Protocol: Mock Recall Exercise

This is a ready-to-use protocol for planning and executing a mock recall exercise, the rehearsal that proves your recall system works before a real recall forces you to find out under pressure. Replace every <<FILL: ...>> placeholder with your own specifics, set your document numbers and dates, and route it through your normal document control, review, and approval before the exercise runs. A worked filled specimen follows the template so you can see how a completed exercise reads. This document is educational and general; confirm every cited regulation against the current source before you rely on it.

Use this protocol alongside the recall and field action plan, which owns the underlying recall procedure this exercise tests, and the recall reconciliation ledger, which is the same instrument you would use in a real action.

Document control header

1. Purpose

This protocol defines how <<FILL: COMPANY NAME>> designs, executes, and scores a mock recall exercise so that the traceability, contact, and communication machinery behind a real recall is proven to work, timed, and improved, without waiting for an actual defect to force the test. A mock recall that always uses an easy lot and never times anything is not a test; this protocol is written to prevent that outcome.

2. Scope

This protocol applies to mock recall exercises run against any product and lot distributed by <<FILL: COMPANY NAME>> at the sites and markets listed in the header. It covers traceability retrieval, contact-list currency, and, where the exercise level selected in section 5 includes it, a simulated notification test. It does not cover an actual recall in progress, which runs under the governing recall plan referenced above, and it is not itself a corrective action; findings from this exercise feed CAPA under <<FILL: CAPA SOP-ID>>.

3. Responsibilities

4. Definitions

• Mock recall exercise: a rehearsal that starts from a real, distributed lot and exercises the traceability, reconciliation, and, optionally, notification machinery a real recall would use, without alerting or contacting real consignees unless the exercise level explicitly calls for a marked test communication.
• Exercise level: the depth to which the rehearsal is carried, from traceability-only through a full simulated notification, set in section 5.2.
• Hard lot: a lot selected because it is difficult to trace, for example one that passed through multiple distributors, split across markets, or partially returned before the exercise, rather than a lot chosen because it is easy to retrieve cleanly.
• Unannounced exercise: an exercise the team executing it does not know is coming until the clock starts, used to test real readiness rather than prepared performance.

5. Exercise design

5.1 Lot selection

Select the lot for each exercise using a documented method, not by habit or convenience.

Rotate the selection criteria across exercises. A program that always exercises the newest, simplest, single-distributor lot is proving that its easiest case works, not that its system works.

5.2 Exercise level

Select one level per exercise and record the rationale. Escalate the level periodically so the program is not permanently stuck at the shallowest option.

5.3 Targets

Set a numeric target for each metric this exercise will measure, before the clock starts. A target set after the result is known is not a target.

6. Procedure

6.1 Pre-exercise approval

1. Complete section 5 and obtain QA approval of the protocol before the exercise date. For an unannounced exercise, approval is limited to QA and the protocol author; the team being tested does not see the protocol in advance.
2. Confirm the observer(s) are named, independent of the team being timed, and briefed on what to record.
3. Where the exercise level includes a simulated notification (Level 4), confirm in advance, in writing, that every consignee in the sample has agreed to receive a clearly marked test communication and will not treat it as a real recall.

6.2 Kickoff and clock start

1. The recall coordinator announces the exercise (for an announced exercise) or triggers it without notice (for an unannounced exercise) by providing the team only the lot number, exactly as a real signal would.
2. The observer starts the clock at that moment and records the start time to the minute.
3. No team member consults this protocol’s targets during execution; the targets exist to score the result afterward, not to pace the work.

6.3 Traceability retrieval

1. The team retrieves the full distribution record for the lot, to the depth specified in section 5.2, from the systems of record (ERP, warehouse management, serialization/EPCIS).
2. The observer records the elapsed time to a complete, exportable distribution list, and separately notes any step that required a manual workaround, an undocumented query, or a person who happened to know something not in a system.
3. Stop the traceability clock when the list is complete and reviewable, not when the team believes it is done; the observer confirms completeness against the known quantity distributed for the lot.

6.4 Reconciliation simulation (Level 2 and above)

1. Using the retrieved distribution list, complete the recall reconciliation ledger as if the exercise were a real recall: quantity manufactured, distributed, in the firm’s own control, and, where records exist for this lot, quantity already returned or destroyed in the ordinary course of business.
2. Record whether the ledger closes within the target tolerance and what, if anything, is unexplained.

6.5 Contact-list verification (Level 3 and above)

1. Draw the sample of consignees per section 5.2.
2. Verify current, reachable contact details for each sampled consignee through a method that does not disclose the exercise, for example a routine account check.
3. Record the number reachable on file versus the sample size.

6.6 Simulated notification (Level 4 only)

1. Issue the clearly marked test communication, using the actual recall letter template with a visible “TEST, NOT A REAL RECALL” marking, to the pre-cleared sample.
2. Capture response-form completion and timing exactly as a real effectiveness check would.
3. Do not extend the test communication beyond the pre-cleared sample under any circumstance; an uncleared consignee receiving a marked test communication is itself a finding to document and prevent, not merely a note.

6.7 Debrief

1. Hold a debrief with the executing team and the observer(s) within <<FILL: e.g. 2 business days>> of the exercise. Discuss what worked, what required a manual workaround, and what surprised the team.
2. Capture every gap as a finding on the findings table in section 8, with a proposed action, an owner, and a due date.
3. Issue the exercise report per section 9.

7. Acceptance criteria

An exercise is a pass only where every target set in section 5.3 that applies to the level exercised is met. A partial result, one target missed, is recorded as a partial pass with the missed target driving a CAPA; it is not rounded up to a pass. A mock recall exercise that passes every target on every run for several cycles in a row is itself a signal to escalate the level or select a harder lot, not evidence the system needs no further testing.

8. Findings table

9. Report contents

On completion, the recall coordinator issues report <<FILL: report number>>, containing: the exercise identification (lot, level, announced or unannounced, date); the targets set against the results achieved for each metric; a narrative of what happened, including any manual workaround the team used; the findings table from section 8; the conclusion, pass, partial pass, or fail, with the basis stated; and the date and level planned for the next exercise. QA approves the report. Every finding rated a partial pass or fail item converts to a tracked CAPA before the report is considered closed.

10. Records and retention

Retain this protocol, the observer’s timing notes, the completed reconciliation ledger (where exercised), the contact-verification record (where exercised), the findings table, the issued report, and evidence of CAPA closure for <<FILL: retention period>>.

11. References

21 CFR Part 7 (FDA recall policy and procedures). 21 CFR 314.81(b)(1) (Field Alert Report) and 21 CFR 600.14 (Biological Product Deviation Report), for the reporting clocks a real recall built on this same traceability machinery would have to meet. EU GMP Chapter 8 (Complaints, Quality Defects and Product Recalls), for the expectation that the recall procedure is evaluated periodically, commonly through a mock recall. ICH Q9(R1), Quality Risk Management, for the risk-based selection of exercise frequency, lot difficulty, and level. ICH Q10, Pharmaceutical Quality System, for the process performance and continual improvement context this exercise feeds.

Confirm the current version and clause numbers of each reference before issue.

12. Revision history

13. Approvals

Filled specimen

The following shows the protocol completed for an example exercise on a solid oral dose product distributed through a wholesaler-heavy network. The company, product, people, and numbers are illustrative; replace them with your own.

Header: Document number MRX-PRT-011, version 1.0, effective 12 August 2026, owner J. Okonkwo (Recall Coordinator), linked plan QA-PLN-031 v2.0, linked report MRX-RPT-011.

Section 5.1, lot selection: Selection method: random draw, weighted toward lots that shipped through more than one wholesaler tier. Lot selected: SB-6620. Why this lot is a fair or hard test: SB-6620 shipped through two primary wholesalers and one secondary/repack distributor, and 8 percent of the lot was returned to stock and re-shipped under a separate transaction six weeks after initial distribution, which makes the retrieval genuinely harder than a single-hop lot. Prior exercises: the last three exercises (MRX-2026-008 through 010) all used single-wholesaler lots; this exercise deliberately breaks that pattern.

Section 5.2, exercise level: Level 3, traceability, reconciliation, and contact-list verification. Rationale: the program has proven Level 1 and 2 competence in the last two cycles; this cycle raises the bar. Unannounced. Depth: retail (pharmacy and hospital level).

Section 5.3, targets: Distribution list within 4 hours; 100 percent traced; reconciliation ledger closes within 1 percent, residual explained; contact-list accuracy 95 percent reachable on a sample of 20 consignees.

Section 6.2 through 6.5, execution: Clock started 10:02 AM without notice to the supply chain team, who were given only the lot number SB-6620. Full distribution list produced at 12:41 PM, elapsed 2 hours 39 minutes, against the 4-hour target: pass. The list showed 41,200 units distributed of 42,000 manufactured, with 3,296 units (8 percent) shown as returned-and-reshipped under a secondary transaction; the team correctly flagged and traced the secondary leg rather than treating the return as a dead end. Reconciliation ledger completed using the returns and quarantine records on file: 41,200 distributed, 3,296 returned-and-reshipped already reconciled to the second shipment, 700 units remaining in firm-controlled stock, ledger closed with 0 unexplained: pass. Contact-list verification on a sample of 20 consignees: 18 of 20 reachable on file at the number and email listed (90 percent), against the 95 percent target: fail.

Section 8, findings table:

Section 9, conclusion: Partial pass. Traceability and reconciliation both met target; contact-list accuracy missed the 95 percent target at 90 percent, driven by two specific stale records with an identified root cause (no refresh cycle), and a system gap in linking return-and-reship legs was surfaced even though the team ultimately traced it correctly. Both findings carry a CAPA with an owner and due date. Next exercise planned for <<FILL: date>> at Level 4, exercising a single-wholesaler lot to isolate the notification mechanics from the traceability complexity tested this cycle.

The value of this specimen is in the partial pass, not a clean pass. The team met the harder retrieval and reconciliation targets on a genuinely difficult lot and still missed contact accuracy by a specific, explainable margin. Recording that honestly, with two real CAPAs, is what makes the next exercise’s Level 4 test meaningful instead of a rehearsed formality.

Common inspection findings this protocol prevents

• A mock recall program that only ever selects easy, single-distributor lots, so a hard multi-hop trace is never actually tested.
• No timing recorded, so “the team can retrieve the distribution list” is asserted with no evidence of how long it actually takes.
• The team being tested knew the exercise was coming and had time to pre-stage the answer, so the result reflects rehearsal rather than real readiness.
• A passing result recorded with no findings table at all, which is itself implausible for any live system and reads as an unexamined exercise.
• Contact-list gaps found during the exercise never converted into a tracked action, so the same stale contacts are found again at the next exercise, or worse, during a real recall.
• The exercise level never escalates past traceability-only, so notification and effectiveness-check mechanics are never rehearsed until a real recall is the first test.

How to adapt this protocol

1. Set your document number, owner, linked plan, and report number in the header.
2. Set your own lot-selection method in section 5.1 so it is documented and defensible, and rotate it deliberately rather than defaulting to the newest or simplest lot.
3. Escalate the exercise level in section 5.2 over successive cycles; do not let the program plateau at Level 1 or 2.
4. Set targets in section 5.3 from your own validated retrieval and reconciliation capability, not from an aspirational number you have never measured.
5. If you run Level 4 exercises, build the consignee pre-clearance step into your account management relationship well before the first test, and never send a marked test communication to a consignee that has not agreed to receive one.
6. Track every finding to a real CAPA, and report the open findings list in the same management review cadence your recall metrics use.
7. Confirm every regulation in section 11 against the current published version before issue.