TL;DR: Whether a project is “human subjects research” is a formal, documented determination made by an authorized official at your institution—usually the IRB or Human Research Protection Program (HRPP) office—not a judgment call an investigator makes alone. The practical process almost always follows the same sequence: run your project through an institutional self-screening tool, submit for an official determination when the answer isn’t obvious, receive a written outcome (not research, research-but-not-human-subjects, exempt, or requires IRB review), and keep that determination on file. This guide walks through that process step by step. For the underlying regulatory definition itself, see Human Subjects Research: Definition Under 45 CFR 46.102; for the review pathways once a project is confirmed to need IRB oversight, see IRB/REC Approval Process.
Why this is a separate step from “applying for IRB approval”
A common mistake is treating IRB submission as the first step in a research project involving people. It isn’t. The Common Rule (45 CFR 46) only applies once an activity clears two conditions at 45 CFR 46.102(l) and 46.102(e)(1): it must be “research” (a systematic investigation designed to develop or contribute to generalizable knowledge) and it must involve a “human subject” (a living individual from whom an investigator obtains data via intervention/interaction, or obtains, uses, studies, analyzes, or generates identifiable private information or identifiable biospecimens). Many projects that feel like they involve real people—program evaluations, case reports, some quality improvement (QI) work, internal institutional assessments—never actually meet that definition, and therefore never enter the IRB review pipeline at all. Making that determination correctly, and early, saves weeks of unnecessary paperwork on projects that don’t need it and prevents a much bigger problem: proceeding with a project that does need review without ever getting it.
Step 1: Understand what’s actually being tested
Before screening a specific project, it helps to know that the determination is really two separate yes/no questions, not one overall impression:
- Is it research? Is the primary intent to develop or contribute to generalizable knowledge—findings intended to apply beyond the specific people, unit, or site involved—rather than to improve a local process or evaluate a specific program for internal use?
- Does it involve a human subject? Will data or biospecimens be obtained through direct intervention or interaction with living individuals, or will identifiable private information or identifiable biospecimens about living individuals be obtained, used, studied, analyzed, or generated?
Both have to be “yes” for the Common Rule to apply. A project can be rigorously systematic and still fall outside human subjects research if it never touches identifiable data about a living person (for example, a purely bibliometric or document-analysis study). A project can also clearly involve real people and still not be “research” if it’s designed only to improve one clinic’s local workflow with no intent to generalize the findings. The full mechanics of this two-part test, including how it plays out for surveys, secondary data, and deceased individuals, are covered in the definition guide — this page focuses on the practical workflow for getting a project through that test.
Step 2: Run the project through your institution’s self-screening tool
Nearly every institution with a Federalwide Assurance runs a Human Research Protection Program that publishes a self-screening worksheet, decision tree, or intake questionnaire for exactly this question — usually available on the IRB office’s website. These tools typically walk through:
- What the project’s primary purpose is (generate generalizable knowledge vs. improve a specific local process)
- Whether any data will be obtained through direct interaction or intervention with living people (interviews, surveys, physical procedures, behavioral manipulation)
- Whether any identifiable private information or identifiable biospecimens will be obtained, even if the investigator never interacts with the person directly (e.g., pulling identifiable records from an existing database)
- Whether the data will ever be reported, analyzed, or shared in a way that includes identifiers
Self-screening is a genuinely useful first pass, and for clearly non-research activities (an internal chart audit used only to fix one clinic’s scheduling, with no intent to publish or generalize) it may be enough on its own. But self-screening is not the same as an official determination, and most institutional policies are explicit that investigators cannot self-certify a “not human subjects research” finding for anything ambiguous — see Step 3.
Step 3: Submit for an official determination when the answer isn’t obvious
If the self-screening tool flags any uncertainty — or if the project involves identifiable data, a comparison condition, an intent to publish or present findings beyond the local setting, or external funding — the next step is to submit the project (often via the same intake form) to the IRB or HRPP office for a formal determination. This is a documented review of the project’s description against the 46.102(l)/(e) test, made by someone formally authorized to make that call on the institution’s behalf. It is not the investigator’s own judgment, and it is not a determination made by reading a reference page like this one. Reviewers typically ask for:
- A short written description of the project’s purpose and design
- What data or specimens will be collected, from whom, and whether they will be identifiable
- Whether results are intended for publication, presentation, or any use beyond the immediate local setting
- Funding source, since some federal or foundation funders impose their own human-subjects requirements independent of the institution’s own Common Rule determination
Turnaround for a determination-only submission (as opposed to a full IRB review) is usually much faster than a full protocol review, since no risk/benefit or consent analysis is being done — only the threshold question of whether the Common Rule applies at all.
Step 4: Understand the four possible outcomes
A determination request typically resolves into one of four outcomes, and knowing the difference between them matters for what happens next:
- Not research. The activity doesn’t meet the 46.102(l) definition (e.g., a genuine QI project intended only to improve local practice). No IRB involvement is required regardless of whether identifiable data is used. See Quality Improvement vs. Human Subjects Research for the specific factors used to make this call, since it’s one of the most common sources of confusion.
- Research, but not human subjects research (NHSR). The activity is genuinely research but never obtains data, biospecimens, or identifiable private information about a living individual (for example, research using only pre-existing, fully de-identified aggregate data, or research on documents/artifacts rather than people). The Common Rule doesn’t apply.
- Human subjects research, but exempt. The activity meets both prongs of the definition but falls into one of the exempt categories at 45 CFR 46.104 (certain educational-setting research, certain benign behavioral interventions, some secondary research on existing identifiable data, and others). Exempt research still requires an institutional determination that it qualifies for exemption — it is not self-determined, and it is a different finding than “not human subjects research.” A common example is a published clinical case report, which many institutions route through an exemption or NHSR pathway depending on how it’s structured; see Case Report IRB Exemption.
- Human subjects research requiring IRB review. The activity meets both prongs and doesn’t qualify for exemption. It proceeds to expedited or full-board review. See IRB/REC Approval Process for what happens from here, including realistic timelines.
Step 5: Get it in writing and keep it on file
Whatever the outcome, the determination should come back as a written record — a determination letter, an approval notice, or a completed and signed worksheet — not a verbal answer from a colleague or a personal reading of the regulation. Keep that document with the project file. It matters for several practical reasons beyond compliance itself: journals and funders increasingly ask for documentation of an IRB determination (including a “not human subjects research” finding) before accepting a manuscript or awarding funds; a determination made at the start of a project doesn’t automatically cover a later expansion of scope (see below); and if a funder or journal later asks “was this reviewed by an IRB,” “no, because it didn’t need to be” is only a defensible answer if there’s a documented determination saying so.
Special case: multi-site and collaborative projects
Determining that a project is human subjects research is only half the picture when more than one institution is involved. OHRP’s Engagement of Institutions in Human Subjects Research guidance (finalized 2008) addresses a related but distinct question: once a project is confirmed to be non-exempt human subjects research, which of the participating institutions are actually “engaged” in it, and therefore need their own Federalwide Assurance and IRB approval (or a reliance arrangement)? Under that guidance, an institution is generally considered engaged when its own employees or agents, for purposes of the project, obtain data about subjects through direct intervention or interaction, obtain identifiable private information about subjects, or obtain subjects’ informed consent. A collaborating site that only receives a fully de-identified dataset after the fact, with no role in recruitment, consent, or data collection, is often not engaged and may not need its own IRB approval — but this is a separate determination from the underlying human-subjects-research question and should be made explicitly, not assumed, especially for federally funded multi-site studies. Institutions using a single IRB of record for multi-site work typically resolve engagement and reliance together through a reliance agreement; see the Common Rule overview for how these pieces fit together.
What happens if you skip this step
Proceeding with data collection before obtaining a determination — on the assumption that a project “obviously” doesn’t need review — is one of the more common sources of IRB noncompliance findings, and it isn’t a purely theoretical risk: it can mean data collected before approval has to be excluded from analysis, a manuscript gets held or rejected at the journal’s compliance check, a funder flags the award for a human-subjects compliance issue, or the institution has to report the lapse to OHRP as a matter of Federalwide Assurance compliance. Getting the determination first, even for a project the investigator is confident doesn’t need review, is substantially cheaper than untangling a compliance problem after data collection has already started.
Project types people frequently get wrong
A few categories account for a disproportionate share of determination confusion, and are worth a second look before assuming an answer:
- Quality improvement and program evaluation — intent to publish alone doesn’t make a project research, but a design that departs from accepted practice, includes a comparison condition, or is explicitly intended to generalize beyond the local setting often does. See the dedicated QI vs. human subjects research guide.
- Case reports and case series — a single patient case report is frequently exempt or NHSR; a case series systematically assembled to identify a generalizable pattern can cross into research. See Case Report IRB Exemption.
- Secondary analysis of existing data — whether this needs review depends heavily on whether the dataset is identifiable and how it was originally collected, not simply on the fact that it’s “already collected.”
- Student course projects and theses — many institutions treat classroom exercises using only classmates or hypothetical scenarios as non-research for pedagogical purposes, but a thesis or dissertation intended to be published or deposited in a repository is usually treated as research and needs its own determination, independent of any course-level exemption.
Frequently asked questions
Who actually makes the human subjects research determination?
An institution’s IRB, HRPP, or research compliance office — or another individual formally delegated that authority under the institution’s written policies. It is not the investigator’s own judgment, a mentor’s informal opinion, or a general reference page. Most institutions require the determination to be documented, even when the answer is “not human subjects research.”
Can I decide on my own that my project doesn’t need IRB review?
Generally no. Nearly all institutional HRPP policies specify that investigators may use a self-screening tool to prepare, but the actual determination — including a “not research” or “not human subjects” finding — must be made and documented by an authorized reviewer, not self-certified by the project team.
Is a “not human subjects research” determination the same as an exemption?
No. An exemption (45 CFR 46.104) presumes the activity is human subjects research and finds that it qualifies for one of the regulation’s exempt categories. A “not human subjects research” (NHSR) finding means the 46.102(l)/(e) two-part test was never met in the first place — the activity isn’t research, doesn’t involve a human subject, or both. Both are documented determinations, but they rest on different findings.
Does the determination cover the whole life of the project?
Only as originally scoped. A material change — adding a new data source, expanding to a new site, adding identifiable data collection that wasn’t in the original description — generally requires going back to the IRB/HRPP office for a new or amended determination rather than relying on the original one.
What if my institution doesn’t have a formal self-screening tool?
Contact the IRB or research compliance office directly and describe the project; most offices will make the determination through a short intake conversation or email exchange even without a published worksheet. Smaller institutions without their own IRB often rely on an external or commercial IRB, or a reliance agreement with a partner institution’s IRB, for this function.
Related CASRAI resources
- Human Subjects Research: Definition Under 45 CFR 46.102 — the regulatory two-part test in full detail
- Quality Improvement vs. Human Subjects Research — the specific factors that distinguish QI from research
- IRB/REC Approval Process — review pathways and timelines once a project needs IRB oversight
- Case Report IRB Exemption
- Exempt Human Subjects Research
- Research (45 CFR 46.102(l))
- Human Research Participant (Human Subject)
- Common Rule (45 CFR 46)
- HRPP Self-Assessment Before Applying for AAHRPP Accreditation
