Critical Appraisal of Randomised Controlled Trials

Vocabulary and key concepts regarding the critical appraisal of randomised controlled trials, including bias types, the PICOS framework, and statistical significance.

Critical appraisal

The process of carefully and systematically examining research to judge its trustworthiness, value, and relevance in a particular context.

Validity

A dimension of critical appraisal that asks what research question the study addresses and how closely it matches the researcher's own research question.

Quality

A dimension of critical appraisal that evaluates how well a study was conducted.

Results/Relevance

Analysis concerned with what the results mean and the probability that they could have occurred by chance.

Predatory journal

A journal that may spread misinformation or publish content for personal interest and money rather than maintaining high quality and trust.

Industry funding

Sponsorship that creates potential for bias in how or where results are published, as the sponsor may have a specific agenda.

PICOS

A framework for framing a research question consisting of Population, Intervention, Comparator, Outcome, and Setting.

Meaningful clinical outcomes

Outcomes that matter to patients, such as mortality, hospitalization, or reduction in heart attack/stroke risk.

Surrogate outcomes

Easy-to-measure markers, such as lipid levels or $HbA1c$, which may suggest a risk but are not definitive evidence of clinical outcomes like cardiovascular death.

Selection bias

Bias resulting from the allocation of comparison groups where one group ends up older, sicker, or with more comorbidities at baseline.

Performance bias

Bias occurring when some patients receive better lifestyle care, treatment, or monitoring than the control group, influencing results.

Detection bias

Biased assessment of outcomes often caused by assessors knowing which participants are in the control group versus the trial group.

Attrition bias

Bias related to the handling of protocol deviations and loss to follow-up, such as when only those who tolerate a drug stay in a trial.

Randomization

A method used to minimize selection bias by balancing groups on all known and unknown characteristics through a genuinely random process.

Blinding

The practice of keeping participants, investigators, or assessors unaware of which intervention is being administered to remove subjectivity in assessing outcomes.

Sample size

The number of participants in a study; it must be large enough to detect a true difference but not needlessly large to avoid ethical concerns.

Gold standard comparator

A fair comparison treatment used in a study where the new treatment and the existing standard are given at comparable, fair doses.

RAMBO framework

A memorable way to link PICO elements to how bias is minimized in methods: Recruitment (Population), Allocation (Intervention), Maintenance (Comparator), and Measurement (Outcome).

P-value

The probability that a result is due to chance; a value lower than $0.05$ is considered statistically significant.

Confidence intervals

Usually expressed as a $95\%$ range around the point estimate, these indicate the precision of a research result.