Safety Study of Airway Clearance Techniques in Infants.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety Study of Airway Clearance Techniques in Infants. (NCT02128867), a randomized na study. The primary outcome is Number of Refluxes.

2. Clinical Question

The registry identifies the study as randomized, with no masking, and gives the model as single group. Those design labels do not fully explain how assignment and comparisons were implemented across the three techniques. In particular, the record does not describe the allocation procedure, whether each infant received one or more techniques, or the order of any treatments. These details are important to interpreting comparisons, so the design should not be reconstructed beyond what is recorded.

3. Trial Design

The registry describes a completed study with randomized allocation and no masking. Its model is listed as single group, while the intervention records name three techniques and the outcome summaries report results for three groups. The record does not explain how the single-group model label relates to the randomized allocation or provide enough detail to resolve that design description. The allocation method, sequence generation, concealment, and any treatment order are not reported in the registry record.

The recorded enrollment is 159. For the primary outcome, the posted group summaries give 50 observations for AAD, 50 for bouncing combined with AAD, and 50 for bouncing. The safety summaries use different at-risk counts: 53 for AAD, 52 for the combined technique, and 54 for bouncing. The registry does not explain these differences in outcome and safety denominators, so they should be read as the counts attached to their respective results rather than reconciled into a single analysis population.

The registry lists the study start as 2013-09 and primary completion as 2015-01; the status is completed. It does not report a phase beyond NA, a detailed protocol, eligibility criteria, participant flow, or the original statistical analysis plan. The record names one secondary outcome, duration of reflux episodes, but does not post a result for it in the available outcome information. It also does not describe whether outcomes were collected at additional time points or under conditions beyond the stated treatment period.

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.

Overall Study

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 Number of Refluxes, measured over 15 minutes.

The primary endpoint is the mean number of reflux episodes counted over 15 minutes of treatment. It is a count outcome: each episode contributes to the number recorded during that defined observation period. The registry reports the mean and standard deviation for each technique, with the unit given as number of refluxes.

A count over a fixed interval provides a direct way to describe how many reflux episodes were observed during treatment. Reporting the mean gives the average count in each group, while the standard deviation describes how much individual observations vary around that average. Neither summary shows the full distribution of counts, such as whether observations cluster at low values or whether a small number of larger counts influence the mean.

The endpoint matches the stated question about whether the techniques provoke or aggravate reflux during treatment. Its scope is nevertheless limited to the recorded 15-minute period. The endpoint does not, by itself, describe reflux outside that period, the duration of episodes, or longer-term outcomes. Duration of reflux episodes is listed as a secondary outcome, but a result for that measure is not reported in the registry record.

7. Results Posted in the Registry

• Bouncing Combined With AAD: mean 1.14 versus 0.48 on control; difference 0.66 (95% CI -0.14 to 1.46), standardized effect d = 0.32; z-test p = 0.106.
• Bouncing: mean 1.62 versus 0.48 on control; difference 1.14 (95% CI 0.02 to 2.26), standardized effect d = 0.40; z-test p = 0.046.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous summary of a count measure: the number of reflux episodes observed during a fixed treatment period. The registry reports a mean and standard deviation for each group. The mean describes the average count, while the standard deviation describes spread around that average. Both are useful summaries, but neither supplies the individual observations or their distribution. Because the outcome is a count, the distribution may not be well described by a symmetric bell-shaped model; the registry record does not report a distributional assessment or a count-specific analysis.

What does a mean difference represent?

The pairwise comparisons express the difference between the mean in a technique group and the mean in the AAD group. A positive estimate in these comparisons means the recorded mean count is higher in the technique group than in AAD. For bouncing combined with AAD, the estimate is 0.66 refluxes; for bouncing, it is 1.14 refluxes. These are mean differences on the original outcome scale, not ratios or percentage changes.

The posted comparisons are recalculated, unadjusted comparisons based on the group summary statistics. An unadjusted comparison does not account for other variables or features of the trial design. The registry record does not report a sponsor analysis for these comparisons, a model specification, or an adjustment set. It lists zero analyses, even though outcome summaries and pairwise comparison estimates are available in the posted results. The way these results relate to any full analysis plan is not reported in the registry record.

How should the confidence intervals be read?

A confidence interval gives a range of estimates compatible with the data and the method used to calculate the interval. In general, a narrower interval indicates greater precision, while a wider interval indicates more uncertainty about the effect estimate. For the combined technique compared with AAD, the 95% interval is -0.141 to 1.461. For bouncing compared with AAD, it is 0.021 to 2.259. The first interval includes zero; the second is above zero but has a lower endpoint close to zero. Intervals describe uncertainty in the estimate and are not a range that contains a known true value with certainty.

What does a p-value tell the reader?

In general, a p-value describes how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the analysis assumptions were correct. It is not the probability that the null hypothesis is true, nor does it quantify the clinical importance of a difference. The posted unadjusted p-values are 0.106 for the combined technique compared with AAD and 0.046 for bouncing compared with AAD. They should be considered alongside the estimated differences and confidence intervals, not used as a stand-alone summary.

The registry does not report whether the original analysis addressed multiple comparisons. When multiple comparisons are considered, the chance of at least one apparently unusual result can increase; adjustment methods are designed to address that issue, though they can also affect precision. This is a general statistical consideration, not a claim about an analysis that the registry did or did not conduct. The two posted pairwise comparisons should therefore be described as unadjusted unless a fuller analysis record says otherwise.

Why can count data require special care?

In general, count outcomes can have many low or zero values and a smaller number of larger values. A mean and standard deviation may not fully describe such a distribution. Over-dispersion occurs when the variability in counts is greater than a chosen count model expects; if present and ignored, standard errors and confidence intervals from that model may be too optimistic. The reported group standard deviations are 1.36, 2.55, and 3.8, but those summaries alone do not establish whether over-dispersion is present. The registry does not report a count-regression model, distribution check, or method for handling over-dispersion.

The registry identifies randomized allocation but also lists a single-group model. It does not describe the allocation sequence, whether treatments were administered in a particular order, whether measurements were repeated within infants, or how dependence among any repeated observations was handled. In general, measurements from the same participant may be correlated, and an analysis that treats them as independent can misstate uncertainty. Whether that issue applies to this study cannot be determined from the registry record.

The summaries also do not show how missing outcome observations were handled or define the analysis population. The outcome counts attached to the three groups are 50 each, while the safety at-risk counts differ. The record does not explain missingness, exclusions, or the relationship between these sets of counts. These omissions limit assessment of how the posted results were derived. The presented estimates remain interpretable as the stated summaries and unadjusted recalculations, but they do not replace a complete protocol and analysis report.

10. Confidence Intervals and Precision

The mean-difference estimate for bouncing combined with AAD compared with AAD is 0.66 refluxes, with a 95% confidence interval from -0.141 to 1.461. This interval spans zero, so the posted data do not pin down the direction of the mean difference with high precision. It also includes values consistent with a higher mean count for the combined technique. The breadth of the interval indicates appreciable uncertainty around the estimate.

For bouncing compared with AAD, the mean difference is 1.14 refluxes, with a 95% confidence interval from 0.021 to 2.259. The interval is above zero, but its lower endpoint is close to zero and its upper endpoint permits a substantially larger difference. Thus, the estimate points toward a higher observed mean count for bouncing, while the range remains broad enough to leave uncertainty about the magnitude.

In general, confidence intervals convey both the direction and precision of an estimate. They are more informative than a binary reading of whether a p-value passes a conventional threshold, because they show a range of effect sizes compatible with the data and calculation method. These intervals are for unadjusted recalculated comparisons. They do not account for multiple comparisons or any adjustments not described in the registry record, and they should not be treated as definitive evidence of equivalence or of a specific clinical impact.

The group standard deviations are larger than the corresponding means in the posted summaries. This highlights variability relative to the average count, but cannot establish the shape of the underlying distributions or the precision of every individual measurement. The record does not post confidence intervals for the group means, nor does it provide participant-level data for independent assessment of the estimates.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations for participants per arm, not totals across arms or total event counts. They use the observed mean difference as a reference and show planning for 100%, 75%, and 50% of that observed effect. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. These are planning calculations, not observed power.

For the comparison of bouncing combined with AAD against AAD, planning for 100% of the observed effect gives 151 participants per arm for 80% power and 202 participants per arm for 90% power. Planning for 75% of the observed effect gives 268 participants per arm for 80% power and 359 participants per arm for 90% power. Planning for 50% of the observed effect gives 602 participants per arm for 80% power and 806 participants per arm for 90% power.

For bouncing compared with AAD, planning for 100% of the observed effect gives 99 participants per arm for 80% power and 132 participants per arm for 90% power. Planning for 75% of the observed effect gives 175 participants per arm for 80% power and 235 participants per arm for 90% power. Planning for 50% of the observed effect gives 394 participants per arm for 80% power and 527 participants per arm for 90% power.

In plain terms, the 80% column gives the estimated per-arm enrollment for a design with an 80% chance of detecting the chosen effect under the stated assumptions; the 90% column gives the estimate for a higher chance of detection. Smaller effects require more participants per arm in these illustrations. The registry does not report the trial's original sample-size assumptions or justification, so these figures should not be read as the trial's design targets.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.660; pooled standard deviation σ = 2.044.
2. n = 2 × (1.96 + 0.842)² × 2.044² / 0.660² = 150.5, rounded up to 151 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.66                      # observed difference in means
sigma <- 2.04353859762912                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(2128867)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 151 (80% power), 202 (90% power)
 75% of observed effect: n per arm = 268 (80% power), 359 (90% power)
 50% of observed effect: n per arm = 602 (80% power), 806 (90% power)
power.t.test: 151.5 
Simulated power at n = 151 per arm: 0.798 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The posted pairwise comparisons use AAD as the reference for bouncing combined with AAD and for bouncing. The record does not post a direct estimate comparing bouncing with the combined technique. It also does not report whether these were the only planned comparisons, whether other outcomes were tested, or whether any multiplicity procedure was used. Because the two reported p-values are unadjusted, they should not be interpreted as if a family-wise or other multiplicity adjustment had been applied.

Multiplicity matters when several hypotheses are tested: in general, the chance of at least one apparently notable finding can rise as the number of comparisons grows. This is a general statistical issue and does not establish that the trial's analysis was inappropriate. The original statistical plan, its hypothesis hierarchy, and any interim-analysis plan are not reported in the registry record. No interim analyses are reported in the record.

Missing-data handling is also not reported in the registry record. The primary outcome summaries list 50 observations in each group; the safety summaries list at-risk counts of 53, 52, and 54. The record does not explain the difference between these counts or state whether any observations were missing, excluded, or unavailable. It does not describe imputation, complete-case analysis, or sensitivity analyses. The status of the 15-minute outcome for participants without a posted observation is therefore unclear.

The registry does not define the analysis population, such as whether results were based on all randomized participants, those who received an intervention, or those with an available outcome. It does not provide a participant flow account or state how departures from assigned treatment were handled. Nor does it explain whether each infant contributed one observation or whether repeated observations were collected. These details affect interpretation of the estimates and their uncertainty; they cannot be inferred from the summaries.

The record lists zero analyses, while also posting outcome summaries and recalculated pairwise comparisons. It does not clarify whether sponsor-conducted inferential analyses were performed but not registered, or whether the posted comparisons are the only inferential results. Accordingly, the available information supports a description of the posted summaries and unadjusted comparisons, but not a complete audit of multiplicity, missingness, or analysis-population choices.

14. Safety Results

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

Other (non-serious) adverse events

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

The safety summaries report zero serious adverse events and zero other adverse events in each of the three groups. The at-risk counts are 53 for AAD, 52 for bouncing combined with AAD, and 54 for bouncing. These are the denominators attached to the safety summaries; they are not the same counts as those shown for the primary outcome.

These figures describe events recorded in the registry for the listed groups. A zero count in a summary does not show that the true risk is zero, and the record does not provide confidence intervals or further detail about safety surveillance, event definitions, or the period during which events were collected. Death counts are not reported in the registry record.

The registry lists no serious adverse-event terms. It does not provide a narrative explanation of how adverse events were sought or classified. The safety results should therefore be read as the posted event counts and at-risk counts, rather than as a complete account of all possible harms or a comparison establishing relative safety.

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

The results describe reflux episode counts during a 15-minute treatment period. The observed mean was 0.48 for AAD, 1.14 for bouncing combined with AAD, and 1.62 for bouncing. The unadjusted recalculated comparison with AAD estimated a mean difference of 0.66 refluxes for the combined technique and 1.14 refluxes for bouncing. The corresponding confidence intervals and p-values indicate uncertainty that differs between the comparisons, but neither set of statistics alone establishes a clinically important effect.

The bouncing comparison has a 95% confidence interval above zero, with an unadjusted p-value of 0.046. The combined-technique comparison has a confidence interval that includes zero, with an unadjusted p-value of 0.106. These are unadjusted findings, and the registry does not report a multiplicity adjustment. The evidence should not be reduced to a claim that one technique is effective or ineffective, or that two techniques are equivalent.

The study question concerns whether airway-clearance techniques provoke or aggravate reflux. The posted primary endpoint can address only the recorded episode count during the stated treatment window. It does not establish what happened outside that window, whether episodes lasted longer, or whether any count difference had a clinical consequence. Although duration of reflux episodes is listed as a secondary outcome, its result is not reported in the registry record.

The design description is incomplete for evaluating causal effects: allocation is described as randomized, while the model is listed as single group, and the record does not explain the relationship between those labels. Details about sequence generation, treatment order, analysis populations, missing data, and the statistical plan are not reported. The safety summaries record zero serious and zero other adverse events in each group, but those counts do not establish absence of risk. Overall, the registry supports a careful account of the posted short-period outcome summaries, not a broad conclusion about the techniques or their effects beyond the measures reported.

16. Why This Trial Matters Statistically

This trial illustrates why statistical interpretation needs more than a comparison of group averages. The outcome is a count observed during a fixed interval, and the posted standard deviations are substantial relative to the means. Means and standard deviations give a compact description, but they do not reveal the distribution of counts, the presence of unusually high observations, or whether a model suited to the outcome was used.

The two unadjusted comparisons also show why estimates, intervals, and p-values should be read together. The bouncing comparison has an estimated difference of 1.14 refluxes and an interval from 0.021 to 2.259; the combined-technique comparison has an estimate of 0.66 and an interval from -0.141 to 1.461. The intervals communicate uncertainty in magnitude and direction. The p-values alone cannot tell readers whether a difference matters clinically or how stable an estimate may be.

Multiplicity is relevant because the registry reports more than one pairwise comparison, but does not report an adjustment or the full analysis plan. A borderline unadjusted result needs particular care when the number and structure of planned tests are unknown. The record's lack of detail about missing data, analysis populations, and the difference between outcome and safety denominators further limits an independent assessment of precision and risk of bias.

The record also presents a design-description challenge: it identifies randomized allocation and a single-group model, without explaining how those descriptions fit the three intervention summaries. This highlights the value of transparent reporting of allocation, treatment exposure, participant flow, outcome collection, and analysis choices. Such details help readers distinguish what the data summarize from what can be inferred about effects.

17. 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 provides limited detail about the trial's conduct and analysis. It identifies randomized allocation but lists a single-group model, and does not explain the relationship between those design labels. The allocation procedure, concealment, treatment order, and whether infants received one or multiple techniques are not reported. This limits assessment of how the three group summaries should be compared.

The primary outcome is reported for 50 observations in each group, while safety at-risk counts are 53, 52, and 54. The registry does not explain the different denominators, provide participant flow, or specify the analysis population. Missing-data procedures and reasons for unavailable outcomes are not reported. The record also does not describe whether repeated measurements were taken or how any within-infant dependence was handled.

The primary outcome is measured over 15 minutes, so the posted summaries address only that observation period. The registry names duration of reflux episodes as a secondary outcome but does not report its result. It does not provide the individual observations, distribution summaries beyond means and standard deviations, or details of measurement procedures. Those omissions prevent assessment of distributional shape and make it difficult to verify the calculations independently.

The pairwise estimates are unadjusted recalculations, and no multiplicity adjustment or original analysis plan is reported. The record does not provide a sample-size justification for the trial itself; the posted planning illustrations are based on observed effects rather than documented original design assumptions. Safety reporting is limited to event counts and at-risk counts, with no event terms or deaths reported. These limitations mean the registry results should be interpreted within their narrow scope and not as a full account of trial methods or outcomes.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Safety Study of Airway Clearance Techniques in Infants. — NCT02128867.
• Registry record last updated 2015-03-17; status last verified 2015-03.
• Results first posted 2015-03-17.
• 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.

21. Record Summary

NCT02128867 is a completed registry study of bouncing, Assisted Autogenic Drainage (AAD), and bouncing combined with AAD in infants from 0 to 12 months. It lists randomized allocation, no masking, and a single-group model. Recorded enrollment is 159; the primary outcome is the mean number of reflux episodes during 15 minutes of treatment.

The posted means are 0.48 for AAD, 1.14 for the combined technique, and 1.62 for bouncing. Unadjusted recalculated comparisons with AAD give mean differences of 0.66 and 1.14, respectively. The registry also lists duration of reflux episodes as a secondary outcome, without a posted result. Safety summaries report zero serious and zero other adverse events for each group.

The record does not report the original sample-size assumptions, detailed analysis plan, missing-data procedures, or analysis populations. Its design labels and differing outcome and safety counts are not further explained.