Asthma Control Test Guided Treatment in Chinese Subjects: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Asthma Control Test Guided Treatment in Chinese Subjects (NCT02868281), a randomized phase 4 study. The primary outcome is Percentage of Participants Who Had an ACT Total Score >=20 or an Improvement of More Than 3 Points in ACT During the 24-week Treatment Period.

2. Clinical Question

3. Trial Design

This was a prospective, multicentre, cluster-randomized, open-label study comparing ACT-guided treatment with usual care in adults with asthma in China. The study used a parallel-group design and is listed as Phase 4. The registry record marks the study as completed. Its recorded start date is 2016-08-26, and its primary completion date is 2019-08-09.

Randomization was by study center rather than by individual participant: each center was assigned to the ACT-guided approach or usual care. The brief description says randomization was to be stratified according to hospital tier, distinguishing Tier 3 from Tier 2 hospitals. Because centers were the randomized clusters, outcomes from participants at the same center may be more alike than outcomes from participants at different centers. An analysis that treats every participant as independently randomized could therefore misrepresent the design.

The registry records enrollment of 530 and says that 528 asthma subjects were planned. It describes the ACT-guided approach as treatment decisions based on ACT scores, while control centers relied on doctors’ subjective judgment. The registry also states that no treatment intervention was provided as an assigned study treatment; the comparison concerned how routine treatment was guided.

The primary outcome was assessed over the 24-week treatment period. Other listed outcomes included daytime and night-time symptom scores, lung function measures, peak expiratory flow, asthma-related quality of life, and time to a specified ACT result. The registry record does not report a fuller account of how centers were selected, how many centers participated, or the details of concealment of cluster allocation before assignment.

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: Total 583 participants were screened and 530 participants were enrolled into the study (53 participants were screen failures).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-To-Treat (ITT) Population consisted of all participants who signed informed consent form, were randomized and who had at least one post-Baseline assessment. Only those participants with data available at the specified data point were analyzed.

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.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Who Had an ACT Total Score >=20 or an Improvement of More Than 3 Points in ACT During the 24-week Treatment Period, measured over Up to Week 24.

The primary endpoint was the percentage of participants who, during the 24-week treatment period, had an ACT total score of at least 20 or an improvement of more than 3 points in ACT. It combines a score threshold with a change-from-score criterion: a participant could meet the endpoint by reaching the specified level of asthma control or by improving by the specified amount. The registry describes the endpoint as a number of participants and reports the comparison using an odds ratio.

The ACT is a self-administered questionnaire with five items, each assessed on a five point categorical scale from 1 to 5. Item scores are summed, giving a total score ranging from 5 to 25. The registry describes a score of at least 20 as denoting well-controlled asthma, a score from 16 to 19 as not well-controlled asthma, and a score of 15 or lower as very poorly controlled asthma. Higher scores indicate improved asthma control, and the questionnaire has a recall period of four weeks.

Using a threshold can make a clinical outcome easier to interpret as a proportion meeting a defined criterion. Combining that threshold with improvement from baseline can also capture participants whose score improves substantially without reaching the threshold. However, because the endpoint combines these routes, the reported result does not by itself show how many participants met each part separately. Separate component results are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does an odds ratio describe?

General: Logistic regression models the odds of a binary outcome. Odds are related to, but not the same as, the probability or percentage of participants with the outcome. An odds ratio compares the odds between groups. It is not a direct measure of the difference in percentages, and it should not be interpreted as though it were a risk ratio. For a common outcome, odds ratios can be notably different from risk ratios.

For the primary endpoint, the outcome is whether a participant met the combined ACT criterion during the specified treatment period. The reported odds ratio is 7.87. To determine whether this means higher odds under ACT-guided treatment or higher odds under usual care, the numerator and denominator groups must be clear. The analysis identifies its groups as OG000 and OG001 but does not map those codes explicitly to the named treatment approaches. Therefore, the direction should not be assigned based on the odds ratio alone.

How should the confidence interval be read?

General: A confidence interval expresses uncertainty in an estimate under the assumptions of the fitted model. A narrower interval indicates greater precision than a wider interval on the same scale. The 95% confidence interval for the reported odds ratio runs from 1.29 to 48.11. Its breadth signals substantial uncertainty about the magnitude of the comparison. Because the interval excludes 1, it is compatible with a difference in odds under the model, but it does not imply that the true value is known precisely or that every value in the interval is equally plausible.

The p-value is 0.027. General: A p-value describes how unusual results at least as incompatible with a null hypothesis would be under that hypothesis and the model assumptions. It is not the probability that the null hypothesis is true, nor does it measure clinical importance. The registry record does not state a multiplicity adjustment for the primary result, so the p-value should be understood as reported rather than as adjusted for the full set of outcomes or analyses.

Why does clustering matter?

General: In a cluster-randomized study, groups such as clinics or study centers, rather than individual participants, are assigned to study conditions. Participants from the same cluster may share clinicians, workflows, and local circumstances. Their outcomes can consequently be correlated. When correlation is ignored, standard errors may be too small and confidence intervals too narrow. Methods for cluster-randomized data account for this dependence, for example through suitable mixed effects or other cluster-aware approaches.

The registry describes each study center as a cluster and reports a mixed effect logistic regression for the primary analysis. That is consistent with using a model capable of representing clustered data, but the registry record does not report the model’s full specification. The number of centers, the estimated degree of within-center correlation, the random-effects structure, and any additional adjustment variables are not reported in the registry record. Consequently, the posted estimate can be described, but its modeling details cannot be independently assessed from the registry record.

What does the mixed effect model add?

General: A mixed effect model includes both population-level terms and terms that represent variation across clusters or other modeled groups. In a logistic model, the outcome is binary and the link between predictors and outcome is expressed on a log-odds scale. Accounting for center-level variation can help reflect the study’s randomization structure. The exact interpretation of a conditional odds ratio can depend on model specification, including how cluster effects are represented.

The registry names the method but does not provide the coefficient coding, covariate list, convergence information, or details of how the model handled missing observations. Those elements are not reported in the registry record. The result should therefore be attributed to the sponsor’s stated model rather than treated as a fully reproducible calculation from the registry information.

Why is there no simple recalculation?

General: Recalculating an effect from arm-level participant counts can be inappropriate when randomization occurred at the cluster level, because such a calculation would treat participants as if they had been independently assigned. The registry explicitly notes that a simple individual-level re-analysis would be misleading and that only the sponsor’s own analysis is shown. No recalculated odds ratio or alternative p-value is presented. The sponsor’s model-based result is the appropriate posted estimate to discuss, subject to the limits of the information recorded about the model.

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("ACT Guided Treatment Group", "Usual Care Group")
n      <- c(240, 262)
events <- c(238, 247)
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  ACT Guided Treatment Group  240  238  99.2  97.0  99.9  97.0  99.8  97.0  99.7
ARM  Usual Care Group  262  247  94.3  90.7  96.8  90.8  96.5  90.8  96.5

11. Analyses Reported by the Sponsor

The primary result is the sponsor’s own analysis: a mixed effect logistic regression reporting an odds ratio of 7.87, a 95% confidence interval from 1.29 to 48.11, and p = 0.027. No simple individual-level recalculation is presented. The registry record explains that the study was cluster-randomized by center and that such a re-analysis would be misleading, so the sponsor’s analysis is the result to use when describing the posted comparison.

This does not mean the model can be fully reconstructed from the registry information. The record does not give the detailed model specification, the number of centers, the covariates, or an explicit mapping of OG000 and OG001 to the named study groups. As a result, the reported estimate and uncertainty can be quoted, but the direction of the comparison cannot be confidently attributed to ACT-guided treatment versus usual care from the analysis codes alone. The wide interval also warrants emphasis: the estimate is not precise, even though its reported interval excludes 1.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple outcomes and analyses, including symptom scores, lung function, peak expiratory flow, quality of life, and time to an ACT result. The primary endpoint itself combines two ways of meeting the criterion: an ACT score of at least 20 or improvement of more than 3 points. The record does not report whether the primary endpoint was divided into those components in the analysis.

Multiplicity arises when several outcomes, time points, or analyses are examined. General: With multiple statistical tests, the chance of at least one apparently unusual result can increase if each test is interpreted separately without adjustment. Some studies address this through a prespecified testing hierarchy or a multiplicity adjustment. The registry record does not report a multiplicity strategy, whether secondary analyses were adjusted, or whether the primary result was protected by a prespecified hierarchy. Thus, p = 0.027 is the reported result, not a result identified in the record as adjusted across all outcomes or analyses.

The registry defines the primary analysis population as participants who signed informed consent, were randomized, and had at least one post-Baseline assessment. It further says that only participants with data available at the specified data point were analyzed. The number of participants missing primary endpoint data, reasons for missingness, and any imputation or sensitivity analyses are not reported in the registry record. Missing outcome data can affect a comparison if availability is related to treatment assignment or outcome status; the registry information does not establish whether that occurred.

The registry record does not report the statistical analysis plan, a sample-size justification, interim analyses, or rules for handling protocol deviations. These omissions limit assessment of whether the reported analysis followed prespecified decisions and how robust the result is to alternative assumptions. They should be treated as unreported details, not as evidence that a particular procedure was or was not used.

13. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: asthma (0.8%), uterine leiomyoma (0.8%), coronary artery disease (0.7%), intervertebral disc protrusion (0.7%), nasal polyps (0.4%).

The registry records serious adverse events for 9 participants at risk in the ACT Guided Treatment Group, with 252 at risk, and for 14 participants at risk in the Usual Care Group, with 278 at risk. It records other adverse events affecting 82 participants at risk in the ACT Guided Treatment Group and 123 participants at risk in the Usual Care Group. These are descriptive counts; they do not establish that a study approach caused an event or that the groups had comparable event risk.

The recorded deaths were 1 in the ACT Guided Treatment Group and 0 in the Usual Care Group. The registry also lists serious-event terms with percentages, including Asthma and Uterine leiomyoma at 0.8%, Coronary artery disease and Intervertebral disc protrusion at 0.7%, and Nasal polyps and Chronic sinusitis at 0.4%. Those listed terms are registry entries, not evidence that the events were caused by the treatment approach.

Safety comparisons need context about event definitions, follow-up, ascertainment, and exposure. Further detail on those matters is not reported in the registry record. The figures should therefore be read as the adverse-event counts recorded for the groups, rather than as a definitive comparative safety analysis.

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

The posted primary result is an odds ratio from a cluster-aware sponsor analysis, with a 95% confidence interval and p-value. It indicates that the fitted model found evidence of a difference in the odds of meeting the combined ACT endpoint under its assumptions. The interval is wide, so the magnitude remains uncertain. Since the analysis codes are not explicitly linked to the named groups, the direction of that comparison cannot be assigned securely from the registry record alone.

The result does not provide the absolute percentage of participants meeting the endpoint in either group, nor the absolute difference between groups. Those primary outcome percentages are not reported in the registry record. It also does not reveal whether participants met the endpoint through an ACT score of at least 20 or through improvement of more than 3 points, because component results are not reported in the registry record.

General: Statistical evidence about a group comparison is not by itself a measure of clinical importance. An odds ratio does not show the likely experience of an individual participant, and the p-value does not quantify the probability that the approach is effective. Interpretation also depends on study conduct, outcome measurement, missing data, the analysis model, and whether the primary analysis was prespecified. The registry record does not report enough detail to resolve all of these considerations.

The trial tested a care-guidance approach in a specific context: adults with asthma in China, with assignment by study center and an open-label comparison. The result should be understood as pertaining to that study and its recorded analysis. It does not establish how the approach performs in other populations or health systems, and the registry record does not provide evidence for conclusions about outcomes beyond those measured in the study.

15. Why This Trial Matters Statistically

This study illustrates why the unit of randomization matters in clinical-trial analysis. Participants were not randomized independently; centers were assigned to the treatment-guidance approach. Analyses that ignore this structure can understate uncertainty because participants within a center may share features that affect their outcomes. The registry identifies a mixed effect logistic regression and explicitly cautions against a simple individual-level re-analysis. That makes the design and the method inseparable when interpreting the primary result.

It also highlights a communication issue with odds ratios. An odds ratio is a relative measure on the odds scale, not a direct percentage difference. Here, the odds ratio has a wide confidence interval, and the registry’s coded group labels are not explicitly mapped to ACT-guided treatment and usual care. These details matter: a numerical estimate can be quoted precisely while its direction or practical magnitude remains difficult to communicate without additional information.

The endpoint combines an ACT threshold and an improvement criterion. That can capture more than one form of a favorable score pattern, but it also means that a single combined estimate cannot show which criterion contributed to the result. The registry does not report the component results or the group-level primary endpoint percentages.

Finally, the record distinguishes a planned enrollment of 528 from recorded enrollment of 530, but does not report the sample-size justification or statistical analysis plan. This limits evaluation of the planning assumptions and prespecified analysis choices. The trial is therefore useful for teaching how to keep design, estimand, model, uncertainty, and reporting completeness in view when reading a registered clinical result.

16. 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 principal trial-specific limitation for statistical interpretation is cluster randomization by study center. The registry record does not report the number of centers or the degree of within-center correlation. Without those details, it is not possible to assess how much the cluster structure influenced precision or to reconstruct the analysis independently. The sponsor’s mixed effect logistic regression is reported, but its full specification is not.

The comparison was open-label, and the approaches differed in how treatment decisions were guided. The registry does not report masking of participants or clinicians, which is consistent with its description of the study as open-label. Subjective judgment in usual care and score-guided decisions in the intervention centers may make care processes and outcome assessment susceptible to differences in behavior or reporting. The registry information does not quantify such effects.

The primary endpoint is a combined criterion, and the registry does not report the outcome frequency in each group or the separate contribution of each component. The reported odds ratio is imprecise, with a wide confidence interval. The group codes in the analysis are not explicitly mapped to named arms, limiting interpretation of direction. Missing-data counts, missingness reasons, and sensitivity analyses are also not reported in the registry record.

Other important details are not reported in the registry record, including a sample-size justification, interim analyses, a statistical analysis plan, multiplicity procedures, and full model diagnostics. The registry records planned enrollment of 528 and enrollment of 530, but the reasons for that difference are not reported. These limitations constrain assessment of precision, robustness, and prespecification; they do not establish that any particular bias or analysis problem occurred.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Asthma Control Test Guided Treatment in Chinese Subjects — NCT02868281.
• Registry record last updated 2021-07-08; status last verified 2021-06.
• Results first posted 2020-08-03.
• 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.

20. Record Summary

NCT02868281 is a completed, Phase 4, open-label, parallel, cluster-randomized study of ACT-guided treatment compared with usual care among adults with asthma in China. The registry records enrollment of 530 and a planned enrollment of 528. Study centers were the randomization clusters, with hospital tier described as a stratification factor.

The primary outcome was the percentage of participants with an ACT total score of at least 20 or improvement of more than 3 points during the 24-week treatment period. The sponsor reported a mixed effect logistic regression odds ratio of 7.87, with a 95% confidence interval from 1.29 to 48.11 and p = 0.027. The registry’s analysis codes are not explicitly mapped to the named groups, and the confidence interval is wide.

The record does not report the primary endpoint percentages, the full model specification, or the missing-data details needed for an independent reconstruction. Because randomization was by center, the sponsor’s cluster-aware analysis is the relevant posted result; an individual-level recalculation is not presented.