A noninferiority trial is a type of clinical study designed to determine whether one treatment is not meaningfully less effective than another established treatment.
Instead of asking whether a new treatment is better, researchers investigate whether any reduction in effectiveness is small enough to fall within a predetermined acceptable limit.
This approach is particularly relevant to Eye Movement Desensitization and Reprocessing (EMDR) research because researchers often compare EMDR with other established psychological treatments, including cognitive behavioral therapy.
Noninferiority trials can also help evaluate different ways of delivering EMDR, such as online versus in-person sessions.
Understanding this study design makes it easier to interpret findings that describe treatments as comparable, similarly effective, or noninferior.
What Does Noninferiority Mean in Clinical Research?
In clinical research, noninferiority means that a treatment has been shown not to be worse than a comparison treatment by more than a specified amount.
Researchers establish this acceptable difference before analyzing the trial results.
For example, imagine researchers want to compare EMDR with an established treatment for post-traumatic stress disorder (PTSD).
Both treatments have evidence supporting their effectiveness, but researchers want to know whether EMDR produces outcomes that are sufficiently close to those of the established treatment.
Rather than requiring EMDR to demonstrate better results, the researchers define the largest difference they would consider acceptable.
If the trial provides sufficiently strong evidence that EMDR does not fall below this limit, the researchers may conclude that it is noninferior to the comparison treatment.
Noninferior does not mean identical. It means that the study has ruled out a difference larger than the predetermined acceptable amount, within the limits of its statistical analysis.
How Is a Noninferiority Trial Different From a Superiority Trial?
Most people are familiar with research that compares treatments to determine which one works better.
This is known as a superiority trial.
A noninferiority trial asks a different question.
| Study design | Main research question |
|---|---|
| Superiority trial | Is one treatment more effective than another? |
| Noninferiority trial | Is the new treatment not meaningfully worse than the established treatment? |
| Equivalence trial | Are the treatments sufficiently similar within a predefined range in both directions? |
Superiority trials
A superiority trial investigates whether one intervention produces better outcomes than another.
For example, researchers might examine whether EMDR reduces PTSD symptoms more than a comparison treatment.
If the evidence demonstrates a sufficiently reliable difference favoring EMDR, the researchers may conclude that EMDR was superior for the outcome studied.
Noninferiority trials
A noninferiority trial does not require the new treatment to perform better.
Instead, the study tests whether the new treatment avoids a clinically important loss of effectiveness.
This can be valuable when the new approach may offer other advantages, such as improved availability or easier access.
Equivalence trials
An equivalence trial investigates whether two treatments have effects that fall within an acceptable range of differences in either direction.
This is different from noninferiority, which primarily rules out the possibility that the new treatment is too much worse than the comparison treatment.
These distinctions matter because the same set of treatment outcomes may lead to different conclusions depending on the study design.
How Does a Noninferiority Trial Work?
A well-designed noninferiority trial involves several important steps.
1. Selecting an established comparison treatment
Researchers begin by identifying a treatment with reliable evidence supporting its effectiveness.
This is called the reference treatment or active control.
For example, a study might compare EMDR with cognitive behavioral therapy for a condition where cognitive behavioral therapy is already an established treatment.
Using an effective comparison treatment helps researchers investigate whether the alternative preserves an acceptable level of treatment benefit.
2. Choosing the main outcome
Researchers identify the outcome they want to compare.
In EMDR research, this might include:
Changes in PTSD symptom severity.
Changes in anxiety or panic symptoms.
Depression symptom scores.
The proportion of participants reaching remission.
Improvements in daily functioning or quality of life.
The outcome needs to be defined clearly because different measures can produce different conclusions.
3. Setting the noninferiority margin
The noninferiority margin is the largest difference researchers are willing to accept between the new treatment and the established treatment.
This margin is a central feature of the study design.
For example, if a PTSD symptom questionnaire uses lower scores to represent fewer symptoms, researchers might decide that a difference of five points is the largest acceptable disadvantage for the new treatment.
The five-point value in this example is hypothetical. A real study would need to justify its margin using relevant clinical and statistical evidence.
4. Randomly assigning participants
Participants are generally assigned at random to the different treatment groups.
Randomization helps reduce the possibility that preexisting differences between participants explain the results.
5. Comparing outcomes statistically
After treatment, researchers compare the groups using the planned statistical methods.
They examine whether the results provide enough evidence to rule out a treatment difference greater than the noninferiority margin.
The conclusion depends on both the estimated treatment difference and the uncertainty around that estimate.
What Is a Noninferiority Margin?
The noninferiority margin determines how much less effective a new treatment could be before researchers would consider the difference clinically important.
It is usually represented by the Greek letter delta.
A smaller margin requires stronger evidence that the treatments produce closely comparable outcomes.
A larger margin allows more difference between treatments.
Why does the margin matter?
Suppose researchers compare two treatments using a symptom scale where lower scores indicate improvement.
They set a hypothetical noninferiority margin of five points.
If the new treatment produces an average score two points higher than the established treatment, that difference may appear acceptable.
However, the average difference alone is not enough to establish noninferiority.
Researchers also need to examine the range of differences that remain plausible after accounting for statistical uncertainty.
For example:
| Hypothetical result | Interpretation |
|---|---|
| Estimated difference: 2 points; 95% confidence interval: −1 to 4 | Supports noninferiority because the entire interval is below the five-point margin |
| Estimated difference: 2 points; 95% confidence interval: −2 to 7 | Inconclusive because the interval includes differences exceeding the margin |
| Estimated difference: 7 points; 95% confidence interval: 6 to 8 | Indicates that the new treatment falls outside the acceptable noninferiority limit |
These are simplified examples in which a positive difference means worse outcomes with the new treatment.
A confidence interval is a statistical range that describes uncertainty around an estimated effect.
The examples show why researchers cannot establish noninferiority by looking only at average scores.
They must consider whether the confidence interval rules out differences that would be considered unacceptably large.
Does Finding No Significant Difference Prove Noninferiority?
No. This is one of the most important distinctions in interpreting treatment research.
A conventional statistical test may find no statistically significant difference between two treatments.
That result does not automatically mean the treatments are equally effective or that one is noninferior to the other.
For example, a study may include too few participants to detect a meaningful difference.
Both treatments could produce similar average scores, but the results might still allow for substantial uncertainty about their actual effects.
A noninferiority trial addresses this by specifying an acceptable difference and testing whether the available evidence rules out a larger disadvantage.
A lack of evidence that treatments differ is not the same as evidence that they are sufficiently similar.
This distinction is particularly important when reading EMDR studies that conclude two treatments produced comparable results.
Unless the study was designed and analyzed appropriately for noninferiority, the absence of a statistically significant difference does not establish it.
What Have Noninferiority Trials Found in EMDR Research?
Researchers have used noninferiority designs to compare EMDR with other psychological treatments.
Several published studies help illustrate how the approach works.
EMDR versus cognitive behavioral therapy for panic disorder
A randomized controlled trial by Horst and colleagues, published in Frontiers in Psychology in 2017, examined EMDR and cognitive behavioral therapy for panic disorder.
The study included 84 participants who were randomly assigned to one of two treatment groups.
Each group received 13 weekly sessions lasting approximately 60 minutes.
Researchers measured panic-related symptoms and quality of life before treatment, immediately afterward, and three months later.
The study used noninferiority testing to determine whether EMDR performed within acceptable limits compared with cognitive behavioral therapy.
What were the findings?
The researchers found that EMDR met the noninferiority criteria for measures of panic-related thoughts and fear of physical sensations.
Noninferiority was also supported for several quality-of-life outcomes.
However, the adjusted analyses for avoidance-related measures were inconclusive.
This means the trial did not establish noninferiority across every outcome.
The findings suggest that EMDR can produce outcomes comparable to cognitive behavioral therapy for several important aspects of panic disorder, while demonstrating why researchers need to report results separately for different measures.
The study was also smaller than originally planned, which is relevant when considering the precision of its conclusions.
EMDR versus cognitive behavioral therapy for recurrent depression
A study by Ostacoli and colleagues, published in Frontiers in Psychology in 2018, investigated EMDR as an additional treatment for recurrent depression.
The study was conducted in Italy and Spain and included 82 participants.
Participants were randomly assigned to receive either EMDR or cognitive behavioral therapy alongside ongoing antidepressant medication and psychiatric care.
Sixty-six participants were included in the analysis of those who completed treatment.
The researchers examined changes in depressive symptoms, remission rates, and quality of life.
Both groups showed improvements over time.
At the end of treatment, 71% of participants who completed EMDR met the study's depression-remission criterion, compared with 48.7% of those who completed cognitive behavioral therapy.
However, the difference in remission rates between groups was not statistically significant.
The researchers described the findings as preliminary and suggested that EMDR could be a viable treatment for recurrent depression.
An important point is that the reported lack of a statistically significant difference does not, by itself, demonstrate noninferiority.
The interpretation depends on the study's planned margin, statistical analysis, and uncertainty around the treatment differences.
This study demonstrates both the usefulness of comparative EMDR research and the importance of interpreting noninferiority claims carefully.
EMDR versus cognitive behavioral therapy for PTSD
A 2025 randomized controlled trial by Khan and colleagues examined EMDR and cognitive behavioral therapy for PTSD in Pakistan.
The study included 80 participants receiving treatment at two mental health rehabilitation centers.
Researchers used the Clinician-Administered PTSD Scale for the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders, commonly called CAPS-5, to assess PTSD symptoms.
Both treatment groups showed substantial reductions in symptom scores.
The researchers reported that EMDR was noninferior to cognitive behavioral therapy and concluded that EMDR was a viable treatment option in the setting studied.
These findings add to research comparing established trauma-focused interventions.
However, evaluating the strength of the noninferiority conclusion also requires examining the predefined margin, confidence intervals, and analysis methods.
The findings of one trial do not establish that EMDR is equivalent or superior to cognitive behavioral therapy for every population or outcome.
Why Might Noninferiority Trials Matter for Online EMDR?
Noninferiority trials are particularly useful when researchers want to compare different ways of providing an established treatment.
For example, researchers may want to investigate whether therapist-guided online EMDR produces outcomes that are sufficiently close to those of in-person EMDR.
Online treatment may offer practical advantages, such as reduced travel requirements or greater access for people living far from treatment services.
A noninferiority trial can investigate whether those advantages come with an acceptable level of clinical effectiveness.
A study protocol involving online and in-person EMDR
A research protocol published by te Brake and colleagues in Trials in December 2025 describes a noninferiority study involving young people aged 12 to 18.
The study is designed to compare online and in-person delivery of EMDR and Acceptance and Commitment Therapy.
Participants are to be randomly assigned to online or in-person treatment.
Researchers plan to examine symptom changes, daily functioning, and the therapeutic relationship.
The study also includes follow-up assessments.
The protocol is an example of how noninferiority methods can be applied to questions about treatment delivery.
Importantly, a published study protocol describes what researchers plan to investigate. It is not a report demonstrating that noninferiority has been established.
The trial's results will need to be evaluated separately when published.
Why Might Noninferiority Matter When Comparing EMDR With Other Treatments?
Noninferiority trials can help answer questions that go beyond whether one therapy works better than another.
Providing more treatment options
If two treatments produce outcomes within an acceptable difference, this may support offering people a choice between established approaches.
Personal preferences, accessibility, and treatment availability may influence which approach someone selects.
Evaluating different treatment formats
Researchers may compare individual, group, online, or other structured treatment formats.
The aim may be to determine whether a more accessible approach preserves enough of the established treatment's benefit.
Examining treatment efficiency
Some studies may investigate whether a treatment involving fewer sessions produces results that remain within an acceptable clinical range.
Noninferiority designs can help evaluate the balance between treatment outcomes and the resources required.
Supporting evidence-based decisions
Research findings can help healthcare services evaluate different treatment options while considering outcomes, availability, cost, and practical implementation.
However, demonstrating noninferiority on a clinical outcome does not automatically establish that a treatment is cheaper, more convenient, or preferable.
Those potential advantages need their own supporting evidence.
What Makes a Noninferiority Trial Reliable?
Several features are important when evaluating the quality of a noninferiority study.
An appropriate comparison treatment
The established treatment needs reliable evidence supporting its effectiveness.
If the comparison treatment performs poorly in the study, two treatments may appear similar without either producing the expected benefits.
A justified noninferiority margin
Researchers need to explain why the chosen margin represents an acceptable difference.
A margin that is too large may allow a meaningfully less effective treatment to be declared noninferior.
An adequate sample size
The study needs enough participants to estimate treatment differences with reasonable precision.
Small samples may produce confidence intervals that are too wide to support a clear conclusion.
Appropriate analysis of participants
Researchers commonly consider both intention-to-treat and per-protocol analyses.
Intention-to-treat analysis generally examines participants according to their original treatment assignments.
Per-protocol analysis focuses on participants who followed the study's treatment requirements closely enough to meet the predefined criteria.
Examining both can help researchers understand how missed sessions, treatment changes, or incomplete participation affect the results.
Consistent treatment delivery
Researchers may also evaluate treatment fidelity, meaning how closely the treatment followed the intended protocol.
This helps establish whether the interventions were delivered as planned.
These features are reflected in the Consolidated Standards of Reporting Trials (CONSORT) guidance for noninferiority and equivalence trials.
What Are the Limitations of Noninferiority Trials?
Although noninferiority trials can provide valuable evidence, their findings require careful interpretation.
Noninferiority does not establish identical effectiveness
A treatment can meet the noninferiority criterion even when its estimated effect is somewhat worse than that of the comparison treatment.
The conclusion depends on the acceptable difference established in the study.
Results apply to specific outcomes
A treatment may be noninferior for symptom reduction but not for another outcome, such as daily functioning or quality of life.
Findings cannot automatically be extended to outcomes the study did not establish.
The margin influences the conclusion
Different studies may choose different acceptable differences.
This means two studies using different margins could reach different conclusions from similar treatment effects.
Incomplete participation can complicate results
If participants miss sessions, change treatments, or discontinue participation, the observed differences between groups may become harder to interpret.
Findings may not apply to every population
Results from one age group, clinical setting, or treatment format may not apply directly to other groups.
These limitations explain why understanding the study design is as important as reading its final conclusion.
How Can Readers Evaluate a Noninferiority Claim in EMDR Research?
When an EMDR study reports that one treatment is noninferior to another, several details can help readers understand the finding.
| Research detail | Why it matters |
|---|---|
| Comparison treatment | Establishes which intervention EMDR was evaluated against |
| Primary outcome | Identifies what researchers were mainly trying to measure |
| Noninferiority margin | Shows the largest acceptable difference between treatments |
| Confidence interval | Indicates the uncertainty around the estimated treatment difference |
| Sample size | Helps assess how precisely the study could estimate outcomes |
| Follow-up period | Shows when the comparison was made |
| Analysis approach | Explains how researchers handled treatment assignments and participation |
| Treatment fidelity | Helps establish whether the treatments were delivered as intended |
A clear research report will explain how these elements contributed to the conclusion.
Readers may also find it useful to distinguish between three statements:
The treatments did not show a statistically significant difference.
The treatments produced similar average improvements.
The new treatment met the study's predefined noninferiority criteria.
These statements are related, but they are not equivalent.
The third statement represents a specific statistical conclusion supported by an appropriately designed noninferiority analysis.
Does a Noninferiority Trial Prove That EMDR Is Effective?
A noninferiority trial can provide evidence about EMDR's effectiveness relative to an established treatment, but the conclusion depends on the quality of the study and its comparison.
If a new intervention is shown to preserve an acceptable amount of the benefit provided by an established treatment, that finding can strengthen confidence in its usefulness.
However, noninferiority alone does not establish that every treatment benefit has been demonstrated independently.
This is why researchers consider previous evidence showing that the comparison treatment works, along with the design and results of the current trial.
EMDR already has substantial research support for PTSD, including randomized controlled trials and systematic reviews.
Therapists also commonly use EMDR for depression, anxiety, grief, addictions, and other distressing experiences.
Noninferiority research adds another perspective by examining how EMDR compares with established interventions or alternative methods of delivery.
It complements rather than replaces other forms of treatment-effectiveness research.
What Questions Could Future Noninferiority Trials Explore?
Future EMDR research could use noninferiority designs to investigate several practical questions.
Researchers may compare therapist-guided online EMDR with in-person EMDR across larger and more diverse participant groups.
They could also examine whether different treatment schedules or delivery formats maintain acceptable clinical outcomes.
Other studies may compare EMDR with established therapies for conditions where more direct treatment comparisons would be useful.
Longer follow-up periods could help determine whether noninferiority findings remain consistent after treatment has ended.
Researchers may also investigate costs, accessibility, treatment completion, and participant preferences alongside clinical outcomes.
These studies could provide a clearer understanding of how different EMDR approaches compare and which treatment formats are useful in different circumstances.
What Is the Main Takeaway?
A noninferiority trial is designed to determine whether a treatment is not meaningfully less effective than an established alternative.
Unlike a superiority trial, it does not require the new treatment to produce better results.
Instead, researchers define an acceptable difference and examine whether the study's findings rule out a larger disadvantage.
In EMDR research, noninferiority trials have been used to compare EMDR with cognitive behavioral therapy for conditions including panic disorder, recurrent depression, and PTSD.
Researchers are also using this approach to investigate differences between online and in-person treatment delivery.
The central lesson is that noninferiority means a treatment has met a predefined standard for acceptable effectiveness—not that two treatments are identical or that no differences exist.
Understanding this distinction helps readers interpret EMDR research more accurately and recognize what comparative treatment studies can and cannot demonstrate.
References
Piaggio, G., Elbourne, D. R., Pocock, S. J., Evans, S. J. W., & Altman, D. G. (2012). Reporting of noninferiority and equivalence randomized trials: Extension of the CONSORT 2010 statement. JAMA, 308(24), 2594–2604.
https://doi.org/10.1001/jama.2012.87802Cuzick, J., & Sasieni, P. (2022). Interpreting the results of noninferiority trials—a review. British Journal of Cancer, 127, 1755–1759.
https://doi.org/10.1038/s41416-022-01937-wHorst, F., Den Oudsten, B., Zijlstra, W., de Jongh, A., Lobbestael, J., & De Vries, J. (2017). Cognitive behavioral therapy vs. eye movement desensitization and reprocessing for treating panic disorder: A randomized controlled trial. Frontiers in Psychology, 8, 1409.
https://doi.org/10.3389/fpsyg.2017.01409Ostacoli, L., Carletto, S., Cavallo, M., et al. (2018). Comparison of eye movement desensitization reprocessing and cognitive behavioral therapy as adjunctive treatments for recurrent depression: The European Depression EMDR Network randomized controlled trial. Frontiers in Psychology, 9, 74.
https://doi.org/10.3389/fpsyg.2018.00074Khan, A., et al. (2025). A comparative exploration of eye movement desensitisation and reprocessing versus cognitive behavioural therapy for post-traumatic stress disorder in Pakistan: Insights from a full-fledged randomised controlled trial. Counselling and Psychotherapy Research, 25(1), e12870.
https://doi.org/10.1002/capr.12870te Brake, E., Scheepers, F., Mous, T., et al. (2025). Comparing the effectiveness of online therapy versus in-person therapy in youth: Protocol for a non-inferiority randomized controlled trial. Trials, 27, 3.
https://doi.org/10.1186/s13063-025-09312-9U.S. Department of Veterans Affairs, National Center for PTSD. Eye Movement Desensitization and Reprocessing for PTSD.
https://www.ptsd.va.gov/professional/treat/txessentials/emdr_pro.asp
