Field notes
Who can do what at a study visit: MAs, vitals, sub-Is, and standard of care
Three gates decide it every time, and the answer is never "because the sponsor said so."
Can a medical assistant take vitals in a clinical trial? Usually yes, if three things are true: vitals are within an MA's scope of practice in your state, the protocol does not require a specific license for them, and the PI has delegated that task to that person on the delegation log after documented training. Blood draws and injections are the same question with harder answers, because state law on MAs varies for both, and protocols sometimes require a nurse or a physician. A study does not make a task legal that was not already legal in the clinic, and the clinic does not make a task delegated that the PI has not signed for. Here is how to work the question for any role and any procedure.
The three gates every procedure passes
I think of it as three doors in a row. A task has to get through all of them before anyone touches a subject.
Gate 1: scope of practice. This is state law plus the person's license or certification. It decides what an MA, an LPN, an RN, a pharmacist, or a phlebotomist may do at all, under what supervision. Medical assistants are the hard case because most states treat them as unlicensed personnel working under a physician's delegation, and the rules on what can be delegated to them differ state to state. The AAMA keeps a state-by-state index of those laws and notes plainly that scope for MAs varies by state. In Virginia, where I coordinated, the Board of Medicine's rule is on the practitioner: delegate patient care only to subordinates who are properly trained and supervised, and never knowingly let a subordinate work outside their scope of practice (18VAC85-20-29). Your state has its own version. Check it, and check with your PI.
Gate 2: the protocol. The FDA's investigator-responsibilities guidance is blunt about this one: when a protocol specifies who performs a task, the protocol wins even where state law would allow someone else. Its own example is a physical exam. If the state lets an NP or PA do physicals under supervision but the protocol says a physician must, a physician must (FDA, 2009, section III.A.1). The same section lists what FDA has found inappropriately delegated on inspections: screening evaluations and eligibility, physical exams, adverse-event evaluation, primary-endpoint assessments, and obtaining consent. Rater-qualification clauses, "vitals must be taken supine after 5 minutes rest," "ECG read by a cardiologist" all live here too.
Gate 3: the PI's delegation, with training behind it. ICH E6(R3) says the investigator may delegate trial-related activities, keeps ultimate responsibility, must ensure the people delegated to are appropriately qualified and informed about the protocol, and must keep a record of who was delegated what (E6(R3) 2.3.1 to 2.3.3; the FDA adopted it as final guidance in September 2025). The FDA guidance adds that the list should describe the tasks, point to the training that qualifies each person, and carry dates of involvement. The training-versus-delegation sequencing question has its own post: trained first, or delegated first.
The order matters. Gate 3 cannot open Gate 1. A PI signing the log does not make a blood draw legal for someone whose state does not allow it. And Gate 1 does not open Gate 3: a licensed RN who has never been trained on this protocol is still not delegated.
Who typically does what
This table is the typical pattern at a small private-practice site. Every cell is governed by your state, your protocol, and your PI, in that order. Treat it as the shape of the conversation, not the answer.
| Task | PI | Sub-I | CRC / nurse coordinator | RN | MA | Phlebotomist | Pharmacist | Imaging tech | Rater |
|---|---|---|---|---|---|---|---|---|---|
| Informed consent | Yes | Yes | Often (per protocol and IRB) | Often | Rarely | No | No | No | No |
| Eligibility determination | Yes | Yes | Assembles; PI or sub-I confirms | Assembles | No | No | No | No | No |
| Physical exam | Yes | Yes | No | Only if licensed and protocol allows | No | No | No | No | No |
| AE causality | Yes | Yes | Records; does not judge | Records | No | No | No | No | No |
| Vital signs | Yes | Yes | Yes | Yes | Usually, within state scope | Sometimes | No | Sometimes | No |
| ECG acquisition | Yes | Yes | Yes if trained | Yes | Often, within state scope | No | No | Yes | No |
| Phlebotomy | Yes | Yes | Yes if trained | Yes | State-dependent | Yes | No | No | No |
| Urine pregnancy test | Yes | Yes | Yes | Yes | Usually | Sometimes | No | No | No |
| IP dispensing / accountability | Yes | Yes | Often (per protocol) | Often | Rarely | No | Yes | No | No |
| IP administration (injection, infusion) | Yes | Yes | Only if licensed | Yes | State- and protocol-dependent | No | Sometimes | No | No |
| Imaging (OCT, fundus photos, DEXA) | Yes | Yes | If certified | If certified | If certified | No | No | Yes | No |
| Efficacy scales (PHQ-9, PASI, EASI) | If qualified | If qualified | If qualified | If qualified | Rarely | No | No | No | Yes |
One note on the last row: "rater" is a role, not a job title. The dermatologist scoring PASI is the rater, and the protocol's rater-qualification clause applies to them regardless of what else they do.
Standard of care, and the sub-I who "just sees the patient"
The recurring question is the person who touches the subject but is not study staff: the practice's MA who rooms every patient and takes vitals, the hospital phlebotomist, the front-desk nurse who happens to give the flu shot. Do they go on the log?
Here is what the guidance actually says, and where it changed. E6(R2) told the investigator to keep a list of qualified persons delegated "significant trial-related duties" (E6(R2) 4.1.5), and the FDA's 2009 guidance uses the same phrase. Neither said what "significant" excluded, so sites listed everyone. E6(R3) 2.3.3 says it out loud: documentation of delegation should be proportionate to the significance of the activity, and where activities are performed as part of clinical practice, delegation documentation may not be required. Section 2.7.1 pairs with it: a qualified physician who is an investigator or sub-investigator holds responsibility for trial-related medical care and decisions, while other qualified healthcare professionals may be involved in the care of participants in line with their normal activities.
Read those two together and the working rule falls out. If the task is something that person does for every patient, exactly the same way, and the protocol adds nothing to it, R3 gives your PI room not to list them. If the protocol changes the task, even slightly, they are doing a trial-related activity and belong on the log. Vitals become trial-related the moment the protocol specifies the position, the rest period, the cuff, or the time relative to dose. A "standard of care" blood draw becomes trial-related when it goes into a sponsor kit with a processing SOP.
What your sponsor will accept is a separate question; many still want everyone listed. That is a sponsor preference, not a regulation, and it is worth settling at the SIV rather than at close-out.
Sub-Is are the same test with a title attached. The R3 glossary defines a sub-investigator as a team member designated and overseen by the investigator to perform significant trial-related procedures or make important trial-related decisions. A physician in the practice who sees a subject for a sinus infection between visits is not doing either. The same physician who signs off eligibility, does the protocol's physical exam, or judges whether that sinus infection is related to the study drug is making study decisions, and needs to be a sub-I on the 1572 and delegated for those tasks before the first time they do it, not after. "Sees the patient" and "makes study decisions" look identical on the schedule. They are not identical on the log.
A new MA starts Monday
The mid-study staffing change is where sites get findings, because the morning is busy and the log is in a binder. Before the new MA takes a single blood pressure on a subject, in this order:
- Scope first. Confirm with the PI that vitals, ECG, and whatever else you plan to hand them are within an MA's scope in your state, under the supervision your site actually provides.
- Protocol second. Read the vitals and ECG sections of each protocol they will touch. If one says "RN or above," they are not doing that one, whatever your state says.
- Credentials on file. CV, MA certificate if they hold one, and GCP training completed and dated.
- Study-specific training, documented. Protocol training with a date and a signature, per study. The FDA guidance explicitly extends training expectations to staff hired after the study began.
- Delegation log, signed by the PI, dated on or before the first task. One line per task, not "all study procedures."
- Sponsor notified if the protocol or the clinical trial agreement requires it, and the signature and initials sheet updated.
Then, and only then, they take the 8:40 vitals on Subject 019.
Delegated, but the GCP lapsed
The trap nobody plans for: the log is clean, the training log is clean, and the coordinator's GCP certificate expired in March. Most sponsors expect GCP renewal on a cycle (three years is common; some ask for less), and the monitor's question is not "were they delegated" but "were they current on the day." Someone delegated in 2024 whose GCP lapsed in March 2026 has, on paper, been doing delegated tasks without current training since March. The fix is cheap and the discovery is expensive, so put every expiry date on one sheet and look at it monthly.
What the morning should already know
Every question above is a question for the visit, not for the close-out. On the day, each line of the visit should carry who is delegated for it and whether that person's training is current, because that is exactly the question a monitor asks about a random visit six months later, and it is a better question for 8:30 on the morning of than for the binder in October. A run sheet that lists the visit's procedures in the right order and, beside each one, the initials of the person delegated for it today, answers the monitor before they ask. If the MA who usually takes vitals is out and the coordinator steps in, the sheet should make that swap visible, and the coordinator should already be on the log for vitals, or the swap is a deviation waiting to be written up.
The three gates do not move for a busy Tuesday. Check your state's scope of practice, read the protocol's staffing language, and make sure the PI has signed for the person before the person touches the subject.