IRB protocol writing.

001The fields that stop people, and what they actually want.

Protocol forms ask for information in a register nobody uses elsewhere, which is why competent researchers stall on them. Each field below has a specific expected answer, and knowing that is most of the work.

• Procedures. Not your method in the abstract, but a sequence: what happens to a participant, in order, from first contact to the end of their involvement, with durations attached.
• Risks. Boards expect you to name them rather than to claim there are none. Even minimal-risk studies carry something, usually discomfort or a confidentiality exposure, and a form claiming zero risk reads as unexamined.
• Risk level. Exempt, expedited or full review is a determination with criteria, and claiming a lower level than your design supports guarantees a return.
• Data handling. Where data lives, who reaches it, how identifiers are separated or destroyed, and when. Vague answers here are among the most frequent revision requests.
• Recruitment. The actual wording participants will see, attached as a document, not a description of how you will approach them.
• Consent. A document written to be understood by your participants that simultaneously agrees with the protocol on every single point.

002The mistake that produces a second round.

Internal contradiction. The protocol says interviews will be recorded and the consent form does not mention recording. The procedures describe collecting names and the risk level claims exempt. The recruitment script promises something the protocol does not deliver. Each of these is a small clerical divergence and each one returns the whole submission.

They happen because the documents are written at different times, often in different sittings, and nobody reads them against each other at the end. That final cross-check takes half an hour and saves weeks, since a review cycle is measured in the board's meeting schedule rather than in your urgency.

So the first thing this desk does with any packet, whether we drafted it or you did, is read every document against every other document and produce a list of the disagreements. That is frequently the entire fix.

003A revision letter is a checklist, not a verdict.

Most first submissions receive one, and reading it as a judgment on your ethics is the wrong frame and leads directly to the wrong response. It is a list of conditions. Answer it as one: point by point, in the board's own order, stating what changed and where it now appears in the revised documents.

Two habits cost extra rounds. Arguing with a point instead of addressing it, which returns the letter with the same point restated. And correcting only the instance the board flagged, when boards routinely cite one example of a problem occurring in several places and expect the pattern fixed. Read each point as a category, correct it everywhere, and say in your response that you have done so.

The structural description of the whole process, from a first draft through to the approval letter, sits at the IRB entry in the lexicon.

004What we draft, and what stays yours.

We draft and revise protocols, consent and assent documents, recruitment materials, instruments and data management plans; check the risk-level claim against what is actually being collected; align every document against every other; and convert a revision letter into a point-by-point response with tracked changes.

We do not submit on your behalf or correspond with a board as though we were you, and we will not help present a study as lower-risk than its design supports, which is the one request here that reliably backfires on the person making it. Where the study belongs to a thesis, the protocol is coordinated with the long-form desk so the methodology chapter and the approved protocol do not drift apart, which otherwise happens reliably.

Questions.

I have never written a protocol. Can you start from nothing?

Yes, and it is the most common order here. Send your board's templates, a description of what you want to study, and your population. The draft comes back complete: procedures in sequence, risks named honestly, data handling specified, with consent and recruitment documents built to agree with it on every point.

How do I know which review level to claim?

It follows from your design rather than from preference, and claiming a lower level than the design supports guarantees a return and costs you a full cycle. The determination turns on identifiers, population vulnerability, and the nature of the procedures. We assess it against your actual protocol and tell you plainly, including when the honest answer is the slower path.

My letter asks for things I do not understand. Can you decode it?

That is a common order and often a short one. Send the letter and your submitted documents. Board language is formulaic once you have read a few hundred of them, and most points reduce to a missing attachment, a contradiction between two documents, or a data handling detail. You get a plain-English reading plus the drafted response.

Can you write the consent form on its own?

Yes, and it is the document that most often causes a return, because it has to be readable by your participants and simultaneously agree with the protocol everywhere. We write it to your board's required elements, check it line by line against the protocol, and flag anything in your design that the consent language cannot honestly describe.

How is protocol work priced?

Per document and per stage, quoted individually and free. A full packet drafted from nothing, a consent rewrite and a revision response are three different jobs at three different prices. Send your board's templates and whatever exists, and an itemized figure comes back within two hours.