Ethics review and consent
The requirements, together
- VALUE: the research must produce knowledge worth having.
- SCIENTIFIC VALIDITY: it must be capable of producing that knowledge. An underpowered or badly designed study is an ethical failure, not only a methodological one.
- FAIR SUBJECT SELECTION: who is asked follows from the scientific question and a fair distribution of risk and benefit, not from who is easy to recruit or unable to refuse.
- FAVOURABLE RISK-BENEFIT: risks minimised, benefits enhanced, and the balance defensible.
- INDEPENDENT REVIEW: by people unaffiliated with the research.
- INFORMED CONSENT: voluntary, informed, and revocable.
- RESPECT FOR PARTICIPANTS: privacy protected, the right to withdraw, welfare monitored, results shared.
Equipoise
Randomizing a patient is defensible only when the expert community is genuinely uncertain which arm is better. That uncertainty is the ethical basis of a trial and is also why a trial should stop when it resolves, which is what interim monitoring is for.
Consent in practice
Consent is a process rather than a form. Comprehension, not signature, is the standard, and readability, time to consider and the ability to ask questions are part of it. Vulnerability, of children, of people who lack capacity, of prisoners, of patients dependent on the investigator for care, calls for additional safeguards rather than exclusion, since exclusion has its own cost in evidence.
Waivers and secondary use
Research on existing records may qualify for a waiver where the risk is minimal, the research could not practicably be done otherwise, and participants' rights are protected. That is a determination for the committee, not the investigator, and the question of what participants would have expected is a legitimate part of it.
Why Mallard raises it
Mallard reads a description for whether human subjects oversight is likely to be required and says so, because discovering after data collection that approval was needed is a common and expensive failure. It is a prompt to ask the committee, never a determination.
What to watch out for
Consent is necessary in most research and is not sufficient in any of it: a study that is scientifically invalid cannot be redeemed by consenting participants, because it exposes them to burden for knowledge that cannot be gained. That makes methodological rigour an ETHICAL requirement rather than only a technical one, which is the framing most consent-focused discussion misses. Note also that consent to treatment and consent to research are different, and a form covering the first does not cover the second.
Clinical example
A quality improvement audit becomes a comparison of two wards with a publication in mind. Whether that requires review depends on the intent and the design, and the question is asked of the committee before the data are analysed, not afterwards.
- What makes clinical research ethical? (opens in a new tab) · JAMA · doi 10.1001/jama.283.20.2701
- World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects (opens in a new tab) · JAMA · doi 10.1001/jama.2013.281053