GMP deviation management
Everything you need to detect, classify, investigate and close deviations the way EU GMP, PIC/S PE 009 and ICH Q10 expect. Free guides, findings and a self-check, plus the practitioner course that teaches the whole method.
6 modules · 3 hours, self-paced · verifiable certificate with a unique id
The first lesson of Module 1, in full. See how the course teaches before you pay.
Start the lessonTen inspection scenarios. Find out which part of your process would generate a finding.
Take the self-checkThe report template and the QA closure checklist, in a working version you can use today.
Open the templatesMapped to the regulations you are inspected against
Referenced for educational context. No regulator endorses or accredits training providers.
Inadequate investigation is consistently a top-three deficiency category in published EU inspection data and EudraGMDP non-compliance statements. It is rarely because sites do not investigate. It is because of these.
'Operator did not follow the procedure' describes what happened. Inspectors want to know why that was possible and why it was not caught.
The event happened to a room, a piece of equipment or a shift. The assessment only looks at the batch number on the form.
Retraining treats the person as the cause. When the same failure recurs with different people, the system is the cause and the CAPA has to change the system.
Contemporaneous notes were never taken. Six weeks later the investigation relies on memory, and the inspector can tell.
A 30-day clock closes records before investigations are finished. Extensions are not documented, so late records look like uncontrolled ones.
Each deviation is closed in isolation. The fifth identical event in a year is still classified as minor and still investigated as a one-off.
Regulators treat training as part of the quality system, not a favour to staff: EU GMP Chapter 2 requires it, requires its effectiveness to be assessed, and requires the records to prove it. The evidence says the same thing in numbers. Below: what the published studies found, what it looks like on your deviation KPIs, and the clause behind it.
41%
fewer quality defects
in the most engaged teams compared with the least engaged, across 112,000 business units. Engagement and development opportunities are measured together. [1]
26%
greater ability to deliver quality products
in organisations with a strong learning culture compared with those without. [2]
37%
higher employee productivity
in the same strong-learning-culture organisations; output per person, not hours. [2]
94%
would stay longer
of employees, at a company that invests in their learning. Retention is the quality metric nobody puts on the dashboard. [3]
The deviation KPIs a trained team moves
These are the numbers a QA head reports at management review and an inspector asks for. No invented percentages here: measure yours before and six months after, and you will have your own.
Overdue deviations
Measured as: % of open records past target
Lower is better
Investigators who know the method finish sooner, and extend properly with a rationale when they cannot. Overdue falls; controlled extension replaces silent lateness.
Repeat deviations
Measured as: % recurring within 12 months
Lower is better
Root causes that can be changed produce CAPA that works. The same failure mode stops coming back with a different name on it.
Right-first-time records
Measured as: % closed without QA rework
Higher is better
Problem statements, scopes and impact assessments written to the standard the first time. QA review becomes review, not rewriting.
Time to closure
Measured as: median days, raised to approved
Lower is better
Containment on day one, a timeline before any analysis, evidence cited by identifier. Less back-and-forth, fewer reopened records.
CAPA effectiveness
Measured as: % of checks passed
Higher is better
Effectiveness checks designed before implementation, measuring recurrence rather than completion. Passed checks mean the fix worked.
Investigator satisfaction and retention
Measured as: survey score up; % turnover in QA
Lower is better
People stay where they are competent and developed. Every QA departure costs recruitment, onboarding and months of reduced capacity.
Be first when the course opens
Leave a work email. You are notified before public launch and get the launch price. Nothing else is sent.
Detect, document, investigate and close deviations the way inspectors expect to see them, with clear ownership and defensible impact assessments.
Hosted on The GMP Training Academy. Launching soon. People on the waitlist are notified first and get the launch price.
What counts as a deviation and what does not
Immediate actions and containment
Classification: minor, major, critical
Investigation tools that produce a defensible root cause
Impact assessment and batch disposition
Writing the report and closing the loop
A fair comparison with the two alternatives most sites use. If the classroom course is genuinely better for you, book it. Most of the time it is not the classroom that is the problem, it is the follow-through.
| This course | Generic GMP e-learning | Two-day classroom course | |
|---|---|---|---|
| Price per person | Launch price to be announced | €40 to €90 | €1,200 to €2,000 plus travel |
| Time away from work | 3 hours, in your own time | 1 to 2 hours | 2 days |
| Written by | Practitioners who run investigations | Content agency | Varies |
| Practice on real records | Yes, three full investigations | No | Sometimes, on the day |
| Regulation cited by clause | Every module | Rarely | Usually |
| Templates you can use tomorrow | Report, timeline, CAPA plan | No | Sometimes |
| Assessment and certificate | Yes, verifiable ID | Completion certificate | Attendance certificate |
| Questions answered by the author | Yes | No | On the day only |
| Access period | 12 months, with updates | Varies | None after the event |
Each guide answers one question a QA reviewer or investigator actually asks, cites the clause, and shows the weak version next to the defensible one.
A plain-language definition of a deviation under EU GMP and PIC/S PE 009, the difference between deviations, incidents and planned deviations, and worked examples of borderline cases.
Read the guideWhy 'human error' is the most cited root cause failure in inspections, what a defensible root cause looks like, and how to use timeline analysis, 5 Whys and Ishikawa without producing a decorative diagram.
Read the guideHow to classify deviations using ICH Q9 principles, what each class should trigger, how classification is commonly done wrong, and a decision table you can adapt.
Read the guideHow to assess product, patient, process, equipment and regulatory impact, why most assessments are scoped too narrowly, and the questions QA should ask before closing.
Read the guideThe structure, language and evidence an inspector expects in a deviation report, the sections that are most often missing, and a section-by-section template.
Read the guideHow to derive corrective and preventive actions from a root cause, why 'retrain' is rarely enough, and how to design an effectiveness check that measures recurrence rather than completion.
Read the guide10 scenario questions, each with the clause behind the answer. About eight minutes. No sign-up. You get a readiness band and the modules that would close your gaps.
An investigation concludes: 'Root cause: operator did not follow SOP step 7.3. CAPA: operator retrained.' The same step was missed by two other operators in the last year. What is wrong?
Answer and explanation in the self-check.
What the inspector wrote, why, and what would have prevented it. The fastest way to learn what gets cited.
Investigation of anomalies: inadequate root cause analysis
Deviation DEV-24-0187 concluded that the root cause was operator error and the CAPA was retraining of the operator concerned. The same failure mode had been recorded in three previous deviations in the preceding fourteen months, each with the same conclusion and the same corrective action. No assessment of contributing factors, procedure design or equipment had been performed.
Read the breakdownDeviation handling: impact on other batches not evaluated
A temperature excursion in cold store CS-02 was recorded as a deviation against batch 4471 only. Twelve other batches of four products were stored in CS-02 during the excursion window. No assessment of impact on these batches was performed and two of them were subsequently certified and distributed.
Read the breakdownDeviation handling: investigation not completed before closure
Of 40 deviations reviewed, 11 had been closed on or before the 30-day target with investigation sections marked 'to be completed'. In 6 cases the investigation was completed after closure and appended without re-approval. No extension requests were on file. The site's deviation metric reported 100% on-time closure for the period.
Read the breakdownThe course was not written from a syllabus. It was written from what regulators publish about what goes wrong, then worked back to the practice that would have prevented it.
EudraGMDP non-compliance statements
Public statements of GMP non-compliance issued by EU national competent authorities where deviation handling and CAPA are cited.
EMA inspection findings and Q&As
EMA GMP/GDP Inspectors Working Group publications and Q&As on Chapter 1 and Chapter 5 expectations for investigations, root cause analysis and CAPA.
National competent authority deficiency reports
Published analyses of inspection deficiencies by chapter and category. Investigation of anomalies and CAPA are consistently in the top categories.
PIC/S and ICH guidance
PE 009, ICH Q9 (R1) and ICH Q10 for the expectations behind the findings.
Site records reviewed in consultancy and audits
Anonymised deviation records, used to build the practice records and the weak-versus-defensible examples.
Tap a term. The full glossary has 30 terms, grouped by stage of the process.
Open the glossaryThe underlying condition that, if removed, would prevent recurrence. Must be something that can be changed; 'operator error' fails this test.
EU GMP 1.4(xiv); ICH Q10 3.2.2
Seat licences for the department or the site, a manager view of progress, and completion records exported for your training matrix. Content can be mapped to your SOP numbers on request.
If your question is not here, the contact page reaches a person, not a ticket queue.
A short breakdown of a real inspection finding, what went wrong and what would have prevented it. Unsubscribe in one click.
6 modules, real records, a template you will use tomorrow, and a certificate your training matrix can reference. Launching soon. People on the waitlist are notified first and get the launch price.