Pragmatic Obstructive Sleep Apnea Weight Loss Trial Assessing Effectiveness and Reach: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pragmatic Obstructive Sleep Apnea Weight Loss Trial Assessing Effectiveness and Reach (NCT05104450), a randomized na study. The primary outcome is Change in Sleep-related Quality of Life.

2. Clinical Question

3. Trial Design

POWER was a randomized, parallel-group trial with single masking, listed as completed. The registry reports enrollment of 696. The groups were lifestyle intervention plus usual care and usual care control. The intervention description emphasizes a pragmatic approach: proactively offering remote video-based and self-directed lifestyle support with remote coaching, rather than requiring the kind of time- and resource-intensive program described in the trial background.

The registry lists a start date of 2022-04-04 and a primary completion date of 2024-08-01. The primary endpoint timeframe was baseline to 12 months post randomization. The main reported outcome was change in FOSQ score, with unadjusted mean changes reported for the groups. The registry also describes an adjusted between-group analysis and reports a mixed-model result.

The study description identifies sleep-related quality of life and weight as co-primary endpoints. The available primary result concerns sleep-related quality of life; a weight result is not reported in the registry record. Listed secondary outcomes include cardiovascular risk scores, sleep symptoms, systolic blood pressure, treatment usage, apnea hypopnea index, longer-term weight change, and short-term sleep-related quality of life. The registry lists 28 outcome measures and 28 analyses, while 26 secondary outcomes are posted in the available results.

The trial’s original sample-size justification, detailed eligibility criteria, masking procedures, and full analysis protocol are not reported in the registry record. The classification “single” indicates the registry’s masking category; who was masked is not reported in the registry record.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The investigators report here the unadjusted mean changes between baseline and 12 month follow-up. The analysis population for adjusted between-group differences includes participants with a measurement at baseline and/or 12 month follow-up; see Statistical Analysis 1 for those results.

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 Change in Sleep-related Quality of Life, measured over baseline to 12 months post randomization.

The primary endpoint is change in sleep-related quality of life, measured with the Functional Outcomes of Sleep questionnaire (FOSQ), between baseline and 12 months post randomization. The registry describes the reported group values as unadjusted mean changes. The FOSQ ranges from 5 to 20, and a higher score indicates a better outcome. It is therefore a patient-reported measure of how sleep-related problems affect functioning, rather than a direct measure of sleep apnea events or body weight.

Sleep-related quality of life is relevant to the trial’s stated aim because the investigators sought to assess whether a pragmatic lifestyle program could improve how patients experience the effects of OSA. It complements physiological outcomes such as apnea hypopnea index and other outcomes such as weight, which the trial description also identifies as important. A change in a questionnaire score can summarize an average group-level shift, but it does not by itself show how every participant changed or whether an observed difference is important to patients.

The registry reports a mean and standard deviation for each group’s change. It also reports the sponsor’s adjusted between-group result from a mixed-model analysis. The registry does not report a threshold for a clinically important FOSQ change, so statistical evidence of a group difference cannot be translated from these results alone into a judgment about clinical importance.

7. Results Posted in the Registry

• Lifestyle Intervention: mean 1.43 versus 0.93 on control; difference 0.50 (95% CI 0.04 to 0.96), standardized effect d = 0.18; z-test p = 0.031.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Cardiovascular Risk Scores

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Lifestyle Intervention: mean 0.12 versus -0.43 on control; difference 0.55 (95% CI -0.15 to 1.25), standardized effect d = 0.14; z-test p = 0.123.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Lifestyle Intervention and the control. The interval gives the range of effects that remain compatible with the data.

PROMIS - Sleep Disturbance Survey Change

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Lifestyle Intervention: mean -3.34 versus -3.02 on control; difference -0.32 (95% CI -1.55 to 0.91), standardized effect d = -0.04; z-test p = 0.611.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Lifestyle Intervention and the control. The interval gives the range of effects that remain compatible with the data.

PROMIS - Sleep Related Impairment Survey Change

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Lifestyle Intervention: mean -5.67 versus -3.13 on control; difference -2.54 (95% CI -3.95 to -1.13), standardized effect d = -0.31; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Lifestyle Intervention than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary results cover several distinct domains, including cardiovascular risk, sleep symptoms and impairment, blood pressure, treatment usage, apnea hypopnea index, and weight. The overall pattern is not uniform across these outcomes: some sponsor-reported comparisons have confidence intervals that exclude zero, while others include zero. This range of findings is compatible with outcome-specific effects and varying precision; it does not establish a single broad effect across all aspects of OSA or health.

The secondary outcomes complement the primary FOSQ result but do not substitute for it. They address related yet different questions, use different measures, and in some cases use different follow-up timeframes. Their results may help describe whether changes in sleep-related quality of life align with other measured outcomes, but the registry does not establish that one outcome mediates or explains another.

Unadjusted secondary findings carry less weight than a prespecified primary result because many outcomes and comparisons create more opportunities for chance findings. The registry lists 28 outcome measures and 28 analyses, and reports 26 secondary outcomes in the available results. A multiplicity adjustment or a prespecified hierarchy for interpreting the secondary tests is not reported in the registry record. Without that information, isolated p-values among secondary outcomes should be treated as exploratory rather than as definitive confirmation.

The secondary comparisons also have outcome-specific sample sizes and uncertainty. Their precision cannot be inferred from the primary endpoint, and the direction of a numerical change is not necessarily favorable unless the measure’s scoring direction is known. The registry’s posted results are best read outcome by outcome, with attention to estimates and confidence intervals, while keeping the primary result distinct from the broader set of secondary analyses.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Cardiovascular Risk Scores
• PROMIS - Sleep Disturbance Survey Change
• PROMIS - Sleep Related Impairment Survey Change
• Systolic Blood Pressure From VA Medical Record
• Treatment Usage Per Day
• Apnea Hypopnea Index 1b Criteria
• Longer-term Change in Weight
• Short-term Change in Sleep-related Quality of Life

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: it is a score measured on a scale, and the registry reports group means and standard deviations. A mean summarizes the average observed change in a group. A standard deviation describes how spread out individual changes are around that group average. The standard deviation is not the uncertainty of the mean or a confidence interval for the between-group effect.

What does the simple recalculation compare?

The recalculated mean difference is an unadjusted contrast between the two posted group means. For the primary outcome, the intervention group’s mean change was 1.43 and the control group’s was 0.93; the recalculated contrast is 0.5 units. This comparison uses the posted summaries without the covariate or repeated-measure structure that may be reflected in a model-based analysis. It is useful as a direct description of the reported group averages, but it is not the sponsor’s primary inferential result.

The recalculation also uses a pooled standard deviation of 2.817 to describe the common spread used in that comparison. Pooling is a statistical approach for combining information about variability across groups under its assumptions. The recalculated interval and p-value depend on the summary-statistic method and should not be mistaken for the sponsor’s model-based result.

What does the mixed-model result represent?

The registry reports the sponsor’s analysis as a Mixed Models Analysis, with a net mean difference of 0.64, a 95% confidence interval from 0.25 to 1.04, and p = 0.001. In general, mixed models can account for correlations among repeated measurements and can estimate group contrasts while incorporating the structure specified in the model. The registry record does not report the model’s full specification, covariates, covariance structure, or estimation details, so those features cannot be characterized more specifically.

The primary result description distinguishes the unadjusted mean changes from the adjusted between-group differences. It says the adjusted analysis population includes participants with a measurement at baseline and/or 12 month follow-up. The exact handling of incomplete measurements, and whether other analysis populations were used, are not reported in the registry record.

How should p-values be read?

A p-value describes how incompatible the observed data are with a specified null hypothesis under the assumptions of an analysis. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The unadjusted recalculation gives p = 0.031; the sponsor reports p = 0.001 for its mixed-model analysis. Because the estimates come from different analyses and differ, the sponsor’s reported analysis takes precedence for interpretation of the trial result.

As a general statistical point, a p-value depends on both the estimated difference and its uncertainty. A small p-value does not mean that every participant experienced improvement, and a larger p-value does not prove that groups are identical. Interpretation should consider the estimate and confidence interval as well as the analysis design.

Why can adjustment change the estimate?

General explanation: adjusted models may account for baseline values, repeated observations, or other prespecified variables, depending on the model. Such adjustment can alter an estimate or its precision compared with a direct comparison of group averages. The registry does not report enough model detail to identify why the sponsor’s estimate differs from the recalculation. The difference should be noted rather than explained by speculation.

The registry’s primary endpoint description gives the questionnaire range and direction: higher FOSQ values indicate a better outcome. That direction helps interpret the sign of a mean change. For other continuous outcomes, scale direction and clinically meaningful thresholds are not consistently stated in the available results, so their numerical direction should not automatically be labeled favorable or unfavorable.

General explanation: over-dispersion matters most for count outcomes when variability exceeds what a simple count model expects; accounting for it can change standard errors and confidence intervals. The posted primary outcome is a continuous questionnaire score, not a count outcome. The registry record does not report an over-dispersion assessment as part of the primary analysis.

The registry does not report the trial’s full statistical analysis plan, the original sample-size justification, any interim analyses, or a multiplicity procedure. Those omissions limit the ability to assess whether the posted analyses follow a prespecified hierarchy or how all planned tests were handled. The reported analysis type and estimates can be described, but unreported details should not be inferred.

11. Confidence Intervals and Precision

The sponsor’s primary estimate is a net mean difference of 0.64 FOSQ units, with a 95% confidence interval from 0.25 to 1.04. A confidence interval expresses uncertainty in an estimated effect: under the model and its assumptions, it gives a range of values compatible with the observed data at the stated confidence level. It is not a range containing 95% of individual participants’ changes, and it does not mean there is a 95% probability that the true effect lies inside this particular interval.

The interval is entirely above zero, which is consistent with the sponsor’s reported p = 0.001 for a test of no between-group difference. The interval also conveys that the estimated effect is not known exactly. Its lower and upper limits represent materially different possible magnitudes, and the registry does not give a threshold for deciding which magnitudes would be clinically important.

The simple recalculation has an estimate of 0.5 units and a 95% confidence interval from 0.045 to 0.955. Its interval is also above zero, but it pertains to the unadjusted calculation rather than the sponsor’s mixed model. Comparing intervals from different methods should not obscure that they answer the contrast using different analyses. Given the discrepancy, the sponsor’s own analysis takes precedence.

Precision also depends on the amount and variability of observed data. The posted primary group summaries include different numbers of participants and standard deviations of 2.73 and 2.89. These describe the available outcome data; they do not explain why some enrolled participants do not appear in the group summaries. Missing-data mechanisms and their impact on precision are not reported in the registry record.

Statistical precision and clinical importance are separate. Even an interval that excludes zero does not establish that the difference is large enough to matter to patients. The FOSQ scale runs from 5 to 20, but the registry does not report a clinically important change threshold. Interpretation should therefore remain anchored to the estimated scale difference and its uncertainty, without assigning an unreported clinical label.

12. 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 illustrations translate an observed mean difference into hypothetical planning requirements under stated assumptions. For these continuous outcomes, the figures are participants per arm, not total events. 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. The trial’s own original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column describes the number of participants per arm that the illustration estimates would be needed to have an 80% chance of detecting the specified effect under those assumptions. The 90% power column describes the corresponding planning requirement for a 90% chance. These are planning targets: greater target power generally requires a larger planned sample when other assumptions remain the same.

Rows for planning for 75% or 50% of the observed effect consider effects smaller than the observed estimate. Smaller effects are harder to distinguish from random variation, so the illustrations require more participants per arm. These scenarios are not claims that such effects will occur or that the observed effect is stable. They show how planning requirements change if the effect worth detecting is smaller than the posted estimate.

These calculations are illustrations based on the observed effect, not the original trial design. They are not observed power, which is not an informative substitute for the estimate and its confidence interval after results are available. The plotted curve represents the same planning relationship across effect assumptions; it should be read as a sensitivity illustration, not as evidence about the trial’s actual operating characteristics.

13. 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 = 0.500; pooled standard deviation σ = 2.817.
2. n = 2 × (1.96 + 0.842)² × 2.817² / 0.500² = 498.3, rounded up to 499 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 <- 0.5                      # observed difference in means
sigma <- 2.81701038716003                     # 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(5104450)
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 = 499 (80% power), 668 (90% power)
 75% of observed effect: n per arm = 886 (80% power), 1,186 (90% power)
 50% of observed effect: n per arm = 1,994 (80% power), 2,669 (90% power)
power.t.test: 499.2 
Simulated power at n = 499 per arm: 0.801 (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.

14. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis is a mixed-model net mean difference of 0.64, with a 95% confidence interval from 0.25 to 1.04 and p = 0.001. The recalculated comparison from the posted unadjusted means is 0.5, with a 95% confidence interval from 0.045 to 0.955 and an unadjusted p-value of 0.031. The discrepancy flag indicates that the two analyses do not match.

The recalculation is a simple, unadjusted analysis of the reported group summaries. The sponsor’s result is identified as a mixed-model analysis and is the analysis reported by the trial. The registry does not provide enough detail about the model specification or adjustment variables to explain exactly why the estimates differ. It would be inappropriate to attribute the discrepancy to a particular modeling decision or data-handling choice without that information.

For interpretation of the trial’s inferential result, the sponsor’s own analysis takes precedence. The recalculation remains useful as a check on the arithmetic relationship between the posted unadjusted means and an unadjusted comparison, but it should not replace or be described as the trial’s adjusted effect. Both estimates point in a positive direction; their different values and uncertainty reflect that they are not the same analysis.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, timeframes, or statistical comparisons. Even if each individual test uses a conventional threshold, conducting many tests increases the chance that at least one result will appear noteworthy by chance. This general statistical issue is relevant because the registry lists 28 outcome measures and 28 analyses and reports 26 secondary outcomes.

A planned hierarchy can distinguish confirmatory outcomes from exploratory ones, and multiplicity procedures can adjust how evidence is evaluated across a family of tests. The registry identifies sleep-related quality of life and weight as co-primary endpoints in the study description, but a weight result is not reported in the registry record. A multiplicity strategy, outcome hierarchy, and any adjustment for multiple testing are not reported in the registry record. Consequently, the secondary p-values should not be treated as if each were the sole confirmatory test.

Missing-data details are also limited. The primary outcome description states that the unadjusted mean changes are reported between baseline and 12 month follow-up. It further says the adjusted between-group analysis includes participants with a measurement at baseline and/or 12 month follow-up. The registry does not report the precise number of missing observations for each analysis, reasons for missingness, imputation methods, assumptions about missing data, or sensitivity analyses.

The group sizes in the posted results differ by endpoint, and they do not match the overall enrollment figure. Those differences can reflect outcome availability or analysis-specific data requirements, but their causes are not reported in the registry record. They should not be interpreted as evidence for a particular dropout pattern or missingness mechanism.

The registry record also does not report the full statistical analysis plan, the original sample-size justification, or interim analyses. These omissions limit assessment of prespecification and the consequences of missing data or multiple testing. They do not erase the reported estimates, but they constrain how strongly the full set of results can be interpreted.

16. 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: cardiovascular (6.0%), gastrointestinal (2.9%), musculoskeletal (2.9%), neurological (2.6%), pulmonary/respiratory (2.0%).

The registry reports serious adverse events in 54 participants at risk of 348 in the lifestyle-intervention group and 40 participants at risk of 348 in the usual-care control group. Other adverse events were reported for 126 participants at risk of 348 in the intervention group and 127 at risk of 348 in the control group. The registry also reports 1 death in the intervention group and 2 in the control group.

These are descriptive counts from the safety record. They do not, on their own, show that an event was caused by the intervention or that the groups had equivalent safety. The listed serious-event categories include cardiovascular, gastrointestinal, musculoskeletal, neurological, pulmonary/respiratory, and renal/urologic events; the registry gives the corresponding percentages as 6.0%, 2.9%, 2.9%, 2.6%, 2.0%, and 1.7%.

The safety counts should be interpreted separately from the primary outcome analysis. The registry does not provide event-level timing, severity details beyond the serious-event classification, attribution, or statistical comparisons in the information. The counts are therefore not a causal safety estimate and should not be used to conclude that the intervention did or did not cause harm.

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

The reported primary outcome is a difference in average change in sleep-related quality of life between randomized groups. The sponsor’s mixed-model estimate is 0.64 FOSQ units, with a 95% confidence interval from 0.25 to 1.04. This is evidence of a group-level difference under the sponsor’s reported analysis. The FOSQ score ranges from 5 to 20, and higher scores indicate a better outcome.

That result does not mean every participant improved, or that the average difference describes any individual’s experience. It also does not establish how important the difference is to patients. A threshold for clinically important change is not reported in the registry record. Statistical evidence that an estimate differs from zero is not the same as evidence that the magnitude is meaningful in everyday life.

The result applies to the endpoint and timeframe reported: sleep-related quality of life from baseline to 12 months post randomization. It does not by itself establish an effect on weight, apnea hypopnea index, cardiovascular risk, or every sleep symptom. Weight is described as a co-primary endpoint, but its result is not reported in the registry record. The available secondary outcomes address some related questions but have their own limitations, including the unreported multiplicity approach.

The intervention was tested as an addition to usual care, compared with usual care alone. The study description presents it as a pragmatic, remotely delivered program with coaching. The available record does not give enough information to determine how participation, intervention exposure, or missing outcome data may have affected results. Generalizing the estimate beyond the population and setting studied requires information about eligibility and implementation that is not reported in the registry record.

The findings should therefore be read as a statistical comparison of group-average FOSQ changes, supported by the sponsor’s reported model, rather than as a guarantee of individual response or a complete assessment of clinical benefit and harm. The estimate, its interval, the absent weight result, and the limitations in reporting all matter to the interpretation.

18. Why This Trial Matters Statistically

This trial raises a useful biostatistical question about the difference between efficacy under intensive conditions and effectiveness when an intervention is proactively offered in a pragmatic setting. The study description emphasizes barriers to translating lifestyle-program findings into population health and describes remote, self-directed support with coaching. That context makes the implementation setting part of the scientific question, rather than treating delivery as incidental.

Statistically, the primary result illustrates why the analysis method matters. The posted group means permit a simple unadjusted comparison, while the sponsor reports a mixed-model estimate. The estimates differ, and the registry does not give the model details needed to account for that difference. Keeping the descriptive calculation distinct from the sponsor’s reported analysis prevents a transparent arithmetic check from being mistaken for the trial’s adjusted result.

The result also demonstrates the importance of precision and clinical interpretation. A confidence interval describes uncertainty around the group contrast, but a statistical difference does not define clinical importance. The registry reports the FOSQ scale range and direction, yet no clinically important change threshold. Statistical interpretation must therefore avoid turning a p-value into a claim about the size of a meaningful benefit.

Finally, the trial’s co-primary framing, broad set of secondary measures, and incomplete reporting of multiplicity and missing-data methods show why outcome structure matters. The primary sleep-related quality-of-life result is informative, but the absence of a posted weight result leaves part of the stated primary question unanswered in the registry record. The case highlights the value of interpreting estimates alongside analysis methods, uncertainty, outcome hierarchy, and reporting completeness.

19. 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.

Several limitations are specific to the available record. The study description names sleep-related quality of life and weight as co-primary endpoints, but only the sleep-related quality-of-life result is reported in the primary results. The registry does not report a clinically important change threshold for the FOSQ, so clinical importance cannot be established from statistical significance alone.

The sponsor’s mixed-model estimate differs from the simple unadjusted recalculation. The sponsor’s result takes precedence, but the model specification, adjustment variables, and reasons for the discrepancy are not reported in the registry record. The posted primary group means are unadjusted, while the sponsor’s estimate is described as a net mean difference; they should not be treated as interchangeable.

The registry describes the adjusted analysis population as participants with a measurement at baseline and/or 12 month follow-up, but does not report detailed missing-data methods, reasons for unavailable measurements, or sensitivity analyses. Outcome-specific group sizes vary, and the record does not explain those differences.

The registry lists multiple outcomes and analyses but does not report a multiplicity strategy or a hierarchy for interpreting secondary findings. The full statistical analysis plan, interim analyses, and original sample-size justification are not reported in the registry record. Eligibility criteria and details identifying who was masked are also not reported in the registry record.

These reporting limits do not invalidate the posted results, but they narrow what can be concluded about robustness, clinical importance, missingness, and the broader collection of outcomes. The results should be interpreted within the boundaries of the endpoint and analysis details that are reported.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Pragmatic Obstructive Sleep Apnea Weight Loss Trial Assessing Effectiveness and Reach — NCT05104450.
• Registry record last updated 2026-06-12; status last verified 2026-05.
• Results first posted 2025-10-21.
• 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.

23. Record Summary

POWER was a completed randomized, parallel-group, single-masked trial comparing a behavioral lifestyle intervention plus usual care with usual care control among patients with obesity and obstructive sleep apnea. The registry reports enrollment of 696 and a primary endpoint of change in FOSQ score from baseline to 12 months post randomization.

The sponsor reported a mixed-model net mean difference of 0.64, with a 95% confidence interval from 0.25 to 1.04 and p = 0.001. A simple unadjusted recalculation differs; the sponsor’s own analysis takes precedence. The record also reports secondary outcomes and safety counts, but the weight result described as co-primary, multiplicity methods, detailed missing-data methods, and full analysis plan are not reported in the registry record.