Comparative Effectiveness of COPD Treatments: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparative Effectiveness of COPD Treatments (NCT03376295), a registered phase not stated study. The primary outcome is The Number of Observed Patients With First Chronic Obstructive Pulmonary Disease (COPD) Exacerbation to Occur After Cohort Entry.

2. Clinical Question

This is a comparative-effectiveness question about outcomes following initiation of two combination therapies. The registry record says LABA-TIO initiators were matched with LABA-ICS initiators using a time-conditional propensity-score approach. Such matching can make groups more comparable on measured characteristics, but it does not by itself establish that treatment groups were alike in all respects. The record does not report the matching variables, the resulting matched-group sizes, or a numerical comparison of the primary outcome. Consequently, the clinical question is clear, but the posted information does not establish the direction or size of any difference.

3. Trial Design

The registry record describes a completed comparative-effectiveness study of COPD treatments sponsored by Boehringer Ingelheim. Its recorded enrollment is 3954. The study start date is 2017-12-01, and the primary completion date is 2017-12-31. The record lists chronic obstructive pulmonary disease as the condition and describes the interventions as LABA-TIO and LABA-ICS.

The comparison is not identified as randomized, and allocation is not stated. Instead, the main analysis is described as using matched patients: people initiating LABA-TIO were matched with people initiating LABA-ICS through a time-conditional propensity-score-matched approach. This is an observational comparison as described in the registry record, not a like-for-like randomized control comparison. Matching is a design and analysis strategy intended to balance measured characteristics; it does not guarantee balance in unmeasured characteristics or eliminate all possible sources of bias.

The primary outcome description refers to an as-treated analysis and states that the event was followed for one year. Separately, the primary outcome timeframe is recorded as “12 Years.” These descriptions do not align, and the registry record does not clarify the discrepancy. The record lists two secondary outcomes: the rate of COPD exacerbations and occurrence of a first hospitalization for community-acquired pneumonia, described as serious pneumonia.

The registry identifies three outcome measures and six analyses, but does not explain how those analyses were allocated across outcomes or which statistical procedures were used. It also does not report a sample-size justification for the study design. Any planning illustration based on an observed effect is distinct from the study’s original design rationale and should not be read as observed power.

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: Population-based propensity score-matched incident new-user cohort study.

Overall Study

Analysis population for the primary outcome, as defined in the registry: The main analysis was on matched patients. The patients with LABA-TIO initiators were matched with initiators of LABA-ICS using a time-conditional propensity score-matched approach.

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: The main analysis was on matched patients. The patients with LABA-TIO initiators were matched with initiators of LABA-ICS using a time-conditional propensity score-matched approach. Baseline measures were presented for matched populations.

6. Primary Endpoint

The registry lists the primary outcome as The Number of Observed Patients With First Chronic Obstructive Pulmonary Disease (COPD) Exacerbation to Occur After Cohort Entry, measured over 12 Years.

The primary endpoint is described as the number of observed patients with a first COPD exacerbation after cohort entry. The event definition combines two types of events: hospitalization for COPD, categorized in the record as a severe exacerbation, or a prescription for oral prednisolone, categorized as a moderate exacerbation. The measure is a participant count, not an event rate or a reported estimate of time until the event.

A first-event endpoint focuses on whether a person experiences an initial qualifying exacerbation during the defined follow-up. It avoids counting repeated episodes as separate primary events, but a count alone does not describe when events occurred or how much follow-up each person contributed. A time-to-first-event analysis can incorporate event timing and censoring, but the registry record does not report a time-to-event estimate, its uncertainty, or the statistical model used to obtain one.

The endpoint is relevant to the stated comparative question because COPD exacerbations are the outcomes the study set out to compare. However, the record’s outcome timeframe is listed as “12 Years,” while the endpoint description refers to one-year follow-up in the as-treated analysis. The record does not resolve this difference. It also does not report the number of patients with an event in either treatment group, so the recorded endpoint definition can be described but the observed treatment contrast cannot be interpreted quantitatively.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical methods that can be described from the registry record are limited. It states that the main analysis used matched patients and that LABA-TIO initiators were matched with LABA-ICS initiators using a time-conditional propensity-score approach. It also characterizes the primary outcome as a count of participants with an event. The record does not report the estimation method, adjustment variables, standard errors, confidence intervals, or testing procedure. General concepts below explain what these approaches mean; they do not establish which additional methods were used in this study.

What does propensity-score matching do?

General explanation: a propensity score represents a person’s estimated probability of receiving a particular treatment based on measured characteristics. Matching people with similar scores is intended to construct groups that are more alike with respect to those measured characteristics. A time-conditional approach indicates that timing is incorporated into the matching strategy, but the registry record does not describe its implementation, the variables included, or the matching criteria.

Matching can improve comparability only for characteristics that were measured and handled appropriately. It cannot balance factors that were not measured, were measured poorly, or were not included in the score. It can also change which patients contribute to the comparison: people who cannot be matched may not be represented in the matched analysis. The record does not report how many patients were matched, whether any were excluded through matching, or whether balance was assessed after matching. These omissions make it difficult to judge the quality and scope of the matched comparison.

What does a participant count convey?

General explanation: a count reports how many participants experienced the event, but does not by itself account for the number of participants in each group or their time under observation. Counts can therefore be difficult to compare when group sizes or follow-up differ. A proportion requires a denominator, while a rate also requires information about time at risk. A time-to-event analysis uses event timing and can address incomplete follow-up through an appropriate method, provided the relevant data and assumptions are available.

The registry identifies the primary measure as the number of observed patients with a first event. It does not report the treatment-group counts or denominators, nor does it provide a time-to-event estimate. The endpoint description mentions one-year follow-up, while the recorded timeframe says “12 Years”; the record does not clarify the intended analysis period. Without those details, no further comparative quantity can be derived from the posted information.

Why do uncertainty estimates matter?

General explanation: a point estimate gives a single summary of an observed contrast, while a confidence interval conveys uncertainty around that estimate under the statistical model. A narrow interval generally reflects greater precision than a wide interval, though precision does not guarantee lack of bias. A significance test addresses compatibility with a specified null hypothesis under its assumptions; it does not measure the size or clinical importance of a contrast. The registry record does not report a primary-outcome estimate, confidence interval, or test result, so neither precision nor statistical evidence for a difference can be evaluated.

What is over-dispersion?

General explanation: over-dispersion occurs when count data vary more than a model assumes. If a count model does not account for this extra variation, its uncertainty estimates can be too small and inferential conclusions can be misleading. Whether over-dispersion is relevant depends on the chosen model and the data structure. The registry record does not identify a count model or report an over-dispersion assessment, so it is not possible to say whether this issue arose in the study analysis.

How should planning illustrations be read?

General explanation: sample-size calculations depend on assumptions about the target effect, variability, allocation, significance level, and analysis. A planning illustration based on a fraction of an observed effect is a hypothetical calculation, not a report of the study’s original design assumptions and not observed power. The registry’s stated illustration assumptions include a two-sided significance level of 0.05, equal allocation, and a normal approximation. The record does not report a sample-size justification for the original study design. Any sample-size illustration should therefore be interpreted as planning context rather than evidence about the certainty of the observed findings.

The registry reports three outcome measures and six analyses, but does not describe the specific statistical plan or identify which method was used for each analysis. The analytic details needed to reproduce or fully assess the primary comparison are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists three outcome measures and six analyses. It does not explain how the analyses relate to each outcome, which result was designated as primary for statistical testing, or whether the analyses were prespecified. A multiplicity adjustment, such as a procedure to account for testing several outcomes or contrasts, is not reported in the registry record. Without that information, the implications of multiple analyses for false-positive risk cannot be assessed. This does not establish that no adjustment was used; it means the record does not report one.

The primary analysis is described as involving matched patients and as-treated follow-up. The record does not report the number of patients in that analysis, the number included in each matched group, or how the analysis handled participants who did not remain under observation. It does not describe missing-data methods, censoring rules, exclusions after cohort entry, or sensitivity analyses addressing incomplete information. These details matter because the observed patients may not represent everyone enrolled, and incomplete follow-up can affect the interpretation of event counts and time-to-event comparisons.

The endpoint description refers to one-year follow-up, but its recorded timeframe is “12 Years.” The registry does not clarify whether this difference reflects a labeling issue or distinct follow-up periods. The secondary outcomes are the rate of COPD exacerbations and the occurrence of a first hospitalization for community-acquired pneumonia, described as serious pneumonia. Their analysis populations, missing-data handling, and multiplicity treatment are not reported in the registry record.

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

The record establishes that the study compared initiators of LABA-TIO and LABA-ICS among matched patients with COPD, with the main question focused on a first COPD exacerbation. It also describes the event definition and names secondary outcomes. These details identify the intended clinical comparison and the way a qualifying primary event was defined.

The record does not report the primary outcome counts in either treatment group, a numerical effect estimate, or uncertainty around a comparison. It therefore does not show whether one group had a higher or lower observed outcome count, how large any contrast was, or how precisely a contrast was estimated. No conclusion about a treatment effect can be drawn from the primary outcome information recorded.

General explanation: observational comparisons can describe associations between treatment initiation and subsequent outcomes. Because treatment was not reported as randomized and allocation is not stated, differences between groups may reflect characteristics related to treatment choice as well as treatment itself. Propensity-score matching can address measured characteristics included in the matching process, but it cannot guarantee that all confounding has been removed. The registry record does not report enough information to assess the matching variables or the success of the matching.

The endpoint description’s one-year follow-up and the “12 Years” timeframe are not reconciled in the record. This uncertainty, together with the absence of group counts, denominators, and analysis details, limits what can be inferred about the observed experience. The record also does not report a statistical estimate for either secondary outcome. The appropriate interpretation is therefore bounded: the study question and outcome definitions are recorded, but comparative results and their precision are not reported in the registry record.

12. Why This Trial Matters Statistically

This study illustrates an important statistical distinction between a randomized treatment comparison and an observational comparative-effectiveness analysis. The registry describes treatment initiators who were matched using a time-conditional propensity-score approach, but does not identify randomized allocation. That distinction affects how results should be interpreted: matching can improve balance on measured characteristics, while residual confounding may remain.

The primary outcome also highlights the difference between a participant count and a time-to-event result. The research question concerns time to a first exacerbation, yet the registered measure is the number of observed patients with a first event. Counts alone do not describe event timing or account for differing follow-up. Interpreting a time-to-first-event question calls for clear reporting of the observation period, event counts, denominators, handling of incomplete follow-up, and an effect estimate with uncertainty.

The registry record’s two follow-up descriptions—one-year follow-up in the outcome description and “12 Years” as the recorded timeframe—show why consistent endpoint reporting matters. An unclear time horizon makes it harder to determine what period the outcome represents and to compare the analysis with other evidence.

Finally, the absence of posted outcome counts and effect estimates limits statistical interpretation even though the study question and design features are identifiable. A clear record of matching methods, group sizes, outcome results, uncertainty, multiplicity, and missing-data handling would allow readers to evaluate both the magnitude and the reliability of a comparison. In this record, those result and analysis details are not reported, so the statistical importance lies chiefly in recognizing the limits of what can be concluded.

13. 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 information recorded for this study. First, the comparison is not described as randomized and allocation is not stated. The matched design may address measured differences between treatment initiators, but the registry record does not identify the matching variables or provide balance diagnostics. The extent to which the groups were comparable after matching cannot therefore be assessed.

Second, the primary endpoint has an unresolved follow-up description. The timeframe is recorded as “12 Years,” while the outcome description says one-year follow-up in the as-treated analysis. The record does not explain the difference. This prevents a confident reading of the intended observation period from the registry information alone.

Third, the record gives no primary-outcome counts by treatment group, matched-group denominators, effect estimate, or confidence interval. It also does not report the statistical model, missing-data handling, censoring approach, or sensitivity analyses. The study’s enrollment of 3954 does not reveal how many patients contributed to the matched analysis or how many experienced the primary event.

Fourth, although three outcome measures and six analyses are listed, the record does not describe how multiplicity was handled or which methods were used for the analyses. Finally, adverse-event fields list zero for serious events, other events, and deaths in both treatment groups, and no serious-event terms are listed. These entries do not establish how adverse events were collected or whether the fields represent complete ascertainment. The record does not report further detail that would support a broader safety interpretation.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Comparative Effectiveness of COPD Treatments — NCT03376295.
• Registry record last updated 2019-07-24; status last verified 2019-07.
• Results first posted 2019-07-08.
• 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.

17. Record Summary

The registry record describes a completed comparative-effectiveness study of LABA-TIO versus LABA-ICS among people with COPD. It lists enrollment of 3954 and identifies Boehringer Ingelheim as the sponsor. The primary question concerns time to a first COPD exacerbation, defined as hospitalization for COPD or prescription of oral prednisolone. The main analysis is described as using time-conditional propensity-score matching of treatment initiators.

The comparison is not described as randomized, and allocation is not stated. The record lists a primary measure counting patients with a first event, as well as secondary outcomes concerning exacerbation rate and first hospitalization for community-acquired pneumonia. It records three outcome measures and six analyses.

The primary outcome timeframe is listed as “12 Years,” while the description refers to one-year follow-up. The registry record does not resolve this discrepancy or report group-specific outcome counts, an effect estimate, or its precision. Matching details, missing-data methods, and multiplicity handling are also not reported in the registry record.