01Three routes, three clocks
The categories describe how much risk a study carries and therefore who has to look at it. Exempt never means unreviewed: somebody still makes that determination, and it is not you. Expedited means a single reviewer, or a small handful, may act without waiting for anybody to convene. Full board means the entire committee, in a scheduled meeting, with your protocol sitting on an agenda.
The difference you actually live with is timing. The first two routes are handled as submissions arrive and answered in a stretch of days or weeks. The third waits for a meeting, and meetings carry submission cut-offs ahead of them, frequently a fortnight or more. Miss a cut-off by one day and your study waits for the next sitting, which is how one absent signature costs a month instead of an afternoon.
02What actually decides the route
Not your subject, and not how careful you feel about it. Boards read a small set of features and the category falls out of them, which is good news, because it means you can predict yours before submitting and build the timeline around it.
- Who the participants are. Children, prisoners, anybody unable to consent for themselves, and at many institutions your own staff or students, each pull a study upward
- What you collect. Identifiers, health information, anything a person could be traced through, and anything sensitive enough to harm somebody if it escaped
- What is done to people. Observation and anonymous survey sit low. An intervention, a concealed purpose, a specimen or a procedure does not
- Where the data comes from. Records already gathered for another purpose are treated quite differently from new information taken from living people
- Who else is involved. A second site, a partner organisation or another university adds an agreement, and usually a review at their end running on their calendar
03The category you claim is a proposal, not a decision
You state one on the form and the board determines the real one. Claiming lower than the facts support is the expensive mistake, because the determination returns asking you to resubmit on the correct route, and the clock restarts against that route's cut-off. Claiming higher than necessary simply spends weeks you were not obliged to spend.
So describe the study plainly and let the determination happen. Where you honestly cannot tell, ask the board office before submitting; the question is routine, costs nothing, and the reply tells you which form and which deadline you are working to. Quality improvement work raises the same question wearing different clothes, since some of it is not research at all under the definition boards use, and having that determination in writing early prevents a long argument later.
04Building the timeline around a meeting date
Once you can predict your route, everything else in the schedule follows. Find the committee's meeting dates and its submission cut-offs for the year, both normally published, then count back through the pieces that must exist before you can file: current training certificate, finalised instruments, signed site permission, consent documents drafted, and your chair's sign-off on the packet.
Then assume one round of revisions, because most first submissions attract one, and put it in the plan rather than in your anxiety. The waiting is not dead time either: recruitment materials, committee scheduling and a literature refresh all proceed happily while a packet sits with reviewers. Where the paperwork itself is the bottleneck, that is precisely what the IRB line is for, and everything in the packet still traces back to the problem you defined at the beginning.
FAQQuestions to control
Does exempt mean I can skip the board?
No. Exempt is a determination the board makes about your study, not a decision you make about your paperwork, and collecting anything before you hold that determination in writing is one of the few errors that can invalidate a completed project. The submission for an exempt route is usually shorter and answered faster, which is the real benefit, and it is a substantial one.
My study only uses records that already exist. Do I still submit?
Almost always yes, and the answer turns on identifiers. Where records are fully de-identified before you ever see them, many boards determine the work exempt or not human subjects research at all. Where you can link a row back to a person, even through a code held elsewhere, the study stays in review. Ask before you request the extract, since how the data is pulled changes the answer.
How long does approval take?
Long enough that you should plan around it rather than hope. The two variables you control are which route your study falls into and how many rounds of revision it takes, and the second one is where candidates lose most of their time. A complete packet answering the board's own template in its own order is the single best thing you can do to shorten the wait.
Can I submit before my chair has finished reading it?
Do not, however tempting the cut-off makes it. Most boards will not accept a student submission without the chair's sign-off, and those that do will return it. A protocol your chair has not read also tends to carry the disagreements between you into the review, where they become the board's questions and cost a full cycle to answer properly.
