Four Findings at a Community Oncology Site
During a routine monitoring visit at a community oncology site enrolling a Phase 3 non-small cell lung cancer trial, you document four findings. Three participants were consented by a board-certified oncologist who is listed as a sub-investigator in the regulatory binder but does not appear on the delegation log and has no documented protocol-specific training. One participant was enrolled and dosed with a baseline absolute neutrophil count of 1.3 × 10^9/L against an inclusion criterion requiring at least 1.5. Investigational product accountability shows 12 tablets unaccounted for across two kits. Two source notes were altered with correction fluid and initialed but not dated, with no reason given. The site has enrolled 19 participants; the PI is the only person with current GCP training.
1. Rank the four findings by severity and state the immediate action for each.
Read the worked answer
Rank by risk to participants first, then integrity of the data, then documentation.
1. The dosed ineligible participant (highest). Notify the sponsor within 24 hours, notify the IRB per its deviation policy, assess the participant clinically, and arrange an unscheduled CBC with repeat ANC. Do not withdraw reflexively-the decision turns on this participant's interests and the protocol's re-dosing criteria, not on administrative tidiness.
2. Twelve unaccounted tablets. Quarantine remaining kits from those two dispensing sessions, recount with the pharmacist, pull dispensing and return records and temperature logs, and interview staff. Report to the sponsor as a potential product-integrity event within 24 hours. In a blinded trial, unreconciled product can mean mis-dosing or diversion and is urgent until proven clerical.
3. Consent by an undelegated sub-investigator. The consent may be substantively valid-the physician is qualified and licensed-but the delegation log and training record are deficient, and the consent process is unauditable. Suspend further consenting by this person until delegated and trained, then verify each of the three consent documents for completeness.
4. Correction fluid on source. A data-integrity noncompliance without immediate safety impact. Recover the original entries if possible, retrain, and require single-line corrections with initials, date, and reason.
2. The participant's ANC later recovers to 1.8 × 10^9/L and the participant asks to continue. How do you handle the decision?
Read the worked answer
Continuation is plausible, but it must be a documented, jointly owned decision rather than a default.
The inclusion threshold exists to protect against neutropenic infection during myelosuppressive therapy. The relevant facts are that the participant received at least one dose below threshold, that the value has since normalized, and that the participant wants to continue. The PI and the sponsor's medical monitor should assess together whether continuing offers this individual a favorable risk-benefit and whether the protocol's re-dosing criteria are met. Record the assessment and its rationale in the deviation file and a clinical note.
Participant preference informs but does not override safety. If continuation is agreed, obtain documented informed consent for the eligibility deviation, since this is material new information relative to what was originally consented.
Reporting: this is a significant deviation-it affected eligibility and potentially safety-so it goes to the sponsor and to the IRB under its prompt-reporting policy, with a root cause.
Root cause matters more than the paperwork. Ask why the ANC was not checked against the criterion: was the value reported after the eligibility check, or was the criterion buried in a worksheet footnote? Fix the eligibility checklist, and add independent second-person verification of all inclusion criteria before first dose. Repeat findings at this site will otherwise recur at the next participant.
3. How do you handle the 12-tablet discrepancy, and what must you avoid doing?
Read the worked answer
Treat it as serious until conclusively proven clerical.
Reconciliation. Build a full ledger per kit: received, dispensed, returned, destroyed, on hand. Cross-check against the Drug Accountability Record, dispensing logs, clinic visit dates, and any participant diary. The most common benign explanation is a dispensing recorded in the medical record but never transcribed to the accountability log-look for that first, because it resolves quickly and cleanly.
If the tablets cannot be located. Escalate as a possible mis-dispensing, tampering, or diversion event. Notify the sponsor within 24 hours, notify the IRB per policy, and consider a for-cause audit. If diversion is plausible, involve the site's research leadership and follow institutional and legal process rather than handling it informally.
Regulatory basis. 21 CFR 312.62 requires the investigator to maintain accountability records; 312.64 requires reporting to the sponsor any deviation likely to affect safety or data integrity.
What to avoid. Do not "write off" the discrepancy by adjusting the accountability log to make it balance. An adjustment without a documented, attributable explanation is a worse finding than the discrepancy itself-it converts a possible clerical error into evidence of concealment-and it destroys the audit trail the sponsor and any inspector will ask for.