Effectiveness of Tiotropium + Olodaterol Versus Inhaled Corticosteroids (ICS) + Long-acting β2-agonists (LABA) Among COPD Patients in Taiwan: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effectiveness of Tiotropium + Olodaterol Versus Inhaled Corticosteroids (ICS) + Long-acting β2-agonists (LABA) Among COPD Patients in Taiwan (NCT05402020), a registered phase not stated study. The primary outcome is Number of Subjects With Event First Moderate or Severe COPD Exacerbations After Index Date.

2. Clinical Question

This is a real-world, non-interventional comparison rather than a randomized treatment assignment. The index date was the first dispensing of either Tio/Olo or an ICS/LABA combined inhaler. The primary outcome was followed from that date until the first qualifying event, for up to 6 years. A moderate exacerbation was defined using an outpatient COPD diagnosis together with a prescription for an oral corticosteroid or an antibiotic for respiratory infections. A severe exacerbation was defined by hospitalization or an emergency room visit with a primary COPD diagnosis.

The question is therefore about recorded exacerbation events in the defined cohorts, not about treatment assignment under randomized conditions. Treatment choice and the characteristics that influence it may differ between cohorts. Propensity score matching was used to make the cohorts more comparable on recorded characteristics, but the full matching specification and outcome results are not reported in the registry record.

3. Trial Design

The study is registered as completed and is described as non-interventional. Its stated aim was to assess the effectiveness and safety profile of Tio/Olo compared with ICS/LABA in people with COPD in Taiwan. The record describes use of Taiwan National Health Insurance (NHI) claims data between 2014 and 2019, alongside Taiwan Cancer Registry and Taiwan Mortality Data for cohort descriptions. Treatment initiation for both cohorts is specified between 1st January 2014 and 31st December 2019.

The primary endpoint is a participant count: the number of subjects with a first moderate or severe COPD exacerbation after the index date. The endpoint timeframe begins at the first dispensing of either combination and continues until the first event, up to 6 years. The cohorts were propensity score matched 1:1. The population description says matching addressed treatment-initiation year and season, prior COPD treatment, and use of respiratory drugs, but that description is incomplete in the registry record. The full covariate list and matching details are not reported in the registry record.

The registry lists 17018 as enrollment. That registry figure does not, by itself, state how many participants contributed to each matched cohort or to the primary-outcome analysis. The record lists 6 outcome measures and 7 analyses. It also lists secondary outcomes concerning triple-therapy escalation and initiation, first hospitalization for community-acquired pneumonia, rescue-medication prescriptions, and COPD exacerbations.

The record does not report the analysis model, allocation method beyond the non-interventional design, or a detailed statistical analysis plan. The primary outcome estimates and uncertainty measures are not reported in the registry record. Adverse-event entries contain zeroes for the listed event and death fields in both cohorts, but the record does not give context sufficient to interpret those entries as a complete safety analysis.

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: All subjects were screened for eligibility prior to participation in the trial. Subjects were not to be allocated to a treatment group if any of the entry criteria were violated. Only subjects that met all inclusion and none of the exclusion criteria were included in the study.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Propensity score matched cohorts: Subjects were 1:1 propensity score matched between cohorts. The two groups were comparable with regards to year of treatment initiation, season of treatment initiation, prior treatment for Chronic Obstructive Pulmonary Disease (COPD), use of respiratory drugs of interest, frequency and severity of COPD exacerbation, prevalence of major comorbidity, Charlson Comorbidity Index, and concomitant medications.

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: Propensity score matched cohorts: Subjects were 1:1 propensity score matched between cohorts. The two groups were comparable with regards to year of treatment initiation, season of treatment initiation, prior treatment for COPD, use of respiratory drugs of interest, frequency and severity of Chronic Obstructive Pulmonary Disease (COPD) exacerbation, prevalence of major comorbidity, Charlson Comorbidity Index, and concomitant medications.

6. Primary Endpoint

The registry lists the primary outcome as Number of Subjects With Event First Moderate or Severe COPD Exacerbations After Index Date, measured over From the index date (the first dispensing of either Tio/Olo or ICS/LABA combined inhaler) until first event, up to 6 years..

The primary endpoint counts participants who experience a first moderate or severe COPD exacerbation after the index date. It combines two operational event definitions. A moderate event requires an outpatient visit with a COPD diagnosis in any field and a prescription for an oral corticosteroid or an antibiotic for respiratory infections. A severe event is a hospitalization or emergency room visit with a primary COPD diagnosis. The index date is the first dispensing of either Tio/Olo or ICS/LABA, and follow-up extends until the first event, up to 6 years.

A first-event endpoint focuses on whether a qualifying event is recorded after treatment initiation, rather than counting every exacerbation. This can make the outcome straightforward to express as a participant count. It also means that the count alone does not describe the number of recurrent events or their timing. A participant with multiple qualifying events contributes to the first-event measure through the earliest one; recurrent-event burden is a different question.

Claims-based definitions make outcomes observable through recorded diagnoses, care settings, and prescriptions. They do not necessarily capture every clinical episode or establish the event with the same detail as direct clinical assessment. The registry record does not report the primary-outcome counts, the cohort denominators for the analysis, or a comparative estimate. Consequently, the endpoint definition can be described, but the comparative outcome cannot be quantified from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Several design and statistical concepts help frame the comparison, while the registry record leaves important analytic details unspecified.

What does non-interventional mean for the comparison?

In a non-interventional study, treatment groups are identified from care or records rather than assigned by the study through randomization. This can support analysis of real-world data, but treatment choice may be related to patient characteristics, disease history, clinician decisions, or other factors that also relate to outcomes. A difference between cohorts can therefore reflect treatment, differences in who received each treatment, or both. The registry record does not report a randomized allocation process; it describes a non-interventional study.

What does propensity score matching contribute?

Propensity score methods use measured characteristics to describe the likelihood of receiving one treatment rather than another. Matching participants across cohorts on that score is intended to improve balance in the measured characteristics used to build it. Here the record states that cohorts were matched 1:1 and names treatment-initiation year and season, prior COPD treatment, and respiratory-drug use among the characteristics considered. The full model, covariate list, matching algorithm, caliper or other matching criteria, and balance diagnostics are not reported in the registry record.

Matching cannot balance characteristics that were not measured, not recorded, or not included in the model. It also does not automatically eliminate bias from inaccurate data or from differences in follow-up and outcome ascertainment. General statistical interpretation should therefore distinguish adjustment for recorded factors from proof that the cohorts are exchangeable.

How should the primary count be understood?

The endpoint is the number of participants with a first qualifying exacerbation. A count of participants with an event is not the same measure as an event rate, because a rate incorporates time at risk and a participant count does not, on its own, describe follow-up duration. The registry defines a follow-up window of up to 6 years, but the distribution of follow-up time and the analysis method for varying follow-up are not reported in the registry record. The first-event definition also differs from a recurrent-event analysis, which would consider repeated exacerbations.

The operational definitions use diagnoses, prescriptions, and care settings in claims data. This can make large-scale record-based outcome assessment feasible. In general, claims definitions can misclassify events if coding or treatment patterns do not correspond perfectly to the clinical event of interest. Whether or how such misclassification was evaluated is not reported in the registry record.

What can confidence intervals and p-values tell readers?

In general, a confidence interval describes the range of effect values that are reasonably compatible with the data and the assumptions of the analysis. Its width conveys precision: a wider interval indicates greater uncertainty about the effect magnitude. A p-value is a measure of how incompatible the data are with a specified null model, under the model's assumptions; it does not measure clinical importance or the probability that a treatment effect is true. No primary-outcome confidence interval or p-value is reported in the registry record, so neither statistical precision nor a test result can be assessed.

What is the role of the planning illustration?

The record describes planning calculations based on the observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. It specifies that the effect to detect is a stated fraction of the observed effect. In general, a sample-size illustration estimates the number of participants or events that a hypothetical study might need under its stated assumptions. It is not an outcome finding, and it is not observed power. The record does not include the observed effect or the resulting planning figures, so those illustrations cannot be independently interpreted from the registry record.

The exact outcome model, handling of censoring, covariate adjustment after matching, standard-error method, and sensitivity analyses are not reported in the registry record. These details matter because they determine how the matched data are used and how uncertainty is quantified.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Tiotropium + Olodaterol Cohort - Propensity Score ", "Inhaled Corticosteroids (ICS) + Long-acting ß2 Ago")
n      <- c(3752, 3752)
events <- c(424, 375)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Tiotropium + Olodaterol Cohort - Propensity Score   3752  424  11.3  10.3  12.4  10.3  12.4  10.3  12.4
ARM  Inhaled Corticosteroids (ICS) + Long-acting ß2 Ago  3752  375  10.0  9.1  11.0  9.1  11.0  9.1  11.0

11. Multiplicity, Missing Data and Analysis Populations

The record lists 6 outcome measures and 7 analyses, including the primary outcome and secondary outcomes addressing treatment escalation, pneumonia hospitalization, rescue-medication prescriptions, and COPD exacerbations. Testing or estimating multiple outcomes can raise a multiplicity question: when many comparisons are considered, the chance of at least one apparently notable result may increase. In general, a prespecified hierarchy or adjustment procedure can help clarify how multiple results should be interpreted. The registry record does not report a multiplicity strategy, a hierarchy among the outcomes, or which analyses were designated confirmatory.

Missing-data handling is not reported in the registry record. The record does not describe missingness in the claims data, how incomplete outcome or covariate information was handled, whether participants could leave the analysis because of data availability, or whether any sensitivity analyses assessed the impact of missing information. These omissions prevent assessment of how missingness might affect the estimates.

The primary endpoint population is described as propensity score matched cohorts, but the number of participants in those cohorts and the analysis denominators are not reported in the registry record. The broader enrollment figure is 17018; it does not specify the matched analysis population. The record also does not report whether all matched participants were included in each outcome analysis or how follow-up ended for people without a recorded event.

These are reporting gaps, not evidence that a particular method was or was not used. Without the analysis populations, missing-data approach, and multiplicity plan, readers cannot fully evaluate the scope of the analyses or the robustness of the reported results.

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

The registry record establishes that the study compared Tio/Olo initiators with ICS/LABA initiators in Taiwan using a non-interventional design and claims-based data. It defines a first moderate or severe COPD exacerbation as the primary endpoint and describes propensity score matching. These details clarify what comparison was intended, but they do not reveal the outcome finding.

No primary outcome count or comparative estimate is reported in the registry record. Thus, the record does not support a statement that one cohort had more or fewer events, nor a conclusion about the magnitude or precision of any difference. It also does not provide the denominator and follow-up information needed to interpret a participant count as a risk or rate. The registry enrollment of 17018 should not be read as the number analyzed in the matched primary-outcome comparison.

Even if an estimate were available, a non-randomized comparison would require consideration of confounding and the quality of claims-based outcome ascertainment. Propensity score matching addresses measured characteristics included in the matching process; it cannot by itself establish causality or guarantee balance on unmeasured factors. The full matching specification and diagnostics are not reported in the registry record.

The listed zeroes in adverse-event fields do not establish that no safety events occurred, because the record does not explain the completeness or interpretation of those entries. Likewise, outcome names and planning assumptions do not constitute outcome results. Any conclusions about comparative effectiveness or safety would require the actual analysis estimates, uncertainty measures, and fuller methodological reporting, none of which are reported in the registry record.

13. Why This Trial Matters Statistically

This study addresses a statistical challenge common to real-world comparative effectiveness research: comparing treatment groups when treatment was not assigned at random. Its use of Taiwan health records makes the setting and data source central to interpretation. The key statistical question is not only whether outcomes differ, but also whether measured cohort characteristics were adequately balanced and whether important sources of confounding remain.

The primary outcome is defined with claims-based criteria that combine diagnosis, prescriptions, and care setting. That definition makes the endpoint operational for record-based analysis, while also raising questions about how well recorded indicators capture the clinical events of interest. The study further focuses on the first qualifying exacerbation, which is distinct from measuring all exacerbations over follow-up.

The registry describes 1:1 propensity score matching and identifies some characteristics considered in matching. It does not report the complete matching model, balance diagnostics, or primary estimates. Those omissions matter statistically: readers need the outcome estimate and its uncertainty to assess the observed comparison, and need methodological details to judge how much confidence to place in it.

The record also lists multiple outcomes and analyses, making the handling of multiplicity and the distinction between primary and secondary findings relevant. Because multiplicity procedures and analysis details are not reported, the strength of any inference cannot be evaluated from the registry record. The study's design and endpoint are informative, but the missing results and analytic detail limit what can be concluded from the registration information alone.

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 it does not report the primary outcome results. Event counts, comparative estimates, uncertainty intervals, and p-values are not reported in the registry record. Without them, the direction and precision of the comparison cannot be assessed.

The study is non-interventional, so treatment groups may differ in ways related to both treatment selection and exacerbation risk. Propensity score matching was used, but the full matching specification and balance diagnostics are not reported in the registry record. The available cohort description is incomplete, and the record does not report which potential confounders were measured or omitted.

The primary outcome depends on claims-based definitions. The registry record does not report validation of the event definitions, the completeness of claims ascertainment, or how changes in follow-up and censoring were handled. It also does not give the matched cohort sizes or outcome-specific denominators, so the enrollment figure cannot be used to infer the analyzed population.

Further details are not reported in the registry record, including the missing-data approach, multiplicity strategy, exact outcome model, and sensitivity analyses. Although the study states an aim to assess safety, the adverse-event fields list zeroes without context on ascertainment or completeness. These limitations do not establish that the analyses were inadequate; they describe what cannot be evaluated from the registry record.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Effectiveness of Tiotropium + Olodaterol Versus Inhaled Corticosteroids (ICS) + Long-acting β2-agonists (LABA) Among COPD Patients in Taiwan — NCT05402020.
• Registry record last updated 2024-08-09; status last verified 2024-08.
• Results first posted 2024-08-09.
• 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.

18. Record Summary

Registered as NCT05402020, this completed non-interventional study compared tiotropium plus olodaterol with inhaled corticosteroids plus long-acting β2-agonists among people with COPD in Taiwan. The record lists enrollment of 17018. The study used Taiwan National Health Insurance claims data between 2014 and 2019, with cohort descriptions also referring to Taiwan Cancer Registry and Taiwan Mortality Data.

The primary endpoint was the number of participants with a first moderate or severe COPD exacerbation after the first dispensing of either treatment combination, followed until the first event for up to 6 years. Cohorts were propensity score matched 1:1. The record lists 6 outcome measures and 7 analyses.

Primary outcome estimates, analysis denominators, and key statistical details are not reported in the registry record. The record therefore describes the study question and endpoint but does not permit an assessment of the comparative result.