WorkflowManufacturing & Quality

Process Validation Protocol & Report Authoring

A process validation protocol for PPQ in which each acceptance criterion carries its source, runs on the plant batches already made before anyone signs, and reaches the client's QA section by section.

See one PPQ protocol drafted, rehearsed and signed ↓
Watch with transcript →
The problem

A criterion copied from the last protocol, first tested in PPQ batch one

A validation engineer builds the PPQ protocol by copying the last one written for a similar product, then editing it against a development report, a risk assessment, an impurity assessment, the client's specification, method reports and equipment records. Criteria end up reading meets specification or report result, the sampling plan belonged to another product, and the batch count is three because it has always been three.

Nobody learns that a criterion cannot be met until a PPQ batch fails it. The miss turns into a deviation inside validation, a send-back from the client's quality team or a question from the investigator, and at a CDMO it moves the client's filing. Review comments travel in marked-up copies that never land on the record.

Built forValidation leadsQA validation reviewersMSAT leadsQC lab heads

The Problem Today

Each new PPQ protocol starts as a copy of the last, and the first test of its criteria against the process is the first PPQ batch.

3
consecutive batches, only a guide for API validation (ICH Q7) published

Typed into the protocol by hand

  • Dev report
  • Risk review
  • Client spec
  • Batch data
  • Criteria nobody can defendMeets specification or report result, with no reason written
  • Client review in marked-up copiesComments arrive by email and never reach the record

The Solution

  1. Each criterion tied to its recordLimits under the day they apply, samples as pins, each criterion with its source
  2. Rehearsed on batches already madeEach criterion runs on the plant batches before anyone signs
  3. Argued over before reviewAs investigator and as client QA, including what the agents wrote
  4. Signed by two, then sent outAuthor and QA validation sign; the client answers each section in fields

Built-in controls

  • Final approval waits while a source is a draft
  • The author never countersigns their own draft
  • Learned rules take effect only with the QA head's signature

What You Get

Criteria the process cannot meet surface on batches already made, and the client's QA and the investigator see the reason behind each limit.

  • A new engineer drafts like a veteranChallenger questions and past rulings come built into each draft
  • PPQ samples kept for monitoringThe same sample points become watch points for continued verification
  • Batch count argued in numbersPer quality attribute, with each sum shown to QA validation
Features

A process validation protocol rehearsed on five plant batches before signing

One drug substance protocol, from the agents' first read of its sources to the client's answers and the report.

demoEvery name, batch number and result in the screens below is sample data.

  1. 01Run sheet

    PPQ sampling plan laid out on the nine working days of one batch

    • Routine and PPQ samples apart. Hollow pins mark routine tests and solid pins the samples PPQ adds, so a reviewer sees at once what validation asks of the lab.
    • Limits under the day they apply. Each critical parameter sits under its stage as a normal-range tag, with an amber dot where plant-scale data never reached the edge.
  2. 02Cake sampling

    Filter-dryer cake sampled at five positions, each criterion with its source

    • Sampling rationale on the pin. The drying study showing that the wall dries slower is cited beside the sample, the rationale EU GMP Annex 15 asks a sampling plan to give.
    • Positions a composite cannot see. Top, wall, center and discharge results sit side by side, so a wet spot hidden inside the release composite shows up before PPQ.
  3. 03Rehearsal

    Each criterion rehearsed on the plant batches already made, before signature

    • Labeled rehearsal, never PPQ. Results from earlier batches stay marked as development data, and a cell with no data reads gray instead of passing.
    • The cause beside the miss. The failing batch is set against its drying time, so the validation lead fixes the process step instead of loosening the limit.
  4. 04Change control

    A process change raised from the rehearsal through change control

    • Raised after reading the batch record. The change cannot be raised until the validation lead has opened the failing batch's record in person.
    • Proposed here, signed off by plant QA. Raising it logs the reason; the plant QA head signs it, and final approval holds for the revised instructions.
  5. 05Challenger

    The PPQ protocol argued over as the investigator and the client's QA would

    • Gaps in what the agents wrote. A five-position cake sample with no wall-to-center criterion is caught in the draft and closed with a limit and its sum.
    • Within-batch consistency in numbers. Meets specification gives way to assay and impurity range criteria derived from each method's repeatability.
    • Unanswerable points go to a person. Mill speed with no plant-scale proven range waits for the MSAT lead, with options to choose between and nothing assumed.
  6. 06Batch count

    Number of PPQ batches justified per quality attribute, with the sums shown

    • Three published approaches side by side. Rationale, target capability and expected coverage each give a count, so three rests on more than custom.
    • Bounds for three, four or five batches. QA validation sees which attributes more batches would help and which are margin problems no batch count can fix.
    • Margin problems routed by name. Palladium and particle size go to the MSAT lead and the QC lab head instead of disappearing into a batch count.
  7. 07QA review pack

    QA validation countersigns after opening the rehearsal and a criteria sample

    • A sampled read of the criteria. One criterion in five goes into the pack with its source passage, so the reviewer checks evidence instead of signing a table.
    • Reading counted per person. A record someone else opened does not count, and the author cannot countersign the draft they wrote.
  8. 08Client portal

    Client approval of each PPQ protocol section, answered in fields

    • Approve, comment or question per section. The client's head of QA answers each section in its own field, which sets that section's status on the CDMO side.
    • A comment goes to a person. Typed words reach the validation lead unread by the agents and hold acceptance until someone opens them.
    • One client's protocols only. The portal shows that client's sections and nothing from other clients who share the plant.
  9. 09PPQ report

    Process validation report drafted criterion by criterion from the protocol

    • Same criteria, executed results. Each report row cross-references its protocol criterion and fills from the executed batch record and the lab results.
    • Watch points for routine monitoring. PPQ samples and attributes with thin margins carry into continued process verification instead of a plan written from scratch.

Validation leads write it; QA and the client see why each limit is set

The people who write, check and approve one PPQ protocol and its report.

Validation leads

Write and sign PPQ protocols and reports for each new product and change

  • A sourced draft to start from
  • Unmeetable criteria found before PPQ
  • Client answers they can act on
QA validation reviewers

Review protocols for the quality unit and countersign as second person

  • Capability bounds per attribute, three to five batches
  • A sampled read of the criteria
  • Rehearsal results kept apart from PPQ
MSAT leads

Own the ranges, the risk assessment and each process decision

  • Ranges quoted only where data exists
  • Process points routed with options
  • Watch points carried into monitoring
QC lab heads

Confirm methods and lab capacity before the first PPQ batch

  • Extra PPQ samples listed per batch
  • Unfinished method reports flagged
  • Capacity confirmed in fields

Read from the development report, batch records and client specification

Process development reportNormal and proven ranges, the scavenger and drying studies, and whether the report is approved or still a draft.
Risk and impurity assessmentsResidual risk after controls, and which impurity is tested at which step against which limit.
SpecificationsThe client's drug substance and intermediate specifications that each criterion is held against.
Batch records and LIMS resultsThe development and clinical batches already made in the plant, read for the rehearsal and the batch count.
Method validation and equipment qualificationWhich methods have an approved validation report and which equipment is qualified before execution.
More screens

Signed twice, and final approval still waits on a draft development report

The rest of the team's screens around one protocol.

Each criterion beside the record behind it80 criteria a batch, 240 in all, each with its source; draft sources are marked, and a proposed rule changes nothing until approved.
What blocks final approval, item by itemSources, instructions, methods, equipment and lab capacity in one view; a draft development report and an unfinished method report block approval.
The protocol document, signed twice and still heldSections in the CDMO's template, with the author's and QA validation's signatures and their meanings; final approval lists the seven items it waits on.
PPQ protocols and reports in flight, by plant and clientEach document with client, plant, state, next date and criteria count; routine status reviews are signed in bulk only after one in five is opened.
Agent steps and signatures in one trailEach action with its time, its why, a signature's meaning and its session; agent runs sit beside people, and the trail downloads as one file.
A looser setting holds until the plant QA head agreesRead time, bulk sampling rate, bound confidence, review clock and retention; tightening applies at once, and the administrator signs no quality record.

demoSample data, from the working solution.

The challenger found a criterion the sampling plan needed and the draft lacked

Eight agents draft and check; people decide and sign.

Source gatherer

Reads the development report, risk and impurity assessments, specifications, instructions and equipment records, and notes each one's approval state.

  • Treats no draft as approved
Parameter mapper

Lays each critical parameter's normal and proven range under the day it applies, with the experiments behind each edge.

  • Extends no range beyond its data
Sampling planner

Places PPQ samples where development showed variability: cake positions, drum positions, after the scavenger.

  • Removes no routine sample
Criteria writer

Writes each criterion with its source and the sum behind it, and applies a learned rule once the plant QA head has signed it.

  • Writes no report-result criterion without a stated reason
  • Applies no rule still waiting for approval
Batch-count analyst

Justifies the number of PPQ batches from residual risk and capability bounds for each attribute.

  • States no count as decided
Rehearsal runner

Runs each criterion on the plant batches already made and labels the result as development data.

  • Calls no rehearsal a PPQ result
  • Fills no cell without data
Challenger

Argues as the inspecting investigator and as the client's QA, and turns on the work of the other agents too.

  • Leaves each point it raised for someone else to settle
Report drafter

Builds the PPQ report from the protocol, fills it from executed batches and writes the hand-off to continued process verification.

  • Writes no conclusion

The plant QA head gives final approval; the client's QA approves each section

Validation lead
Sees
The rehearsal, the challenger's points and the source behind each criterion
Signs
The protocol and report as author, and raises any change the rehearsal calls for
QA validation reviewer
Sees
The review pack: rehearsal result, challenger points and a sample of criteria
Signs
The countersignature as second person, on a draft they did not write
Plant QA head
Sees
The list of what still blocks final approval
Signs
Change control, learned rules and final approval of the protocol and report
Client head of quality assurance
Sees
The sections sent for review, each criterion with its reason
Signs
Approval of each section under the quality agreement

Process validation protocol questions from validation and QA leads

What is a process validation protocol?

A process validation protocol is the approved plan for process performance qualification: how each PPQ batch is run and sampled, the criteria it must meet, and how many batches are made. FDA's guidance and ICH Q7 expect the quality unit to approve it before execution.

How many PPQ batches does a drug substance process need?

No fixed number. ICH Q7 treats three consecutive successful batches as a guide, and EU GMP Annex 15 asks for a count justified by quality risk management. Here residual risk sits beside three published approaches and confidence bounds per attribute, for QA validation to review.

What goes in an API process validation protocol template?

FDA's guidance lists it: manufacturing conditions, critical parameters with ranges, a sampling plan with its rationale, criteria for each significant step, statistics for variability within and between batches, deviation handling, and equipment, method and training status.

Can a failing batch be left out of the evaluation?

A batch with a criterion not met stays in. Leaving out any other plant batch needs a written reason and a QA signature from someone other than the requester; it clears the signatures and withdraws the client's copy.

Who approves the protocol, and when can PPQ start?

The validation lead signs as author, QA validation countersigns, and the plant QA head approves with the client's QA under the quality agreement. Execution waits for that approval, blocked while a source is a draft or the batch instructions are unapproved.

What does the process validation report contain?

It follows the protocol criterion by criterion, filled from executed batch records and lab results, with each deviation and its effect on the conclusion, and hands PPQ sample points to continued process verification. The author proposes the conclusion; QA validation, then the plant QA head, sign it.

Bring one product's
development report.

Send a development report, its risk assessment and the plant batch records; we will rehearse the criteria they support.