Preference Attributes of ELLIPTA Dry Powder Inhaler (DPI) and HANDIHALER DPI in Subjects With Chronic Obstructive Pulmonary Disease (COPD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Preference Attributes of ELLIPTA Dry Powder Inhaler (DPI) and HANDIHALER DPI in Subjects With Chronic Obstructive Pulmonary Disease (COPD) (NCT02786927), a randomized phase 4 study. The primary outcome is Percentage of Participants Preferring the Inhaler Based on the Number of Steps Needed to Take the COPD Medication.

2. Clinical Question

The study asked whether more participants with chronic obstructive pulmonary disease (COPD) preferred the ELLIPTA dry powder inhaler or the HANDIHALER dry powder inhaler based on the number of steps needed to take medication. It also recorded preferences concerning how easy it was to tell how many doses were left and the size of the inhaler. The primary question concerned device preference, not whether either device improved COPD or was better at delivering active treatment.

Participants used both inhalers in sequence, with the order determined by randomization. The study was placebo-only: neither inhaler contained active treatment, and participants were to continue their current COPD medications as prescribed. After the two treatment periods, they completed an inhaler preference questionnaire. The posted registry information describes the question and planned analysis, but the primary percentage preferring each inhaler and the statistical test result are not reported in the registry record. Consequently, the record does not show which inhaler participants preferred on the primary measure, or how precisely any preference was estimated.

3. Trial Design

This was a Phase IV, multicenter, randomized, open-label crossover study. The registry lists 214 participants enrolled and describes the planned enrollment as approximately 211. In a crossover design, participants receive each study device in different periods, rather than being assigned to use only one device throughout the study. Here, the randomized sequences were ELLIPTA followed by HANDIHALER, or HANDIHALER followed by ELLIPTA. Questionnaire versions were also represented in the listed study groups.

Participants were to self-administer the inhalation once daily for 5-9 days in each period. The preference questionnaire was completed at Visit 3 or at an Early Withdrawal visit, and the primary endpoint timeframe is listed as up to 18 days. The registry describes the study as placebo-only: neither inhaler contained active treatment. Participants continued their prescribed COPD medications and followed up with their regular physician for COPD care during the study.

Open-label means that the study was not masked; the registry lists masking as none. This matters for an outcome based on participants’ stated preferences, since participants knew which device they were using. Randomization of sequence helps assign the order of device use, but it does not make the experience blinded. The record does not report further details about the randomization procedure, handling of possible period or carryover effects, or the sample-size justification.

The registry lists three outcome measures: the primary preference measure based on steps needed, and secondary preference measures based on how easy it was to tell how many doses were left and on inhaler size. The posted design information does not establish that the study evaluated clinical efficacy or comparative effects of active COPD medications.

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.

Before assignment: A total of 227 participants with diagnosis and documented history of chronic obstructive pulmonary disease (COPD) were screened for this open-label, cross-over study, of which 13 were screen failures, 2 were randomized by error and did not receive treatment and 212 participants entered the study treatment phase.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Modified ITT Population

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.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Preferring the Inhaler Based on the Number of Steps Needed to Take the COPD Medication, measured over Up to 18 days.

The primary endpoint was the percentage of participants preferring an inhaler based on the number of steps needed to take COPD medication. It was assessed with an inhaler preference questionnaire at Visit 3 or at an Early Withdrawal visit. The endpoint therefore concerns participants’ reported preference after experiencing both devices, rather than a clinical measure such as lung function, symptoms, or exacerbations.

Preference can be useful when a study aims to understand how people experience features of a device. Asking about the number of steps focuses the question on a specific attribute rather than asking only for a general impression. In a crossover study, participants can report a preference after using each device, which makes the comparison relevant to their own experience of both devices. That is a general feature of crossover preference studies, not evidence that this particular endpoint is free from bias.

The registry record does not report the percentage preferring ELLIPTA, the percentage preferring HANDIHALER, the number expressing no preference, or a confidence interval. It also does not report the primary test statistic or p-value. The endpoint’s definition and analysis approach are described, but the posted results do not permit a numerical interpretation of the preference finding.

Because neither inhaler contained active treatment in this study, the endpoint should be read as a device-preference question in the context of placebo use. It does not by itself quantify medication effectiveness or establish that a device changes health outcomes.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint was a percentage: the proportion of participants expressing a preference based on the number of steps needed to take medication. The registry states that this response was assessed using an inhaler preference questionnaire at Visit 3 or an Early Withdrawal visit. The analysis population was the Modified Intent-to-treat Population. The record describes this as participants from the ITT Population who completed at least one question from the five preference questions. The ITT Population is described as all randomized participants who received one dose of at least one study inhaler. These definitions are important because the analysis population was not simply everyone enrolled.

What does the Cochran-Mantel-Haenszel test do?

In general, the Cochran-Mantel-Haenszel test evaluates an association between categorical variables while accounting for specified strata. Here, the registry says the test was adjusted for study inhaler use sequence and questionnaire version. Sequence refers to which inhaler was used first; questionnaire version identifies which version was completed. An adjusted test can account for differences associated with those categories when assessing the preference comparison. The record does not provide the test statistic, the resulting p-value, or a detailed model specification.

Adjustment for sequence is pertinent to a crossover study because the order of device use may be relevant to how a preference is reported. Questionnaire version is also explicitly included as an adjustment factor. These statements identify what the registered analysis accounted for, but they do not establish that all possible sources of variation were addressed. The registry does not report whether the analysis evaluated period effects, carryover, or other aspects of the crossover structure.

Why is the crossover structure relevant?

In general, a crossover design exposes participants to more than one study condition, with the sequence assigned according to the study design. This can make each participant’s experience of both devices relevant to the preference response. The sequence is important because the experience of using a device first may differ from the experience of using it second. The registered test adjustment for sequence acknowledges this design feature. However, the record does not report details that would allow a reader to assess every assumption underlying the crossover comparison.

This trial was open-label, with no masking. In general, an unmasked design is especially relevant when outcomes are based on participant reports, since participants know which device they are using. That does not show that any particular response was influenced by awareness; it identifies a consideration when interpreting a subjective preference endpoint. The study compared placebo devices, not active medication effects.

What do the analysis populations mean?

The registry’s ITT definition includes randomized participants who received one dose of at least one study inhaler. The Modified Intent-to-treat Population then requires at least one response among the five preference questions. This requirement means that a participant with no completed preference questions would not meet the stated modified-population definition. It also means that the primary analysis population may differ from the number enrolled or randomized. The record does not report how many participants were excluded by these criteria, the final modified-population size, or the number with a response to the primary question specifically.

In general, the way missing questionnaire responses are handled can affect the analysis population and the interpretation of a percentage. The registry describes a minimum questionnaire-completion requirement, but does not report how missing answers within an otherwise eligible questionnaire were treated, whether any imputation was used, or how incomplete responses were represented. Those details are not reported in the registry record.

How should planning illustrations be read?

General planning illustrations based on an observed effect are not the same as the trial’s original sample-size justification or an estimate of achieved power. The stated assumptions for such illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a fraction of the observed effect. For outcomes other than time-to-event outcomes, those illustrations describe participants per arm. These are planning scenarios, not observed trial results and not evidence that the study had a particular probability of detecting an effect. The registry record does not report the trial’s own sample-size justification or its design assumptions.

In general, over-dispersion occurs when observed variation is greater than a model assumes; if ignored, uncertainty can be understated. That is a broad statistical concept, not a reported feature of this study. The registry describes a categorical preference analysis using the Cochran-Mantel-Haenszel test and does not report an over-dispersion assessment. No additional distributional assumption or alternative model should be inferred from the posted description.

The primary results, confidence interval, p-value, and effect estimate are not reported in the registry record. Consequently, the method can be described, but its numerical result and precision cannot be evaluated from the posted information.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists three outcome measures: the primary preference endpoint based on steps needed, and secondary preference endpoints based on dose-counter readability and inhaler size. It also lists three analyses. The primary analysis description specifies adjustment for inhaler use sequence and questionnaire version. That adjustment concerns the design factors named in the record; it is not the same as an adjustment for testing multiple endpoints.

The registry does not report whether a multiplicity procedure was used to control the chance of false-positive findings across the primary and secondary outcomes, nor does it describe a testing hierarchy or how secondary outcomes were to be interpreted. The reported significance level in the planning assumptions does not, by itself, establish how multiplicity was handled in the trial’s own analysis. Multiplicity methods are not reported in the registry record.

The primary analysis population was the Modified Intent-to-treat Population: participants in the stated ITT Population who completed at least one of the five preference questions. The ITT Population was defined as randomized participants who received one dose of at least one study inhaler. The registry does not report the number of participants in either population or the number completing the primary question.

The record does not describe how missing responses to individual questions were handled, whether any participants were excluded because of missing data, or whether missing values were imputed. It also does not report sensitivity analyses for missing questionnaire data. While the stated population definition requires at least one completed preference question, that alone does not explain how incomplete responses were treated in each endpoint analysis. Further details about missing-data handling, multiplicity, and analysis-population counts are not reported in the registry record.

11. 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 3% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: sepsis (0.5%), escherichia urinary tract infection (0.5%), acute kidney injury (0.5%), chronic obstructive pulmonary disease (0.5%).

The registry lists serious adverse event summaries separately for ELLIPTA and HANDIHALER. In the ELLIPTA group, 208 participants were listed as at risk and 0 as affected by a serious adverse event. In the HANDIHALER group, 212 were listed as at risk and 2 as affected. The registry also lists 0 affected by other adverse events in each group. These are descriptive counts in the posted safety summaries; they do not establish that a device caused or prevented an event.

The serious-event terms listed are sepsis, Escherichia urinary tract infection, acute kidney injury, and chronic obstructive pulmonary disease, each shown as 0.5%. The registry record does not provide further detail about event timing, severity beyond the serious-event classification, or attribution. The participant counts at risk differ between the device summaries, and the record does not explain the reasons for those denominators. Because this was a crossover study, these figures should not be treated as if they were necessarily counts from two independent groups of participants.

Deaths are listed as null rather than as a numerical count. That means death information is not reported in the registry record in these summaries; it should not be interpreted as zero deaths. The safety counts do not answer the primary preference question, and the preference results do not establish comparative safety.

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

This study was designed to compare preferences for two inhaler devices after participants with COPD had used each one. Its primary focus was a specific device attribute: the number of steps needed to take medication. The registry also identifies questions about dose-counter readability and inhaler size. These are preference outcomes, not measures of disease control or treatment efficacy.

The posted record does not report the primary preference percentages, the secondary preference results, or the statistical test result. It is therefore not possible to say from this record which inhaler participants preferred, whether a preference was statistically distinguishable from a specified comparison value, or how precise an estimated preference might be. A missing result cannot be interpreted as no difference, and the design description cannot substitute for an observed result.

The study used placebo in both devices, and participants continued their usual prescribed COPD medications. Accordingly, the study description does not establish comparative effects of active treatment delivered through the devices. A preference response about operating steps does not demonstrate better symptom control, fewer exacerbations, or any other clinical benefit. Such outcomes are not reported as results in the posted information.

In general, a preference result can describe participants’ reported experience in the conditions studied. Its interpretation depends on who contributed responses, how incomplete responses were handled, and the uncertainty around the estimate. Here, the record identifies a Modified Intent-to-treat Population and a questionnaire-based method, but does not report the analysis-population size, outcome estimate, or uncertainty. The open-label design is also relevant as a general consideration for self-reported preferences, because participants knew which device they were using.

The serious adverse event summaries should be read as descriptive listings, not as a comparison establishing causality. They do not fill the gap in preference results or show whether either device has a safety advantage. Overall, the record establishes the study’s intended question, crossover design, and planned primary test; it does not provide the outcome information needed to quantify the answer.

13. Why This Trial Matters Statistically

This trial is statistically interesting because its primary question is about a subjective device preference, while its design exposes participants to both devices. The crossover arrangement makes the order of use an important feature of the comparison, and the registry states that the primary Cochran-Mantel-Haenszel analysis adjusted for inhaler use sequence. Questionnaire version was also included in the adjustment. This makes the registered method more informative than a description that simply compares raw preferences without identifying these design factors.

The study also illustrates why a clear analysis-population definition matters. The registry distinguishes the ITT Population from the Modified Intent-to-treat Population and requires at least one completed preference question for inclusion in the latter. Enrollment alone therefore does not tell a reader how many participants contributed to the primary outcome. Without population counts and response counts, readers cannot assess the extent to which the endpoint data represent those enrolled.

Open-label status is another important feature. In general, self-reported preferences can be shaped by participants’ direct experience and awareness of the device. Such awareness does not make a preference measure invalid, but it is part of the context for interpreting responses. The study’s placebo-only setup also separates device preference from the effects of active treatment, a distinction that helps define what the question can answer.

Finally, this record shows the limits of interpreting a methods description without posted results. It states the endpoint, population, and test, but omits the primary estimate and its uncertainty. Statistical methods explain how a comparison was intended to be made; they do not reveal what the data showed. The listed safety summaries and secondary endpoints likewise do not resolve the unreported primary preference result.

14. 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.

The primary limitation for interpreting this record is that no numerical preference results are reported. The primary percentages, confidence interval, p-value, and primary analysis population size are not reported in the registry record. Results for the secondary preference outcomes are also not reported. Without these details, the magnitude and precision of any observed preference cannot be assessed.

The study was open-label and used participant-reported preferences. In general, awareness of the device may be relevant to responses, and the record does not describe any approach to reduce or evaluate that influence. The questionnaire versions are mentioned, and the primary analysis adjusted for version, but the registry does not provide the questionnaire wording, scoring details, or evidence about the measurement properties of the versions.

The crossover design means sequence and period are relevant considerations. The registry says that the primary analysis adjusted for study inhaler use sequence and questionnaire version, but does not report whether period effects, carryover, or other crossover assumptions were assessed. Additional details about the randomization procedure are not reported. The record also does not report how missing answers within the questionnaire were handled, whether sensitivity analyses were conducted, or how many participants completed the primary question.

The brief summary describes approximately 211 participants as planned, while enrollment is listed as 214. The sample-size justification and the trial’s own power assumptions are not reported in the registry record. Planning illustrations based on observed effects should not be read as proof of achieved power, and the record does not provide the observed preference effect needed to evaluate such illustrations independently.

Finally, placebo was used in both devices, and participants continued their existing COPD medications. The study’s preference question therefore does not establish comparative effects of active medication or clinical outcomes. Safety listings are brief, and death information is null rather than reported as a count.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Preference Attributes of ELLIPTA Dry Powder Inhaler (DPI) and HANDIHALER DPI in Subjects With Chronic Obstructive Pulmonary Disease (COPD) — NCT02786927.
• Registry record last updated 2018-10-15; status last verified 2018-09.
• Results first posted 2017-06-22.
• 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.

18. Record Summary

This completed Phase IV study enrolled 214 participants with COPD and compared preference for the ELLIPTA and HANDIHALER dry powder inhalers in a randomized, open-label crossover design. Participants were scheduled to use each placebo device once daily for 5-9 days and then complete an inhaler preference questionnaire. The primary endpoint concerned preference based on the number of steps needed to take medication.

The primary analysis used a Cochran-Mantel-Haenszel test adjusted for inhaler use sequence and questionnaire version, in a Modified Intent-to-treat Population defined by inhaler exposure and questionnaire completion. The registry lists two secondary preference outcomes concerning dose-counter readability and inhaler size. It does not report numerical results for the primary or secondary outcomes, nor the primary confidence interval or p-value. Serious adverse event summaries list 0 affected in the ELLIPTA group and 2 in the HANDIHALER group; death information is not reported as a numerical count.