GCP · Blog
Back to journal

Principal Investigator Responsibilities: The Non-Delegable Accountability Boundary Under ICH E6(R3) and 21 CFR 312

A new PI usually meets the role as a checklist: screen subjects, sign consent, report adverse events, keep records. Every line item looks delegable, and most of the day-to-day work genuinely is. That framing is also where PIs get into trouble, because it hides the one thing the regulations refuse to let you hand off. The PI role is a boundary of personal, non-transferable accountability. You may delegate tasks to sub-investigators and coordinators; you may never delegate the supervision, the subject-safety judgment, the control of the investigational product, or the commitments your signature on Form FDA 1572 makes. This article draws that line and maps each duty to the regulation that pins it to you.

GCP 10 min read
A

Aileen

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

On this page · 10 sections
  1. 01 At a glance
  2. 02 What a principal investigator actually is
  3. 03 The accountability boundary: delegable tasks vs. non-delegable duties
  4. 04 The four pillars GCP makes personal
  5. 05 Form FDA 1572: what your signature legally commits you to
  6. 06 Records, the investigator site file, and source-data duties the PI owns
  7. 07 Supervision in practice: the delegation log, training, and how a BIMO inspector tests it
  8. 08 Where PIs get cited
  9. 09 PI supervision and inspection-readiness checklist
  10. 10 Sources

At a glance

  • The principal investigator (PI) role is an accountability boundary, not a task list. You can delegate the work; you cannot delegate the responsibility for it.
  • ICH E6(R3) §2.3.1 lets you delegate trial-related activities but pins the ultimate responsibility and oversight to you personally, scaled to the risk of each activity.
  • Signing Form FDA 1572 is a binding set of commitments under 21 CFR 312.53(c)(1): personally conduct or supervise, protocol adherence, informed consent, IRB review, and adverse-event reporting.
  • 21 CFR 312.60, 312.61, 312.62, and 312.64 attach personal duties for subject protection, drug control, recordkeeping, and reporting that survive any delegation log.
  • An FDA bioresearch monitoring (BIMO) inspector tests your supervision through the delegation of authority log, training records, and whether you “retained control and knowledge of the study” (CP 7348.811, Part III.C.3.c).
  • The most common citations are not exotic: inadequate supervision, consent-process gaps, and investigational-product accountability lapses.

A new PI usually meets the role as a checklist: screen subjects, sign consent, report adverse events, keep records. Every line item looks delegable, and most of the day-to-day work genuinely is. That framing is also where PIs get into trouble, because it hides the one thing the regulations refuse to let you hand off. The PI role is a boundary of personal, non-transferable accountability. You may delegate tasks to sub-investigators and coordinators; you may never delegate the supervision, the subject-safety judgment, the control of the investigational product, or the commitments your signature on Form FDA 1572 makes. This article draws that line and maps each duty to the regulation that pins it to you.

What a principal investigator actually is

In a clinical trial run under an FDA Investigational New Drug (IND), an “investigator” is the individual under whose immediate direction the drug is administered or dispensed, and the principal investigator is the responsible leader of the team at a site. A sub-investigator (research fellow, resident, another physician) assists the PI but does not carry the PI’s signature-level accountability. A sponsor-investigator is the person who both initiates and conducts a trial, which means they carry the sponsor’s obligations and the investigator’s obligations at the same time, a real double burden worth naming early.

ICH E6(R3) §2.7.1 requires that a qualified physician (or, where appropriate, a qualified dentist or other qualified healthcare professional per local requirements) who is an investigator or sub-investigator have responsibility for trial-related medical care and decisions. That is the clinical anchor of the role: the medical-care judgment lives with a qualified person on the investigator side, not with a coordinator and not with the sponsor.

The accountability boundary: delegable tasks vs. non-delegable duties

The cleanest statement of the boundary is in ICH E6(R3) §2.3.1: the investigator may delegate trial-related activities to other persons or parties, but retains the ultimate responsibility and should maintain appropriate oversight of those activities to protect participant rights, safety and well-being and the reliability of data. The same section tells you how much oversight: it should be proportionate to the importance of the data and the risk to participant safety. ICH E6(R3) §2.3.3 then requires you to keep a record of the persons and parties to whom you delegated activities. That record is the delegation log, and it is the artifact that proves supervision when an inspector arrives.

So the line is not “important vs. trivial.” It is “the doing vs. the answering for it.” Below is the operational mapping practitioners need.

DutyDelegable?Anchor
Conducting study procedures, data entry, IP dispensingYes, with oversightICH E6(R3) §2.3.1; §2.10.2
Conducting the informed consent discussionYes, to delegated site staffICH E6(R3) §2.8.5, §2.8.6
Overall supervision of delegated staffNoICH E6(R3) §2.3.1; 21 CFR 312.53(c)(1)(vi)(c)
Ensuring delegates are qualified, trained, informedNo (you own this)ICH E6(R3) §2.3.2
Personal conduct or supervision of the investigationNo21 CFR 312.53(c)(1)(vi)(c); 312.60
Protecting rights, safety, welfare of subjectsNo21 CFR 312.60
Control of the investigational drugNo21 CFR 312.60; 312.61
Maintaining the delegation record itselfNo (your record)ICH E6(R3) §2.3.3

Two points sharpen the table. First, ICH E6(R3) §2.10.2 explicitly allows you to delegate investigational-product management to a pharmacist or another individual, but says that person works under the oversight of the investigator/institution, so the control never leaves you. Second, ICH E6(R3) §2.3.2 makes the qualification and training of your delegates your responsibility, which means a poorly trained coordinator is a PI finding, not a coordinator finding.

The four pillars GCP makes personal

Four duties recur across both ICH E6(R3) and 21 CFR 312 as personal to the signatory.

Protocol adherence. ICH E6(R3) §2.5.2 requires the investigator to comply with the protocol, GCP and applicable regulatory requirements, and §2.5.4 says you should follow the protocol and deviate only where necessary to eliminate an immediate hazard to participants, informing the sponsor promptly when you do. The same emergency-deviation logic appears in 21 CFR 312.66, which lets you act without prior IRB approval only to eliminate apparent immediate hazards.

Subject safety and welfare. 21 CFR 312.60 makes the investigator responsible for protecting the rights, safety, and welfare of subjects under the investigator’s care. ICH E6(R3) §2.7.2(d) lets you delegate safety-reporting activities to qualified site staff but holds that you retain the overall responsibility for the safety of participants and for compliance with reporting requirements. Delegation of the paperwork never moves the accountability.

Informed consent oversight. ICH E6(R3) §2.8.5 allows the consent discussion to be conducted by the investigator or delegated site staff, and §2.8.1 requires you to comply with applicable regulatory requirements and adhere to GCP and the ethical principles of the Declaration of Helsinki. Under FDA’s IND framework, 21 CFR 312.60 requires the investigator to obtain the informed consent of each subject in accordance with 21 CFR part 50. (Part 50, the FDA informed-consent rule, and part 56, the IRB rule, sit behind the 1572 but are not yet in this corpus; treat the specific element-by-element consent requirements as governed by Part 50/56, which are out of scope here, and rely on the E6 and 312.60 duties above for the PI’s personal obligation.)

Investigational product control. ICH E6(R3) §2.10.1 places responsibility for investigational-product management, including accountability, handling, dispensing, administration and return, with the investigator/institution. 21 CFR 312.61 is blunter: you may administer the drug only to subjects under your personal supervision or the supervision of a sub-investigator responsible to you, and you must not supply it to anyone not authorized to receive it. This is why drug-accountability lapses land on the PI.

Form FDA 1572: what your signature legally commits you to

Form FDA 1572 is the “signed investigator statement” required by 21 CFR 312.53(c)(1) before the sponsor may let you begin. It is not a formality; each commitment in §312.53(c)(1)(vi) triggers a specific operating obligation. The breakdown:

1572 commitment (per 21 CFR 312.53(c)(1)(vi))What it triggers
Conduct the study per the current protocol; change only after notifying the sponsor, except to protect subjectsProtocol adherence and deviation control (ICH E6(R3) §2.5.2, §2.5.4)
Comply with all obligations of clinical investigators in this partThe full 312.60 to 312.69 duty set
Personally conduct or supervise the investigationNon-delegable supervision (21 CFR 312.60; 312.61)
Ensure informed consent (part 50) and IRB review (part 56) are metConsent and IRB duties; IRB assurance under 21 CFR 312.66
Report adverse experiences in accordance with 312.64Safety reporting (21 CFR 312.64(b))
Has read and understands the investigator’s brochureQualification to use the IP (ICH E6(R3) §2.1.2)
Ensure associates, colleagues, and employees know their obligationsTraining and supervision of delegates (ICH E6(R3) §2.3.2)

Read that list as a single sentence: by signing the 1572 you personally promise to run the study by the protocol, to supervise it yourself, to protect subjects through consent and IRB review, to report safety events, and to make sure everyone you delegate to understands their obligations. The 1572 is the legal hook; 312.60 through 312.66 are the duties it hangs.

Note one tension worth flagging rather than smoothing. ICH E6(R3) §2.3.1 frames delegation expansively (“may delegate trial-related activities to other persons or parties”) and scales oversight to risk. 21 CFR 312.53(c)(1)(vi)(c) frames the same point more rigidly, committing you to “personally conduct or supervise” the investigation, and 312.61 narrows drug administration to your personal supervision or that of a sub-investigator responsible to you. They align in spirit, but the FDA wording is harder-edged on personal supervision of drug handling. For a US IND site, honor the stricter 312.61 standard for the investigational product even where E6’s risk-proportionate language might read as more permissive.

Records, the investigator site file, and source-data duties the PI owns

ICH E6(R3) §2.12.1 makes you responsible for the integrity of data under your responsibility, and §2.12.2 requires adequate source records that are attributable, legible, contemporaneous, original, accurate and complete. ICH E6(R3) §2.12.11 requires you to maintain the trial records (the essential documents that make up the investigator site file) and to have control of all essential records generated by the site. On the FDA side, 21 CFR 312.62(b) requires you to prepare and maintain adequate and accurate case histories that document, for each subject, that informed consent was obtained prior to participation, and 21 CFR 312.62(a) requires adequate records of drug disposition including dates, quantity, and use by subjects.

Retention is personal too. 21 CFR 312.62(c) requires the investigator to keep these records for two years after a marketing application is approved for the indication, or two years after the investigation is discontinued and FDA is notified if no approval is sought. The investigator site file is not a binder a coordinator owns; it is your evidentiary record, and its gaps are your findings.

Supervision in practice: the delegation log, training, and how a BIMO inspector tests it

FDA’s bioresearch monitoring program for clinical investigators, Compliance Program 7348.811, is the inspection playbook, and it tells you exactly how supervision is tested. Under Part III, Section C.3.c, the inspector documents “whether the authority for the conduct of the various aspects of the study was contracted and/or delegated properly so that the clinical investigator retained control and knowledge of the study,” and if there are concerns about appropriate delegation, the inspector obtains the delegation of authority log, the qualifications and training of the person who performed the delegated task, and evidence of the clinical investigator’s oversight. That is the operational test: control and knowledge retained, demonstrated through the log and training records.

The same program reaches consent and IP. Under Part III, Section G.1.b, the inspector reviews the process by which consent was obtained and determines “whether the person obtaining informed consent was delegated this task by the clinical investigator” and whether the investigator or delegated personnel was available to answer subject questions. Under 21 CFR 312.68, you must permit an authorized FDA officer to access, copy, and verify the records and reports you made under 312.62. Inspection access is not optional and not delegable.

Practical translation for inspection readiness: your delegation log must match reality (every person performing a study task is listed, with dates that bracket their actual involvement), your training records must support each delegated task, and your own documented oversight (review of CRFs, signed source data, monitoring-visit follow-up, deviation review under ICH E6(R3) §2.5.3) must be visible. The log is not paperwork; it is the proof of the one duty you cannot delegate.

Where PIs get cited

The Form FDA 483, issued at the close of an inspection when deviations from the regulations are observed (CP 7348.811, Part III.A.6), clusters around a few recurring PI failures:

  • Inadequate supervision: a delegation log that does not reflect who actually did the work, untrained delegates, or no documented PI oversight. This maps straight to the 312.53(c)(1)(vi)(c) “personally conduct or supervise” commitment and ICH E6(R3) §2.3.1.
  • Consent-process gaps: consent obtained by an undelegated person, consent obtained after study procedures began, or the wrong (non-current) consent version, all of which the inspector probes under CP 7348.811 Part III.G.1.
  • Investigational-product accountability lapses: drug-accountability records that do not reconcile receipt, dispensing, and return, in tension with 21 CFR 312.62(a) and the personal-supervision rule of 312.61.

None of these are exotic. They are the predictable failure modes of treating a non-delegable duty as if it were delegable.

PI supervision and inspection-readiness checklist

  • Maintain a delegation of authority log that lists every person performing a study task, with start and end dates, signed by you (ICH E6(R3) §2.3.3).
  • Keep training records that support each delegated task and confirm your delegates are qualified and informed (ICH E6(R3) §2.3.2).
  • Document your own oversight: CRF review, source-data sign-off, deviation review under ICH E6(R3) §2.5.3, and follow-up on monitoring findings.
  • Confirm consent is conducted by the investigator or delegated staff, with the current IRB-approved version, before any study procedure (ICH E6(R3) §2.8.5; 21 CFR 312.62(b)).
  • Reconcile investigational-product accountability records (receipt, dispensing, return) and confirm administration occurs under your personal supervision (21 CFR 312.61, 312.62(a); ICH E6(R3) §2.10.4).
  • Verify the investigator site file is complete and under your control (ICH E6(R3) §2.12.11) and that records meet the 312.62(c) retention period.
  • Be ready to permit FDA access to all records and reports under 21 CFR 312.68.

The throughline is simple to state and hard to live: software, coordinators, and SOPs can enable compliance, but they cannot assume the PI’s accountability. ICH E6(R3) and 21 CFR 312 enable a well-supervised, compliant trial; they do not certify one, and they never relocate the responsibility off the signatory. Related siblings on this site, including the informed consent process, protocol deviation handling, the investigator site file and essential documents, and the delegation of authority log, go deeper on the artifacts named here.

Sources

A

Written by

Aileen

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