Preference on the Feedback Mechanisms of Dose Delivery Confirmation With the Breezhaler® Device Compared to the Ellipta® Device in Patients With Chronic Obstructive Pulmonary Disease (COPD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Preference on the Feedback Mechanisms of Dose Delivery Confirmation With the Breezhaler® Device Compared to the Ellipta® Device in Patients With Chronic Obstructive Pulmonary Disease (COPD) (NCT02551224), a randomized na study. The primary outcome is Patient's Preference on the Feedback Mechanisms of Dose Delivery Confirmation Using the Breezhaler® and Ellipta® Devices.

2. Clinical Question

The study addresses the experience of using device feedback, rather than a comparison of active drug treatments. Its primary measure is a preference questionnaire administered after device use, with the primary outcome timeframe listed as 6 hours. The questionnaire description gives three prompts: confidence that the full dose was received, certainty that no drug remained in the device, and the extent to which the device helped the patient know that all medication was received. Each prompt is described on a scale from 1 to 5, with higher response options expressing greater confidence, certainty, or extent.

The clinical question is therefore about patient-reported perceptions following exposure to both devices. It is not a question about changes in COPD symptoms, lung function, exacerbations, or the comparative effectiveness of active medications. Those outcomes are not reported as study endpoints in the registry record. The secondary question concerns comfort of the mouthpieces when making a tight seal with the lips.

3. Trial Design

The registry record describes a randomized, open-label crossover study, with 100 participants enrolled. Participants were assigned to one of two device sequences: a single placebo dose through the Breezhaler® followed by a single placebo dose through the Ellipta®, or the reverse order. The record describes the two sequences as having half of the patients assigned to each. Questionnaires were completed after use of each device.

A crossover design means that participants receive both conditions in sequence, so each person can contribute experience with each device. Randomizing the order can help balance order effects across the sequences. Because masking was listed as none, participants and study personnel were not blinded to the device being used. The interventions were device-use experiences with placebo, not active medication comparisons.

The study started in 2015-09, had primary completion in 2016-03, and is listed as completed. The primary outcome timeframe is 6 hours. The registry identifies the primary analysis population as the full analysis set. Participant flow, exclusions from that set, the exact questionnaire administration schedule, and further details of the crossover analysis are not reported in the registry record.

The listed primary and secondary outcomes are continuous measures summarized as means and standard deviations. The registry also reports a Wilcoxon (Mann-Whitney) analysis for each posted outcome. It does not describe whether the analysis accounted for the paired, within-participant structure of the crossover or the sequence of device use. The original statistical analysis plan, the trial’s own sample-size justification, and any interim analyses are not reported in the registry record. The planning illustrations associated with the posted results are separate from the trial’s original design assumptions.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

First Intervention

Second Intervention (5 Min After First )

Analysis population for the primary outcome, as defined in the registry: Full analysis set

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: A total of 100 subjects were randomized, all of whom were included in the FAS and the safety set

6. Primary Endpoint

The registry lists the primary outcome as Patient's Preference on the Feedback Mechanisms of Dose Delivery Confirmation Using the Breezhaler® and Ellipta® Devices, measured over 6 hours.

The primary endpoint is the patient’s reported perception of the devices’ dose-delivery confirmation feedback. The registry describes a mean score on a scale from 1 to 5, where the response options range from not at all confident or not at all to extremely confident or a very great deal. It lists three questionnaire prompts: confidence that the full dose was received, certainty that no drug remained in the device, and how much the device helped the patient know that all medication had been received.

These questions target the feedback experience itself. They are relevant to the stated study question because the devices offer feedback mechanisms that patients may use to judge whether dose delivery is complete. The outcome is patient-reported rather than a direct measurement of medication remaining in the device, inhaled dose, or clinical response. It describes participants’ perceptions, not an objectively verified dose-delivery result.

A mean is a familiar summary of responses on a numeric scale, but it compresses the pattern of answers. It does not show how many participants selected each response, whether responses clustered near the middle or ends of the scale, or whether particular questionnaire items differed. Although the registry lists three prompts, it does not report in its description whether the primary score is calculated separately for each item or combined across them. The listed timeframe is 6 hours; further details about timing are not reported in the registry record.

7. Results Posted in the Registry

• Ellipta: mean 3.60 versus 4.30 on control; difference -0.70 (95% CI -0.96 to -0.44), standardized effect d = -0.74; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Comfort of the Mouth Pieces for Performing a Tight Seal With the Lips.

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Ellipta: mean 3.90 versus 4.30 on control; difference -0.40 (95% CI -0.61 to -0.19), standardized effect d = -0.52; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Ellipta than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry posts one secondary endpoint, mouthpiece comfort when performing a tight seal with the lips. Across the primary and secondary outcomes, the overall descriptive pattern favors higher scores for Breezhaler® than for Ellipta®. That direction is consistent across the device-feedback and mouthpiece-comfort measures, but the measures address related rather than identical aspects of device experience.

The secondary outcome can add context to the primary finding because it concerns another patient-reported feature of device use. It does not independently validate the primary measure, establish a mechanism, or show that comfort explains the feedback ratings. The registry does not report participant-level associations between the primary and secondary responses, so the relationship between the measures cannot be assessed from the posted summaries alone.

Unadjusted secondary results generally carry less confirmatory weight when multiple outcomes or comparisons may have been examined. In general, repeated testing can make apparently strong results more likely by chance unless the analysis plan accounts for multiplicity. The registry record does not report whether a multiplicity strategy was specified, whether all intended secondary outcomes are posted, or whether the secondary endpoint was treated as confirmatory. The secondary finding should therefore be interpreted as supporting context rather than as an independent definitive conclusion.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Comfort of the Mouth Pieces for Performing a Tight Seal With the Lips.

9. Statistical Interpretation

10. Statistical Methods Explained

What does a mean difference describe?

A mean is the arithmetic average of observed scores. A mean difference subtracts one group’s average from the other’s. Its sign depends on the order of subtraction, so the same contrast can be positive in one convention and negative in the reverse convention. The primary recalculation is expressed as Ellipta® minus Breezhaler®: the estimate is -0.7. The sponsor reports a net mean difference of 0.75, using a direction that aligns with the higher posted Breezhaler® mean. A reader should attend to the named device order and the stated direction, not just the sign.

The outcome is a score on a scale from 1 to 5. Treating a scale score as continuous and summarizing it by a mean is common, but it assumes that averages are useful descriptions of the response pattern. The mean and standard deviation do not reveal the full distribution, such as the number of responses at each scale point or whether responses are concentrated near a ceiling. Those response distributions are not reported in the registry record.

Why does the crossover structure matter?

In a crossover study, the same participant uses both devices. Their responses to the two devices may be related because they come from the same person. A paired analysis uses that within-person relationship when estimating the contrast; an analysis that treats the device groups as independent does not. Using the pairing can affect the estimated uncertainty, because the relationship between a person’s scores under each device contributes information about the comparison.

The recalculated values are simple, unadjusted calculations from the posted summary statistics. They do not have participant-level paired data available and do not account for the study’s randomized sequence. The posted summaries give the same number of observations for each device, but do not show how the individual participant scores correspond across devices. The registry’s reported method is Wilcoxon (Mann-Whitney), a rank-based comparison commonly used for independent groups. The registry does not explain how that method was adapted, if at all, for this crossover design. Therefore, the posted sponsor result and the unadjusted recalculation should not be treated as identical analyses.

What does the Wilcoxon (Mann-Whitney) label mean?

In general, the Wilcoxon rank-sum test, also called the Mann-Whitney test, compares the rank ordering of observations between two groups. It does not directly test a difference in means in the same way as a standard two-sample t-test. Its interpretation can depend on the shapes and spreads of the outcome distributions, not only on their central values. The registry labels the reported parameter as “Mean Difference (Net),” but does not provide additional detail on how that parameter was estimated alongside the rank-based method. The analysis output should consequently be read as the registry’s stated net difference and interval, with the methodological detail left unresolved.

How should the p-value be read?

In general, a p-value describes how incompatible the observed data, or more extreme data, would be with a specified null model. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. The primary sponsor analysis reports p < 0.001, while the summary-statistic recalculation reports p <0.001. These values come from different analysis descriptions and should not be treated as interchangeable. A small p-value does not address whether questionnaire scores are important to patients, whether the analysis handled the crossover structure, or whether all planned analyses were reported.

The p-values shown are unadjusted. If multiple outcomes or comparisons are assessed, conducting multiple tests can increase the chance of at least one apparently strong result under a setting where there is no real effect. In general, multiplicity procedures can control that familywise chance or otherwise account for a testing strategy. Whether such procedures were planned or used is not reported in the registry record.

What does a confidence interval contribute?

In general, a confidence interval expresses the uncertainty around an estimated contrast under the assumptions of the analysis. A narrower interval indicates more precision under those assumptions; a wider interval indicates less precision. It is not a range guaranteed to contain the true effect, nor does it account for biases or design features omitted from an analysis. Here the recalculated interval is based on posted summary statistics and an unadjusted comparison, while the sponsor interval is part of its reported analysis. The two should be interpreted with their distinct analysis bases in mind.

Over-dispersion, in general, means that the variability in data exceeds what a model assumes. If a model understates variability, standard errors can be too small and confidence intervals too narrow. The registry does not report a model-based dispersion assessment for these outcomes, so over-dispersion is not established as a feature of this trial’s analysis. More broadly, when questionnaire scores are bounded and have a limited set of possible values, means and standard deviations alone may not capture every aspect of their distribution.

11. Confidence Intervals and Precision

The recalculated primary mean difference is -0.7 for Ellipta® minus Breezhaler®, with a 95% confidence interval from -0.961 to -0.439. This interval expresses uncertainty around the unadjusted summary-based estimate under the assumptions used for that calculation. Its values are below zero in this sign convention, corresponding to a lower mean score for Ellipta® in the posted summaries. The interval does not measure the importance of the score difference to patients, and it does not correct for features not represented in the calculation.

The sponsor’s reported primary analysis gives a 95% confidence interval from 0.51 to 0.99 for its net mean difference of 0.75. Its positive orientation is consistent with the Breezhaler® mean being higher than the Ellipta® mean. The sponsor’s interval and the recalculated interval arise from distinct analysis descriptions and orientations. They should not be combined into one estimate or treated as interchangeable measures of uncertainty.

In general, the width of a confidence interval reflects statistical precision, conditional on the analysis and its assumptions. A confidence interval describes a range of values compatible with the observed data under the method; it is not a guarantee that the true contrast lies inside it. It also does not account for possible missing-data bias, measurement limitations, or an analysis that fails to reflect the crossover structure.

The posted means and standard deviations give a compact summary, but the registry record does not include the individual-level paired scores, response frequencies, or additional details needed to evaluate the full distribution. Consequently, the precision of the recalculated interval should be understood as precision for that unadjusted calculation, rather than a complete account of uncertainty in the crossover comparison.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning estimates are participants per arm, not total event counts. They are illustrations based on the observed effect and are not the trial’s original design calculation. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and a target effect equal to a stated fraction of the observed effect. The original trial’s own sample-size assumptions and justification are not reported in the registry record.

For 100% of the observed primary effect, the illustration gives 29 participants per arm for 80% power and 39 participants per arm for 90% power. For 75% of the observed effect, it gives 51 participants per arm for 80% power and 68 participants per arm for 90% power. For 50% of the observed effect, it gives 114 participants per arm for 80% power and 153 participants per arm for 90% power.

In general, the 80% power column means the number of participants per arm estimated to give an 80% chance of detecting the specified effect under the stated assumptions, if that effect is real. The 90% column uses a higher target chance of detection and therefore requires more participants under these illustrations. As the target effect becomes smaller, it is harder to distinguish from random variation, so more participants are needed. These figures are not observed power. They do not indicate the probability that the trial’s result is correct, and they do not replace the unreported design assumptions for the crossover trial.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.700; pooled standard deviation σ = 0.942.
2. n = 2 × (1.96 + 0.842)² × 0.942² / -0.700² = 28.4, rounded up to 29 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.7                      # observed difference in means
sigma <- 0.942045646452442                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(2551224)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 29 (80% power), 39 (90% power)
 75% of observed effect: n per arm = 51 (80% power), 68 (90% power)
 50% of observed effect: n per arm = 114 (80% power), 153 (90% power)
power.t.test: 29.4 
Simulated power at n = 29 per arm: 0.799 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The sponsor’s primary analysis is reported as a Wilcoxon (Mann-Whitney) comparison with a net mean difference of 0.75, a 95% confidence interval from 0.51 to 0.99, and p < 0.001. The registry lists the difference in a positive orientation, consistent with the higher Breezhaler® mean in the posted group summaries. The recalculated comparison is expressed in the reverse direction, Ellipta® minus Breezhaler®, and gives -0.7 with a 95% confidence interval from -0.961 to -0.439 and p <0.001.

The recalculation is an unadjusted analysis from the posted means and standard deviations. It is not a reconstruction of the sponsor’s rank-based analysis or a paired crossover analysis. The registry does not provide the participant-level data or enough methodological detail to reconcile the methods beyond recognizing the difference in orientation and analysis approach. The two analyses point to the same ordering of device means, but their estimates and intervals should be read as outputs from distinct methods.

The registry record does not flag a discrepancy between the sponsor’s result and the recalculation. The sponsor’s result remains the reported trial analysis, while the recalculation offers a transparent, summary-statistic comparison with its assumptions made explicit. Neither analysis establishes clinical importance from statistical significance alone.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary endpoint and one posted secondary endpoint, and reports analyses for both. The primary outcome description lists three questionnaire prompts, but the record does not clarify whether these were analyzed separately, combined into one score, or handled in another way. The number and organization of all planned tests are therefore not fully clear from the registry record.

Multiplicity arises when more than one hypothesis or outcome is tested. In general, the more tests conducted, the greater the chance that at least one will appear statistically notable even when there is no underlying difference. Analysts may use a prespecified hierarchy, adjust significance thresholds, or interpret secondary analyses as exploratory. Whether any such approach was planned or used is not reported in the registry record. The posted p-values should be read as unadjusted.

The primary analysis population is listed as the full analysis set. The registry does not report how that set was defined, whether any participants were excluded from an analysis, how incomplete questionnaires were handled, or whether missing values were imputed. It also does not report whether the analysis followed an intention-to-treat approach or a per-protocol approach. These details matter because exclusions and missing responses can affect which participants contribute to each estimate.

Further information about the statistical analysis plan, handling of the crossover sequence and within-participant pairing, any interim analyses, and the original sample-size justification is not reported in the registry record. Without those details, the posted results can be described, but the full prespecified analysis strategy and its implementation cannot be assessed.

16. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry lists 100 participants at risk for serious adverse events under each device. It reports zero serious adverse events for Breezhaler® and zero for Ellipta®. No serious adverse-event terms are listed. These are the reported serious-event counts in the registry; they do not establish that no other safety outcomes occurred or were assessed.

The record also lists one other adverse event for Breezhaler® and zero for Ellipta®. The definition and description of that event are not reported in the registry record. Death counts are not reported in the registry record. Because participants were assigned to sequences and used both devices, the device-specific safety figures should be read within the crossover context and not as a comparison of independent treatment groups. The posted counts are descriptive and do not support broader conclusions about safety beyond the recorded events.

17. What the Result Does, and Does Not, Mean

The posted results describe participants’ questionnaire ratings after using two inhaler devices with placebo. The primary summaries show a higher mean score for Breezhaler® than for Ellipta®, and the sponsor’s reported analysis supports a difference in that direction. The secondary mouthpiece-comfort measure follows the same overall direction. These findings concern reported perceptions during the study, not the delivery of active medication or clinical outcomes in COPD.

A higher average questionnaire score does not by itself establish how important the difference is to patients. The registry does not report a threshold for a meaningful score difference, a validation analysis for the questionnaire, the full distribution of responses, or the proportion of participants preferring one device. The three primary prompts address confidence, certainty, and perceived help from the device, but the scoring relationship among those prompts is not specified in the registry record.

The results do not demonstrate differences in symptom control, lung function, exacerbations, medication adherence, or long-term device use. Those outcomes are not reported in the registry record. Placebo use also means that the trial’s comparison is focused on the devices and their feedback, rather than the effects of active treatment.

Statistical significance and a confidence interval help describe the evidence and its precision under an analysis. They do not establish causality beyond the study design, clinical importance, or applicability to people with COPD who differ from the enrolled population, including those not naïve to DPI devices. The open-label design also means device identity was not masked. The appropriate reading is a comparison of reported device perceptions in this study, with the analysis and generalizability limitations kept in view.

18. Why This Trial Matters Statistically

This trial is statistically interesting because it uses a randomized crossover design to compare subjective ratings after participants have used both devices. In principle, a crossover comparison can make use of within-person information: each participant can serve as their own reference, which may help separate device-related variation from stable differences between people. That advantage depends on an analysis that reflects the paired structure and considers sequence and period effects. The registry record does not report enough detail to determine how those features were handled.

The posted results also illustrate why the direction of a contrast matters. The recalculation subtracts Breezhaler® from Ellipta®, while the sponsor’s net mean difference is reported in the direction consistent with the higher Breezhaler® mean. Reading the labels and orientation alongside the estimate prevents a negative sign from being mistaken for evidence favoring the opposite device.

Finally, the study highlights the distinction between statistical evidence and practical meaning. A questionnaire can quantify participants’ perceptions, while a confidence interval describes uncertainty around a contrast. Neither alone establishes the clinical importance of a score difference. The result is most informative when considered as evidence about device feedback perceptions, alongside the limits of the outcome measurement and the registry’s incomplete description of the analysis plan.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several limitations are specific to the study question and the posted record. The study is open-label, so participants knew which device they were using while rating its feedback. The outcomes are self-reported perceptions, and the registry does not report a validation study, a meaningful-difference threshold, or the distribution of responses across the 1-to-5 scale. The description lists three primary prompts but does not clarify how their scores contribute to the reported mean.

The crossover design makes within-participant dependence central to interpretation. The recalculation uses posted means and standard deviations and does not account for paired observations, sequence, or period. The registry reports a Wilcoxon (Mann-Whitney) method but does not explain how the analysis addressed the crossover structure. Participant-level results and the detailed statistical analysis plan are not reported in the registry record.

The enrolled population is described as patients with COPD naïve to DPI devices. Participant characteristics beyond that description are not reported in the registry record, limiting assessment of how broadly the result may apply. The study used placebo doses and therefore does not compare active medication outcomes. The record also does not report missing-data handling, exclusions from the full analysis set, multiplicity procedures, or the original sample-size justification. These omissions limit assessment of potential selection, analysis, and reporting effects.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Preference on the Feedback Mechanisms of Dose Delivery Confirmation With the Breezhaler® Device Compared to the Ellipta® Device in Patients With Chronic Obstructive Pulmonary Disease (COPD) — NCT02551224.
• Registry record last updated 2017-12-18; status last verified 2017-11.
• Results first posted 2017-10-09.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

The registry record describes a completed, randomized, open-label crossover study of 100 patients with COPD who were naïve to DPI devices. Participants used Breezhaler® and Ellipta® in randomized order, with a single placebo dose through each device and questionnaires after use. The primary outcome assessed perceptions of dose-delivery confirmation; the secondary outcome assessed mouthpiece comfort.

The posted mean scores are higher for Breezhaler® than Ellipta® for both reported outcomes. The sponsor’s primary analysis reports a net mean difference of 0.75 with a 95% confidence interval from 0.51 to 0.99 and p < 0.001. The summary-statistic recalculation gives an unadjusted Ellipta®-minus-Breezhaler® difference of -0.7, with a 95% confidence interval from -0.961 to -0.439 and p <0.001.

The results concern patient perceptions of device feedback and mouthpiece comfort. The registry does not report enough detail to assess the complete crossover analysis strategy, missing-data handling, multiplicity procedures, or original sample-size assumptions.