-z19
Method, Sampling Plan and Sample Number
[9] ISO 3951-2, Sampling procedures for inspection by
Designation, Acceptance Criteria, and Test results and variables — part 2: general specification for single
Statistics, are proposed, which should be able to make sampling plans indexed by acceptance quality limit
the test plan description clear and complete with (AQL) for lot-by-lot inspection of independent quality
rationales to justify how the test plan is set based on characteristics, 2013. Available from: https://www.iso.
science, risk, statistics and regulatory requirement con- org/standard/57491.html
siderations. In addition, the subsection contents are [10] ANSI/ASQ Z1.9, Sampling procedures and tables for
provided with analysis, which should be able to help inspection by variables for percent non-conforming, 2003.
Available from: https://asq.org/quality-resources/z14-z19
process validation practitioners develop appropriate
[11] EU Guidelines for good manufacturing practice for
and justifiable Detailed Test Plans. medicinal products for human and veterinary use,
 Except for the validation approach (IQ/OQ/PQ annex 15: qualification and validation; March 2015.
framework) and Detailed Test Plan used for produc- Available from: https://ec.europa.eu/health/sites/
tion equipment’s Equipment Validation OQ/PQ and health/files/files/eudralex/vol-4/2015-10_annex15.pdf

You might also like