GCP · Blog
Back to journal

Drug Accountability in Clinical Trials: Closing the Custody Chain Before the Inspector Opens the Log

Most accountability guidance hands you a template and calls it done: subject ID, lot, quantity, dispenser initials. That misreads the problem. Inspection findings almost never come from a missing field. They come from a handoff that nobody reconciled, a returned blister card that was logged as received but never re-counted, a temperature excursion that was noticed but not quarantined. Accountability is a custody-and-reconciliation chain across five links: receipt, storage, dispensing, return, and destruction. Your log is the contemporaneous record that the chain stayed closed at every transfer. Treat it that way and inspections stop being a fishing expedition.

GCP 11 min read
A

Aileen

Aileen writes practical guidance for clinical trial teams at GCP Blog.

On this page · 9 sections
  1. 01 At a glance
  2. 02 Accountability is a custody chain, not a log: the five links
  3. 03 Who owns each link: the RACI
  4. 04 The reconciliation equation
  5. 05 What the log must capture vs what inspectors test
  6. 06 Upstream of the log: packaging, labeling, and storage
  7. 07 When the count does not close: discrepancy triage
  8. 08 The recurring findings list, mapped to prevention
  9. 09 Sources

At a glance

  • Drug accountability is not a log you maintain; it is the evidence that custody of the investigational product (IP) stayed unbroken from receipt to final disposition. The log is proof, not the deliverable.
  • Under ICH E6(R3) §2.10.1, responsibility for IP management, including accountability, rests with the investigator/institution, even when a pharmacist holds the keys and the sponsor supplies the system.
  • The reconciliation must close at every handoff: received = dispensed + returned + destroyed + on-hand. A findings list is built from un-reconciled transfers, not from a missing column.
  • When a BIMO inspector opens the log, they compare amounts shipped, received, dispensed, used, returned, and destroyed (FDA 7348.811, Part III.L). They are testing the math, not admiring the template.
  • Storage excursions and labeling defects can invalidate IP before a single unit is dispensed; quarantine, do not dispense, and document contemporaneously.
  • When the count will not balance, triage and CAPA. Never backfill a number to make a row reconcile; a reconstructed log is worse than a discrepancy you investigated.

Most accountability guidance hands you a template and calls it done: subject ID, lot, quantity, dispenser initials. That misreads the problem. Inspection findings almost never come from a missing field. They come from a handoff that nobody reconciled, a returned blister card that was logged as received but never re-counted, a temperature excursion that was noticed but not quarantined. Accountability is a custody-and-reconciliation chain across five links: receipt, storage, dispensing, return, and destruction. Your log is the contemporaneous record that the chain stayed closed at every transfer. Treat it that way and inspections stop being a fishing expedition.

Think of every unit of IP as having an unbroken provenance from the moment it lands on your loading dock to the moment it is destroyed or shipped back. Each link is a transfer of custody, and each transfer is where the chain can silently break.

ICH E6(R3) §2.10.4 requires that the investigator/institution and/or a pharmacist or other appropriate individual maintain records of the product’s delivery, the inventory, the use by each participant, and the return to the sponsor and destruction or alternative disposition of unused product. The same provision specifies those records include dates, quantities, batch/serial numbers, expiration dates where applicable, and the unique code numbers assigned to IP and participants. Read that as a spec for what the chain must capture at each link, not as a one-time form to fill in.

The five links:

  1. Receipt. IP arrives with shipping records. You verify quantity, lot, condition, and that storage conditions held in transit.
  2. Storage. The product sits under controlled conditions until dispensed.
  3. Dispensing. A unit leaves inventory and is assigned to a subject.
  4. Return. Unused or partially used product (the returned blister card with three tablets left) comes back and is re-counted.
  5. Destruction or return-to-sponsor. The unit leaves the site permanently, either destroyed on site under an authorized procedure or shipped back.

A clean log with every field populated still fails if the returned card was never re-counted against what was dispensed. That is the gap most guides never close.

The single most consequential sentence in the corpus is ICH E6(R3) §2.10.1: responsibility for IP management, including accountability, handling, dispensing, administration, and return, rests with the investigator/institution. The sponsor may facilitate (forms, computerized systems, distribution), but facilitation is not delegation of accountability.

When the investigator delegates IP management to a pharmacist or another individual, ICH E6(R3) §2.10.2 requires that the delegated individual operate under the oversight of the investigator/institution. Delegation moves the task, not the duty. Under US regulation, 21 CFR 312.61 requires the investigator to administer the drug only to subjects under the investigator’s personal supervision or that of a responsible subinvestigator, and not to supply IP to any person not authorized to receive it. That is the legal floor under “who may touch the product.”

Custody linkPI / investigatorStudy pharmacist / dispenserMonitor (CRA)Sponsor
ReceiptAccountable for the record existing (E6 §2.10.1)Responsible: verifies and logs receiptVerifies receipt vs shipmentSupplies IP and shipping records (312.57(a))
StorageAccountable; oversight of conditionsResponsible: maintains conditions, quarantines excursionsReviews excursion handlingDefines storage conditions (E6 §3.15.2)
DispensingPersonal supervision (312.61)Responsible: dispenses per protocol, logsSource-verifies dispensing recordsn/a at site
ReturnAccountable for re-countResponsible: re-counts returned productVerifies returns reconcileReceives returned product
Destruction / returnAccountable; authorizes per procedureResponsible: executes, documentsConfirms disposition documentedAuthorizes destruction or receives returns (312.59)

The monitor’s verification is not a separate link; it is the sponsor’s check that each link held. 21 CFR 312.62(a) puts the recordkeeping duty squarely on the investigator: maintain adequate records of the disposition of the drug, including dates, quantity, and use by subjects, and on termination, return unused supplies to the sponsor or otherwise dispose of them under 312.59.

The reconciliation equation

This is the part the listicles skip. Accountability is arithmetic that must close at every transfer:

received = dispensed + returned + destroyed + on-hand

Rearranged for a running check at any moment: on-hand = received − dispensed − destroyed + returned (where “returned” is product returned to inventory pending destruction or shipment back). The point is not the algebra; it is that the equation must balance every time custody changes, not only at study close-out.

Worked example, one subject across one visit cycle:

EventReceivedDispensedReturnedDestroyedOn-handBalance check
Shipment of lot A123 arrives (30 bottles)300003030 = 0+0+0+30 ✓
Subject 014 dispensed 1 bottle (Visit 2)301002930 = 1+0+0+29 ✓
Subject 014 returns bottle, 4 of 30 tablets unused (Visit 3)3011029 (+1 returned)counted: 26 taken, 4 returned, reconciles to dose log ✓
Returned bottle destroyed on site per SOP301012930 = 1(−1 returned-then-destroyed)+0+1+29… re-derive ✓

The discipline is the re-count at return. ICH E6(R3) §2.10.4 ties dispensing records to “the use by each participant (including documenting that the participants were provided the doses specified by the protocol).” If the subject was dosed 26 and returned 4, your record must show 26 used and 4 accounted for at destruction. A returned card logged only as “returned” with no tablet count is an un-reconciled handoff, and that is exactly what an inspector reconstructs.

What the log must capture vs what inspectors test

Build the log from the regulatory source of each field, then test it the way an inspector will. FDA 7348.811 Part III.L directs the investigator to compare the amount of IP shipped, received, dispensed, used, and returned or destroyed, and to review records documenting shipment, receipt, disposition, return, and destruction. The inspector opens the log to run that comparison; design the fields so the comparison closes.

FieldSourceGCP basisCommon error
Date of each transferShipping/dispensing/return recordE6 §2.10.4 (dates)Back-dated or batch-entered, not contemporaneous
Quantity (received / dispensed / returned / destroyed)Physical count at each eventE6 §2.10.4; 312.62(a)Returned product not re-counted, only logged
Batch / lot and serial numberShipment labelE6 §2.10.4 (batch/serial numbers)Lot not traced through to the dispensed unit
Expiration dateProduct labelE6 §2.10.4 (expiration dates)Expired unit dispensed; no expiry check at dispense
Subject unique codeRandomization / enrollmentE6 §2.10.4 (unique code numbers)Dispensing not linked to a specific subject
Dispenser identityDelegation log312.61 (personal supervision)Dispensed by an unauthorized, undelegated person
Storage condition at receiptTemperature recordE6 §2.10.5; 7348.811 Part III.LExcursion in transit not assessed before use
Disposition (return vs destruction)Disposition record/SOP312.59; 312.62(a)Destruction undocumented or unauthorized

The log captures fields; the inspector tests transitions between fields. A field-complete log that does not let you reconcile received against dispensed-plus-returned-plus-destroyed is a finding waiting to be written.

Upstream of the log: packaging, labeling, and storage

Accountability can be defeated before dispensing. The corpus locates these duties on the sponsor side, but they shape what the site must check on receipt and storage.

ICH E6(R3) §2.10.5 requires that IP be stored as specified by the sponsor and in accordance with applicable regulatory requirements. The conditions themselves are set upstream: ICH E6(R3) §3.15.2 requires the sponsor to determine acceptable storage temperatures, storage conditions (for example, protection from light), and shelf life, and to inform all involved parties (monitors, investigators, pharmacists, storage managers) of those determinations. The same provision requires that IP be packaged to prevent contamination and unacceptable deterioration during transport and storage, and that labeling comply with applicable regulatory requirements.

Operationally: when a fridge logs a temperature excursion, the affected units are quarantined, not dispensed, until the sponsor confirms the product remains usable. A unit dispensed during an unassessed excursion is an accountability defect even if every quantity in your log balances, because the chain captured custody but not integrity. FDA 7348.811 Part III.L instructs inspectors to determine how IP was stored on arrival and whether it was stored under appropriate conditions, and to verify that unused IP was appropriately labeled; if it is not appropriately labeled, the inspector collects a copy or photograph of the label. Storage and labeling are accountability inputs, not a separate compliance silo.

A scoping caveat for honesty: this corpus does not contain a dedicated IP packaging and labeling regulation. The packaging, labeling, and storage claims above anchor to the ICH E6(R3) sponsor IP provisions. Detailed labeling-content rules (for example, region-specific clinical-trial labeling annexes) are outside what these sources establish, so this guide does not assert them.

When the count does not close: discrepancy triage

A count that will not balance is not a crisis; an undocumented count that will not balance is. Work the discrepancy as an investigation, never as an editing problem.

A practical decision path:

  1. Recount physically. Most off-by-one discrepancies are a miscount or a returned card not yet entered, not a missing unit.
  2. Trace the last reconciled transfer. Identify the handoff where received last equaled dispensed + returned + destroyed + on-hand. The break is downstream of that point.
  3. Classify. Miscount and arithmetic error, a dispensing not logged, a return not re-counted, a destruction not recorded, or a genuine loss/diversion.
  4. Document the gap as found. Record the actual numbers, the date you discovered the discrepancy, and your investigation. ICH E6(R3) §2.10.4 frames these as contemporaneous records of inventory and use; a contemporaneous note of a discrepancy is a maintained record, a silently corrected number is not.
  5. CAPA for systemic causes. If returns are routinely not re-counted, fix the return-handling step, not the single row.
  6. Escalate genuine loss. A real shortfall of IP, especially a controlled substance, escalates to the sponsor and per protocol. 21 CFR 312.61 prohibits supplying IP to anyone not authorized to receive it, which is the lens for any unexplained disappearance.

What not to do: do not backfill. A reconstructed quantity that makes the log close is a fabricated record. An honest discrepancy with a documented investigation survives an audit; a too-perfect log that an inspector traces to a backfilled entry does not.

The recurring findings list, mapped to prevention

FDA 7348.811 Part III.L is, in effect, the inspector’s checklist; read backward, it is your prevention list.

What the inspector does (7348.811 Part III.L)The recurring deficiencyThe contemporaneous habit that prevents it
Reviews shipment, receipt, disposition, return, destruction recordsRecords exist but do not reconcile across linksReconcile at every transfer, not only at close-out
Compares shipped, received, dispensed, used, returned, destroyedUsed vs returned never reconciled per subjectRe-count returned product against the dose log
Checks shipping records for dates, batch/lot, shipping conditionsLot not traceable to the dispensed unitCarry batch/lot through to each dispensing row (E6 §2.10.4)
Determines how IP was stored on arrival and whether conditions heldExcursion noticed but product still dispensedQuarantine on excursion; dispense only after sponsor assessment (E6 §2.10.5)
Determines who is authorized to dispense or administerDispensed by an undelegated personKeep the delegation log current; enforce 312.61 supervision
Verifies unused IP is appropriately labeledMislabeled or unlabeled returnsCheck label and expiry at receipt and at dispense (E6 §3.15.2)
Reviews disposition where not returned to sponsorDestruction undocumented or unauthorizedDocument on-site destruction per SOP; authorize per 312.59

A note on alignment across sources: ICH E6(R3) and 21 CFR Part 312 do not pull in different directions here. ICH E6(R3) §2.10.4 and 21 CFR 312.62(a) both require investigator-side records of disposition with dates, quantities, and use by subjects; FDA 7348.811 Part III.L is the inspection lens that tests those same records. The duties reinforce each other. The practical implication is that satisfying the ICH recordkeeping standard contemporaneously is what lets the BIMO comparison close, and what 312.62(a) calls “adequate records” is the same record an inspector reconciles.

Adjacent disciplines feed this chain without being part of it. A temperature excursion is also a protocol deviation, so it belongs in your deviation handling as well as your quarantine record. Dispensing records are source data, so they fall under monitoring and source data verification. And the disposition and accountability records are essential documents that live in the trial master file. Treat those siblings as connected, not duplicated.

Accountability is not the log. It is the closed chain the log records. Build the log from the regulatory source of each field, reconcile at every handoff, quarantine on doubt, and investigate rather than backfill. Software and well-built logs enable this discipline; they do not relieve the investigator/institution of the responsibility that ICH E6(R3) §2.10.1 places there, and they do not, on their own, make a site compliant.

Sources

  • ICH E6(R3) Good Clinical Practice (version r3) — International Council for Harmonisation. https://www.ich.org/page/efficacy-guidelines
  • 21 CFR Part 312 Investigational New Drug Application (version 2026-04) — U.S. Food and Drug Administration.
  • FDA Compliance Program 7348.811, Bioresearch Monitoring: Clinical Investigators and Sponsor-Investigators (version 2020) — U.S. Food and Drug Administration. https://www.fda.gov/media/75927/download
A

Written by

Aileen

Aileen writes practical guidance for clinical trial teams at GCP Blog.