This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Comparative Study of the ANNE™ One System to Diagnose Obstructive Sleep Apnea (NCT04643782), a na na study. The primary outcome is Sensitivity and Specificity of Obstructive Sleep Apnea (OSA) Diagnosis Using ANNE Sleep System Compared to Polysomnography.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT04643782 |
| Phase | NA |
| Condition | Obstructive Sleep Apnea of Adult |
| Allocation | Not applicable (single group) |
| Design model | Single Group |
| Masking | None (open label) |
| Enrollment | 287 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Northwestern University |
| Results posted | Yes |
2. Clinical Question
Population
Adults with obstructive sleep apnea were the population of interest. The registry record does not report detailed eligibility criteria, participant characteristics, or the number meeting criteria for moderate to severe disease.
Intervention
Participants wore the ANNE Sleep system during an attended polysomnography study for one night. The device was evaluated as a way to determine the Apnea-Hypopnea Index and diagnose obstructive sleep apnea.
Comparator
Polysomnography (PSG) was the comparator and reference test described in the primary outcome. The two methods were used concurrently during the same night; the registry record does not describe further details of the PSG procedure or how disagreements were adjudicated.
Primary question
How well did diagnoses of moderate to severe obstructive sleep apnea using the ANNE Sleep system agree with diagnoses using PSG, as assessed by sensitivity and specificity? The registry record defines these measures but does not report their estimates or confidence intervals.
This is a diagnostic comparison rather than a trial of treatment. Its central statistical task is to assess how often the device identifies cases and non-cases relative to PSG. Sensitivity concerns correctly identified true cases; specificity concerns correctly identified true non-cases. Both depend on the classifications made by the comparator and the threshold used to define moderate to severe disease. The registry record identifies the diagnostic target but does not report the threshold, the number of cases and non-cases, or the resulting classification counts. Therefore, the research question can be described, but the magnitude and precision of the observed diagnostic performance cannot be assessed from the registry record.
3. Trial Design
The study is listed as completed, with a single-group, open-label, multi-center design. It enrolled 287 participants, and the intervention was the ANNE Sleep device. The registry record lists the condition as obstructive sleep apnea of adult. The planned procedure was for participants to wear the device during an attended PSG study for one night. The primary outcome also specifies that the two methods were worn or used concurrently over one night.
This is not a randomized comparison between separate treatment arms. The device and PSG are compared within the diagnostic assessment, rather than by assigning participants to independent groups receiving different interventions. The registry record lists allocation as NA and masking as none. It does not report whether test interpretation was blinded, whether readings were reviewed independently, or whether the same personnel interpreted both methods.
The registry record gives a study start date of 2021-05-11 and a primary completion date of 2021-11-17. It identifies Northwestern University as the sponsor. Although the study is described as multi-center, the number and locations of participating centers are not reported in the registry record. The protocol's sample-size justification, recruitment approach, and detailed eligibility criteria are also not reported in the registry record.
The enrollment figure describes the study record, but the number contributing data to the primary diagnostic comparison is not reported. The serious and other adverse-event entries each list 225 participants at risk. The record does not explain how that figure relates to enrollment or to the population included in the diagnostic analysis. No statistical analysis entries are listed, so the registry record does not document an analysis population or analysis procedure for the primary endpoint.
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
| Stage | ANNE Sleep System |
|---|---|
| STARTED | 287 |
| COMPLETED | 225 |
| NOT COMPLETED | 62 |
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.
| Characteristic | Unit | Single Arm Study |
|---|---|---|
| Participants analysed | 225 | |
| Age, Continuous | years | 52.7 (SD 14.0) |
| Sex: Female, Male: Female | Participants | 127 |
| Sex: Female, Male: Male | Participants | 98 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 19 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 202 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 4 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 1 |
| Race (NIH/OMB): Asian | Participants | 20 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 1 |
| Race (NIH/OMB): Black or African American | Participants | 26 |
| Race (NIH/OMB): White | Participants | 164 |
| Race (NIH/OMB): More than one race | Participants | 10 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 3 |
| Region of Enrollment: United States | Participants | 225 |
6. Primary Endpoint
The registry lists the primary outcome as Sensitivity and Specificity of Obstructive Sleep Apnea (OSA) Diagnosis Using ANNE Sleep System Compared to Polysomnography, measured over 1 night.
| Element | Registry entry |
|---|---|
| Result type | Mean |
| Measure of spread | 95% Confidence Interval |
| Unit | percentage of true cases |
| Endpoint classification | Continuous Endpoint |
The primary endpoint concerns sensitivity and specificity of the ANNE Sleep system for diagnosing moderate to severe obstructive sleep apnea compared with PSG. The registry record says the diagnosis is based on determining a patient's Apnea-Hypopnea Index using the device, with the methods used concurrently during one night. It defines sensitivity as true positives divided by true positives plus false negatives, and specificity as true negatives divided by true negatives plus false positives.
These measures answer distinct diagnostic questions. Sensitivity describes the proportion of comparator-defined cases that the device classifies as cases. Specificity describes the proportion of comparator-defined non-cases that the device classifies as non-cases. Considering both is useful because a diagnostic method may classify cases and non-cases differently; a single combined measure can obscure that distinction.
The endpoint is reported as a mean with a 95% confidence interval and a unit described as percentage of true cases. However, the registry record does not provide the actual sensitivity or specificity values, their confidence intervals, the diagnostic threshold, or the underlying counts of true positives, false negatives, true negatives, and false positives. Without those results, the endpoint's direction and precision cannot be evaluated. The endpoint describes diagnostic performance relative to PSG; it does not by itself establish how the device performs against an independent or different reference standard.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record identifies a diagnostic accuracy question and defines the two primary measures, but it does not report the numerical results. In particular, no sensitivity estimate, specificity estimate, or corresponding 95% confidence interval is given. There are also no diagnostic classification counts or analysis entries. Consequently, the posted information does not establish the observed level of agreement between ANNE Sleep and PSG, nor how precisely any estimate was obtained.
In general, sensitivity and specificity should be interpreted together and in relation to the reference standard and case definition. Sensitivity concerns the device's classification of PSG-defined cases, while specificity concerns its classification of PSG-defined non-cases. Neither measure is interchangeable with overall agreement, and neither alone describes all aspects of diagnostic performance. Here, the registry record names PSG and moderate to severe obstructive sleep apnea as the comparison framework, but does not state the threshold used to classify disease or provide the counts needed to see how participants were distributed across the diagnostic categories.
A 95% confidence interval is, in general, a way to express statistical uncertainty around an estimate under the assumptions of the analysis. A narrower interval indicates greater precision than a wider interval, while the point estimate alone does not communicate that uncertainty. Because neither interval nor point estimate is reported, no conclusion about the precision of the diagnostic estimates can be made from the registry record. The number enrolled does not substitute for the number with usable paired diagnostic results; that analysis denominator is not reported.
The study's single-group structure means that the diagnostic methods are compared in the same participant setting, not as randomized arms. That makes the central comparison paired in intent, but the registry record does not report the statistical method used to analyze paired results or whether every participant had interpretable measurements from both methods. Masking is listed as none; the record does not clarify what this means for test interpretation. These missing details limit assessment of possible measurement and interpretation bias.
The record reports 287 enrolled participants and lists 225 at risk for each adverse-event category. Those figures describe different registry entries, and the reason for the difference is not reported. They should not be treated as interchangeable denominators or used to infer the diagnostic analysis sample. No sample-size justification or completed statistical analysis is reported. Any planning illustration based on an observed effect should be understood as a planning exercise, not as observed power or a substitute for the trial's actual results. The registry record does not report a diagnostic effect estimate from which readers can evaluate such planning assumptions.
Overall, the available record supports a description of the study question, design, and outcome definitions, but not a quantitative judgment about the device's diagnostic accuracy. The absence of posted estimates is not evidence that the estimates were favorable, unfavorable, or equivalent to PSG; it means that those results are not reported in the registry record.
9. Statistical Methods Explained
What do sensitivity and specificity measure?
In general, diagnostic accuracy is assessed by comparing each test's classification with a reference classification. For this study, the registry record names PSG as the comparator and defines sensitivity and specificity for the diagnosis of moderate to severe obstructive sleep apnea. Sensitivity focuses on comparator-defined cases: it is the fraction classified as positive by the device among those classified as cases by the reference. Specificity focuses on comparator-defined non-cases: it is the fraction classified as negative by the device among those classified as non-cases by the reference. The registry's definitions use the corresponding true-positive, false-negative, true-negative, and false-positive counts.
These measures are not the same as the fraction of all participants classified correctly. A study can have different sensitivity and specificity values, and each describes a separate aspect of classification. The registry record does not report the required counts or the resulting estimates, so neither measure can be interpreted numerically for this study.
Why does the reference standard matter?
In general, a diagnostic comparison measures performance relative to its chosen reference standard. A result indicates how often the test agrees with that reference under the study's classification rules; it does not automatically establish agreement with every possible definition of disease. The registry record identifies PSG as the comparison method and specifies concurrent use over one night. It does not report the PSG scoring details, the AHI threshold for moderate to severe disease, or any adjudication procedure. Those details matter because changing the reference classification or threshold can change which observations count as cases and non-cases.
How should a confidence interval be read?
In general, a confidence interval describes uncertainty around an estimate arising from sampling variation and the analysis assumptions. The point estimate gives the central estimated performance, while the interval conveys the range of values compatible with the data under the specified method. Intervals are interpreted in the context of the measure, study design, and underlying data; they do not remove the possibility of bias or correct limitations in measurement. The registry record specifies a 95% confidence interval for the primary outcome but does not report its endpoints. Therefore, the uncertainty of the estimates cannot be evaluated.
What is important about concurrent testing?
In general, testing the same participants with both methods can support a paired diagnostic comparison, because each participant contributes information to both classifications. The registry describes ANNE Sleep use during an attended PSG study, with concurrent use over one night. However, the record does not state whether all enrolled participants completed both tests, whether results were technically interpretable, or how incomplete pairs were handled. It also does not report the statistical procedure used for paired diagnostic data. The design description alone does not establish that all enrolled participants contributed to the endpoint analysis.
What do the design labels imply?
The registry labels the study single-group, open-label, multi-center, and lists masking as none. In general, an open-label design means the record does not describe masking of study participants or personnel, but the details of who knew which readings are not stated. Lack of masking can be relevant to diagnostic studies if knowledge of one test's result affects interpretation of another. That is a possible methodological concern, not a claim that such influence occurred in this study. The record does not describe reading order, independent interpretation, or blinding safeguards.
There is no randomized allocation between separate arms. Rather, the device is compared with PSG within the study's diagnostic assessment. Therefore, treatment-arm concepts such as a randomized treatment effect do not describe the primary question. The appropriate focus is the agreement and classification performance of the device relative to PSG, while recognizing that the numerical estimates and analysis details are not reported.
What can planning calculations say?
In general, sample-size calculations illustrate how much information a study might require under stated assumptions. The registry's planning description refers to a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effects expressed as fractions of an observed effect. It also distinguishes participants per arm for ordinary outcomes from total events for time-to-event calculations. Those are planning assumptions, not descriptions of this single-group diagnostic design or proof of observed statistical power. Sample-size illustrations are not observed power. The protocol's own sample-size justification is not reported in the registry record, and the record does not provide the diagnostic estimates needed to assess planning based on an observed effect.
In general, over-dispersion refers to variation greater than a statistical model expects. When present, it can make uncertainty estimates too optimistic if the model does not account for it. The registry record does not describe a count-based outcome model or report over-dispersion, so this concept is not evidence about the analysis in this study. More broadly, the registry's missing analysis details mean that model assumptions, handling of paired observations, and precision calculations cannot be independently assessed.
10. Multiplicity, Missing Data and Analysis Populations
The registry record lists one primary outcome measure and no analysis entries. It does not report a multiplicity strategy, such as whether sensitivity and specificity were treated as co-primary measures, whether one was designated as primary within the endpoint, or how any additional analyses were controlled. With no posted estimates or analysis details, the effect of any multiplicity approach cannot be evaluated. The presence of a single listed outcome measure does not establish that only one statistical test or analysis was conducted.
Missing-data procedures are not reported in the registry record. The record does not state how missing device readings, incomplete PSG measurements, uninterpretable recordings, or absent paired results were handled. It also does not report the number of participants with complete data for the primary outcome. Therefore, enrollment of 287 should not be assumed to equal the analysis population.
Analysis populations are not described. The registry record does not identify an intention-to-diagnose, per-protocol, or other analysis set, and no statistical analysis entries are listed. The adverse-event entries identify 225 participants at risk for serious adverse events and 225 at risk for other adverse events, but do not define the diagnostic analysis population or explain how those counts relate to enrollment. These omissions limit the ability to assess selection, missingness, and the denominator used for the accuracy estimates.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Single Arm Study | 0 | 225 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Single Arm Study | 0 of 225 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Single Arm Study | 0 of 225 |
The registry lists one study group. For serious adverse events, 225 participants are recorded as at risk and 0 as affected; deaths are listed as 0. For other adverse events, 225 participants are recorded as at risk and 0 as affected. No serious adverse-event terms are listed.
These entries should be read as registry reporting, not as a complete characterization of all possible safety outcomes. The registry record does not explain how the 225 participants at risk relate to the 287 enrolled, whether all participants were monitored for the same duration, or what definitions and collection procedures were used. It also does not provide event narratives or a broader description of adverse-event ascertainment. The zero affected counts therefore describe the posted entries; they do not establish that no unrecorded events occurred or that the device has no risks. The study's primary question concerns diagnostic performance, and these safety figures do not answer that question.
12. What the Result Does, and Does Not, Mean
The study is designed to compare ANNE Sleep system diagnoses of moderate to severe obstructive sleep apnea with PSG diagnoses during the same night. Its primary measures are sensitivity and specificity, which describe distinct aspects of how device classifications correspond to the reference classification. The record also states that 287 participants were enrolled and that the study was completed.
However, the registry record does not report the sensitivity or specificity estimates, their 95% confidence intervals, the underlying diagnostic counts, or the analysis population. It is therefore not possible to determine from these registry entries how often the device identified PSG-defined cases or non-cases, or how precise those estimates were. A completed study and a defined outcome do not, on their own, reveal the outcome results.
Even if numerical estimates were available, they would need to be interpreted in light of the reference standard, diagnostic threshold, paired-data methods, and potential missing or uninterpretable results. Those details are not fully reported in the registry record. The record also lists the study as open-label with no masking; it does not explain how readings were interpreted or whether any safeguards were used.
The comparison is not a randomized treatment evaluation. It does not estimate a treatment effect, and its diagnostic endpoint alone would not establish outcomes beyond classification relative to PSG. The appropriate conclusion is limited: the registry describes a study intended to evaluate diagnostic accuracy, but the posted results do not permit a quantitative conclusion about that accuracy. This absence of reported estimates should not be interpreted as evidence for or against the device's performance.
13. Why This Trial Matters Statistically
Diagnostic studies require a different statistical reading from trials that compare treatments. Here, the central question is how classifications from a device correspond to those from PSG, not whether participants assigned to one intervention have better outcomes than participants assigned to another. The registry's sensitivity and specificity definitions make that distinction explicit: one measure concerns reference-defined cases and the other concerns reference-defined non-cases.
The single-group design and concurrent use of the device and PSG are important features of that question. In general, paired observations can support a within-participant comparison, while still requiring clear definitions, complete paired data, and appropriate analysis. The registry record does not provide the analysis details or diagnostic results needed to judge those elements. It does identify the comparison framework and timeframe, which helps clarify what the study intended to evaluate.
The record also illustrates why an outcome definition is not a result. A named endpoint and a confidence-interval format indicate what was intended to be measured, but without the estimates, intervals, and denominators, readers cannot assess magnitude or precision. The completed status and enrollment count do not fill that gap. Transparent reporting of the estimates and analysis population would be needed to understand the observed diagnostic performance.
Finally, the registry identifies limitations relevant to interpretation, including no masking and unreported analysis procedures. These do not prove that the results are biased; they mark questions that cannot be resolved from the registry record. Separating what the design intended to assess from what the posted results actually establish is essential when evaluating diagnostic evidence.
14. Limitations
- The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
- Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
The principal limitation of the registry record is that it does not report the primary diagnostic estimates or their confidence intervals. It also omits the classification counts and the number of participants with usable paired measurements. Without these data, the diagnostic performance and its precision cannot be assessed.
The record does not report the AHI threshold defining moderate to severe obstructive sleep apnea, detailed PSG scoring procedures, or methods for resolving disagreements between the device and PSG. Eligibility criteria and participant characteristics are also not described in detail. These omissions make it difficult to understand the case mix and the exact classification rules used for the primary outcome.
The study is described as open-label and masking is listed as none, but the record does not specify who interpreted each test or whether readings were independent. It does not describe missing-data handling, analysis populations, or the statistical method used for the diagnostic comparison. Although the study is multi-center, the number and locations of centers are not reported.
The adverse-event entries list 225 participants at risk, whereas enrollment is 287; the reason for that difference is not reported. The registry also does not provide the protocol's sample-size justification. These limitations concern the available record and should not be read as evidence about unreported study procedures or results.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Comparative Study of the ANNE™ One System to Diagnose Obstructive Sleep Apnea — NCT04643782.
- Registry record last updated 2023-02-14; status last verified 2023-01.
- Results first posted 2023-02-14.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
NCT04643782 is a completed, single-group, open-label, multi-center diagnostic study sponsored by Northwestern University. It enrolled 287 participants and evaluated the ANNE Sleep system against PSG for diagnosing moderate to severe obstructive sleep apnea. The planned comparison used the methods concurrently during one night.
The primary outcome was sensitivity and specificity, reported in the registry as means with 95% confidence intervals. The registry record defines both measures but does not report their estimates, confidence intervals, diagnostic counts, or analysis entries. It lists 225 participants at risk and 0 affected for serious adverse events, 225 at risk and 0 affected for other adverse events, and 0 deaths. The relationship between the safety denominator and enrollment is not reported.