This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Impact of Initiating Tiotropium Alone Versus Initiating Tiotropium in Combination With Fluticasone Propionate/Salmeterol Xinafoate Combination (FSC) on Chronic Obstructive Pulmonary Disease-related Outcomes in Patients With Pre-existing Exacerbations (NCT01381406), a registered phase not stated study. The primary outcome is Incidence Rate Per 100 Person Years of Hospitalization or Emergency Department (ED) Visit Related to Exacerbation of Chronic Obstructive Pulmonary Disease (COPD).
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT01381406 |
| Phase | Phase not stated |
| Condition | Pulmonary Disease, Chronic Obstructive |
| Allocation | Not Stated |
| Design model | Not Stated |
| Masking | Not Stated |
| Enrollment | 3333 |
| Arms in the primary outcome | n/a |
| Lead sponsor | GlaxoSmithKline |
| Results posted | Yes |
2. Clinical Question
Population
The target population was managed-care enrollees aged >40 years with COPD, at least one COPD-related exacerbation at baseline, and newly initiating tiotropium during the enrollment period. COPD was identified in claims using specified ICD-9 code groups. The registry describes the population as having at least one pharmacy claim for tiotropium and defines the index prescription as the first qualifying claim in the enrollment period. The reported enrollment is 3333, but the registry record does not give treatment-group sizes or clarify the number included in each outcome analysis.
Intervention
The treatment strategy of interest was initiation of tiotropium with fluticasone propionate/salmeterol combination therapy, abbreviated TIO + FSC. Combination status was based on pharmacy claims during the 30-day combination assessment period after the index tiotropium prescription. The description includes an FSC claim on the same date as tiotropium, or claims whose days supply overlapped within 30 days of the index prescription.
Comparator
The comparison group initiated tiotropium without FSC during the combination assessment period. This is a retrospective cohort comparison, not a randomized allocation. Treatment classification comes from claims, and the registry does not establish that the two groups were comparable in their underlying health or other factors that could affect exacerbation rates.
Primary question
The primary question is how the incidence rate of hospitalization or emergency department visits related to COPD exacerbation compares between the two treatment cohorts. The rate is expressed per 100 person years, with person time used to account for differing follow-up lengths. The registry record does not report the observed rate in either cohort, a between-group estimate, or its uncertainty. Therefore, it describes the question and outcome definition, but not the magnitude or precision of a comparative result.
3. Trial Design
Despite its trial registration, the study is described as a retrospective cohort analysis of administrative claims, rather than a randomized clinical trial. The registry lists enrollment of 3333 and a completed status. It records a start of 2008-07 and a primary completion of 2010-06. These registration dates should not be confused with the claims observation window: the brief description says claims from Jan 1, 2003 through Sep 30, 2007 were used, with an enrollment period for index tiotropium prescriptions from Jan 1, 2004 to Aug 31, 2006.
Eligible managed-care enrollees were identified through claims. The index date was defined by the first chronologically occurring tiotropium claim during the enrollment period. A 1-year pre-index period was used to characterize the period before initiation. Treatment cohorts were defined according to claims for FSC during the 30-day combination assessment period after the index prescription. The registry describes a TIO-alone cohort and a TIO + FSC cohort. It also notes that enrollees adding FSC after the assessment period were handled separately in the study description, but the available text is truncated before the full rule is given.
After the index date, the description distinguishes the combination assessment period from a 1-year follow-up period, while also stating that clinical and economic outcomes were measured over a variable-length follow-up and that the primary endpoint could be collected over a maximum period of 4 years. The registry record does not report the operational details needed to reconcile these follow-up descriptions or specify the follow-up available for each cohort.
The nonrandomized design means treatment assignment could reflect clinical characteristics, prior utilization, prescribing choices, or other factors. Whether and how such differences were controlled is not reported in the registry record. The record names adjusted mean monthly costs and hospitalization and emergency-room visit incidence as secondary outcomes. It does not report the statistical model used for those outcomes or the primary rate comparison.
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: Participants were treated with tiotropium bromide (TIO) or tiotropium bromide plus fluticasone propionate/salmeterol xinafoate combination (TIO + FSC) by their practitioners, and the database included information about their healthcare encounters.
Overall Study
| Stage | Tiotropium Bromide (TIO) | TIO Plus FSC/Salmeterol Xinafoate in Combination (TIO + FSC) |
|---|---|---|
| STARTED | 2481 | 852 |
| COMPLETED | 2481 | 852 |
| NOT COMPLETED | 0 | 0 |
Analysis population for the primary outcome, as defined in the registry: Managed care enrollees (aged \>40 years) with at least one COPD-related exacerbation at baseline and newly initiating therapy with TIO with or without the addition of FSC during the study enrollment period was the target population. The date of TIO-alone therapy or TIO+FSC add-on date was the index date.
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.
| Characteristic | Unit | Tiotropium Bromide (TIO) | TIO Plus FSC/Salmeterol Xinafoate in Combination (TIO + FSC) | Total |
|---|---|---|---|---|
| Participants analysed | 2481 | 852 | 3333 | |
| Age, Continuous | Years | 66.1 (SD 10.9) | 64.6 (SD 11.0) | 65.7 (SD 10.9) |
| Sex: Female, Male: Female | Participants | 1166 | 399 | 1565 |
| Sex: Female, Male: Male | Participants | 1315 | 453 | 1768 |
| Number of health care encounters, as a measure of resource utilization: Home oxygen therapy | number of encounters | 801 | 248 | 1049 |
| Number of health care encounters, as a measure of resource utilization: Intensive care unit stays for COPD | number of encounters | 81 | 29 | 110 |
| Number of health care encounters, as a measure of resource utilization: Severe exacerbation of COPD | number of encounters | 360 | 175 | 535 |
| Number of health care encounters, as a measure of resource utilization: Moderate exacerbation of COPD | number of encounters | 2274 | 747 | 3021 |
| Number of health care encounters, as a measure of resource utilization: Hospitalization for COPD | number of encounters | 467 | 228 | 695 |
| Number of health care encounters, as a measure of resource utilization: Emergency department visit for COPD | number of encounters | 570 | 199 | 769 |
| Mean number of health care encounters, as a measure of resource utilization: SABA containers | number of encounters | 2.4 (SD 5.0) | 2.2 (SD 5.4) | 2.4 (SD 5.1) |
| Mean number of health care encounters, as a measure of resource utilization: Oral corticosteriod prescriptions | number of encounters | 2.1 (SD 2.9) | 1.9 (SD 2.3) | 2.0 (SD 2.8) |
| Mean number of health care encounters, as a measure of resource utilization: Number of physician visits for COPD | number of encounters | 3.3 (SD 3.6) | 2.9 (SD 3.4) | 3.2 (SD 3.6) |
6. Primary Endpoint
The registry lists the primary outcome as Incidence Rate Per 100 Person Years of Hospitalization or Emergency Department (ED) Visit Related to Exacerbation of Chronic Obstructive Pulmonary Disease (COPD), measured over Data were collected over a maximum period of 4 years.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | exacerbations per 100 person years |
| Endpoint classification | Event-Rate Endpoint |
The primary endpoint is the incidence rate of hospitalization or emergency department visits related to COPD exacerbation, expressed as exacerbations per 100 person years. The registry describes severe exacerbations as those with a primary diagnosis of COPD. It describes moderate exacerbations using several claims-based criteria: an emergency department visit with a primary COPD diagnosis; a physician visit with a COPD diagnosis and an oral corticosteroid prescription; a physician visit with a COPD diagnosis code and an antibiotic for respiratory infection; or physician administration of nebulized albuterol within 3 days of an office visit.
An incidence rate relates the number of qualifying events to the amount of time people were observed. This is useful when follow-up duration varies, because a simple event count does not account for how long each person contributed observation time. Person years are the exposure-time denominator in the rate definition. In general, a rate comparison describes event occurrence relative to observed time; it is not the same quantity as the proportion of participants who experienced an event.
Claims-based endpoint definitions make it possible to classify events using recorded diagnoses, visits, and prescriptions across a large administrative database. They also mean the endpoint depends on the completeness and coding of those claims. The registry record does not report validation of the case definition, the number of events, the person-time denominator, the cohort-specific rates, or a comparative estimate with a confidence interval. The endpoint’s definition is therefore available, but its observed value and precision are not reported in the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record identifies a clinically relevant comparison: the rate of COPD exacerbation-related hospitalization or emergency department visits among people initiating tiotropium alone versus people initiating tiotropium with FSC. It gives the endpoint definition and specifies a rate per 100 person years. It does not give the observed rates, the estimated difference or ratio between cohorts, a confidence interval, or a p-value. The direction, magnitude, and statistical precision of the primary comparison therefore cannot be described from the registry record.
This absence of numerical outcome results matters for interpretation. An endpoint definition says what investigators intended to count; it does not show how often the events occurred. Without rates and a comparative estimate, it is not possible to tell whether the cohort results were close together or far apart, or whether the estimate was precise or uncertain. The enrollment listing of 3333 is not a substitute for the treatment-group denominators, the number of events, or the amount of person-time contributing to the primary analysis. Those details are not reported in the registry record.
The design also limits what a comparison could establish. This was a retrospective cohort study based on administrative claims, and the registry states that treatment cohorts were formed according to prescriptions during a defined assessment period. Unlike random allocation, claims-based treatment classification does not itself balance the groups on prognostic factors. In general, differences in underlying health, treatment history, health-care use, or prescribing decisions could be associated with both cohort membership and outcomes. Whether the analysis adjusted for such factors, and which factors were considered, is not reported in the registry record.
The rate uses person years to account for variable observation time. That is an appropriate conceptual distinction from comparing raw event counts, but it does not by itself resolve differences in patient characteristics or follow-up patterns. Interpretation would also depend on the event definition, how recurrent events were handled, and the analysis model. These methods are not reported in the registry record. No conclusion about treatment effect, clinical importance, or comparative benefit should be drawn from the endpoint description alone.
The registry lists adjusted mean monthly costs per COPD patient and incidence rates of hospitalization and emergency-room visits as secondary outcomes. Numerical results for those outcomes are not reported in the registry record either. Thus the record supports a description of the intended observational comparison and its outcome framework, but not a quantitative assessment of its findings.
9. Statistical Methods Explained
The registered outcome is a rate, so the basic statistical idea is to relate event counts to exposure time. The registry defines incidence rate as the number of exacerbations divided by person years and reports the unit as exacerbations per 100 person years. In general, exposure time is useful when observation periods differ: a person observed longer contributes more time at risk than someone observed for a shorter period. A rate is not a risk or proportion, because its denominator is time rather than the number of people.
Why use person years?
For a cohort with variable-length follow-up, comparing raw counts alone can be misleading because the cohorts may contribute different amounts of observation time. A rate puts event counts in relation to accrued follow-up. The registry says outcomes were measured over a variable-length follow-up and describes a maximum period of 4 years for the primary endpoint. It does not report the total person years by cohort, how follow-up ended for each participant, or the precise rules for censoring. The practical interpretation of the rate therefore depends on details that are not reported in the registry record.
What does a confidence interval add?
In general, a confidence interval expresses the uncertainty around an estimated quantity under the assumptions of the statistical method. For a rate comparison, an interval around the comparative estimate would help show which effect sizes are reasonably compatible with the data and model. A narrow interval indicates greater precision than a wide interval, though precision alone does not establish that an estimate is free from bias. The registry record does not report a primary estimate or confidence interval, so the precision of the comparison cannot be assessed.
Why might over-dispersion matter?
In general, count outcomes can vary more than a simple statistical model expects. This is called over-dispersion. If it is present and the analysis does not account for it, standard errors and confidence intervals may be unreliable. Recurrent exacerbations also mean that a participant may contribute more than one event, so the independence assumptions used by some methods may not fit the data. The registry record does not state the model, whether recurrent events were included, or whether over-dispersion or within-person correlation was assessed. These are general considerations, not claims about the analysis used in this study.
What can adjustment address?
In general, regression adjustment can account for measured differences between cohorts by including relevant covariates in a model. It cannot automatically remove confounding from factors that were not measured, were measured poorly, or were modeled inadequately. Because this study classified cohorts through claims rather than random allocation, adjustment choices would be important to interpreting an association. The registry record does not report which covariates were used, whether the primary comparison was adjusted, or how model assumptions were evaluated. The secondary outcome title refers to adjusted mean monthly costs, but this does not establish the adjustment strategy for the primary endpoint.
How should planning calculations be read?
In general, sample-size calculations based on an observed effect are planning illustrations: they ask what study size might be required to detect an effect of a specified size under stated assumptions. They are not observed power, and they do not change the uncertainty in the study that generated the effect estimate. The record describes a two-sided significance level of 0.05, equal allocation, a normal approximation, and planning for stated fractions of the observed effect. It also distinguishes event totals for time-to-event calculations from participants per arm for other calculations. However, numerical planning estimates and the observed effect on which such illustrations would rely are not reported in the registry record. Such calculations cannot be interpreted as the study’s original design justification.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists 3 outcome measures and 3 analyses. It identifies a primary endpoint and names secondary outcomes concerning adjusted mean monthly costs and incidence rates of hospitalizations and emergency room visits. The record does not report whether the analyses were prespecified in a detailed statistical analysis plan, whether any multiplicity adjustment was used, or how the primary and secondary comparisons were prioritized beyond their labels. When several outcomes or comparisons are examined, the chance of at least one apparently notable finding can increase; in general, multiplicity procedures or a clear hierarchy can help define how results should be interpreted. No such procedure is reported in the registry record.
Missing-data methods are not reported in the registry record. In claims-based research, missingness may include incomplete enrollment periods, gaps in observable claims, or unavailable covariates; these are general possibilities, not descriptions of what occurred in this study. The record does not say how such cases were identified, whether participants with incomplete data were excluded, or whether any imputation or sensitivity analysis was conducted. It also does not report outcome-specific analysis denominators.
The population description identifies managed-care enrollees aged >40 years with COPD-related exacerbation history and newly initiating tiotropium, but the analysis populations for the primary and secondary outcomes are not specified. The record does not provide group-specific participant counts, event counts, exclusions, or follow-up completeness. The enrollment listing of 3333 alone does not establish the size or composition of any particular analysis population. These omissions limit assessment of selection, missingness, and the consistency of the analyses across outcomes.
11. What the Result Does, and Does Not, Mean
The record establishes that the primary outcome was defined as a rate of COPD exacerbation-related hospitalizations or emergency department visits, with event time represented through person years. It also describes the claims-based rules used to classify moderate and severe exacerbations and the approach used to form treatment cohorts. These details clarify the intended comparison, but they do not report the outcome values.
Because the primary rates and comparative estimate are not reported in the registry record, the magnitude and direction of the observed association cannot be stated. No conclusion about whether one cohort had a higher or lower rate is supported by the available outcome information. There is also no confidence interval or p-value with which to characterize statistical uncertainty. The registry’s enrollment listing does not fill this gap: enrollment is not a substitute for event counts, cohort-specific person time, or an estimate from the primary analysis.
Even if numerical results were available, this observational comparison would describe an association rather than establish that treatment assignment caused any difference. In general, treatment choices in routine care may be related to clinical factors that also predict outcomes. The record does not report whether adjustment addressed those factors, and statistical adjustment cannot guarantee removal of unmeasured confounding.
The rate scale addresses differing follow-up duration in principle, but does not by itself make cohorts comparable or show the probability that an individual experienced an event. The endpoint relies on administrative claims and specified diagnosis, visit, and prescription criteria, so its interpretation depends on how accurately those records capture qualifying exacerbations. The registry does not report validation information or detailed analytic handling of repeated events.
Accordingly, the record is informative about the study question and measurement framework, but does not support a quantitative claim about comparative outcomes. It should not be read as evidence that a treatment does or does not work. The numerical primary and secondary results, their precision, and the analysis details needed to assess them are not reported in the registry record.
12. Why This Trial Matters Statistically
This study illustrates a common statistical challenge in clinical research: comparing treatment groups assembled from routine-care records rather than random assignment. Such data can address questions about outcomes in managed-care populations, while leaving potential for systematic differences between groups. Understanding the cohort definitions, time windows, and adjustment methods is essential to judging what a comparison can mean.
Its primary endpoint also illustrates why rates and follow-up time must be read together. The registry defines exacerbation incidence relative to person years and describes variable-length follow-up. This moves beyond raw event totals, which can be difficult to compare when observation time differs. At the same time, a rate alone does not show how many people had events, how recurrent events were modeled, or how precisely the comparison was estimated. Those complementary results and methods are not reported in the registry record.
The endpoint definition reflects the use of multiple claims signals to identify exacerbations, including diagnoses, visits, prescriptions, and treatment administration. Such operational definitions make large-scale analysis feasible, but the resulting outcome is tied to what is recorded and how the criteria are applied. This is a useful reminder that statistical interpretation depends not only on the model, but also on the quality and meaning of the underlying measurement.
Finally, the record shows why registration details should be distinguished from an evaluable result. A stated enrollment, outcome definition, and set of analyses do not reveal the estimated treatment-group contrast. Without results and uncertainty measures, readers can describe the research question and design but cannot assess the observed comparative finding. That distinction is central to transparent interpretation of observational clinical evidence.
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.
The registry record does not report the primary outcome rates, comparative estimate, confidence interval, or p-value. It also does not give treatment-group sizes, event counts, or cohort-specific person-time. This prevents assessment of the observed difference and its precision, regardless of the enrollment listing.
The study is described as a retrospective cohort based on administrative claims, and allocation is not described as randomized. Cohort assignment depended on claims during a 30-day assessment period. The record does not report how measured confounding was handled, which covariates were included, or whether unmeasured confounding was considered. It also does not report model diagnostics or sensitivity analyses.
The primary outcome uses claims-based definitions for moderate and severe exacerbations. Validation of these definitions, coding completeness, and the handling of repeated events are not reported in the registry record. The follow-up description includes a 1-year period and variable-length follow-up, with a maximum period of 4 years for data collection. The available description does not fully explain how these periods relate to the analysis or how follow-up ended for each person.
The brief study description is truncated after beginning to describe people who added FSC after the assessment period. The complete handling of those enrollees is therefore not reported in the registry record. Missing-data procedures, analysis populations, multiplicity methods, and the detailed statistical plan are also not reported. Adverse-event fields list zero participants at risk and zero affected for both groups, while deaths are not reported; these fields do not provide a basis for evaluating comparative safety in this claims-based outcome analysis.
14. Related Tutorials
15. Related Statistical Calculators
16. Sources
- ClinicalTrials.gov: Impact of Initiating Tiotropium Alone Versus Initiating Tiotropium in Combination With Fluticasone Propionate/Salmeterol Xinafoate Combination (FSC) on Chronic Obstructive Pulmonary Disease-related Outcomes in Patients With Pre-existing Exacerbations — NCT01381406.
- Registry record last updated 2017-06-14; status last verified 2017-05.
- Results first posted 2011-10-21.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
17. Record Summary
The registry record describes a completed retrospective cohort study of managed-care enrollees with COPD who newly initiated tiotropium alone or with FSC. The reported enrollment is 3333. Claims from Jan 1, 2003 through Sep 30, 2007 were described, with qualifying index prescriptions in an enrollment period from Jan 1, 2004 to Aug 31, 2006. Cohorts were classified using claims during the 30-day combination assessment period.
The primary outcome was the incidence rate of COPD exacerbation-related hospitalization or emergency department visits per 100 person years. The registry describes claims-based definitions for moderate and severe exacerbations and notes variable-length follow-up, with a maximum period of 4 years. Secondary outcomes included adjusted mean monthly costs and hospitalization and emergency-room visit incidence rates.
Numerical outcome results, confidence intervals, group sizes, event counts, and detailed analysis methods are not reported in the registry record. The design is observational, so any reported association would require consideration of confounding and claims-based measurement.