CAPA examples are most useful when they show the whole record — not just the summary. The examples below show every section the way it would appear in a defensible CAPA log. For the underlying process, see the CAPA process guide.
CAPA record structure (used in every example)
- Header — CAPA number, type (Corrective / Preventive), source, dates, owner.
- Problem or risk statement.
- Risk assessment — severity, occurrence, detection; classification.
- Containment (corrective only).
- Root cause analysis — occurrence and detection.
- Action plan — actions, owners, dates, evidence.
- Implementation evidence.
- Verification of implementation.
- Effectiveness check — criteria, data, disposition.
- Read-across and closure.
Example 1 — Loose structural fastener (automotive Tier 1, Corrective)
Header: CAPA-2026-0142, Type: Corrective, Source: Customer complaint OEM-CR-88321, opened 2026-03-04, owner: J. Reyes (QE, Chassis Line 2), target close 2026-06-04.
Problem statement: Three returned chassis cross-member assemblies from OEM customer over 7 days, all with loose M10 bracket bolts at position 3. Affected lot range 22041-22047. Detected by OEM final assembly torque audit.
Risk assessment: Severity 8 (safety-related, customer-facing), Occurrence 4 (one production day, 1,800 units exposed), Detection 7 (sampling-based audit missed the lot). Classification: High; CAPA priority 1.
Containment: Sort and retorque at OEM (1,840 units, completed 2026-03-05), warehouse sort (560 units, completed 2026-03-05), WIP sort (220 units, completed 2026-03-04), stop-ship on 2 open shipments lifted after sort. Closed 2026-03-06.
Root cause: Method 8D. Occurrence: DC tool had two active rundown strategies and operator manually selected the lower-torque variant after changeover; no interlock between part barcode and tool program. Detection: End-of-line torque audit sampling-based at 1/50 units, affected lot fell between audit points.
Action plan:
- A1 — Add hardware interlock between part scan and DC tool program; owner M. Patel; due 2026-04-15; evidence: PLC program rev 12, sign-off record.
- A2 — Move audit from 1/50 sampling to 100% torque verification via DC tool data with auto-alarm; owner J. Reyes; due 2026-04-15; evidence: SCADA configuration record, control plan rev D.
- A3 — Update PFMEA and control plan; owner J. Reyes; due 2026-04-22; evidence: PFMEA rev 7, control plan rev D.
Verification of implementation: Confirmed 2026-04-23 by R. Lin (QM). Interlock observed on floor (3 part-scan events with wrong program correctly blocked), SCADA alarm tested, PFMEA and control plan retrieved at new rev levels.
Effectiveness criteria (set when plan written): Zero recurrence over 90 days, zero failed torque verifications in DC tool data, 100% scan-program lockout function rate.
Effectiveness check (2026-07-22): 22,000 units produced, zero recurrence, zero failed verifications, 348 scan-program lockout events all functioning. Effectiveness verified.
Read-across: Chassis Line 1 (same DC tool model, different part) reviewed — interlock retrofitted under CAPA-2026-0167. Skid Line (different fastener, different DC tool) reviewed — no action.
Closure: Approved 2026-07-25, J. Reyes / R. Lin / D. Owens (QD).
Example 2 — Wrong-revision label on shipped product (medical device CM, Corrective)
Header: CAPA-2026-0098, Type: Corrective, Source: Customer complaint MDC-CR-441, FDA risk class II, opened 2026-02-12, owner: A. Singh (QE, Packaging).
Problem statement: Customer received 2,200 cartons labeled with previous-revision UDI. Affected lot 2026-02-LOT-018.
Risk assessment: Severity 9 (regulatory, UDI traceability), Occurrence 3 (single lot), Detection 8 (no in-line UDI verification). High priority; reported to FDA per 21 CFR 803 review (determined non-reportable, documented).
Containment: Recall of shipped lot (2,200 units, completed 2026-02-15), full inventory audit at customer and CM warehouse, hold on all open shipments to customer pending verification.
Root cause: Occurrence: Label master updated in ERP, print server cached previous PDF; cache not invalidated by master change. Detection: No verification scan between print and apply; operator visually confirmed application but not data.
Action plan:
- A1 — ERP workflow change: label-master change invalidates print server cache; owner IT/QA; due 2026-03-01.
- A2 — Add barcode verification scanner at labeler comparing printed UDI to work order, reject on mismatch; owner Eng; due 2026-03-15.
- A3 — Revise WI-PKG-204 to include scan step; update DHF; owner A. Singh; due 2026-03-22.
Effectiveness criteria: 100% scan match for 90 days, zero label-related complaints in same period.
Effectiveness check (2026-06-22): 100% scan match across 41,000 cartons, zero complaints. Verified.
Read-across: All 4 packaging lines reviewed; scanner deployment extended to remaining 3 lines under CAPA-2026-0103.
Example 3 — Preventive action from PFMEA review (precision machining)
Header: CAPA-2026-0211, Type: Preventive, Source: Quarterly PFMEA review, opened 2026-04-05, owner: L. Chen (ME).
Risk statement: PFMEA RPN for coolant concentration on Cell 4 = 280 (above the 200 action threshold). Failure mode: coolant drift accelerates tool wear; effect: out-of-spec surface finish; current control: manual refractometer check 1/shift.
Risk assessment: No nonconformance occurred. Trend on tool wear and surface finish reject rate showed slow increase over 8 weeks. Lessons learned applied from CAPA-2025-0411 (coolant excursion on Cell 1).
Action plan: Install inline conductivity sensor with auto top-up on Cell 4, mirroring Cell 1 retrofit; owner L. Chen; due 2026-05-30; evidence: install record, calibration cert, control plan rev B.
Implementation: Sensor installed 2026-05-22, calibrated, control plan revised.
Verification: Confirmed 2026-05-28 by J. Reyes (QE).
Effectiveness criteria: Tool wear trend returns to or improves on Cell 1 baseline over 60 days; zero coolant alarms unresolved beyond 15-minute SLA.
Effectiveness check (2026-07-28): Tool wear trend matches Cell 1 baseline (within 4%), 11 alarms in window, all resolved within SLA. Verified.
Read-across: Cells 2 and 3 added to FY27 capital plan; PFMEA RPN recalculated.
What auditors look for in CAPA records
- Risk assessment present and proportional to action priority.
- Both occurrence and detection root causes stated separately.
- Effectiveness criteria written when the plan was written, with measurable data.
- Verification and effectiveness as two separate signatures on different dates.
- Read-across performed and documented, not just claimed.
- For medical device, the link to MDR / 803 review even when non-reportable.
FAQ
What does a complete CAPA record look like?
It includes header, problem or risk statement, risk assessment, containment (corrective only), root cause (occurrence and detection), action plan with owners and dates, implementation evidence, verification, effectiveness check, and read-across. See the three worked records on this page for the level of detail.
How is a CAPA example different from a corrective action example?
CAPA adds an explicit risk assessment and a type field (Corrective or Preventive) to the corrective action structure. The rest of the record is identical. See CAPA vs corrective action.
Do I need to file a CAPA for every NCR?
No. Most quality systems use a screening step: NCRs below a defined severity, frequency, or risk threshold are dispositioned without opening a CAPA. The screening criteria themselves should be documented and reviewed periodically.
Download the CAPA-ready template
Free Word and PDF. Same structure as the records below.
Get the template