---
title: "Predetermined Change Control Plan (PCCP) Template"
type: white-paper
publisher: Ketryx
source: "https://www.ketryx.com/assets/pccp-template"
content: text extracted from PDF (layout/tables/figures not preserved)
---

# Predetermined Change Control Plan (PCCP) Template

*Source: [https://www.ketryx.com/assets/pccp-template](https://www.ketryx.com/assets/pccp-template)*

---

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX Instructions: To use this template effectively, please note the following color-coded guidance: Blue Text: Indicates context, instructions, common considerations, and questions pertaining to the section.

This

text

should

be

deleted

prior

to

finalizing

the

document. Red Text: Indicates content that should be replaced with product- and manufacturer-specific information

for

the

ease

of

use

of

this

template.

Red

text

also

indicates

the

presence

of

hypothetical

medical

device

information.

This

information

should

not

be

present

in

your

PCCP. Follow these steps to complete the PCCP template:

1)

Read

through

the

template:

Understand

the

contents

of

each

section

and

how

it

could

apply

to

your

product.

The

example

medical

device

content

provided

could

give

you

an

idea

for

what

information

you

may

want

to

include

in

your

PCCP. 2) Customize the Template: Replace all red text with specific information about your medical device and

organization.

3)

Describe

the

ML-DSF:

Provide

details

about

the

ML-DSF

units

in

your

device,

including

an

architecture

diagram

and

a

description

of

how

they

fulfill

the

intended

use

of

the

product.

4)

Outline

Modifications:

Clearly

describe

each

planned

modification,

ensuring

that

they

are

specific,

verifiable,

and

appropriate

for

a

PCCP.

Address

the

questions

provided

in

the

template

to

ensure

a

comprehensive

description.

5)

Develop

the

Modification

Protocol:

Use

the

template

sections

on

Data

Management

Practices,

Re-Training

Practices,

Performance

Evaluation,

and

Update

Procedures

to

develop

a

detailed

protocol

for

implementing

and

assessing

modifications. 6) Assess the Impact: Evaluate the impact of each modification and the collective impact of all modifications

on

the

safety

and

effectiveness

of

the

device.

Example

Medical

Device

Description

Version 3 | 1

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX The example provided in this template is a wearable medical device that incorporates a Machine Learning

Device

Software

Function

(ML-DSF).

The

ML-DSF

is

capable

of

predicting

and

diagnosing

healthcare

conditions

using

sensor

data.

It

features

a

core

algorithm

that

can

be

updated

remotely

and

a

personal

ML-DSF

algorithm

that

continuously

learns

from

the

user's

real-world

data.

The

modifications

outlined

in

this

template

include

weekly

core

algorithm

re-training

DISCLAIMER

The

example

provided

in

this

template,

involving

a

wearable

medical

device

with

a

Machine

Learning

Device

Software

Function

(ML-DSF),

is

for

ILLUSTRATIVE

PURPOSES

ONLY.

It

is

intended

to

demonstrate

how

to

apply

the

template

to

a

hypothetical

medical

device

and

is

not

representative

of

any

specific

product

or

manufacturer.

Users

of

this

template

should

replace

the

example

with

details

specific

to

their

own

medical

device

and

organization. Version 3 | 2

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX VAL-XXXXX Predetermined Change Control Plan Template

Version 3 | 3

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX Table of Contents 1. Introduction 3 2. Machine Learning Device Software Functions 4 3. Description of Modifications 4 4. Modification Protocol 6 5. Impact Assessment 10 6. Authorization 11 7. Revision History 12 Version 3 | 4

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX 1. Introduction 1.1.

Purpose This document serves as the Predetermined Change Control Plan (PCCP) for Medical Device Name developed by Name of Organization. This document describes the modifications that will be

made

to

the

Machine

Learning

Device

Software

Function

(ML-DSF)

for Medical Device Name and how the modifications will be controlled. This PCCP shall be included and discussed in

General

Device

Characteristics

of

the

marketing

submission

and

included

as

a

standalone

section

as

a

“Predetermined

Change

Control

Plan”.

1.2.

Related Materials

Document Title Source standard POL-0001 Quality Manual per ISO 13485 SOP-001 Data Management Operations per Good Machine Learning Practices SOP-002 Software and Algorithm Lifecycle Processes per IEC 62304 SOP-003 Verification and Validation per IEC 62304 SOP-004 Risk Management per ISO 14971 WOR-001 Re-training DNN Models PLAN-001 Data Specifications per Good Machine Learning Practices ARC-XXXX Software Architecture Description per IEC 62304 SDS-XXXX Software Design Specification and Detailed design per IEC 62304 REQ-003 Core Module Performance requirements per IEC 62304 VVP-001 Core Algorithm V&V protocol RRR-001 Re-training Report PLAN-001 Cloud deployment and monitoring LBL-001 User labeling and field notifications Version 3 | 5

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX 2. Machine Learning Device Software Functions The Medical Device Name software system is composed of several units, including units with Machine

Learning

Device

Software

Functions

(ML-DSFs).

See ARC-XXXX for further details. [Insert an architecture diagram that describes the ML-DSF in context with the rest of

the

device

design.] With Ketryx, an architecture diagram could be generated automatically as you define and

develop

relationships

between

software

configuration

items. The ML-DSF units undergoing modification directly impact the intended use of the product. This plan

shall

control

such

changes

to

ensure

the

device

maintains

original

device

planned

performance. The following planned device modifications described have been reviewed as changes which would

otherwise

require

a

new

submission

per Deciding When to Submit a 510(k) for a Software Change

to

an

Existing

Device

(FDA,

October

2017). 3. Description of Modifications Defining your description of modifications for a PCCP can be efficiently achieved by utilizing a system

like

Ketryx

to

define

your

Requirements.

The

Ketryx

platform

enables

clear

documentation

and

management

of

each

modification's

requirements,

ensuring

compliance

and

facilitating

verification

and

validation

processes

essential

for

FDA

approval. The ML-DSF incorporated into the wearable device system can predict and diagnose healthcare conditions

by

comparing

the

user’s

Real-world

EKG

sensor

data

with

personalized

health

(physiological)

data.

The

ML-DSF

core

EKG

algorithm

“off-the

shelf”

is

trained

on

representative

sample

data

which

can

be

updated

through

cloud-deployed

changes. See SDS-XXXX for further details. For the modification, consider: ● Is the modification described specific? ● Can the modification be verified and validated? ● Is the modification implemented automatically or manually? ● Will the modification be implemented across all devices on the market or locally? ● Are the modifications appropriate for a PCCP? ○ Does the modification improve the safety or effectiveness of the device? Version 3 | 6

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX ○ Is the modification related to quantitative measures of ML-DSF performance

specifications?

○ Is the modification related to device inputs of the ML-DSF? ○ Is the modification related to the device’s use and performance within a specific

subpopulation?

● Is the modification in alignment with the intended use of the device? 3.1. Modification #1 - Weekly Core Algorithm Re-training The device contains a core ML-enabled algorithm for assessing a user’s healthcare

state

using

ECG

data

and

personal

health

history.

This

machine

learning

model

is

trained

on

sample

data

which

may

drift

from

population

data

over

time.

The

modification

proposed

would

be

no

change

to

the

overall

architecture

or

functionality,

but

instead

re-training

of

the

machine

learning

model

using

more

recently

acquired

data

to

maintain

the

same

level

of

performance

over

time.

This

core

algorithm

change

can

be

deployed

remotely

and

updated

weekly.

This

is

a

change

in

quantitative

measures

of

ML-DSF

performance

specifications…

3.2. Modification #2 - Core Algorithm Retraining This modification leverages real world data from new sources to improve the level

of

performance.

This

is

considered

a

change

in

quantitative

measures

of

ML-DSF

performance

specifications…

3.3. Modification #3 - Core Algorithm Re-training Real world data shows that current ML-DSF needs architecture modifications to account

for

new

sources

of

data

relevant

to

the

training

process.

Version 3 | 7

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX

4. Modification Protocol Ketryx is equipped to efficiently manage evidence generated for the modification protocol of a PCCP. The

platform

supports

comprehensive

verification,

validation,

and

traceability

management

and

facilitates

the

documentation

and

control

of

data

handling,

model

re-training,

and

performance

evaluations

necessary

for

satisfying

this

section

of

your

PCCP. Learn more automatic documentation generation in Ketryx: https://go.ketryx.com/4bBIv1b A comprehensive and lifecycle approach to ML-DSF modification shall be used to ensure the continued

safety

and

effectiveness

of

the Medical Device Name software system. All changes are

subject

to

quality

system

processes

outlined

in POL-0001 Quality Manual (ISO 13485) and software

processes

described

in SOP-002 Software and Algorithm Lifecycle Processes (IEC 62304). See Appendix A from the FDA’s PCCP ML-DSF Guidance for example questions that may be considered to develop the Modification Protocol Section. 4.1. Data Management Practices 4.1.1. Method A SOP-001 Data Management Operations describes how data shall be collected,

synthesized,

controlled,

quality

controlled

and

assured,

labeled,

enhanced,

cleaned

and

sequestered. PLAN-001 Data Specifications Wearable

Personalization describes the specific data requirements suitable

for

this

modification

to

be

met.

[Collection Protocols, Assurance of Data Quality, Reference Standard Determination,

Sequestration

of

Test

Data

Sets] 4.1.2. Method B [Overview of Method B and associated supporting documents] [Collection Protocols, Assurance of Data Quality, Reference Standard Determination,

Sequestration

of

Test

Data

Sets]

4.2. Re-Training Practices 4.2.1. Method C SOP-002 Software and Algorithm Lifecycle Processes (IEC 62304) describes how algorithm re-training is controlled and monitored. For this specific

machine

learning

model WOR-001 Re-training DNN Models shall be

used

and

a

re-training

report

shall

be

produced

per

the

instruction.

Version 3 | 8

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX This report identifies the training objectives, focus, implementation details and

overall

results

of

the

activity

including

anomalies. RRR-001 Re-training

Report documents the most recent re-training of the machine learning

model.

All

anomalies

are

managed

per

the

software

problem

resolution

process

described

in SOP-002 Software and Algorithm Lifecycle

Processes

(IEC

62304). A new report shall be issued for each re-training

activity. [Discuss Re-training Objectives and Focus, Re-training Implementation] 4.2.2. Method D [Overview of Method D and associated supporting documents] [Discuss Re-training Objectives and Focus, Re-training Implementation] 4.3. Performance Evaluation 4.3.1. Method E The core algorithm shall undergo performance evaluation as part of any new

product

introduction

(NPI)

cycle.

Any

change

to

the

software,

whether

it

is

anticipated

to

directly

or

indirectly

affect

the

core

algorithm

performance,

shall

undergo

performance

testing

per

VVP-001

Core

Algorithm

V&V

protocol. The performance requirements and assessment metrics

are

documented

in REQ-003 Core Module Performance requirements. [Triggers to Initiate Performance Evaluation, Assessment Metrics and Elements,

Statistical

Analysis

Plans,

Performance

Targets,

Additional

Testing

Needs] 4.3.2. Method F [Overview of Method F and associated supporting documents] [Triggers to Initiate Performance Evaluation, Assessment Metrics and Elements,

Statistical

Analysis

Plans,

Performance

Targets,

Additional

Testing

Needs] 4.4. Update Procedures 4.4.1. Method G The Medical Device Name software system shall undergo field changes per the process outlined in PLAN-001 Cloud deployment and

monitoring

at

least

once

per

month

if

an

approved

build

candidate

is

available.

This

process requires that users are notified per LBL-001 User labeling

and

field

notifications. All approved build candidates are required to

undergo

software

verification

and

validation per SOP-003 Verification and

Validation.

Version 3 | 9

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX [Software Verification and Validation, When and How Updates will be Implemented,

Communication

and

Transparency

to

Users,

Device

Monitoring

Plan,

Marketing

considerations] 4.4.2. Method H [Overview of Method H and associated supporting documents] [Software Verification and Validation, When and How Updates will be Implemented,

Communication

and

Transparency

to

Users,

Device

Monitoring

Plan,

Marketing

considerations]

Version 3 | 10

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX Table 1 - Modification to Modification Protocol Component Traceability Modification Protocol Component Modification Data Management Practices Re-training Practices Performance Evaluation Update Procedures Modification #1 - Core Algorithm Re-training SOP-001 DataOps PLAN-001 Data Specifications Wearable Personalization Method A (See Section 4.1.1) SOP-002 Software and Algorithm Lifecycle Processes (IEC 62304)* WOR-001 Re-training DNN Models RRR-001 Re-training Report** Method C (See Section 4.2.1) REQ-003 Core Performance requirements VVP-001 Core Algorithm V&V protocol Method E (See Section 4.3.1) SOP-003 Verification and Validation PLAN-001 Cloud deployment and monitoring LBL-001 User labeling and field notifications Method G (See Section 4.4.1) Modification #2 Method B (See Section 4.1.2) Method D (See Section 4.2.2) Method F (See Section 4.3.2) Method H (See Section 4.4.2) Modification #3 Method B (See Section 4.1.2) Method D (See Section 4.2.2) Method F (See Section 4.3.2) Method G (See Section 4.4.1) * Ketryx can quietly enforce software procedures during the software lifecycle process so developers can focus on algorithm modifications. Learn more about traceability in Ketryx: https://go.ketryx.com/3QLFeEG ** Ketryx can automatically generate documentation from test execution results, which can include test setup details such as algorithm training. Learn more automatic documentation generation in Ketryx: https://go.ketryx.com/4bBIv1b Version 3 | 11

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX 5. Impact Assessment The impact of each change, as well as their collective impacts, are assessed using

the

risk

management

process

described

in SOP-004 Risk Management (ISO

14971). The Impact Assessment for each modification should address the following questions: ● How does the version of the device with each modification implemented compare

to

the

version

of

the

device

without

any

modifications

implemented? ● What are the benefits and risks, including risks of social harm, associated with

each

individual

modification?

● How do the activities proposed within the Modification Protocol continue to reasonably

ensure

the

safety

and

effectiveness

of

the

device?

For marketing submissions: ● How does the implementation of one modification impact the implementation

of

another?

○ How does the individual modification impact other device software functions,

as

well

as

device

hardware?

If the ML-DSF is a multiple function device product ● Determine if additional information is necessary for submission per Multiple Function Device Products: Policy and Considerations. 5.1. Modification #1 Impact - Core Algorithm Re-training [Discussion of the clinical benefits of this modification] [Discussion of the risks associated with this modification, See ISO 14971:2019 Cl.

5.1-5.5,

6.0] [Discussion of the risk mitigations, with references to specific risk management file

artifacts

and

requirements,

See

ISO

14971:2019

Cl.

7.1-7.2,

7.5-7.6] [Discussion of the overall impact on the medical device system, residual risk and labeling

considerations,

See

ISO

14971:2019

Cl.

7.3,

7.4,

8.0]

5.2. Modification #2 Impact - Core Algorithm Re-training [Discussion of the clinical benefits of this modification] Version 3 | 12

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX [Discussion of the risks associated with this modification, See ISO 14971:2019 Cl.

5.1-5.5,

6.0] [Discussion of the risk mitigations, with references to specific risk management file

artifacts

and

requirements,

See

ISO

14971:2019

Cl.

7.1-7.2,

7.5-7.6] [Discussion of the overall impact on the medical device system, residual risk and labeling

considerations,

See

ISO

14971:2019

Cl.

7.3,

7.4,

8.0]

5.3. Modification #3 Impact - Core Algorithm Re-training [Discussion of the clinical benefits of this modification] [Discussion of the risks associated with this modification, See ISO 14971:2019 Cl.

5.1-5.5,

6.0] [Discussion of the risk mitigations, with references to specific risk management file

artifacts

and

requirements,

See

ISO

14971:2019

Cl.

7.1-7.2,

7.5-7.6] [Discussion of the overall impact on the medical device system, residual risk and labeling

considerations,

See

ISO

14971:2019

Cl.

7.3,

7.4,

8.0]

5.4. Collective Impact Assessment [Describe the collective impact of implementing all modifications] 6. Authorization Has this PCCP been previously authorized by the FDA? ☐ Yes ☐ No If yes, list the premarket approval and the document number and version submitted. Premarket approval: K1234567 Document and version: VAL-XXXXX version 0.1 ☐ N/A

Version 3 | 13

Category: Validation Title: Predetermined Change Control Plan Template Owner: Owner Name Version 0.1 State Draft Effective Date XXX. XX,2024 Document ID VAL-XXXXX 7. Revision History Version Changes 1.0 ● New Document Version 3 | 14