Why English? | Reporting CONSORT Noninferiority and Equivalence Trials Through Margins, Hypotheses, Populations, Confidence Intervals and Interpretation

ORIENTATION

What this guide owns

Current fact check

The 2012 update replaced the 2006 extension and aligns noninferiority and equivalence reporting with CONSORT 2010. The current CONSORT 2025 statement supersedes CONSORT 2010 for core reporting and advises continued use of existing specialised extensions until they are updated.

Canonical boundary

This article owns reporting of randomised noninferiority and equivalence trials. The existing Gold Standard article owns the broader design theory of noninferiority margins, assay sensitivity and constancy; the CONSORT 2025 article owns generic trial reporting.

Worked example

Every chapter returns to a fictional trial comparing a shorter reading-support programme with an established programme, using a prespecified comprehension-score margin of 3 points where higher scores are better. It is fictional, low-stakes and educational; it does not report a real school, clinical decision or guaranteed outcome.

Warm up with eduKate’s guides to explaining scientific evidence and asking better questions in class.

CHAPTER 1 OF 13

1. Identify the Design in the Title and Abstract

Back to contents

Why this move matters

Name noninferiority or equivalence, the comparator, allocation and primary outcome so readers do not infer superiority.

Write the claim as an inequality

Name noninferiority or equivalence, the comparator, allocation and primary outcome so readers do not infer superiority. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The abstract says the shorter programme was tested for noninferiority to an established programme. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

A generic randomised-trial label hides the different hypothesis and decision boundary. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Identify the Design in the Title and Abstract”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 2 OF 13

2. State the Clinical or Educational Rationale

Back to contents

Why this move matters

Explain the expected secondary advantage—such as lower burden, easier delivery or fewer harms—and why accepting some loss of efficacy might be reasonable.

Write the claim as an inequality

Explain the expected secondary advantage—such as lower burden, easier delivery or fewer harms—and why accepting some loss of efficacy might be reasonable. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The shorter programme reduces class time, but the report does not assume that convenience compensates for any loss. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

A secondary advantage must be documented rather than used as a marketing phrase. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “State the Clinical or Educational Rationale”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 3 OF 13

3. Define and Justify the Noninferiority or Equivalence Margin

Back to contents

Why this move matters

Give the margin on the outcome scale, its direction, evidence base, stakeholder reasoning and relation to the active control effect.

Write the claim as an inequality

Give the margin on the outcome scale, its direction, evidence base, stakeholder reasoning and relation to the active control effect. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: Three points is justified using prior trials, measurement meaning and a conservative preservation argument. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

A margin chosen after results are known cannot support the original claim. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Define and Justify the Noninferiority or Equivalence Margin”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 4 OF 13

4. Write the Hypotheses and Decision Rules Explicitly

Back to contents

Why this move matters

State null and alternative hypotheses, one-sided or two-sided error, confidence level and every possible conclusion.

Write the claim as an inequality

State null and alternative hypotheses, one-sided or two-sided error, confidence level and every possible conclusion. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The protocol states that the lower confidence limit must exceed minus 3 points for noninferiority. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

A nonsignificant superiority test is not a noninferiority test. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Write the Hypotheses and Decision Rules Explicitly”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 5 OF 13

5. Describe Participants, Settings and the Active Control

Back to contents

Why this move matters

Report eligibility, setting, comparator dose or delivery, adherence supports and evidence that the active control should work here.

Write the claim as an inequality

Report eligibility, setting, comparator dose or delivery, adherence supports and evidence that the active control should work here. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The established programme is delivered at its usual intensity by trained teachers. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

A weakly delivered control can make an ineffective new programme look noninferior. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Describe Participants, Settings and the Active Control”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 6 OF 13

6. Specify Outcomes, Time Points and Estimands

Back to contents

Why this move matters

Define the primary outcome, direction of benefit, measurement time, analysis metric, intercurrent events and target population.

Write the claim as an inequality

Define the primary outcome, direction of benefit, measurement time, analysis metric, intercurrent events and target population. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The estimand concerns the 12-week comprehension score under the stated handling of programme discontinuation. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

Changing the time point or intercurrent-event strategy can change the practical meaning of the margin. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Specify Outcomes, Time Points and Estimands”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 7 OF 13

7. Report Sample Size and Error Assumptions

Back to contents

Why this move matters

Give the margin, anticipated difference, variability, power, one-sided alpha or confidence level, attrition and planned analysis population.

Write the claim as an inequality

Give the margin, anticipated difference, variability, power, one-sided alpha or confidence level, attrition and planned analysis population. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: Planning assumes no average difference and tests sensitivity to score variability and withdrawal. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

High power against a generous margin does not make the margin appropriate. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Report Sample Size and Error Assumptions”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 8 OF 13

8. Preserve Randomisation, Blinding and Trial Quality

Back to contents

Why this move matters

Report sequence generation, concealment, masking, adherence, contamination, protocol deviations and similarity to trials supporting the control effect.

Write the claim as an inequality

Report sequence generation, concealment, masking, adherence, contamination, protocol deviations and similarity to trials supporting the control effect. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: Allocation is concealed and outcome assessors remain unaware of programme assignment. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

Poor conduct often biases groups toward similarity, which can favour a noninferiority claim. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Preserve Randomisation, Blinding and Trial Quality”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 9 OF 13

9. Report Intention-to-Treat and Per-Protocol Analyses

Back to contents

Why this move matters

Define both populations, exclusions, missing-data handling and whether conclusions agree, while linking each analysis to the estimand.

Write the claim as an inequality

Define both populations, exclusions, missing-data handling and whether conclusions agree, while linking each analysis to the estimand. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The report presents both analyses and explains why two students are excluded only from the per-protocol set. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

Neither population is automatically conservative in every setting. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Report Intention-to-Treat and Per-Protocol Analyses”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 10 OF 13

10. Display Confidence Intervals Against the Margin

Back to contents

Why this move matters

Give effect direction, point estimate, confidence interval, margin and a figure or table that makes superiority, noninferiority, equivalence and inconclusive regions visible.

Write the claim as an inequality

Give effect direction, point estimate, confidence interval, margin and a figure or table that makes superiority, noninferiority, equivalence and inconclusive regions visible. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: A lower limit of minus 1.2 lies above the minus 3 margin but below zero, supporting noninferiority without superiority. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

Reporting only a p-value or saying no difference hides the decision geometry. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Display Confidence Intervals Against the Margin”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 11 OF 13

11. Report Harms and the Claimed Secondary Advantage

Back to contents

Why this move matters

Present adverse events, burden, acceptability, cost or convenience with denominators and uncertainty rather than assuming the new option is better.

Write the claim as an inequality

Present adverse events, burden, acceptability, cost or convenience with denominators and uncertainty rather than assuming the new option is better. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: Class time saved is measured, while withdrawals and unwanted effects are reported by group. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

A noninferior primary outcome does not establish superior safety, cost or acceptability. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Report Harms and the Claimed Secondary Advantage”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 12 OF 13

12. Discuss Assay Sensitivity and Constancy

Back to contents

Why this move matters

Explain why the active control would have been effective, whether current conduct resembles historical evidence and how uncertainty affects the conclusion.

Write the claim as an inequality

Explain why the active control would have been effective, whether current conduct resembles historical evidence and how uncertainty affects the conclusion. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The discussion compares adherence and participant severity with the trials used to justify the margin. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

Constancy across eras, settings and populations is an assumption, not an observable guarantee. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Discuss Assay Sensitivity and Constancy”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

CHAPTER 13 OF 13

13. Interpret Noninferiority Without Overclaiming

Back to contents

Why this move matters

State exactly what boundary was met, preserve uncertainty and avoid translating one endpoint into universal equivalence.

Write the claim as an inequality

State exactly what boundary was met, preserve uncertainty and avoid translating one endpoint into universal equivalence. State the direction of benefit, the outcome scale and the noninferiority or equivalence margin before reading the result.

Work the confidence interval against the margin

Worked example: The conclusion says the shorter programme was noninferior on the prespecified 12-week score, not equivalent in all respects. Mark zero, the prespecified margin and both confidence-limit positions, then classify the result without substituting a p-value slogan.

Keep both analysis populations visible

Compare intention-to-treat and per-protocol results, explain exclusions and deviations, and investigate disagreement instead of selecting the friendlier analysis.

Protect assay sensitivity and interpretation

Regulatory, clinical and school decisions may weigh margins and secondary advantages differently. A noninferiority conclusion depends on the active control, conduct, constancy assumptions, margin and estimand—not merely on a nonsignificant difference.

Student, parent and career transfer

Make one evidence card for “Interpret Noninferiority Without Overclaiming”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical trials, education research, regulation, health technology assessment, biostatistics and evidence appraisal without promising a grade, admission, funding decision or job.

PRACTICE LAB

Build a noninferiority-margin decision trail

Student route

1. Choose one original report inside this guide’s scope.
2. Copy its objective and underline actor, action, context, evidence and intended decision.
3. Draw the specific reasoning chain used in this article.
4. Mark each step reported, inferred, contested or unknown.
5. Rewrite the conclusion so it preserves the narrowest important limitation.

Parent or teacher route

Ask the learner to explain CONSORT Noninferiority and Equivalence’s central reporting method twice: once to a classmate and once to a method specialist. Praise a precise correction and a well-stated unknown; both show disciplined attention to evidence.

Early-career route

Create a one-page CONSORT Noninferiority and Equivalence audit with source link, date checked, scope, method logic, decisive wording, contrary evidence, limit and next action. A colleague should be able to continue without guessing.

Continue with eduKate’s guides to reading course and career information and giving useful feedback in group projects.

USEFUL NEXT READING

Connect this narrow owner to its neighbour

Read the CONSORT 2025 trial-reporting guide for a related but different reader job. Return here when the question is specifically how to report “not unacceptably worse” without turning absence of evidence into equivalence. This boundary keeps one canonical owner per intent.

For English practice across research, school and career communication, visit the English at work and in public life guide.

FREQUENTLY ASKED QUESTIONS

Questions students and parents often ask

Is “no significant difference” the same as noninferiority?

No. Noninferiority needs a prespecified margin and a confidence interval assessed against it.

What is the margin?

The largest prespecified loss considered acceptable for the stated outcome and context.

Why report two analysis populations?

Protocol deviations and nonadherence can affect noninferiority conclusions; consistency and discrepancies are informative.

Can a trial show both noninferiority and superiority?

Yes, if the prespecified analysis and confidence interval also support superiority.

Does noninferior mean identical?

No. It means the result excludes a loss larger than the chosen margin under the stated assumptions.

Which CONSORT version applies now?

Use the current CONSORT 2025 core with the existing specialised extension until an aligned update is published.