Sample Pull Prompt

A Sample Pull Prompt is a governed MES instruction that tells an operator—at a precisely defined point in a batch or device build—what material to sample, how much, how many containers, which container/label to use, and where to send or store it. It links the act of sampling to the batch/eDHR context (lot, equipment, line, recipe step) and to downstream testing or retention, creating an unbroken chain of identity and custody.

In ISA-88 terms, prompts can be associated with Unit Procedures, Operations, or Phases (e.g., at blend endpoint or during compression start-up). At ISA-95 Level 3, the prompt executes in MES, registers or reserves container IDs, spawns labels, and transacts with LIMS for test registration and with QMS for any sampling deviations or late/missed pulls.

Pharmaceutical in-process controls and sampling are mandated by 21 CFR 211.110, which expects scientifically sound procedures for sampling and testing during processing. For medical devices, 21 CFR 820.250 requires appropriate statistical techniques—including sampling plans—for process controls and acceptance activities. Dietary supplements must maintain reserve samples per 21 CFR 111.83, and food manufacturers operate under preventive controls programs that hinge on defined sampling and verification activities.

Because the prompt, the operator response, and subsequent sample lifecycle are electronic, they fall under 21 CFR Part 11 and EU GMP Annex 11 expectations for validated computerized systems, audit trails, and electronic signatures. A robust Sample Pull Prompt ensures that sampling is timely, traceable, and defensible, supporting ICH-style control strategies and quality risk management by preventing missed or misidentified samples from undermining release decisions.

Design prompts where the sampling obligation naturally lives in the procedure. In ISA‑88, anchor the prompt in the Operation or Phase that establishes the sample’s critical context (e.g., at the end of wet granulation when moisture is within target; at the first-article tablets during compression start-up; at tank transfer initiation for content uniformity checks). This avoids ambiguous timing and strengthens genealogy links.

Under ISA‑95, the Level 3 MES should publish the sample event (who, what, when, where, why) to enterprise systems. That includes: LIMS registration (test plan, priority, due time), ERP/WMS (if retain storage locations are managed), and QMS (if a sampling deviation occurs). Clear data contracts—lot/equipment IDs, recipe/step identifiers, sample type codes, and container instance IDs—ensure deterministic handoffs and reduce reconciliation burden.

• Attach prompts to S88 Phases with explicit preconditions (e.g., Unit status ‘agitation complete’ AND ‘temperature within band’).
• Define exception paths: if a prerequisite fails (e.g., hold tag), either suppress the prompt or route to deviation.
• Use ISA‑95 Master Data to standardize sample type codes, container types, and analyses to ensure consistent LIMS mapping.

Encode the statistical and regulatory requirements inside the prompt: the sampling plan (n, units per sample, locations, frequency), acceptance criteria linkage, and timing windows. For pharmaceuticals, in-process sampling plans should support control of critical quality attributes per 21 CFR 211.110, while medical device acceptance activities should rely on statistically valid sampling per 21 CFR 820.250. Dietary supplement reserve samples must be retained per 21 CFR 111.83, which can be enforced as a dedicated prompt at packaging completion.

Trigger patterns

• Event-based: At Phase end (e.g., blend complete) or state transition (filling line first article).
• Time-based: Every X minutes/hours (e.g., environmental swabs each 4 hours on a long run).
• Count/quantity-based: Every Y units or Z kg produced.
• Conditional: Trigger only if a control limit is approached or exceeded (pre-emptive confirmation).
• Composite: X locations per lot combined into one composite for analysis.

Each trigger should include a grace window (e.g., must be sampled within 15 minutes of event) and consequence handling if missed (automatic deviation, hold on the batch segment, or increased sampling until recovery). Prompts should also encode replacement logic for compromised samples and capture reason codes if a sample is intentionally skipped (with QA authorization).

Mislabeling or identity gaps are among the most common sampling defects cited by inspectors. Sample Pull Prompts should drive container creation with unique identifiers and barcoded/encoded labels, ensure label–batch–material linkage is scanned/verified at the point of pull, and capture custody transfers (operator to courier to lab, or to controlled retain storage). If retain samples are required, storage location, environmental conditions, and expiry/re-test dates should be set at creation and monitored thereafter.

• Generate sample container IDs at prompt execution; require scan-back to associate the physical container to the batch step.
• Enforce label content: product lot, sample type, date/time, step/phase ID, container # of #, storage condition, hazard indicators if applicable.
• Record every custody event: who handed off, who received, where (room/freezer), and at what time.
• Prevent duplicate or orphaned containers by disallowing print without registration and blocking registration without print verification.

Electronic prompts, responses, and label prints are GxP data. Systems must implement secure, computer-generated audit trails for creation, modification, and deletion of sample events and metadata; capture contemporaneous timestamps; and attribute actions to uniquely identified, authorized users per 21 CFR Part 11 and EU GMP Annex 11. Controls should include enforced reason codes for changes, prohibition of backdating, clock synchronization, and periodic audit trail review procedures.

• Signature requirements: dual signatures for witnessed pulls (e.g., high-risk operations) and QA e-signature for any deferral/skip.
• Contemporaneous recording: record pull time at the point of action; if delayed entry is allowed, capture late-entry rationale.
• Metadata immutability: freeze core identity fields post-creation; corrections via controlled amendment workflows with audit trails.
• Attributable scans: barcode/QR scans bind the user/device/location to the event; avoid manual ID rekeying unless justified.

A Sample Pull Prompt should either pre-allocate or auto-register tests in LIMS so the lab queue, methods, and due dates are known immediately. Result statuses then feed back to MES to drive holds/releases or additional sampling when out-of-trend conditions occur. If a sample is late, broken, or misidentified, the prompt should invoke QMS deviation workflows, and if necessary, CAPA. These closed loops sit squarely in ISA‑95 Level 3 to Level 4 integration patterns.

• Synchronous LIMS registration: return of LIMS Sample ID stored on the batch record for traceability.
• Rules to place WIP on electronic hold until critical sample results are ‘pass’ (or provision to continue under conditional release).
• Automatic triggering of re-sampling if lab rejects a non-viable sample (e.g., insufficient quantity).
• QMS hooks for deviations (missed/late pull), investigations (root cause), and effectiveness checks after remediation.

Well-instrumented prompts enable proactive management. Track timeliness, completeness, labeling quality, and impact on cycle time and release. Use dashboards to focus on bottlenecks (e.g., swab turnaround) and escape routes (e.g., mislabeled retains). Integrate these KPIs into management review and continuous improvement programs to reduce defects and shorten release timelines without eroding control.

Treat Sample Pull Prompts as configurable functionality within a validated MES. Apply GAMP 5 risk-based validation: scale rigor by impact to product quality and data integrity. Document User Requirements (URS) for triggering logic, label content, security, audit trail, and integration behavior; map to functional specifications and configuration; and verify through risk-based test protocols. Negative testing (e.g., attempt pull outside window; print without registration) is critical to show robust controls.

• Traceability: URS → specs → tests (including interface tests to LIMS and printers).
• Security: role-based access for creating, executing, deferring, or cancelling prompts; two-person e-signature when required.
• Data migration/Change control: governed workflows for modifying sampling plans or label templates; impact assessment and regression testing.
• Periodic review: re-verify prompt timing vs. process capability and control strategy; retire obsolete sample types.

• Pharma solid dose: stratified blend sampling and tablet weight/content checks at defined turret counts; retain samples at packaging completion.
• Sterile fill: first-article fills, periodic container closure integrity (CCIT) pulls, and environmental swabs at risk-based frequencies with tight time windows.
• Biotech liquids: in-process bioreactor and buffer samples at phase-specific times; cold-chain custody and stabilized containers.
• Medical devices: statistically justified in-process acceptance sampling (820.250) and component lot verification; serialized unit trace-back to sample pulls.
• Dietary supplements/food: allergen verification swabs after changeover; reserve samples per 111.83; HACCP/Preventive Controls verification.
• Long campaigns: rolling time-based prompts across shifts with handover logic; ensure no duplicate or missed prompts at shift boundaries.

Edge cases include composite sampling (combine defined increments across locations), destructive samples (automatic yield adjustment), and constrained environments (BSL or potent compound areas) where scanning and label printing must be intrinsically safe or staged via pass-throughs. For cold-chain or moisture-sensitive samples, prompts should assert container type and precondition verification (e.g., pre-chilled tubes, desiccated vials) before allowing execution.

• Ambiguous timing: prompts fired before the state is truly ‘ready’. Control with explicit preconditions (sensor/state confirmation) and clear windows.
• Label/content drift: uncontrolled edits to label templates. Govern via change control and locked templates; verify at PQ.
• Offline sampling: manual pulls when systems are down lead to lost genealogy. Provide controlled paper contingencies with rapid back-entry and reconciliation paths.
• Missed prompts during rework or re-batching: ensure prompts re-evaluate when material re-enters a step.
• Identity break at handoff: custody transfer not recorded. Enforce scan-to-receive at lab or retain storage.
• Time sync issues: unsynchronized clocks causing sequence confusion. Implement clock drift monitoring and NTP controls.
• Excess or insufficient sample quantity: failure to account for lab rework/retain splits. Encode quantity buffers and split logic.

Auditors frequently trace a sample from prompt to LIMS result to batch disposition; any gap suggests a data integrity failure. Proactive exception dashboards and routine audit trail reviews reduce the likelihood of undetected defects and late-cycle surprises.

V5 models Sample Pull Prompts as first-class MES objects tied to ISA‑88 phases. Each execution binds lot/equipment context, spawns unique container IDs and labels, and registers samples to LIMS with immediate feedback of LIMS IDs into the eBMR/eDHR. If a prompt is late or skipped, V5 opens a deviation in QMS, can place the affected batch segment on electronic hold, and routes for QA e-signature per Part 11/Annex 11 controls.

• Configurable trigger logic (event/time/quantity/conditional) with grace windows and escalation.
• Barcoded labels and scan-to-verify custody; cold-chain and retain storage mapping into WMS locations if used.
• Closed-loop status: LIMS reject causes auto-resample prompt; pass/fail feeds QA Disposition Steps.
• Audit trail and review-by-exception dashboards highlighting late/missed pulls, identity mismatches, and lab rejections.