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 ClinicalTrials.gov record.
1. Trial at a Glance
RESPIRE 1 was a randomized, parallel, quadruple-masked phase 3 trial in bronchiectasis. The registry reports 416 enrolled participants, four arms, and two primary statistical comparisons for the registered endpoint of time to first exacerbation event within 48 weeks.
| Feature | RESPIRE 1 |
|---|---|
| Trial name | RESPIRE 1 |
| NCT identifier | NCT01764841 |
| Phase | Phase 3 |
| Condition | Bronchiectasis |
| Brief title | Ciprofloxacin Dry Powder for Inhalation in Non-cystic Fibrosis Bronchiectasis (Non-CF BE) |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 416 |
| Arms | 4 |
| Interventions | Ciprofloxacin DPI (BAYQ3939); Placebo |
| Lead sponsor | Bayer |
| Sponsor type | Industry |
| Status | Completed |
2. Clinical Question
The statistical question was whether ciprofloxacin dry powder for inhalation, compared with pooled placebo, was associated with a difference in the time to first exacerbation event within 48 weeks in participants with bronchiectasis.
Population
Participants enrolled in the phase 3 RESPIRE 1 trial for bronchiectasis, described in the registry brief title as non-cystic fibrosis bronchiectasis.
Intervention
Ciprofloxacin DPI (BAYQ3939), evaluated in two dosing schedules represented in the reported analyses: 28 Days on/Off and 14 Days on/Off.
Comparator
Pooled placebo.
Primary question
Does either ciprofloxacin DPI schedule change the time to first qualifying exacerbation event within 48 weeks compared with pooled placebo?
3. Trial Design
Reported comparison groups
Ciprofloxacin DPI 28 Days on/Off
- Compared with pooled placebo in the primary time-to-event analysis.
- Compared with pooled placebo in the secondary exacerbation-frequency analysis.
Ciprofloxacin DPI 14 Days on/Off
- Compared with pooled placebo in the primary time-to-event analysis.
- Compared with pooled placebo in the secondary exacerbation-frequency analysis.
4. Endpoints
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Time to First Exacerbation Event Within 48 Weeks | Time to first exacerbation was defined as the time from randomization until the visit at which the first qualifying exacerbation is recorded by the investigator. Exacerbation events are defined as exacerbations with systemic antibiotic use and presence of fever or malaise / fatigue and worsening of at least three signs/symptoms. | Up to Week 48 | Time-to-event |
| Number of Participants With Exacerbation Events With Worsening of at Least Three Signs/Symptoms Over 48 Weeks | Number of participants with exacerbation events characterized by worsening of at least three signs/symptoms. | Up to Week 48 | Binary registry outcome; analyzed using Poisson regression in the posted statistical analysis |
The distinction between the two endpoint structures is important. The primary endpoint asks when the first qualifying event occurred, whereas the secondary analysis concerns the frequency of exacerbation events over 48 weeks. Those questions require different statistical summaries.
5. Statistical Methodology
Primary endpoint: Cox proportional-hazards model
For both primary comparisons, the registry states that the hazard ratio for time to first exacerbation event within 48 weeks was calculated using a Cox proportional-hazards model. The comparison was made between each ciprofloxacin DPI schedule and pooled placebo.
A hazard ratio below 1 indicates a lower estimated instantaneous event hazard in the ciprofloxacin comparison group relative to pooled placebo, under the fitted model. It is not a direct ratio of cumulative probabilities of having an exacerbation by Week 48.
Wald-type testing
The registry reports a Wald-type test for the primary endpoint P-values. The registry describes these as Wald-type tests.
A Wald test evaluates the estimated treatment effect relative to its estimated standard error. Conceptually, a test statistic can be represented as:
For a Cox model, the exact parameterization is typically based on the log hazard ratio. The key point is that the test compares the estimated effect with the amount of statistical uncertainty around it.
Secondary endpoint: Poisson regression
The secondary analyses used Poisson regression with adjustment for over-/under dispersion to analyze the number of exacerbation events over 48 weeks. The reported effect measure was the incidence rate ratio.
A rate ratio below 1 indicates a lower estimated event rate in the ciprofloxacin comparison group. This differs from the primary hazard ratio because it addresses event frequency rather than the time until the first event.
Covariate adjustment
The secondary Poisson analyses are explicitly described as including adjustment for over-/under dispersion and are identified in the registry data as involving covariate adjustment. The ClinicalTrials.gov record does not identify the individual covariates used, so this page does not infer them.
Intention-to-treat principle
The registry analysis notes identify intention-to-treat analysis as an associated concept, while the formal analysis-population field specifies the full analysis set as participants who were randomized. This anchors the reported efficacy comparisons to randomized participants rather than restricting the primary comparison to those who completed treatment.
6. Primary Results: Time to First Exacerbation
The primary endpoint was Time to First Exacerbation Event Within 48 Weeks, measured up to Week 48. Both reported primary analyses used the full analysis set and compared a ciprofloxacin DPI schedule with pooled placebo.
Ciprofloxacin DPI 28 Days on/Off vs Pooled Placebo
Hazard ratio for time to first exacerbation
97.5% two-sided CI: 0.5027–1.0690 · P = 0.0650
Analysis: Cox proportional-hazards model with Wald-type test
The estimated hazard ratio of 0.7331 means that the fitted Cox model estimated the instantaneous hazard of a first qualifying exacerbation to be about 73.31% as large in the Cipro 28 comparison group as in pooled placebo. Expressed as a relative model-based interpretation, this corresponds to an estimated hazard that is about 26.69% lower.
That does not mean that 26.69% fewer participants necessarily experienced an exacerbation, nor does it mean that each individual participant had a 26.69% reduction in probability of an exacerbation. The hazard ratio describes a relative time-to-event quantity under the fitted Cox model.
The 97.5% two-sided confidence interval of 0.5027–1.0690 communicates the statistical precision of the estimated hazard ratio. Because the interval extends above 1, the registry-reported interval is compatible with an estimated hazard ratio greater than 1 as well as below 1.
The P-value of 0.0650 is evidence from the reported Wald-type test against the null hypothesis represented by the model-based test. It is not a measure of the magnitude of the treatment effect, and it does not quantify the probability that the null hypothesis is true.
As with other Cox-model analyses, interpretation of a single hazard ratio also depends on the proportional-hazards assumption. The ClinicalTrials.gov record does not report a formal assessment of that assumption, so the estimate should be understood as the reported Cox-model summary rather than as proof that hazards were proportional at every time point.
Ciprofloxacin DPI 14 Days on/Off vs Pooled Placebo
Hazard ratio for time to first exacerbation
97.5% two-sided CI: 0.3568–0.7971 · P = 0.0005
Analysis: Cox proportional-hazards model with Wald-type test
The estimated hazard ratio of 0.5333 means that the fitted Cox model estimated the instantaneous hazard of a first qualifying exacerbation to be about 53.33% as large in the Cipro 14 comparison group as in pooled placebo. Put another way, the model-based estimated hazard was approximately 46.67% lower in the Cipro 14 group.
This is a relative time-to-event estimate. It does not mean that 46.67% of participants avoided exacerbation, that the median time to exacerbation increased by 46.67%, or that every participant received the same proportional reduction in hazard.
The 97.5% two-sided confidence interval of 0.3568–0.7971 describes uncertainty around the estimated hazard ratio. Its values remain below 1 throughout the reported interval, indicating that the reported estimate and its stated confidence interval are on the lower-hazard side of the null value.
The P-value of 0.0005 is the result of the reported Wald-type test. It describes the statistical evidence against the relevant null hypothesis under the specified testing framework; it does not describe the size, clinical importance, or individual-level probability of benefit.
The same Cox-model cautions apply here: censoring is inherent to time-to-event analysis, and interpretation of a single hazard ratio relies on the model assumptions. The ClinicalTrials.gov record does not report a separate diagnostic assessment of proportional hazards or the detailed censoring pattern.
| Primary comparison | Estimate | 97.5% two-sided CI | P-value | Analysis |
|---|---|---|---|---|
| Cipro 28 vs pooled placebo | HR 0.7331 | 0.5027–1.0690 | 0.0650 | Cox proportional-hazards model; Wald-type test |
| Cipro 14 vs pooled placebo | HR 0.5333 | 0.3568–0.7971 | 0.0005 | Cox proportional-hazards model; Wald-type test |
7. Secondary Results: Exacerbation Frequency
The posted secondary statistical analyses address the number of participants with exacerbation events with worsening of at least three signs/symptoms over 48 weeks. The registry states that Poisson regression with adjustment for over-/under dispersion was used to analyze the number of exacerbation events over 48 weeks. The P-value was analyzed using a Wald-type test together with the incidence rate ratio.
Ciprofloxacin DPI 28 Days on/Off vs Pooled Placebo
Incidence rate ratio
97.5% two-sided CI: 0.6264–1.1848 · P = 0.2944
Analysis: Poisson regression with adjustment for over-/under dispersion
The incidence rate ratio of 0.8615 indicates that the estimated exacerbation event rate in the Cipro 28 group was approximately 86.15% of the corresponding estimated rate in pooled placebo. Expressed comparatively, that is an estimated rate about 13.85% lower.
The rate ratio is not a hazard ratio. The primary analysis asks about the time until the first event, whereas the Poisson analysis addresses the frequency of events over the specified period.
The 97.5% two-sided CI of 0.6264–1.1848 is relatively broad and crosses 1. It therefore includes rate-ratio values below and above the null value.
The P-value of 0.2944 is the reported Wald-type testing result. It should not be interpreted as a 29.44% probability that the treatment has no effect, nor should it be used as a substitute for examining the estimated rate ratio and its confidence interval.
Ciprofloxacin DPI 14 Days on/Off vs Pooled Placebo
Incidence rate ratio
97.5% two-sided CI: 0.5237–1.0256 · P = 0.0382
Analysis: Poisson regression with adjustment for over-/under dispersion
The incidence rate ratio of 0.7329 means that the estimated exacerbation event rate in the Cipro 14 group was approximately 73.29% of the estimated rate in pooled placebo. This corresponds to an estimated rate approximately 26.71% lower under the reported model.
Again, this is an event-rate comparison rather than a statement about the probability that an individual participant experienced at least one exacerbation. Multiple events can contribute to a frequency-based analysis, whereas the primary time-to-event endpoint is defined by the first qualifying event.
The 97.5% two-sided confidence interval of 0.5237–1.0256 extends slightly above 1. This means that the interval includes rate-ratio values on both sides of the null value, despite the reported P-value of 0.0382.
The P-value of 0.0382 is the reported Wald-type testing result for the Poisson-regression comparison. It does not measure effect size and does not provide a probability that the estimated rate ratio is correct.
| Secondary comparison | Effect measure | Estimate | 97.5% two-sided CI | P-value |
|---|---|---|---|---|
| Cipro 28 vs pooled placebo | Incidence rate ratio | 0.8615 | 0.6264–1.1848 | 0.2944 |
| Cipro 14 vs pooled placebo | Incidence rate ratio | 0.7329 | 0.5237–1.0256 | 0.0382 |
8. How the Two Primary Analyses Differ
The two primary comparisons answer the same registered endpoint question but evaluate two different ciprofloxacin DPI schedules against pooled placebo. That creates an important statistical distinction: the numerical treatment estimates should be interpreted separately rather than averaged or combined.
| Feature | Cipro 28 vs pooled placebo | Cipro 14 vs pooled placebo |
|---|---|---|
| Endpoint | Time to first exacerbation event within 48 weeks | Time to first exacerbation event within 48 weeks |
| Analysis population | Full analysis set; randomized participants | Full analysis set; randomized participants |
| Model | Cox proportional-hazards model | Cox proportional-hazards model |
| Test | Wald-type test | Wald-type test |
| HR | 0.7331 | 0.5333 |
| 97.5% two-sided CI | 0.5027–1.0690 | 0.3568–0.7971 |
| P-value | 0.0650 | 0.0005 |
It is tempting to focus only on the smaller hazard ratio, but the confidence interval and testing result are equally important. The Cipro 14 estimate is farther below 1 and its reported confidence interval remains below 1, while the Cipro 28 confidence interval extends above 1. Those are descriptive differences in the registry-reported estimates; they do not by themselves constitute a formal statistical comparison between the two ciprofloxacin schedules.
9. Statistical Methods Explained
Why was a Cox model used for the primary endpoint?
The registered primary endpoint is explicitly a time-to-event endpoint: time from randomization until the first qualifying exacerbation. A Cox proportional-hazards model is designed for this structure because it uses both event times and censored follow-up rather than reducing the entire follow-up period to a simple event/no-event indicator.
This is particularly useful when participants can have different amounts of observed follow-up. A participant who has not yet experienced the event at the end of observed follow-up can still contribute information about the time during which they remained event-free.
What does HR 0.5333 mean?
It means the fitted model estimated the instantaneous hazard of a first qualifying exacerbation in the Cipro 14 group to be about 53.33% of the corresponding hazard in pooled placebo. The complementary descriptive statement is an estimated hazard about 46.67% lower.
It does not mean that 46.67% of participants benefited, nor that the cumulative probability of an exacerbation was exactly 46.67% lower.
Why is the confidence interval important?
An effect estimate is only one point on a statistical scale. The confidence interval shows the uncertainty surrounding that estimate under the specified model and sampling framework. For Cipro 28, the reported interval is 0.5027–1.0690; for Cipro 14, it is 0.3568–0.7971.
The fact that these intervals have different relationships to the null value of 1 is important information that is not captured by the point estimates alone.
Why is the P-value not an effect-size measure?
A P-value is influenced by the size of the estimated effect, its statistical precision, and the testing framework. A smaller P-value does not necessarily mean a larger clinical effect, just as a larger P-value does not establish that the estimated effect is exactly zero.
For RESPIRE 1, the primary reported P-values are 0.0650 and 0.0005. Their interpretation must remain tied to the corresponding hazard ratios, confidence intervals, and Wald-type testing procedure.
Why is a rate ratio different from a hazard ratio?
The hazard ratio summarizes the relative instantaneous hazard for the time-to-first-event analysis. The incidence rate ratio from the Poisson model compares estimated event rates over the analysis period.
For example, the Cipro 14 secondary estimate of 0.7329 indicates an estimated event rate about 26.71% lower than pooled placebo. That does not imply that the time to first exacerbation has the same relative effect.
Why adjust a Poisson model for over- or under-dispersion?
A basic Poisson model makes a variance assumption that may not adequately describe observed count data. The registry specifically states that the RESPIRE 1 secondary analysis used Poisson regression with adjustment for over-/under dispersion. That adjustment addresses the relationship between variability and the mean count so that the inferential procedure better reflects the dispersion structure assumed by the analysis.
Why does randomization matter for interpretation?
Randomization establishes the treatment groups through the trial's allocation process rather than through investigators choosing treatment according to participant characteristics. The reported efficacy analysis uses the full analysis set consisting of randomized participants, preserving the randomized comparison as the primary analytical framework.
10. Confidence Intervals and the Null Value
For both the hazard ratio and incidence rate ratio, the null value is 1. A ratio of 1 corresponds to equal estimated hazards or rates between the comparison groups.
HR 0.7331
The point estimate is below 1, but the 97.5% confidence interval extends from 0.5027 to 1.0690.
HR 0.5333
The point estimate is below 1, with a 97.5% confidence interval from 0.3568 to 0.7971.
IRR 0.8615
The point estimate is below 1, while the 97.5% confidence interval is 0.6264–1.1848.
IRR 0.7329
The point estimate is below 1, while the 97.5% confidence interval is 0.5237–1.0256.
These intervals should not be read as ranges in which an individual patient's treatment effect must fall. They quantify uncertainty around the corresponding population-level model estimate under the statistical framework used for the analysis.
11. Primary vs Secondary Evidence
| Evidence component | Question addressed | Reported method | Effect measure |
|---|---|---|---|
| Primary | Time to first exacerbation within 48 weeks | Cox proportional-hazards model; Wald-type test | Hazard ratio |
| Primary | Time to first exacerbation within 48 weeks | Cox proportional-hazards model; Wald-type test | Hazard ratio |
| Secondary | Number of exacerbation events with worsening of at least three signs/symptoms over 48 weeks | Poisson regression with over-/under-dispersion adjustment; Wald-type test | Incidence rate ratio |
| Secondary | Number of exacerbation events with worsening of at least three signs/symptoms over 48 weeks | Poisson regression with over-/under-dispersion adjustment; Wald-type test | Incidence rate ratio |
The duplication in the table is intentional: the same endpoint has two separately reported treatment comparisons. Keeping those comparisons distinct prevents a common analytical error—treating multiple treatment contrasts as though they were a single estimate.
12. Safety Results
The ClinicalTrials.gov record reports serious adverse events by treatment comparison group using affected participants over participants at risk.
| Group | Serious adverse events | At risk |
|---|---|---|
| Ciprofloxacin DPI 28 Days on/Off | 29 | 141 |
| Ciprofloxacin DPI 14 Days on/Off | 23 | 136 |
| Pooled Placebo | 32 | 137 |
These figures describe serious adverse events in the registry-reported affected/at-risk format. They should not be treated as equivalent to the efficacy estimates because the safety endpoint, denominator structure, and analytical purpose are different.
13. Trial Timeline
Trial start
The registry data identify May 2, 2013 as the trial start date.
Primary completion
The registry data identify March 9, 2016 as the primary completion date.
Results posted
The ClinicalTrials.gov record is marked completed and reports eight outcome measures and four statistical analyses.
14. Multiplicity: Why Two Primary Comparisons Deserve Attention
The statistical analyses posted on ClinicalTrials.gov contain two primary endpoint comparisons: Cipro 28 versus pooled placebo and Cipro 14 versus pooled placebo. Both concern the same registered primary endpoint but represent separate treatment contrasts.
When more than one confirmatory hypothesis is tested, the interpretation of the overall type I error depends on the prespecified testing strategy. The ClinicalTrials.gov record identifies the hypothesis type as superiority, but they do not provide an alpha-allocation or multiplicity-adjustment procedure for these two primary comparisons.
15. What the Results Do and Do Not Establish
What the HR describes
A relative model-based comparison of the instantaneous hazard of the first qualifying exacerbation over the analyzed follow-up.
What the HR does not describe
It is not an individual-level probability, a cumulative risk ratio, or a statement that every participant experienced the same proportional change.
What the IRR describes
A relative comparison of estimated exacerbation event rates over 48 weeks under the reported Poisson model.
What the IRR does not describe
It is not the same estimand as time to first exacerbation and should not be substituted for the primary hazard ratio.
16. Statistical Interpretation of the Four Reported Estimates
| Estimate | Descriptive interpretation | Key caution |
|---|---|---|
| HR 0.7331 | Estimated instantaneous hazard was about 26.69% lower in Cipro 28 than pooled placebo under the Cox model. | 97.5% CI includes 1; proportional-hazards interpretation applies. |
| HR 0.5333 | Estimated instantaneous hazard was about 46.67% lower in Cipro 14 than pooled placebo under the Cox model. | Still a model-based relative measure, not an individual probability or absolute risk difference. |
| IRR 0.8615 | Estimated exacerbation event rate was about 13.85% lower in Cipro 28 than pooled placebo under the Poisson model. | 97.5% CI includes 1; event frequency is a different estimand from time to first event. |
| IRR 0.7329 | Estimated exacerbation event rate was about 26.71% lower in Cipro 14 than pooled placebo under the Poisson model. | 97.5% CI extends above 1 despite the reported P-value of 0.0382. |
These translations are simple consequences of the reported ratios: for a ratio below 1, subtracting the ratio from 1 expresses the complementary relative reduction. They do not introduce a new statistical estimate.
17. Limitations
- Limited registry detail: the ClinicalTrials.gov record does not provide the complete four-arm enrollment allocation, so this page does not reconstruct individual arm sample sizes from the total enrollment.
- Multiplicity: two primary comparisons are reported, but the ClinicalTrials.gov record does not identify a multiplicity-adjustment or alpha-allocation strategy.
- Cox-model assumptions: the primary hazard ratios come from Cox proportional-hazards models. The ClinicalTrials.gov record does not report a formal assessment of proportional hazards.
- Censoring information: the registry data do not provide the underlying event and censoring times, preventing reconstruction of a valid Kaplan-Meier curve or independent survival-function estimates.
- Secondary endpoint structure: the registry's outcome description and the posted statistical method use different levels of description—the outcome is presented as a participant-level measure while the analysis text explicitly models the number of exacerbation events. The analysis interpretation therefore follows the posted Poisson-regression description.
- Safety inference: serious adverse-event counts are provided by arm, but the ClinicalTrials.gov record does not include formal statistical comparisons for safety.
- Covariate detail: the secondary analysis is identified as involving covariate adjustment, but the ClinicalTrials.gov record does not specify the individual covariates.
- External generalization: this page is restricted to the trial information from the registry and does not add baseline characteristics, subgroup results, long-term outcomes, or other findings from publications.
18. Why This Trial Matters Statistically
RESPIRE 1 provides a compact example of how a single clinical question can require more than one statistical estimand. The primary endpoint is time to the first qualifying exacerbation, while the secondary analysis addresses the frequency of exacerbation events over 48 weeks.
| Concept | How it appears in RESPIRE 1 |
|---|---|
| Randomization | The trial uses randomized allocation. |
| Parallel design | The registry identifies the design model as parallel. |
| Quadruple masking | The registry identifies masking as quadruple. |
| Time-to-event analysis | Time to first exacerbation is the registered primary endpoint. |
| Cox regression | Used to calculate the primary hazard ratios. |
| Hazard ratio | Primary treatment effect measure for time to first exacerbation. |
| Wald / z-test | Used for the reported primary P-values. |
| Poisson regression | Used for the secondary exacerbation-frequency analyses. |
| Rate ratio | Secondary effect measure for exacerbation event frequency. |
| Over-/under-dispersion | The secondary Poisson analysis explicitly adjusted for dispersion. |
| Confidence intervals | 97.5% two-sided intervals accompany all four posted statistical estimates. |
| ITT-related analysis | The reported analyses use the full analysis set of randomized participants and identify intention-to-treat analysis as an associated concept. |
19. Primary Endpoint vs Event-Rate Analysis
A useful statistical lesson from RESPIRE 1 is that time and frequency are not interchangeable.
Suppose two treatment groups had exactly the same number of exacerbation events but those events occurred at different times. A time-to-event analysis could distinguish the groups because it incorporates when events occurred. Conversely, a count/rate analysis focuses on how many events occurred over the observation period.
That is why the trial's two sets of estimates should be read as complementary rather than redundant:
The primary analysis asks how treatment assignment relates to the timing of the first qualifying event.
The secondary analysis asks how treatment assignment relates to the frequency of exacerbation events over 48 weeks.
20. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The registry-reported Cox-model estimates for the primary endpoint were 0.7331 and 0.5333, with corresponding 97.5% two-sided confidence intervals of 0.5027–1.0690 and 0.3568–0.7971. The secondary incidence rate ratios were 0.8615 and 0.7329.
Clinical interpretation
Clinical meaning requires considering the endpoint definition, the magnitude and precision of the effect estimate, the analysis population, and the distinction between first-event timing and event frequency. The ClinicalTrials.gov record does not provide enough additional information to construct a broader clinical benefit-risk assessment.
This separation is important. Statistical significance, effect magnitude, endpoint meaning, and clinical importance are related but distinct concepts. A well-reported trial analysis keeps those concepts separate rather than collapsing them into a single conclusion.
21. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
22. Related Statistical Calculators
Use these tools to explore the statistical methods represented in the RESPIRE 1 analysis:
23. Sources
- ClinicalTrials.gov: NCT01764841 — RESPIRE 1.
- PubMed: PubMed record for PMID 29371383.
- PubMed: PubMed record for PMID 28495619.
Continue through the Clinical Biostats statistical library
Explore tutorials and statistical calculators covering survival analysis, regression, confidence intervals, hypothesis testing, and clinical-trial methodology.
24. Record Summary
RESPIRE 1 provides a useful statistical teaching case because its reported analyses distinguish between two related but different questions: time to the first qualifying exacerbation and frequency of exacerbation events over 48 weeks. The primary endpoint was analyzed with Cox proportional-hazards models and Wald-type tests, producing hazard ratios of 0.7331 for Cipro 28 versus pooled placebo and 0.5333 for Cipro 14 versus pooled placebo. The secondary analyses used Poisson regression with adjustment for over-/under dispersion and produced incidence rate ratios of 0.8615 and 0.7329, respectively.
The most important statistical lesson is that these estimates should not be collapsed into a single measure of treatment effect. The hazard ratio describes the relative timing of the first qualifying event, while the incidence rate ratio describes event frequency. Confidence intervals provide information about precision, P-values describe evidence under the specified testing procedure rather than effect magnitude, and the model assumptions determine how the reported ratios should be interpreted.