Self-Management of Continuous Positive Airway Pressure Settings: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Self-Management of Continuous Positive Airway Pressure Settings (NCT03536572), a randomized na study. The primary outcome is Treatment Adherence.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group trial with single masking. Its listed enrollment was 270, and its status is completed. The registry record identifies the sponsor as the VA Office of Research and Development. The study start date is 2018-10-01, and the primary completion date is 2023-03-31. The phase is listed as NA.

The two groups received related but distinct approaches to education and support. The Sleep Apnea Self-Management Program involved protocol-based sleep apnea and CPAP education and support. The Individualized Pressure Adjustment group received additional education and support intended to allow participants to adjust PAP pressures. The record does not report further detail about how the program was delivered, the pressure-adjustment procedures, or the masking procedures. It also does not identify which parties were masked.

The primary outcome was treatment adherence, defined as the amount of time CPAP was used on average per night over the 2 months outcome timeframe. Results are posted for 96 participants in the self-management group and 113 in the individualized pressure-adjustment group. Those are the group sizes associated with the reported outcome; the reason they differ from the listed enrollment is not reported in the registry record. The record reports one outcome measure and no secondary outcomes. Although the study is marked completed, the registry record does not describe the original sample-size justification, any interim analyses, or the planned statistical analysis in detail.

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 Treatment Adherence, measured over 2 months.

The primary endpoint measures average CPAP use per night over the 2 months outcome timeframe. Its unit is hours per night, and the reported summary is a mean with a standard deviation for each group. The underlying concept is treatment adherence: how much time participants used CPAP on an average night during the specified outcome period.

Average use is a continuous measure. It preserves information about the amount of device use rather than reducing each participant’s experience to a binary category such as adherent or not adherent. Comparing group means addresses whether average nightly use differed between the two assigned approaches. A mean is a useful summary when the research question concerns average use, though it does not describe every participant’s experience or the full shape of the distribution.

The registry record specifies the outcome timeframe and measure but does not report how nightly device-use data were collected, how the 2 months were aggregated for each participant, or how incomplete records were handled. It also does not define a threshold for clinically meaningful adherence or identify a separate adherence category. The posted outcome therefore supports interpretation of group averages and their uncertainty, not a conclusion about the number of individuals meeting an unreported threshold.

7. Results Posted in the Registry

• Individualized Pressure Adjustment: mean 2.20 versus 2.10 on control; difference 0.10 (95% CI -0.32 to 0.52), standardized effect d = 0.06; z-test p = 0.641.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous, so the registry reports a mean and standard deviation in each group. The mean describes the group’s average nightly CPAP use. The standard deviation describes how much individual observations typically vary around that average. A standard deviation is not the uncertainty in the mean: uncertainty in a group mean also depends on the number of observations contributing to it.

What does the mean difference represent?

The comparison is expressed as the mean in the Individualized Pressure Adjustment group minus the mean in the Sleep Apnea Self-Management Program group. The posted means are 2.2 and 2.1 hours per night, respectively, and the reported mean difference is 0.1 hours per night. A positive difference therefore points toward a higher group average in the individualized pressure-adjustment group. This is a contrast between group means, not a statement that every participant had more use or that each participant’s use changed by that amount.

What does the p-value tell the reader?

The unadjusted p-value is 0.641. In general, a p-value describes how unusual data at least as incompatible with a specified null model would be if that model and its assumptions were correct. It is not the probability that the null hypothesis is true, nor does it measure the size or practical importance of an effect. A larger p-value does not prove that the groups are identical. The registry record does not report a detailed analysis plan, so it is not possible to determine from the record whether this was the sole planned test or how the analysis was selected.

Why does variability matter?

General statistical explanation: when individual observations vary substantially, a modest difference in group averages can be difficult to estimate precisely. The posted standard deviations, 1.5 and 1.6 hours per night, describe variation in the observed outcomes. The comparison’s confidence interval incorporates uncertainty in estimating the difference, rather than only reporting the observed point estimate. The reported group sizes also matter to that uncertainty. Neither the mean nor the standard deviation alone captures the distribution’s full shape or reveals whether a small number of observations strongly influence the average.

What assumptions underlie a mean-based comparison?

General statistical explanation: conventional methods for comparing means rely on assumptions about the outcome data and the sampling or assignment process. The particular method, any distributional checks, and any robust or alternative analyses are not reported in the registry record. The record identifies the trial as randomized, but that alone does not tell the reader which statistical model was used for the posted comparison or whether it adjusted for other variables.

The registry record lists a mean difference, confidence interval, and unadjusted p-value. It does not identify covariate adjustment, stratification, repeated-measures modeling, or a method for handling missing outcome data. Those methods should not be inferred from the existence of a result. No discrepancy is flagged between the posted result and a recalculated result, but the record does not post a separate analysis narrative that would explain the estimation procedure.

Why can the interval include effects in both directions?

General statistical explanation: a confidence interval represents a range of effect estimates reasonably compatible with the observed data under the method’s assumptions. When the interval includes zero, the data do not distinguish clearly between a positive difference, no difference, and a negative difference at the corresponding confidence level. This does not mean that all values in the interval are equally likely, or that the interval assigns a probability to the true effect after the data are observed.

More broadly, the posted analysis addresses one continuous outcome at one stated timeframe. The registry record reports one outcome measure and no secondary outcome measures, and it lists zero posted analyses. It does not report whether any planned comparisons were omitted, whether interim analyses occurred, or how analysis populations were defined. Such omissions limit what can be inferred about the full analysis strategy; they do not warrant filling in undocumented methods.

10. Confidence Intervals and Precision

The point estimate for the difference in mean nightly CPAP use is 0.1 hours per night, with a confidence interval extending from -0.321 to 0.521 hours per night when expressed to three decimal places. The interval includes zero and spans values on both sides of no difference. It therefore indicates substantial uncertainty about the direction and size of the group difference, rather than pinpointing a single effect.

A confidence interval is a range generated by a statistical procedure that, over repeated use under its assumptions, would cover the true parameter at its stated confidence level at the intended frequency. The confidence level for this reported interval is not stated in the registry record. The interval should not be read as assigning a probability that the true difference lies within these particular endpoints. Nor does an interval that includes zero prove equivalence: an equivalence conclusion requires a prespecified margin and an analysis designed to assess that question, neither of which is reported in the registry record.

The interval’s breadth reflects the precision of the estimate, which depends on the observed variability and the number of participants contributing outcome data, as well as the statistical method. The group standard deviations are 1.5 and 1.6 hours per night, and the posted outcome group sizes are 96 and 113. The registry record does not explain why those sizes differ from the listed enrollment of 270 or report how missing outcome data were handled.

Precision and statistical significance are related but distinct. The p-value is 0.641, while the interval communicates a range of compatible mean differences in the outcome’s original units. Neither statistic alone determines clinical importance. The registry record does not identify a threshold for a meaningful difference in nightly use, so the interval cannot be judged against such a threshold from the posted information.

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 planning figures are participants per arm, not total events. This is a continuous-outcome comparison, and the planning display gives the number of participants needed in each arm. The distinction matters: a per-arm figure describes the planned size of each group, whereas a total-events figure applies to time-to-event planning and counts events across groups.

The planning illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its reference and considers detecting the stated fraction of that observed effect. For 100% of the observed effect, the estimates are 3,796 participants per arm for 80% power and 5,081 participants per arm for 90% power. For 75% of the observed effect, the estimates are 6,748 participants per arm for 80% power and 9,033 participants per arm for 90% power. For 50% of the observed effect, they are 15,182 participants per arm for 80% power and 20,324 participants per arm for 90% power.

In plain terms, the 80% power column describes the planning sample size associated with an 80% chance of detecting the specified effect under the stated assumptions; the 90% column uses a higher target chance of detection. Smaller target effects need more participants because they are harder to distinguish from random variation. These are planning illustrations based on the observed effect, not observed power, and they are not the trial’s original design justification. The original design assumptions are not reported in the registry record. The curve represents the same planning relationship between target effect size and required sample size; it does not report additional observed trial outcomes.

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 = 0.100; pooled standard deviation σ = 1.555.
2. n = 2 × (1.96 + 0.842)² × 1.555² / 0.100² = 3,795.3, rounded up to 3,796 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 <- 0.1                      # observed difference in means
sigma <- 1.55490497092235                     # 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(3536572)
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 = 3,796 (80% power), 5,081 (90% power)
 75% of observed effect: n per arm = 6,748 (80% power), 9,033 (90% power)
 50% of observed effect: n per arm = 15,182 (80% power), 20,324 (90% power)
power.t.test: 3796.3 
Simulated power at n = 3,796 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

Multiplicity concerns the increased chance of a false-positive finding when many outcomes, timepoints, or comparisons are tested. The registry record reports one outcome measure and no secondary outcomes, which limits the number of outcome measures visible in the posted results. It does not report whether the investigators planned additional outcomes or comparisons that were not posted, how many statistical tests were planned, or whether any adjustment for multiple testing was specified. A multiplicity strategy is therefore not reported in the registry record, and the single posted p-value should not be taken as evidence that no other analyses were contemplated.

Missing-data handling is also not reported in the registry record. The listed enrollment is 270, while the primary outcome summaries include 96 participants in the self-management group and 113 in the individualized pressure-adjustment group. The record does not explain the difference, identify reasons that outcome data may be absent, state whether the displayed group sizes are the full randomized groups, or describe any imputation or sensitivity analysis. It is not appropriate to infer attrition, exclusions, or a particular analysis population from these figures alone.

The registry record does not state whether the primary comparison followed an intention-to-treat principle, used a per-protocol population, or relied on another analysis population. It also does not report a prespecified statistical analysis plan, interim analyses, or methods for adjusting for baseline characteristics. These details matter when judging whether the posted comparison follows the intended analysis and how robust it may be to incomplete data or analytic choices.

These gaps do not invalidate the reported summaries, but they constrain their interpretation. The result can be described as the posted unadjusted comparison of the reported outcome groups. Broader claims about the full randomized population, missingness, confirmatory status, or multiplicity control are not supported by information reported in the registry record.

14. 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.

The registry reports zero serious adverse events in the Sleep Apnea Self-Management Program group, among 96 participants at risk, and zero in the Individualized Pressure Adjustment group, among 113 participants at risk. The posted figures also report zero other adverse events in each group and zero deaths in each group. No serious adverse-event terms are listed.

These are counts recorded in the registry, not evidence that no adverse events of any kind occurred. A count of zero means that no events in the corresponding reported category are posted for the stated group. The record does not provide event definitions, a collection window, severity detail beyond the serious-versus-other categories, or an explanation of how safety information was solicited and reviewed. It also does not report whether any participants had incomplete safety follow-up.

Safety counts should be interpreted alongside those reporting limitations and the group sizes. A zero count cannot establish that an event is impossible or exclude uncommon harms. The registry record does not provide comparative safety analyses or estimates of uncertainty for these counts. These posted figures describe what was reported, without supporting a broader conclusion about safety.

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

The posted results describe average CPAP use at 2 months in two randomized groups. The mean was 2.1 hours per night for the Sleep Apnea Self-Management Program and 2.2 hours per night for Individualized Pressure Adjustment. The estimated difference was 0.1 hours per night, and the unadjusted p-value was 0.641. The confidence interval includes values below and above zero, so the estimate remains uncertain in both direction and magnitude.

The finding does not establish that the groups are equivalent, nor does it establish that one approach improves adherence. A non-small p-value is not proof of no effect, and the interval does not test equivalence against a specified margin. The registry record does not report a clinically meaningful difference threshold against which the estimate could be judged.

The result concerns one measure over one follow-up timeframe. It does not show effects on other outcomes, because no secondary outcome measures are posted. It also does not indicate whether use persisted after the 2 months timeframe, explain how the group averages arose, or describe individual responses. The record does not report details of data collection or missing-data handling that could clarify these matters.

The sample-size figures are planning illustrations derived from the observed effect, not evidence about the probability that the trial detected a true effect. The trial’s original sample-size justification is not reported in the registry record. Likewise, the reported safety counts are limited to the categories and reporting available in the record.

Accordingly, the narrow statistical conclusion is that the posted group means are close, while the estimated difference is imprecise enough to remain compatible with effects in either direction. That description is not medical advice and does not determine which approach should be used. The record’s reporting limitations and the specific population, intervention descriptions, outcome, and timeframe should remain part of any interpretation.

16. Why This Trial Matters Statistically

This trial raises a useful statistical question about evaluating a behavioral and educational approach using a continuous adherence outcome. Average nightly CPAP use retains more detail than a simple yes-or-no adherence classification, but its interpretation depends on the distribution of use, the quality and completeness of measurement, and the precision of the group comparison. The registry reports means, standard deviations, group sizes, and a mean difference, enabling readers to consider both the estimated contrast and its uncertainty.

The result also illustrates why a p-value alone is not an adequate summary. The unadjusted p-value is 0.641, while the confidence interval spans values on either side of zero. The point estimate is small, but the interval leaves uncertainty about the underlying difference. A reader should distinguish failure to reject a no-difference hypothesis from evidence of equivalence, which would require an explicit margin and relevant analysis not reported in the registry record.

The planning illustration highlights how strongly required sample size depends on the effect chosen for detection. Its estimates increase as the target fraction of the observed effect becomes smaller, and the higher power target also requires more participants per arm. These planning figures are not the trial’s original design rationale and are not observed power.

Finally, the record illustrates the importance of transparent reporting beyond the headline outcome. The posted outcome counts differ from the listed enrollment, but the reason and missing-data approach are not described. The analysis population, multiplicity strategy, and detailed statistical plan are also not reported. Such information helps readers assess precision, robustness, and the scope of conclusions without having to infer undocumented methods.

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 gives limited detail about the trial’s conduct and analysis. It identifies randomized allocation, single masking, and a parallel model, but does not describe the masking procedures or state who was masked. The intervention descriptions identify the broad content of the two groups but do not provide additional detail about delivery, adherence to the programs, or how individualized pressure adjustments were undertaken.

The listed enrollment is 270, whereas primary outcome results are reported for 96 participants in the self-management group and 113 in the individualized pressure-adjustment group. The registry record does not explain the difference, characterize missing outcome data, or describe the analysis population. Because missing-data methods are not reported, the posted results cannot establish how the difference between listed enrollment and outcome counts affects interpretation.

The record also does not provide the original sample-size justification, a detailed statistical analysis plan, interim-analysis information, covariate adjustment details, or a multiplicity strategy. The confidence interval’s confidence level is not stated in the registry record. The registry reports one primary outcome and no secondary outcomes, limiting the range of posted results available for interpretation.

The primary endpoint is average use over 2 months, so the record does not establish whether that pattern continued beyond the measured timeframe. It does not report a threshold for meaningful adherence or a clinically meaningful mean difference. Finally, safety reporting is limited to the posted counts; event definitions and safety follow-up details are not reported in the registry record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Self-Management of Continuous Positive Airway Pressure Settings — NCT03536572.
• Registry record last updated 2024-09-19; status last verified 2024-08.
• Results first posted 2024-09-19.
• 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

The registry record describes NCT03536572, “Self-Management of Continuous Positive Airway Pressure Settings,” a completed randomized, single-masked, parallel-group trial with listed enrollment of 270. The study concerns Veterans with obstructive sleep apnea and compares protocol-based sleep apnea and CPAP education and support with additional education and support intended to allow individualized PAP pressure adjustment.

The primary outcome is treatment adherence, measured as average CPAP use in hours per night over 2 months. The posted group means are 2.1 hours per night among 96 participants in the self-management group and 2.2 among 113 in the individualized pressure-adjustment group. The mean difference is 0.1 hours per night, with an unadjusted p-value of 0.641 and a confidence interval that includes zero.

The registry reports no secondary outcome measures and posts zero serious adverse events, zero other adverse events, and zero deaths in both groups. Details of the original sample-size rationale, analysis plan, missing-data handling, and analysis populations are not reported in the registry record.