Site SOPs, Delegation and Training Records

Part of Clinical Research Coordinator (CRC) Training: The Regulatory Framework.

About 5 minutes. Free, in your language, no account needed.

Why this matters

Three quiet documents cause a large share of all audit findings: the standard operating procedure nobody follows, the delegation log nobody updates, and the training record nobody completes.

None of them treat a participant. All of them decide whether your site's work can be defended.

Standard operating procedures

A standard operating procedure, or SOP, is a written description of how your site does a recurring task. Typical site SOPs cover informed consent, source documentation, investigational product handling, adverse event identification and reporting, delegation and training, equipment calibration, and archiving.

An SOP is useful when it describes what your site actually does. It becomes a liability the moment it describes something better than what your site does, because an auditor will hold you to your own written standard.

Three rules:

1. Write what you do. If your consent SOP says two staff members review every form and only one does, either change the practice or change the SOP.
2. Version and date them. Each SOP needs a version number, an effective date, and an owner.
3. Train people on them and record it. An SOP nobody has been trained on is not an SOP; it is a document.

Small sites sometimes have no SOPs of their own and work to the sponsor's manuals. That can be acceptable, but it should be a deliberate, documented decision rather than an accident.

The delegation of authority log

The delegation log records who the principal investigator has authorized to perform each trial task, and for what period. It is the single most examined document at many inspections.

What it must show:

• Each person's name, role and signature or initials, so that entries in other records can be attributed to them.
• The specific tasks each person is authorized to perform. "All study duties" is a weak entry; itemized tasks are better.
• A start date, and an end date when someone leaves the study.
• The principal investigator's signature and date.

How it goes wrong, in order of frequency:

• A new coordinator starts work before being added.
• Someone leaves and is never given an end date, so the log claims they are still active.
• Tasks appear in the records that nobody on the log is authorized to do.
• The log is signed by the PI once, at the beginning, and never again as staff change.

The test an auditor applies is simple. They take an entry from a source record or a consent form, find the initials, and check the log. If the person was not delegated for that task on that date, it is a finding — even if the work was done perfectly and by an expert.

Training records

Every person on the delegation log must be trained, and the training must be documented.

Three kinds of training normally apply.

Good Clinical Practice training. Sponsors commonly require this to be current and to reflect the applicable standard, which today means ICH E6(R3). Renewal intervals of two or three years are common, but the interval is set by sponsor or institutional policy, not by ICH.

Protocol training. Specific to this study, and repeated whenever the protocol is amended. Record the protocol version each person was trained on.

System and procedure training. The electronic data capture system, the randomization system, the laboratory manual, the site's own SOPs.

A training record should show who was trained, on what, on which version, on what date, and by whom or by what method. A signature on a sheet with no version and no date proves nothing.

Curricula vitae and licenses

The site file also holds a signed and dated curriculum vitae for the investigator and normally for other key staff, and a current professional license or registration for anyone whose role requires one. Both go stale. Licenses expire and CVs stop reflecting what a person does. Put renewal dates in the same calendar you use for ethics committee expiry dates.

In practice

A monitoring visit at Northside Clinic goes badly for ten minutes.

Marcus, the monitor, finds a screening assessment signed with initials he does not recognize. Nadia explains that it was Tunde, a nurse who covered for two weeks in August while she was away.

Tunde is not on the delegation log. He is a competent nurse who did the assessment correctly. That is not the point, and Nadia does not argue it. The record cannot show that the person who performed a trial task was authorized and trained to perform it.

They document it as a deviation. Dr. Osei reviews the affected participants' data. The site adds a step to its cover arrangements: no one covers a study task until they are added to the log, trained, and their training is filed. Nadia writes it into the site's SOP because the same gap will otherwise reappear the next time someone takes leave.

Marcus checks it at the next visit and finds a named backup already trained and delegated in advance.

Competence is not authorization. The log is what turns one into the other.

Key takeaways

• A standard operating procedure must describe what the site actually does, because an auditor will hold the site to its own written standard.
• SOPs need a version, an effective date, an owner, and documented training for the people expected to follow them.
• The delegation of authority log records who may perform each trial task and when, with start and end dates and the principal investigator's signature.
• The commonest delegation failures are staff working before being added, leavers never given an end date, and tasks appearing in records that nobody was delegated to perform.
• Training records must show person, topic, version, date and method, and GCP renewal intervals such as two or three years come from sponsor or institutional policy rather than from ICH.