Skip to content
All lessons
Cross-cutting · Medical Ethics

Research Ethics & the IRB

How the USMLE tests research ethics: IRB approval before any human-subjects study begins, the Belmont Report's three principles, informed consent and equipoise, and the next-best-step calls on placebo use, vulnerable populations, and conflicts of interest.

11 min readHigh yield

Research ethics governs experiments involving human subjects, and the Institutional Review Board (IRB) is its gatekeeper. Modern rules grew directly out of abuse: the Tuskegee Syphilis Study (1932–1972), in which penicillin was deliberately withheld from Black men with syphilis, prompted the National Research Act (1974) and the Belmont Report (1979). Earlier, Nazi experiments produced the Nuremberg Code (1947) and later the WMA's Declaration of Helsinki (1964).

Every institution conducting human research convenes an IRB (≥5 members of mixed background, including at least one scientist, one non-scientist, and one member unaffiliated with the institution). Its mandate: protect subjects' rights, safety, and welfare by weighing scientific value against risk. On the exam the recurring rule is simple — no human-subjects research may begin until the IRB approves the protocol, and the investigator may not self-declare a study 'exempt.'

High-yield
  • IRB approval is required BEFORE enrolling any subject — the most common next-best-step answer
  • Belmont Report — 3 principles: Respect for persons (autonomy/consent), Beneficence (favorable risk–benefit), Justice (fair subject selection)
  • Informed consent: capacity, full disclosure of risks/benefits/alternatives, voluntary, and the right to withdraw at any time without penalty to clinical care
  • Equipoise — genuine uncertainty over which arm is superior — is required to ethically run an RCT
  • Placebo is ethical only when no proven effective therapy exists; otherwise use an active control or add-on design
  • Vulnerable populations (children, prisoners, pregnant women, cognitively impaired, economically disadvantaged) require extra safeguards
  • DSMB monitors ongoing trials and can stop early for harm, overwhelming benefit, or futility
  • Conflict of interest (financial ties, per-patient payments to investigators) must be disclosed and managed by the IRB; separately, subject compensation must not be excessive (undue inducement)
  • The investigator submits to the IRB, which determines exempt/expedited/full review — never the investigator alone
Clinical correlation

Vignette: A resident designs a study randomizing clinic patients to a new antihypertensive vs standard therapy and begins enrolling them during routine visits. Next best step?Stop; obtain IRB approval before enrolling any subject. Good intentions and even 'minimal risk' do not exempt a study from prior review.

Vignette: A trial participant says she no longer wishes to continue and fears her doctor will be angry. Next step?Honor the withdrawal immediately, without penalty, and reassure her that her regular clinical care is unaffected.

Vignette: During a research brain MRI, a clinically significant incidental mass is found. Next step?Disclose the incidental finding to the participant and arrange appropriate clinical follow-up.

Belmont principleCore ideaHow it's applied
Respect for personsHonor autonomy; protect those with diminished autonomyInformed consent; child assent + parental permission; safeguards for vulnerable groups
BeneficenceMaximize benefit, minimize harmFavorable risk–benefit ratio; DSMB oversight; stop harmful trials
JusticeDistribute burdens and benefits fairlyEquitable subject selection; don't exploit the poor or vulnerable; ensure access to resulting benefits
Mnemonic

Belmont = 3 principles. Recall them as the three questions an IRB asks about a protocol:

  • Respect for persons'Did the subject choose freely?' (consent, assent, withdrawal)
  • Beneficence'Do the benefits outweigh the harms?' (risk–benefit, DSMB)
  • Justice'Is the burden shared fairly?' (who is recruited, who benefits)

Historical chain: Tuskegee → National Research Act (1974) → Belmont Report (1979). Abuse drove the rules — the same reasoning tests the why behind each safeguard.

Clinical correlation

Vignette: A pharmaceutical sponsor offers an investigator $5,000 per patient enrolled. Concern / next step? → This per-capita payment is a financial conflict of interest — it creates an incentive to over-enroll or enroll ineligible patients and can bias judgment. Disclose to the IRB and participants; the IRB manages or limits such arrangements. (Excessive payments to subjects would instead raise undue inducement.)

Vignette: Investigators want to enroll 8-year-olds in a leukemia drug trial. Requirement?Parental (guardian) permission PLUS the child's assent. Assent may be waived only when the trial offers a prospect of direct benefit available only through the research.

Vignette: An emergency trial studies a drug in unconscious cardiac-arrest patients who cannot consent. How is this permitted?Exception from informed consent for emergency research — life-threatening condition, current treatments unproven/unsatisfactory, narrow therapeutic window, plus community consultation and public disclosure.

IRB review levelRiskExamples
ExemptMinimal; the IRB (not the investigator) confirms exemptionAnonymous surveys, de-identified existing data, standard educational research
ExpeditedNo more than minimal riskMinor protocol changes; small blood draws; noninvasive procedures
Full boardGreater than minimal risk, or vulnerable subjectsNew drug/device trials; children/prisoners; sensitive identifiable data

Practice Medical Ethics now

Board-style questions, spaced-repetition flashcards, and a Socratic AI tutor — free to start.