Alice PDx User/Validation Extended Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Alice PDx User/Validation Extended Trial (NCT00841906), a randomized na study. The primary outcome is This Study Will Compare the Apnea Hypopnea Index Recorded by the Alice PDx to the Physiological Data Recorded by Its Predicate Device the Alice 5 System and Validate Its Equivalence..

2. Clinical Question

The trial also addressed usability: whether participants could apply basic leads using the participant manual and whether sleep technicians could apply advanced leads using the health care provider manual. Questionnaires were intended to assess the manuals and setup process. The posted primary result, however, focuses on AHI rather than questionnaire-based usability. It therefore does not by itself answer all of the trial’s stated questions about user experience, technician setup, or data quality.

3. Trial Design

This was a randomized, parallel-group trial with no masking. The registry lists an enrollment of 32 and a completed status. The recorded start was 2008-09, and primary completion was 2009-06. Philips Respironics was the sponsor.

The two intervention labels distinguish Alice PDx with written instructions from Alice PDx with written and verbal instructions. However, both arm descriptions say participants would receive little or no instruction about setup. The difference between the labels and the matching descriptions leaves the actual instruction contrast unclear in the registry record. The primary physiological comparison is separate from this instruction-group comparison: participants completing the overnight portion each had an Alice 5 System night and an Alice PDx night. The registry reports 30 observations for each device in the primary outcome results.

The primary outcome was mean AHI during the lab night. A secondary objective compared the Alice PDx Good Study Indicator (GSI) measurement between home and laboratory settings; the posted secondary results include 27 observations for each setting. The stated study aims also included questionnaires, event-by-event comparison, and assessing the ability to record good-quality data. The registry record does not provide posted results for each of these aims.

The original sample-size justification, planned analysis details beyond the reported test, and any interim-analysis plan are not reported in the registry record. The sample-size figures associated with the observed results are planning illustrations, not documentation of the trial’s original design.

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: All participants that completed the overnight portion of the study were analyzed, they each had an Alice 5 System night and an Alice PDx night.

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 This Study Will Compare the Apnea Hypopnea Index Recorded by the Alice PDx to the Physiological Data Recorded by Its Predicate Device the Alice 5 System and Validate Its Equivalence., measured over Lab Night.

The primary endpoint is the mean apnea-hypopnea index (AHI) recorded during the lab night. AHI is defined in the registry as the number of apnea and hypopnea events per hour of sleep, and is used as an index of sleep-apnea severity. The comparison asks whether Alice PDx records an AHI similar to the value recorded by the predicate Alice 5 System.

For each device, the posted results give the mean and standard deviation. The Alice 5 System mean is 13.16 events, with a standard deviation of 24.69; the Alice PDx mean is 20.24 events, with a standard deviation of 24.88. These are group summaries, not a description of each participant’s paired device readings. The registry says that participants in the primary analysis had a night with each device, but does not report the individual paired differences.

Agreement or equivalence between measurement systems is not established merely by comparing means or by finding no statistically significant difference. Interpretation also depends on the analysis method, the uncertainty around the comparison, and a prespecified acceptable margin. An equivalence margin and a registry-reported equivalence decision are not reported in the registry record.

7. Results Posted in the Registry

• Alice PDx: mean 20.24 versus 13.16 on control; difference 7.08 (95% CI -5.46 to 19.62), standardized effect d = 0.29; z-test p = 0.269.

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.

The Secondary Objective of This Study is to Compare the Measurements of the Alice PDx When Patients Complete the Set-up of the Device at Home and When Patients Are Set up by a Sleep Technician in the Sleep Laboratory.

• 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:
  • Alice PDx in the Lab: mean 70.30 versus 88.90 on control; difference -18.60 (95% CI -32.04 to -5.16), standardized effect d = -0.74; z-test p = 0.007.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Alice PDx in the Lab than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry posts one secondary outcome: a comparison of Alice PDx GSI measurements for home and laboratory settings. The direction of the posted summaries favors a higher average GSI in the home setting, while the unadjusted recalculation indicates a difference between settings. No sponsor-reported analysis is listed for this endpoint. These secondary data concern the device’s GSI measure, not the primary AHI comparison.

The endpoint is relevant to the study’s data-quality objective, but it addresses a different question from whether Alice PDx and Alice 5 System record equivalent AHI values. A difference in average GSI across settings could reflect the setting or measurement circumstances; the registry does not report enough detail to identify a cause. It should not be interpreted as direct evidence for or against primary endpoint equivalence.

Secondary findings generally carry less weight when they are unadjusted and when the registry does not document a prespecified testing hierarchy. In general, considering multiple outcomes can increase the chance of an apparently notable finding arising by chance. The record does not report a multiplicity adjustment. The secondary result is therefore best considered supportive or exploratory context, rather than a confirmatory answer to the primary question.

More broadly, the registry’s stated aims include usability, physiological measurement comparison, and data quality, but only limited secondary outcome results are posted. The available secondary information does not provide a complete picture across all those aims. It adds context about setting-related GSI measurements without resolving the statistical discrepancy in the primary AHI analysis.

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

• The Secondary Objective of This Study is to Compare the Measurements of the Alice PDx When Patients Complete the Set-up of the Device at Home and When Patients Are Set up by a Sleep Technician in the Sleep Laboratory.

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports group means and standard deviations for AHI, alongside a named sponsor analysis. A mean describes the arithmetic average of the recorded values. A standard deviation describes how spread out individual values are around that average; a relatively large standard deviation signals considerable variation in the observed measurements. The summaries alone do not show the distribution’s shape or the relationship between the two device readings for each participant.

What does the recalculated mean difference represent?

The recalculated estimate is the Alice PDx mean minus the Alice 5 System mean: 7.08 events. It is an unadjusted comparison computed from the posted group summaries. It does not incorporate participant-level data, covariates, or a paired analysis. Although each participant in the primary analysis had a night with each system, the summary statistics do not provide the individual within-participant differences needed to assess how those readings correspond. The recalculation is therefore not a substitute for an analysis that uses the trial’s underlying data and design.

What does a confidence interval communicate?

A confidence interval expresses uncertainty around an estimated effect under the assumptions of the method used to construct it. A narrower interval indicates greater precision; a wider interval leaves a broader range of effects compatible with the estimate and method. The recalculated interval runs from -5.463 to 19.62. It includes zero and extends in both directions, so the recalculated mean difference is not estimated precisely enough to establish a clear direction of the mean contrast. The confidence level for this interval is not reported in the registry record.

How should the p-values be read?

A p-value describes how compatible data are with a statistical null model under the assumptions of a particular test. It is not the probability that the null hypothesis is true, and it does not measure clinical importance, agreement, or equivalence. The recalculated unadjusted p-value is 0.269. The registry’s reported Wilcoxon (Mann-Whitney) p-value is < 0.001. These values arise from results that are discrepant, and the registry does not provide sufficient detail to explain the difference. As a general statistical principle, different tests can target different features of a distribution and rely on different assumptions; the reported method and the recalculated mean contrast should not be treated as interchangeable.

Why does the design structure matter?

In general, when the same participants contribute observations under both measurement conditions, their paired values may be correlated. An analysis that accounts for pairing can use within-participant contrasts and may estimate uncertainty differently from an analysis that treats groups as independent. The registry states that participants completing the overnight portion had a night with each device, but the sponsor’s reported test is named Wilcoxon (Mann-Whitney), and the registry does not describe how pairing was handled. No further assumptions about the sponsor’s method can be made from the record.

Over-dispersion, in general, means that observed variability is greater than a model expects. Large variability can make estimates less precise and can make a mean sensitive to unusually high or low observations. The posted standard deviations are large relative to the means, but the registry does not give individual data or distribution plots in text. It is not possible from these summaries alone to determine the extent or cause of skewness, outliers, or other distributional features.

What is the role of the secondary analysis?

The secondary outcome compares the mean GSI measurement for home and laboratory settings. A mean difference and an unadjusted recalculation are available for that outcome, but no sponsor-reported analysis is listed. The GSI is described as a firmware function used to assess whether recorded data are good quality. It is distinct from AHI: a difference in GSI does not directly establish equivalence of AHI measurements or settle the primary comparison.

Multiple outcomes and comparisons can raise multiplicity concerns. In general, testing many hypotheses increases the chance that at least one result appears statistically notable by chance if results are considered without adjustment. The registry does not report a multiplicity adjustment or a hierarchy specifying how secondary findings should be interpreted. As a result, unadjusted secondary p-values, where available, should be regarded as exploratory rather than treated as definitive confirmation.

The registry’s analysis description is incomplete for independent evaluation: a confidence level for the recalculated interval, the exact sponsor model and its handling of repeated observations, missing-data procedures, and a full prespecified statistical plan are not reported in the registry record. These omissions limit how far statistical interpretation can go beyond the posted summaries.

11. Confidence Intervals and Precision

The recalculated primary mean difference is 7.08 events, with an interval from -5.463 to 19.62. The interval is broad and includes zero, indicating that the available summary-based estimate is imprecise. It is compatible with values on either side of zero under the recalculation method, rather than locating the mean contrast narrowly. The confidence level for this interval is not reported in the registry record.

Precision concerns how narrowly an effect is estimated, not whether a result is important or whether devices are equivalent. The primary standard deviations are 24.69 for Alice 5 System and 24.88 for Alice PDx, showing substantial spread in the recorded values relative to the estimated mean contrast. Group-level means and standard deviations cannot show whether participants with higher AHI on one device also tended to have higher AHI on the other, nor how close each participant’s two readings were.

The registry’s sponsor-reported p-value is < 0.001, while the recalculated mean-based result is different. The registry does not report an interval for the sponsor analysis or enough methodological detail to derive its precision from the posted record. The sponsor result should take precedence given the discrepancy, but the reported p-value alone does not convey the size or precision of the sponsor’s estimated effect. An equivalence margin is also not reported, so the interval cannot be judged against a stated acceptable range.

For the secondary GSI comparison, the recalculated interval is also available in the posted results, but the registry lists no sponsor-reported analysis. As with the primary outcome, precision must be considered separately from whether a comparison is statistically notable. The secondary estimate does not supply the missing equivalence criterion for the primary device comparison.

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 planning figures express participants per arm for 80% power and participants per arm for 90% power. They are based on the observed effect and use the stated planning assumptions: two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. The calculation uses the stated fraction of the observed effect as the effect to detect. These are planning illustrations, not observed power.

The 80% power column indicates the number of participants per arm the planning calculation estimates would be needed to have an 80% chance of detecting the chosen effect under those assumptions, if that effect and model were appropriate. The 90% column uses a higher target chance of detecting the same chosen effect, which generally requires more participants per arm. The rows represent planning for 100%, 75%, and 50% of the observed effect. The smaller the assumed effect, the more participants the calculation requires. For example, the primary comparison’s planning estimates rise as the assumed effect fraction decreases.

These estimates should not be read as the trial’s original recruitment target or as evidence that the trial was adequately powered. The registry’s original sample-size justification and original design assumptions are not reported in the registry record. The planning figures are calculated from the observed effect, and observed effects can be unstable, particularly when estimates are imprecise. They do not validate the effect estimate or resolve the discrepancy between the sponsor’s reported test and the recalculation.

The outcomes in this record are continuous, so the figures refer to participants per arm rather than total events. The calculation assumptions are a simplified illustration and do not document how repeated measurements, potential pairing, missing observations, or the sponsor’s reported test would affect a design-specific sample-size analysis.

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 = 7.080; pooled standard deviation σ = 24.785.
2. n = 2 × (1.96 + 0.842)² × 24.785² / 7.080² = 192.4, rounded up to 193 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 <- 7.08                      # observed difference in means
sigma <- 24.7851820650969                     # 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(841906)
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 = 193 (80% power), 258 (90% power)
 75% of observed effect: n per arm = 343 (80% power), 458 (90% power)
 50% of observed effect: n per arm = 770 (80% power), 1,031 (90% power)
power.t.test: 193.3 
Simulated power at n = 193 per arm: 0.797 (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 registry reports a sponsor analysis of the primary comparison using the Wilcoxon (Mann-Whitney) method, with p < 0.001. It does not report the sponsor’s effect estimate or confidence interval. A recalculation from the posted means and standard deviations gives a different result: a mean difference of 7.08 events, an interval from -5.463 to 19.62, and an unadjusted p-value of 0.269.

These results are discrepant. The recalculation is a simple, unadjusted comparison of the available summary statistics; it is not a reproduction of the sponsor’s participant-level analysis. The registry does not report enough detail to identify why the results differ, including how the sponsor analysis handled the fact that participants had a night with each device. In keeping with the interpretation of discrepant results, the sponsor’s own analysis takes precedence. The recalculation is useful as a transparent check on the posted summary data, but it should not replace or overrule the sponsor-reported result.

Even with precedence given to the sponsor analysis, a p-value alone does not establish equivalence. The record does not provide an equivalence margin, the sponsor’s estimate and interval, or a clear equivalence decision. Those omissions prevent a full assessment of the stated equivalence objective.

15. Multiplicity, Missing Data and Analysis Populations

The record includes a primary AHI comparison and a posted secondary GSI comparison, while the study description also mentions questionnaires, event-by-event comparisons, and data-quality assessment. It does not report a statistical hierarchy for these outcomes or a multiplicity adjustment. In general, when multiple hypotheses are tested, the chance of at least one apparently notable result can increase if each result is considered on its own without adjustment. Without a prespecified approach, secondary findings should be interpreted cautiously rather than as independent confirmation.

The primary analysis population is described as participants who completed the overnight portion; the posted results list 30 observations for each device. The secondary comparison lists 27 observations for each setting. The registry does not explain why the numbers differ, whether observations were missing, or how missing data were handled. It also does not state whether any participants were excluded from an analysis after enrollment or provide a complete accounting of the analysis populations.

The record says participants in the primary analysis each had an Alice 5 System night and an Alice PDx night. However, the reported sponsor test is Wilcoxon (Mann-Whitney), and the registry does not say how the repeated observations were handled. Details about the sponsor’s model, assumptions, data transformations, and treatment of within-participant dependence are not reported in the registry record.

The original sample-size rationale, a complete statistical analysis plan, and any interim-analysis plan are not reported in the registry record. Nor is an equivalence margin or an explicit rule for declaring equivalence reported. These omissions matter because they limit evaluation of which analyses were planned, how missing observations could affect results, whether secondary tests were exploratory, and what criterion the investigators intended to use for equivalence.

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 non-serious adverse event terms

The registry reports serious adverse events for the two instruction groups. There were 0 serious adverse events among 18 participants at risk in the written-instructions group and 0 among 14 participants at risk in the written-and-verbal-instructions group. Deaths were 0 in each group. For other adverse events, the record reports 1 among 18 at risk in the written-instructions group and 0 among 14 at risk in the written-and-verbal-instructions group.

These are the posted event counts and at-risk denominators, not a basis for estimating comparative safety with precision. The registry does not provide details about the nature of the other adverse event or an analysis comparing the groups. The trial’s enrollment and short study description do not establish that uncommon harms could be assessed reliably. The arm descriptions also do not clearly distinguish the instruction conditions, which further limits interpretation of a safety comparison by group. The record reports the observed counts; they should not be read as proof that no unreported or less common event could occur.

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

The posted record documents a randomized study that evaluated Alice PDx usability and compared its AHI measurements with those recorded by Alice 5 System. The primary outcome summaries show variability in both device measurements, and the simple unadjusted recalculation yields an imprecise mean difference. The sponsor-reported statistical test gives a different result and takes precedence because of that discrepancy.

The registry states equivalence as an objective, but an objective is not the same as a demonstrated result. A formal equivalence conclusion requires a defined acceptable margin and an analysis that evaluates the estimate against that margin. Neither the margin nor a clear equivalence decision is reported in the registry record. The reported p-value, whether sponsor-reported or recalculated, cannot by itself establish equivalence.

The trial also concerns usability, home setup, technician setup, questionnaires, event detection, and the GSI data-quality indicator. The posted results do not answer every one of these questions. The instruction-group descriptions are not clearly differentiated, and the registry does not provide detailed questionnaire results or a full account of event-by-event findings. The secondary GSI comparison addresses a different outcome and does not settle the AHI comparison.

Accordingly, the record supports a cautious description of the reported measurements and the discrepancy between analyses, not a broad claim about device validation or clinical performance. The statistical information is limited by missing analysis details, unclear handling of the repeated device nights, absent equivalence criteria, and incomplete reporting of missing data and the full analysis plan. These limits should remain central when interpreting the posted results.

18. Why This Trial Matters Statistically

This trial illustrates why agreement between measurement devices is not established by a comparison of group averages alone. When each participant contributes readings from both devices, the relationship between those readings can be important. Group means and standard deviations do not show whether the devices agree closely for individual participants, and the registry does not post participant-level paired differences.

It also illustrates the difference between a superiority-style test and an equivalence question. A p-value addresses compatibility with a statistical null model; it does not state whether two devices are sufficiently close for a defined purpose. Equivalence requires an acceptable margin and an analysis designed to evaluate that margin. Those details are not reported in the registry record, despite equivalence being part of the stated objective.

The discrepancy between the sponsor-reported Wilcoxon (Mann-Whitney) result and the unadjusted recalculation from summary means and standard deviations is another important statistical issue. Readers need effect estimates, uncertainty intervals, and enough method detail to understand why analyses differ. Here, the sponsor analysis takes precedence, but the record does not provide the estimate and interval needed to assess its magnitude and precision.

Finally, the trial combines usability and measurement questions, while the posted results address only part of that scope. Clear reporting of populations, missing data, outcome hierarchy, analysis of repeated measurements, and equivalence criteria would make the evidence easier to evaluate. These are general principles for interpreting device-comparison studies, and they are especially relevant to the limits of this registry record.

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.

The instruction-group descriptions are ambiguous: the titles differ, but both descriptions say participants would receive little or no instruction concerning setup. The registry therefore does not clearly establish what practical contrast the randomization created. This limits interpretation of any comparison by instruction group.

The primary outcome was collected on separate device nights in the same participants, according to the registry description. Yet the listed sponsor method is Wilcoxon (Mann-Whitney), and no explanation is given for handling the repeated observations. The summary statistics do not include paired differences or participant-level data. This is a trial-specific limitation for interpreting the primary device comparison.

The registry reports a sponsor p-value but no sponsor effect estimate or interval. The summary-based recalculation is discrepant and cannot be reconciled from the available methods description. An equivalence margin and a formal equivalence conclusion are also absent. Consequently, the posted record does not permit a complete assessment of the central equivalence claim.

The primary outcome population is described by completion of the overnight portion, while the secondary outcome has fewer observations per setting. Reasons for these analysis counts, missing-data handling, and the full analysis populations are not reported in the registry record. The study description names questionnaires, event-by-event assessment, and GSI quality assessment, but the record does not provide complete results for all these aims. These reporting gaps restrict interpretation beyond the outcomes and summaries that are posted.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Alice PDx User/Validation Extended Trial — NCT00841906.
• Registry record last updated 2018-11-19; status last verified 2018-10.
• Results first posted 2016-11-29.
• 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

The registry describes NCT00841906, a completed randomized, unmasked, parallel-group study sponsored by Philips Respironics. It lists enrollment of 32, with a recorded start of 2008-09 and primary completion of 2009-06. The study assessed Alice PDx usability and compared its AHI measurements with those from Alice 5 System.

For the primary endpoint, the posted summaries include 30 observations for each device. The sponsor-reported Wilcoxon (Mann-Whitney) result is p < 0.001. An unadjusted recalculation from the posted summary statistics is discrepant, so the sponsor’s analysis takes precedence. The registry does not report an equivalence margin or a clear equivalence decision.

A secondary GSI comparison between home and laboratory settings is also posted. Missing-data methods, multiplicity handling, a complete statistical plan, and explanations for analysis-population differences are not reported in the registry record.