This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
QUARTZ is a recruiting observational study in asthma with planned enrollment of 8,000 participants. The registry describes three cohorts and a primary endpoint focused on severe exacerbations during follow-up, summarized using annualized asthma exacerbation rates and rate ratios.
| Feature | QUARTZ |
|---|---|
| Study name | QUARTZ: Comparative Effectiveness of Switching From Albuterol to Albuterol/Budesonide Versus Adding Low-dose ICS or LTRA to Albuterol |
| NCT identifier | NCT07787117 |
| Status | RECRUITING |
| Condition | Asthma |
| Study start | 2026-08-03 |
| Primary completion date | 2026-09-30 |
| Lead sponsor | AstraZeneca |
| Sponsor type | INDUSTRY |
| Enrollment | 8,000 |
| Number of arms | 3 |
| Intervention field | None (observational study) (other) |
2. Clinical Question
The registry describes a comparative-effectiveness question in people with asthma: how do severe exacerbation rates observed among the Airsupra cohort compare with rates among each comparator cohort?
Population
People with the condition listed in the registry as asthma who enter the observational study and are followed according to the registry-defined follow-up period.
Comparative exposure
The study title describes switching from albuterol to albuterol/budesonide, with comparator strategies involving adding low-dose ICS or LTRA to albuterol.
Study structure
The registry identifies the study as observational, with three arms or cohorts.
Primary question
How does the occurrence of severe asthma exacerbations compare between the Airsupra cohort and each comparator cohort during follow-up?
3. Trial Design
The registry characterizes QUARTZ as an observational study rather than a randomized interventional trial. That distinction is central to the statistical interpretation. The primary endpoint is explicitly framed as a comparison of observed severe exacerbation rates among cohorts.
Airsupra cohort
- The registry identifies Airsupra as the cohort used for comparison with each comparator cohort.
- The primary endpoint includes absolute rates and rate ratios comparing Airsupra with each comparator cohort.
Comparator cohorts
- The study title identifies two comparator strategies.
- These are adding low-dose ICS or LTRA to albuterol.
- The primary endpoint compares Airsupra with each comparator cohort.
4. Endpoint and Follow-Up Definition
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Severe exacerbations during follow-up | Annualized asthma exacerbation rates (AAER) by cohort, overall and by setting (IP, ED/urgent care, outpatient); plus absolute rates and rate ratios comparing Airsupra with each comparator cohort. | Follow-up starts the day after index and ends at the earliest of death, data end, insurance disenrollment, loss to follow-up, or switch to the alternative rescue treatment, up to 36 months; add-on maintenance therapy does not censor follow-up. |
The registry describes the purpose of the endpoint as describing and comparing the occurrence of severe exacerbations observed among the Airsupra cohort and each comparator cohort. This endpoint is fundamentally a rate-based outcome rather than a simple binary outcome such as whether a participant ever experienced an exacerbation.
Why the follow-up rule is statistically important
Each participant contributes observation time until a prespecified censoring event or the end of the allowed follow-up. Consequently, the denominator for an annualized rate is person-time rather than simply the number of enrolled participants.
The rule that add-on maintenance therapy does not censor follow-up is also important. Participants remain under observation for the primary endpoint despite that treatment change, whereas switching to the alternative rescue treatment is explicitly listed as a reason for ending follow-up.
5. Statistical Methodology
Annualized asthma exacerbation rate
The primary endpoint is expressed as an annualized asthma exacerbation rate, or AAER. Conceptually, a rate is the number of observed exacerbation events divided by the amount of follow-up time during which those events could occur.
Annualization expresses the observed event frequency on a common yearly time scale, allowing cohorts with different amounts of observed follow-up to be compared using rates rather than raw event counts alone.
Absolute rates
The registry specifies absolute rates for the Airsupra cohort and each comparator cohort. Absolute rates are useful because they describe the observed burden of severe exacerbations directly rather than only expressing a relative comparison.
Rate ratios
The registry also specifies rate ratios comparing Airsupra with each comparator cohort. A rate ratio compares two event rates over their respective observed person-time.
A rate ratio below 1 would indicate a lower observed exacerbation rate in the Airsupra cohort relative to the comparator cohort; a value above 1 would indicate a higher observed rate.
Event-setting analysis
The endpoint is also summarized by setting: inpatient (IP), emergency department/urgent care, and outpatient. This separates the overall exacerbation burden into clinically distinct settings rather than collapsing all events into a single category.
Person-time and censoring
The registry-defined follow-up creates a variable observation window. Follow-up ends at the earliest of death, data end, insurance disenrollment, loss to follow-up, or switch to the alternative rescue treatment, with a maximum of 36 months. Correct rate analysis therefore depends on correctly accounting for the amount of time each participant remains observable.
6. Statistical Methods Explained
Why use an annualized rate instead of simply counting exacerbations?
Participants may contribute different amounts of observable follow-up. A raw count does not distinguish one participant observed briefly from another observed for a much longer period. An annualized rate incorporates the amount of person-time contributing to the observation.
What does a rate ratio measure?
A rate ratio describes the relative frequency of events per unit of observed person-time. For example, a rate ratio of 0.80 would mean that the estimated event rate in the numerator cohort is 80% of the comparator rate, corresponding to a 20% lower estimated rate. Such an interpretation concerns event rates, not the probability that an individual patient will experience an exacerbation.
Why report absolute rates as well as rate ratios?
Relative measures can make two cohorts appear substantially different even when their absolute event burden is small. Conversely, a modest relative difference can correspond to a meaningful absolute difference when the underlying event rate is high. Reporting both gives readers two complementary views of the endpoint.
Why separate inpatient, ED/urgent care, and outpatient events?
The same overall exacerbation rate can arise from different distributions of event settings. Breaking the endpoint down by IP, ED/urgent care, and outpatient setting provides information about where the observed severe exacerbations occurred.
Why does switching to the alternative rescue treatment end follow-up?
The registry explicitly defines switching to the alternative rescue treatment as a censoring event. Once a participant switches, the observation period under the original comparative framework ends. This makes the follow-up definition part of the estimand: the rate describes events observed before the earliest prespecified censoring event.
Why does add-on maintenance therapy not censor follow-up?
The registry explicitly states that add-on maintenance therapy does not censor follow-up. Participants therefore remain eligible to contribute observed exacerbation time despite that maintenance-treatment change, subject to the other specified censoring rules.
7. Planned Analysis
The planned primary analysis centers on severe exacerbations during follow-up. The registry specifies annualized asthma exacerbation rates by cohort, both overall and by setting, together with absolute rates and rate ratios comparing the Airsupra cohort with each comparator cohort.
For an endpoint of this type, the principal analysis would ordinarily treat the number of exacerbation events as a count accumulated over observed person-time. A rate-based model or equivalent person-time comparison can then estimate the relative rate between cohorts while accounting for unequal follow-up. The registry does not identify a specific regression distribution, link function, offset specification, or variance estimator for the primary endpoint.
The follow-up definition is therefore an important component of the planned analysis. Death, data end, insurance disenrollment, loss to follow-up, and switching to the alternative rescue treatment terminate observation, while add-on maintenance therapy does not. These rules determine which person-time contributes to the numerator and denominator of the rate comparison.
| Planned quantity | Statistical role |
|---|---|
| Annualized asthma exacerbation rate | Describes severe exacerbation frequency per unit of observed follow-up time. |
| Absolute rate | Provides the observed event burden for each cohort. |
| Rate ratio | Compares the exacerbation rate in the Airsupra cohort with each comparator cohort. |
| Setting-specific rate | Describes rates separately for IP, ED/urgent care, and outpatient settings. |
8. Understanding Rate Ratios
A rate ratio compares the frequency of exacerbation events per unit of observed person-time between two cohorts. If the Airsupra-to-comparator rate ratio were less than 1, the observed exacerbation rate would be lower in the Airsupra cohort. If it were greater than 1, the observed rate would be higher.
A rate ratio is not a probability, odds ratio, risk ratio, hazard ratio, or absolute risk difference. Its denominator is the comparator cohort's event rate rather than the number of participants who experience an event.
Suppose two cohorts each experienced the same number of exacerbations but one cohort contributed substantially more person-time. Their crude event counts would be identical, but their rates would not be. The annualized rate therefore incorporates exposure time into the comparison.
When formal confidence intervals are eventually reported, they will describe statistical uncertainty around the estimated rate or rate ratio under the specified analysis framework. A confidence interval should not be interpreted as the range of individual patient outcomes.
If a formal hypothesis test is subsequently reported, its p-value would address evidence against a specified null hypothesis under the statistical model. It would not measure the magnitude or clinical importance of the rate difference. Absolute rates and the rate ratio remain necessary for interpreting the size of an observed association.
9. Censoring and the Analysis Population
QUARTZ's primary endpoint is particularly sensitive to how follow-up is defined because the outcome is a rate accumulated over time. The registry specifies a clear set of events that terminate follow-up:
| Follow-up rule | Statistical consequence |
|---|---|
| Follow-up starts the day after index | The risk-observation period begins after the index date rather than on the index date itself. |
| Death | Observation ends when the participant dies. |
| Data end | Observation ends when available data end. |
| Insurance disenrollment | Observation ends when the participant leaves the observable insurance population. |
| Loss to follow-up | Observation ends when follow-up is no longer available. |
| Switch to alternative rescue treatment | Observation ends when the participant switches to the alternative rescue treatment. |
| Add-on maintenance therapy | Does not censor follow-up. |
| Maximum duration | Follow-up extends up to 36 months. |
These rules are more than administrative details. They define the observation window used to calculate the annualized event rate. In observational research, informative loss to follow-up or differences in treatment switching can also affect comparability between cohorts, making careful description of censoring especially important.
10. Observational Comparisons and Confounding
Because QUARTZ is described in the registry as observational, the three cohorts are not presented as randomized treatment groups. This changes the interpretation of any eventual rate difference.
In a randomized trial, random assignment is designed to balance measured and unmeasured baseline factors in expectation. In an observational cohort comparison, the choice or receipt of a rescue-treatment strategy can be associated with patient characteristics, disease severity, healthcare utilization, prescribing patterns, or other factors. If such characteristics also affect exacerbation rates, a crude comparison of rates can combine treatment-associated differences with differences in the underlying populations.
Association
An observed difference in annualized exacerbation rates describes an association between cohort membership and the observed outcome under the study's follow-up definition.
Causal effect
A causal interpretation requires assumptions and analytic methods addressing the non-randomized nature of cohort assignment. The ClinicalTrials.gov record does not specify such an adjustment strategy.
This distinction is especially important when interpreting a future rate ratio. A rate ratio can be estimated precisely and still represent an association affected by confounding. Statistical precision and causal validity are different properties.
11. Endpoint Decomposition by Setting
The primary endpoint is intentionally multidimensional. The registry calls for overall annualized asthma exacerbation rates and rates by three settings: inpatient, emergency department/urgent care, and outpatient.
| Setting | Role in the endpoint |
|---|---|
| Overall | Total annualized asthma exacerbation rate for each cohort. |
| IP | Annualized rate of severe exacerbations occurring in the inpatient setting. |
| ED/urgent care | Annualized rate of severe exacerbations occurring in the emergency department or urgent-care setting. |
| Outpatient | Annualized rate of severe exacerbations occurring in the outpatient setting. |
Setting-specific estimates should not automatically be treated as independent confirmatory outcomes. They are components of the same broader severe-exacerbation endpoint, and the statistical interpretation of multiple comparisons depends on the prespecified analysis plan.
12. Multiplicity and Multiple Comparisons
The registry specifies comparisons of the Airsupra cohort with each comparator cohort. It also specifies overall and setting-specific annualized rates. This creates multiple reported quantities even though the registry identifies one primary endpoint.
Multiplicity matters because each additional statistical test can increase the chance of observing an apparently unusual result by chance when many hypotheses are examined. Formal interpretation therefore depends on whether the comparisons are intended to be confirmatory, descriptive, or part of a prespecified hierarchy.
The registry's single primary endpoint contains several planned summaries: overall rates, setting-specific rates, and Airsupra-versus-comparator rate ratios. Their statistical roles should be distinguished rather than treating every reported number as an independent confirmatory test.
The registry does not specify an alpha-allocation strategy, multiplicity-adjustment procedure, or formal testing hierarchy for these comparisons. Those details should therefore not be inferred from the endpoint description alone.
13. Missing Data and Follow-Up Completeness
For a person-time rate analysis, incomplete observation does not necessarily mean that a participant must be excluded. The registry instead defines several circumstances that end follow-up, including insurance disenrollment and loss to follow-up.
The statistical consequence is that each participant's contribution is bounded by the time during which the study can observe the outcome. If follow-up differs systematically between cohorts, the resulting person-time and event counts may also differ systematically.
14. What the Primary Endpoint Does — and Does Not — Mean
An annualized asthma exacerbation rate summarizes the number of observed severe exacerbation events relative to the amount of observed follow-up time. It is a measure of event frequency across the cohort.
The annualized rate is not the percentage of participants who experienced an exacerbation. A participant can contribute more than one event, and the rate incorporates person-time rather than simply counting affected individuals.
A rate ratio compares event frequencies between cohorts after accounting for observed follow-up time. It does not directly state the proportion of individuals who benefit or the probability that a particular patient will experience an exacerbation.
An observed cohort difference is not automatically a randomized treatment effect. Causal interpretation requires assumptions about comparability and confounding that are separate from the numerical calculation of a rate ratio.
15. Design Features That Matter Statistically
| Design feature | Statistical relevance |
|---|---|
| Observational study | Cohort comparisons are not based on random assignment as described in the registry. |
| Three arms | The endpoint includes comparisons of the Airsupra cohort with each comparator cohort. |
| 8,000 enrollment | Defines the planned size of the observational study population. |
| Follow-up up to 36 months | Places an upper bound on the observation period for the primary endpoint. |
| Person-time follow-up | Supports annualized event-rate comparisons when participants contribute different amounts of observable time. |
| Switching as censoring | Defines the point at which observation ends for participants who switch to the alternative rescue treatment. |
| Maintenance therapy does not censor | Allows continued contribution of follow-up after add-on maintenance therapy. |
| Setting-specific rates | Allows the overall endpoint to be examined by IP, ED/urgent care, and outpatient setting. |
16. Limitations and Interpretation Issues
- Observational design: cohort comparisons can be affected by confounding because the study is not described as randomized.
- Treatment switching: switching to the alternative rescue treatment ends follow-up, making the censoring rule an important part of the estimand.
- Variable follow-up: death, data end, insurance disenrollment, loss to follow-up, and treatment switching can result in different amounts of observable person-time.
- Unspecified adjustment: the registry endpoint description does not specify a particular confounding-adjustment model, propensity-score method, weighting strategy, or regression framework.
- Unspecified multiplicity control: the registry describes multiple comparisons and setting-specific summaries but does not identify an alpha-spending or multiplicity-adjustment procedure.
- Rate versus individual risk: annualized rates and rate ratios describe event frequency and should not be interpreted as individual-level probabilities.
- Setting-specific interpretation: inpatient, ED/urgent care, and outpatient rates describe components of the endpoint and may not have the same statistical precision.
- Incomplete follow-up: insurance disenrollment and loss to follow-up can reduce observable person-time and may matter particularly if their occurrence differs between cohorts.
17. Why This Trial Matters Statistically
QUARTZ is a useful teaching case because it illustrates a different statistical structure from the randomized survival trials commonly encountered in clinical-trial reporting. Its primary endpoint is based on recurrent event frequency over person-time rather than time to a single first event.
| Concept | How it appears in QUARTZ |
|---|---|
| Observational cohort comparison | Three cohorts are compared without a randomized allocation described in the registry. |
| Person-time | Follow-up continues until prespecified censoring events or up to 36 months. |
| Recurrent events | The endpoint concerns asthma exacerbation rates rather than simply whether an individual experienced one event. |
| Annualized rate | Exacerbation frequency is expressed on an annualized basis. |
| Rate ratio | The Airsupra cohort is compared with each comparator cohort using rate ratios. |
| Censoring | Death, data end, insurance disenrollment, loss to follow-up, and switching to the alternative rescue treatment end follow-up. |
| Setting-specific analysis | Rates are described overall and by IP, ED/urgent care, and outpatient setting. |
| Confounding | Observational cohort differences require careful distinction between association and causal effect. |
| Multiplicity | Multiple cohort and setting-specific comparisons require attention to the role of each analysis. |
18. ClinicalTrials.gov Record Status
Study start
The registry lists August 3, 2026 as the study start date.
Recruiting
The ClinicalTrials.gov record identifies QUARTZ as recruiting, with planned enrollment of 8,000 participants.
Primary completion date
The registry lists September 30, 2026 as the primary completion date.
19. A Practical Framework for Reading the Future Results
When formal results are posted, the most informative first step will be to read the absolute annualized rates for all three cohorts. These establish the underlying event burden before the relative comparisons are considered.
The next step is to examine the Airsupra-versus-comparator rate ratios and their confidence intervals. A rate ratio summarizes relative event frequency, while the confidence interval indicates the precision of that estimate. The magnitude of the absolute rate difference should be considered alongside the relative measure.
The setting-specific results should then be examined separately. An overall rate difference could arise from differences concentrated in one setting or distributed across inpatient, ED/urgent care, and outpatient events.
Finally, the observational design should remain central to interpretation. Even a narrow confidence interval around a rate ratio describes the precision of the estimated association under the statistical analysis; it does not by itself eliminate confounding or establish a causal treatment effect.
20. Sources
- ClinicalTrials.gov: QUARTZ, NCT07787117.
Continue through the Clinical Biostats trial library
Explore additional clinical-trial analyses, statistical tutorials, and quantitative methods across the Clinical Biostats site.
21. Record Summary
QUARTZ is a recruiting observational study in asthma with planned enrollment of 8,000 participants and three cohorts. Its primary endpoint is severe exacerbations during follow-up, measured as annualized asthma exacerbation rates overall and by inpatient, emergency department/urgent care, and outpatient setting, together with absolute rates and rate ratios comparing the Airsupra cohort with each comparator cohort.
The statistical structure is therefore centered on recurrent event rates, person-time, rate ratios, and censoring rules rather than a conventional randomized time-to-first-event comparison. The observational design also makes the distinction between an observed association and a causal treatment effect particularly important. The registry currently does not contain posted formal statistical analyses or numerical outcome estimates, so the primary statistical story at this stage is defined by the endpoint, follow-up rules, cohort comparisons, and planned rate-based framework.