Vulnerable Populations Under the Common Rule

The Common Rule (45 CFR 46) does not treat all research subjects the same. Under §46.111(b), an IRB cannot approve a study unless it determines that, wherever subjects are “likely to be vulnerable to coercion or undue influence,” additional safeguards have been built into the protocol. For three specific populations — pregnant women, human fetuses, and neonates; prisoners; and children — those safeguards aren’t left to case-by-case IRB judgment alone. They’re spelled out in dedicated regulatory subparts with specific, mandatory requirements. This guide is the categorical reference: which populations trigger additional protection, which regulatory subpart (if any) governs each one, and exactly what that subpart requires an IRB and investigator to do before enrollment can proceed.

This is a companion to, not a replacement for, two other CASRAI resources that cover adjacent ground from a different angle. Undue Influence and Vulnerable Populations in Research focuses on the mechanics of coercion, undue influence, and undue inducement — the payment-structure and power-relationship questions that can arise with any subject, vulnerable-category or not. Vulnerable Populations in Research is the glossary-level definition. This page is the practical, category-by-category compliance reference: for each population, what protection is legally required, not just why the concept exists.

The Regulatory Basis: 45 CFR 46.111(b)

Every IRB approval decision runs through the general criteria at 45 CFR 46.111(a) first — risks minimized, risks reasonable relative to benefits, equitable subject selection, informed consent obtained and documented, and (where appropriate) adequate provisions for monitoring and privacy. Section 46.111(b) adds a further condition that applies specifically to vulnerable subjects:

“When some or all of the subjects are likely to be vulnerable to coercion or undue influence, such as children, prisoners, individuals with impaired decision-making capacity, or economically or educationally disadvantaged persons, additional safeguards have been included in the study to protect the rights and welfare of these subjects.”

Two things about this text drive how it actually gets applied in IRB review:

The list is illustrative (“such as”), not exhaustive. An IRB can and does treat other subjects as vulnerable under §46.111(b) even when they don’t appear in the regulatory text and aren’t covered by a dedicated subpart.
Only three of the named categories — pregnant women/fetuses/neonates, prisoners, and children — have a dedicated numbered subpart spelling out mandatory, specific additional protections. The rest are addressed through the same general “additional safeguards” requirement, evaluated study-by-study, without a fixed regulatory checklist.

Categories at a Glance

Category	Regulatory basis	Core additional protection required
Pregnant women, human fetuses, neonates	Subpart B, §§46.201–46.207	Risk-to-fetus conditions; restrictions on inducements to terminate pregnancy; father’s consent where research offers direct benefit to the fetus; separate viable/nonviable-neonate pathways
Prisoners	Subpart C, §§46.301–46.306	IRB must include a prisoner representative and a non-affiliated majority; research restricted to four permitted categories under §46.306; parole-related advantages can’t be so large they impair voluntary choice
Children	Subpart D, §§46.401–46.409	Protocol classified into one of four risk/benefit categories (§§46.404–46.407); dual-consent structure — parental/guardian permission plus the child’s own assent under §46.408
Individuals with impaired decision-making capacity	§46.111(b) (no dedicated subpart)	IRB-determined additional safeguards case-by-case — e.g., capacity assessment, surrogate/legally authorized representative consent, simplified consent materials
Economically or educationally disadvantaged persons	§46.111(b) (no dedicated subpart)	IRB-determined additional safeguards — e.g., scrutiny of payment size/structure, plain-language consent, attention to whether participation is effectively non-voluntary given the subject’s circumstances
Deferential vulnerability (students, employees, patients recruited by an instructor/employer/treating clinician)	§46.111(b), research-ethics literature (no dedicated subpart)	IRB-determined safeguards addressing the power relationship itself — e.g., third-party recruitment, no penalty for non-participation, recruiter not also the evaluator/supervisor

Subpart B: Pregnant Women, Human Fetuses, and Neonates

Subpart B (§§46.201–46.207) applies once a subject falls into one of its precisely defined categories — pregnancy (implantation to delivery), fetus, neonate, and the viable/nonviable distinction for neonates specifically. The additional protections layer on top of, not instead of, the general §46.111 criteria:

§46.204 (pregnant women/fetuses): requires prior animal and non-pregnant-subject studies; risk to the fetus must either arise solely from interventions offering direct benefit to the woman or fetus, or be minimized to no more than minimal; no inducements to terminate a pregnancy are permitted; research staff cannot be involved in determining neonatal viability or pregnancy-termination decisions; and where the research offers direct benefit to the fetus, both the pregnant woman and the father must generally consent (with defined exceptions — father unreasonably unavailable, unascertainable, or the pregnancy resulting from rape or incest).
§46.205 (neonates): splits into neonates of uncertain viability (research must offer a reasonable possibility of enhancing survival to viability, with risk minimized to that end) and nonviable neonates (permitted only where vital functions won’t be artificially maintained, the research won’t hasten death, and the purpose is knowledge not obtainable another way). Once a neonate is determined viable, Subpart D governs instead.
§46.206 (placenta, dead fetus, fetal material): requires compliance with applicable state/local law on procurement and use, plus consent controls on identifiable information.
§46.207: a narrow research-not-otherwise-approvable pathway, structured like the parallel Subpart D provision below.

Full statutory detail: 45 CFR 46 Subpart B.

Subpart C: Prisoners

Subpart C (§§46.301–46.306) applies based on confinement status at the time of the research — “prisoner” under §46.303(c) covers anyone involuntarily confined or detained, including pretrial detainees and those held under civil-commitment alternatives to incarceration, not only convicted inmates. Its additional protections are structural, not just procedural:

IRB composition (§46.304): a majority of the reviewing board, excluding prisoner members, must have no affiliation with the prison(s) involved beyond board membership, and at least one member must be a prisoner or a qualified prisoner representative.
Additional review duties (§46.305): the IRB must find that any parole-related or other advantage to participating isn’t so large it impairs a prisoner’s ability to weigh risk against benefit in the constrained prison environment, that risks are commensurate with those accepted by non-prisoner volunteers, and that subject selection within the institution is fair and not left to correctional staff discretion.
Permitted research categories (§46.306): even after composition and duty requirements are met, prisoner research is restricted to four categories — study of incarceration/criminal behavior, study of prisons as institutions, research on conditions particularly affecting prisoners as a class (requires HHS Secretary consultation and Federal Register notice), and research intended to improve the individual subject’s health.

Full statutory detail: 45 CFR 46 Subpart C.

Subpart D: Children

Subpart D (§§46.401–46.409) applies to persons who haven’t reached the legal age of consent for the relevant procedures under the jurisdiction where the research occurs — a legal-capacity test, not a fixed number, so the same chronological age can trigger Subpart D in one state and not another. Every protocol must be classified into one of four IRB-determined categories before approval:

§46.404: no greater than minimal risk.
§46.405: greater than minimal risk, but justified by prospect of direct benefit to the subject.
§46.406: greater than minimal risk, no prospect of direct benefit, but likely to yield generalizable knowledge about the subjects’ disorder or condition — requires the risk be only a minor increase over minimal risk, among other findings.
§46.407: a narrow escape valve for research that doesn’t fit the categories above, requiring HHS Secretary review, expert consultation, and public comment — rarely invoked.

On top of the risk classification, §46.408 requires a dual-consent structure: the child’s own assent (affirmative agreement, not mere non-objection) wherever the IRB judges the child capable of providing it, plus parental or guardian permission — one parent’s permission suffices for §46.404/§46.405 research, but both parents must generally consent for §46.406/§46.407 research, subject to defined exceptions. The IRB can waive assent in narrow circumstances but cannot waive parental permission on that same basis.

Full statutory detail: 45 CFR 46 Subpart D.

Vulnerability Without a Dedicated Subpart

Three further categories appear in Common Rule guidance and research-ethics practice but have no numbered subpart of their own. For these, the IRB determines what “additional safeguards” means on a study-by-study basis under the general §46.111(b) standard, rather than following a fixed statutory checklist:

Individuals with impaired decision-making capacity — named directly in §46.111(b). Typical safeguards include a formal capacity assessment tied to the specific decision at hand (capacity isn’t all-or-nothing), consent from a legally authorized representative where the subject can’t consent independently, simplified or supplemented consent materials, and re-assessment where capacity may fluctuate.
Economically or educationally disadvantaged persons — also named directly in §46.111(b). Safeguards commonly focus on whether a payment amount or structure could function as an undue inducement given the subject’s financial circumstances, and on whether consent materials are genuinely comprehensible given literacy or educational background, rather than merely translated or simplified in form only.
Deferential vulnerability — a situational vulnerability arising from a power or authority relationship between researcher and subject (an instructor recruiting their own students, an employer recruiting employees, a treating clinician recruiting their own patients) rather than from an inherent trait of the subject. See Deferential Vulnerability for the full treatment; typical safeguards include third-party recruitment, explicit assurance that declining carries no penalty, and separating the recruiter role from any evaluative or supervisory role over the subject.

See Vulnerable Populations in Research for the glossary-level synthesis of how these categories relate to the three codified subparts.

Determining Which Protections Apply to a Protocol

In practice, an investigator or research-administration reviewer working through a new protocol should check, in order:

Does the subject population fall into a Subpart B, C, or D category as legally defined — not as a matter of general topic, but by the subject’s actual status at the time of the research (pregnant, confined, under the local age of consent)? If yes, the applicable subpart’s specific findings and consent structure are mandatory, not discretionary.
Does the population fall into one of the §46.111(b) categories without a dedicated subpart — impaired decision-making capacity, economic or educational disadvantage, or a deferential/power-relationship dynamic with the research team? If yes, the IRB must document specific additional safeguards, tailored to the actual source of vulnerability, before approval.
Is more than one category present at once — e.g., an incarcerated pregnant subject, or a cognitively impaired child? Where categories overlap, all applicable protections apply cumulatively; one subpart’s requirements don’t substitute for another’s.
Does the protocol’s recruitment or payment design independently raise undue-influence or inducement concerns, separate from subject-category vulnerability? This is evaluated under §46.111(b) and §46.116(a)(8) regardless of which population category applies — see Undue Influence and Vulnerable Populations in Research for that analysis specifically.

Frequently Asked Questions

Do all four vulnerable-population categories in 45 CFR 46.111(b) have their own subpart?

No. Only pregnant women/fetuses/neonates (Subpart B), prisoners (Subpart C), and children (Subpart D) have a dedicated, numbered subpart with specific mandatory requirements. Individuals with impaired decision-making capacity and economically or educationally disadvantaged persons are named in §46.111(b) but are addressed through the same general “additional safeguards” requirement, decided by the IRB study-by-study.

Can an IRB treat a population as vulnerable even if it isn’t named in 45 CFR 46 at all?

Yes. The §46.111(b) list is introduced with “such as,” making it illustrative rather than exhaustive. Deferential vulnerability — arising from a power relationship like instructor/student or employer/employee — is a recognized example that IRBs apply under the general standard even though it doesn’t appear in the regulatory text by name.

Does Subpart D apply to every study that includes a subject under 18?

Subpart D turns on the legal age of consent for the specific procedure in the jurisdiction where the research occurs, not a universal age of 18 — the age at which a minor can independently consent to particular procedures varies by state and by procedure type, so the same chronological age can trigger Subpart D in one state and not another.

What happens if a neonate’s viability status changes during a study?

Once a neonate is determined viable, it is no longer covered by Subpart B’s neonate provisions (§46.205) and instead falls under Subpart D (or Subpart A generally, if Subpart D doesn’t independently apply) as a child subject.

Is a prisoner representative required on every IRB, or only when reviewing prisoner research?

Only when the IRB is reviewing research covered by Subpart C. §46.304 requires that board (or, where more than one IRB reviews the same project, at least one of them) to include a prisoner or qualified prisoner representative and a non-affiliated majority — it is not a standing requirement for every IRB regardless of what it reviews.