FFICM Research and Development 3 — Questions and Answers
Question 1: Which ethical principle underpins the requirement that ICU research participants (or their surrogates) must be told about risks, benefits, and alternatives before enrolling?
- Beneficence
- Justice
- Autonomy (Correct answer)
- Non-maleficence
Correct answer: Autonomy
Autonomy requires that individuals have the right to make informed decisions about participation in research, forming the ethical basis for informed consent.
Question 2: A multicentre ICU trial reports an odds ratio of 0.72 (95% CI 0.55–0.94) for 30-day mortality with a new intervention. What is the correct interpretation?
- The intervention increases mortality by 28%
- The intervention is associated with a statistically significant reduction in the odds of mortality (Correct answer)
- The result is not statistically significant because the OR is less than 1
- The confidence interval crossing 0.72 confirms significance
Correct answer: The intervention is associated with a statistically significant reduction in the odds of mortality
An OR of 0.72 with a 95% CI that does not cross 1.0 (0.55–0.94) indicates a statistically significant 28% reduction in the odds of mortality.
Question 3: When designing a crossover trial in critically ill patients, which major limitation must be considered compared to a parallel-group design?
- Higher statistical power per patient
- Carryover effects from the first treatment period (Correct answer)
- Reduced need for randomisation
- Elimination of inter-patient variability
Correct answer: Carryover effects from the first treatment period
Carryover (residual) effects occur when the first treatment continues to influence outcomes during the second treatment period, potentially biasing results in crossover designs.
Question 4: In critical care, surrogate outcomes (e.g., PaO2/FiO2 ratio) are used in trials. What is the main risk of using surrogate rather than patient-centred outcomes?
- They are always harder to measure
- Improvement in the surrogate may not translate to benefit in patient-important outcomes (Correct answer)
- They require larger sample sizes
- They violate the Declaration of Helsinki
Correct answer: Improvement in the surrogate may not translate to benefit in patient-important outcomes
Surrogates may not capture the full spectrum of patient benefit or harm; a drug improving oxygenation does not necessarily reduce mortality or improve quality of life.
Question 5: A researcher calculates the number needed to treat (NNT) from a trial where 30% of controls and 20% of treatment patients died. What is the NNT?
- 5
- 10 (Correct answer)
- 3
- 20
Correct answer: 10
NNT = 1 / absolute risk reduction = 1 / (0.30 − 0.20) = 1 / 0.10 = 10.
Question 6: Which type of bias is introduced in a retrospective ICU cohort study when sicker patients are less likely to have complete records due to early death?
- Hawthorne effect
- Survivorship bias (Correct answer)
- Observer bias
- Lead-time bias
Correct answer: Survivorship bias
Survivorship bias occurs when those who did not survive (or complete follow-up) are systematically excluded, making the surviving cohort appear healthier than the true population.
Question 7: In a Bayesian adaptive platform trial for critical care, what does the posterior probability represent?
- The probability of an event given no prior information
- The updated probability of a hypothesis after incorporating prior knowledge and observed data (Correct answer)
- The p-value converted to a probability scale
- The likelihood ratio of positive to negative findings
Correct answer: The updated probability of a hypothesis after incorporating prior knowledge and observed data
The posterior probability combines the prior distribution (pre-existing belief) with the likelihood of the observed data to yield an updated probability of the hypothesis being true.
FFICM Research and Development 3
Which ethical principle underpins the requirement that ICU research participants (or their surrogates) must be told about risks, benefits, and alternatives before enrolling?