AirFLO2 Treatment for Hypoxia and/or Tachypnea in Patients With COVID-19: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for AirFLO2 Treatment for Hypoxia and/or Tachypnea in Patients With COVID-19 (NCT04649775), a randomized na study. The primary outcome is Oxygenation as Measured by P:F Ratio (PaO2:FiO2).

2. Clinical Question

The study was designed to compare a device intervention with standard care in a hospitalized population identified by respiratory rate and oxygenation criteria. The P:F ratio is a continuous measure, with higher values described in the record as indicating better oxygenation. The question is therefore about measured oxygenation over the stated time points, not about every possible clinical consequence of device use.

The registry also lists secondary questions concerning dyspnea, cough and respiratory symptoms, transfer to higher levels of respiratory support or intensive care, intubation, length of hospitalization, and survival to discharge. These are distinct outcomes and should not be treated as interchangeable evidence about oxygenation. The record lists those outcomes but does not report analyses for them.

3. Trial Design

This was a randomized, parallel-group, unblinded controlled trial sponsored by Duke University. The registry lists the study phase as NA and enrollment as 11. The study start date is 2021-04-09, the primary completion date is 2021-11-08, and the current status in the record is terminated. The reason for termination is not reported in the registry record.

Patients in the intervention group were to receive the AirFLO2 device and training by video, with coordinator reinforcement on device use and self-application. Clinical data were to be recorded before and after the intervention. Questionnaires were planned at baseline and at serial time points after intervention, although the remaining detail of that description is not reported in the registry record. The control group was to receive standard of care, with clinical data recorded and questionnaires performed at baseline and at the end of the study.

Random allocation describes how participants were assigned in principle; it does not establish that the groups were balanced on every characteristic, particularly in a small trial. The record does not report the randomization sequence, allocation concealment, or stratification methods. It does report that masking was none, so the trial was unblinded. Further details about who knew treatment assignments, or how that knowledge was addressed during outcome measurement and analysis, are not reported in the registry record.

The registry identifies oxygenation measured by the P:F ratio as the primary outcome and gives baseline, 24 hours, and 48 hours as its time frame. It notes that 48-hour data were not collected for two participants in the standard-of-care control arm. Although the record lists nine secondary outcomes, it lists zero analyses. No numerical primary or secondary outcome results are reported.

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

Analysis population for the primary outcome, as defined in the registry: Data not collected at 48 hours on two participants in the Standard of Care - Control arm.

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 Oxygenation as Measured by P:F Ratio (PaO2:FiO2), measured over Baseline, 24 hours, 48 hours.

The primary endpoint is oxygenation measured by the P:F ratio, also written as PaO2:FiO2. It is a continuous measure, and the registry describes the reported summary as a mean with standard deviation. The specified time points are baseline, 24 hours, and 48 hours. The record states that a higher P:F ratio indicates better oxygenation and gives a range of 20 to 500.

Using a continuous measure can retain gradations in oxygenation rather than reducing observations to a simple category. A mean describes the central value across participants, while a standard deviation describes how spread out individual measurements are around that mean. These summaries can help readers understand both the group-level measurement and participant variability, but they do not by themselves show whether groups differ or how precisely a difference is estimated.

Repeated measurements make it possible to describe oxygenation at more than one point in time. Interpretation depends on having the relevant observations and on a defined comparison, such as a between-group contrast or a change from baseline. The registry does not report P:F ratio means, standard deviations, a contrast, or an inferential result. It therefore describes the endpoint but does not establish what the observed oxygenation measurements were or how they compared between groups.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary measure represent?

The P:F ratio is the registered continuous primary measure, expressed as PaO2:FiO2. The record specifies a mean and standard deviation as its summary and says that higher values indicate better oxygenation. A mean condenses a set of measurements into a central value. A standard deviation describes the spread of the measurements around that mean. Neither statistic alone describes the uncertainty in a comparison between groups. The registry gives a range of 20 to 500, but it does not report observed values or summaries at any of the specified time points.

How would a group comparison be interpreted?

For a randomized two-arm study, a between-group estimate describes how far apart the groups are on an outcome, according to the chosen statistical definition. An estimate should be read alongside its confidence interval. In general, a confidence interval represents a range of values compatible with the data and the method used; a narrower interval indicates greater precision, while a wider interval indicates greater uncertainty. A p-value, in general, measures how surprising data at least as inconsistent with a specified null hypothesis would be under the statistical model. It does not measure the probability that the null hypothesis is true, nor does it communicate the magnitude or practical importance of an effect. No estimate, interval, or p-value for the primary outcome is reported in the registry record.

Why do the time points matter?

The primary outcome is specified at baseline, 24 hours, and 48 hours. Repeated measurements can be used to describe the course of an outcome over time, but an analysis needs to state what is being compared: values at a particular time, change from baseline, or a pattern across time. These approaches answer related but not identical questions. A model for repeated observations may account for the fact that measurements from the same participant are related. The registry record does not report the planned statistical model, the contrast used, or whether the analysis would compare follow-up values, changes, or trajectories.

The record notes that 48-hour measurements were not collected for two participants in the standard-of-care control arm. Missing observations matter because the available measurements may not fully represent the randomized groups. The consequences depend on why measurements are missing and on the analysis assumptions. For example, complete-case analysis uses only participants with the required observations; other approaches may use a model or imputation. This is general statistical context, not a description of what was done in this trial. The missingness reason and handling method are not reported in the registry record.

How should the design information be used?

Random allocation is intended to make treatment groups comparable on average with respect to factors that could affect outcomes. In a study with enrollment of 11, chance differences between groups can remain, and the record does not provide baseline characteristics or a balance assessment. The parallel design means participants were assigned to one group rather than described as crossing between interventions. The trial was unblinded; the record does not report methods to limit possible effects of knowledge of assignment on measurement, participant responses, or analysis.

General statistical reasoning also distinguishes the analysis population from the number enrolled. An intention-to-treat analysis generally retains participants in their randomized groups, while a per-protocol analysis uses a defined set of participants meeting protocol criteria. The registry record does not report an analysis population or identify which participants were included in any outcome analysis. With zero analyses listed, no population-specific estimate can be assessed.

What do the planning assumptions mean?

The registry describes a sample-size planning illustration based on the observed effect, using a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is described as a stated fraction of the observed effect. For outcomes other than time-to-event outcomes, the illustration is expressed as participants per arm; time-to-event illustrations use the total number of events. These are planning assumptions, not evidence about the trial's actual design or conduct. The record does not report the trial's own sample-size justification, and a planning illustration is not observed power. Because the registry does not report the primary effect estimate or its variability, no numerical planning result or precision assessment can be interpreted from the record.

More generally, power calculations depend on assumptions about the effect size, variability, allocation, significance threshold, and analysis. A calculation can describe the sample size needed under those assumptions, but it cannot establish whether an observed effect is true or clinically meaningful. Power is a design concept and should not be used after the fact as a substitute for an effect estimate and confidence interval. These general principles do not fill in the missing trial-specific analysis details.

10. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is measured at baseline, 24 hours, and 48 hours, and the registry lists nine secondary outcomes. Considering multiple outcomes or multiple time points can raise a multiplicity question: if many statistical tests are conducted, the chance of at least one apparently notable result may increase. In general, a protocol may identify a primary comparison in advance, specify a hierarchy, or adjust inference for multiple tests. The registry record does not report a multiplicity strategy, planned contrasts, or whether any adjustment was intended. It lists zero analyses, so there are no reported inferential results to evaluate for multiplicity.

The record states that 48-hour data were not collected on two participants in the standard-of-care control arm. It does not report why these measurements were absent, whether other outcome data were missing, or how missing observations were handled. It also does not report whether all enrolled participants contributed to the primary outcome analysis. The analysis population, any exclusions, and any imputation or sensitivity analysis are not reported in the registry record.

These omissions limit interpretation in different ways. Missing-data methods affect which observations contribute and what assumptions support an estimate. Analysis-population definitions determine which randomized participants are represented in the estimate. Multiplicity procedures affect how evidence is interpreted across several tests. Without the method and the results, it is not possible to assess these features or determine whether the listed outcomes were analyzed as planned.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

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

The adverse-event figures list five participants at risk in the AirFLO2 Device Intervention group and six in the Standard of Care - Control group. The record reports zero serious adverse events, zero other adverse events, and zero deaths in each group. These are counts recorded for the participants listed as at risk; they are not rates or comparative estimates.

Zero reported events in a small group does not establish that an event cannot occur or that the groups have equivalent safety. The precision of an event-rate estimate depends on the number observed and the number at risk, and a zero count can still be compatible with uncertainty about the underlying event frequency. The record does not report follow-up duration for safety assessment, definitions used for adverse-event categories, or a detailed event-by-event account. It lists no top serious-event terms.

These safety entries should be read as registry reporting, not as a comprehensive safety evaluation. They do not resolve whether all relevant events were captured or whether the data were assessed using a prespecified comparative method. The figures alone cannot establish a causal safety difference between the randomized groups.

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

The record documents a randomized, parallel-group, unblinded trial of AirFLO2 versus standard of care among patients admitted with COVID-19 and tachypnea and/or hypoxia. It identifies oxygenation measured by the P:F ratio as the primary outcome and specifies baseline, 24 hours, and 48 hours as measurement times. It also describes the intervention and control procedures, lists secondary outcomes, and records adverse-event counts.

Those design details are not a report of treatment effect. The registry contains no numerical primary outcome values, no estimated between-group comparison, and no statistical analysis results. It therefore does not show whether oxygenation differed between groups, how large any difference was, or how precisely any difference was estimated. The higher-is-better direction for the P:F ratio describes the measure, not an observed result.

The record also does not report analyses for the listed secondary outcomes. The existence of a registered outcome does not by itself show that complete data were collected or that an analysis was performed. Similarly, the adverse-event entries of zero in both groups should be interpreted as reported counts for the listed participants at risk, not as proof of equal risk or absence of all possible harms.

The trial status is terminated, but the reason is not reported in the registry record. Enrollment was 11, so the study is small; without outcome estimates, the degree of statistical uncertainty cannot be quantified. The record does not report the statistical analysis plan, sample-size justification, or missing-data method. The central limitation is thus not a result suggesting benefit or lack of benefit, but the absence of posted comparative outcome analyses from which such a conclusion could be evaluated.

In general, a treatment comparison is best interpreted using an effect estimate, its confidence interval, and a clear account of the population and analysis. These elements allow readers to distinguish the direction and size of an estimate from its precision. They are not reported for the primary outcome in this registry record, so no clinical or statistical conclusion about the comparative effect is warranted from the posted results.

13. Why This Trial Matters Statistically

This trial is statistically notable because the record contains a randomized design and a defined continuous primary outcome, yet reports no outcome analyses. That contrast illustrates the distinction between registering a question and presenting evidence that answers it. Design details can clarify what investigators intended to compare, but without group summaries and an estimate of the comparison, readers cannot judge the magnitude or precision of a result.

The endpoint has repeated time points, and the record identifies missing 48-hour data for two control participants. This makes transparent reporting of the analysis especially important. Readers need to know whether the comparison uses a particular follow-up time, change from baseline, or a model of measurements over time, and how missing observations are addressed. Those methods are not reported in the registry record.

The trial also lists several secondary outcomes. When a study considers multiple outcomes, clear specification of the primary comparison and any multiplicity approach helps readers understand which findings were confirmatory and how to interpret multiple tests. The registry does not describe such a plan, and it lists zero analyses. This means the registered secondary questions cannot be treated as reported evidence.

Finally, the small enrollment and the absence of estimates emphasize why statistical interpretation should focus on uncertainty rather than on a binary reading of event counts or study status. The registry's safety counts are descriptive, and termination does not explain the reason for the study's status. Together, these details make the record useful for understanding the study question and reporting gaps, but not for estimating comparative effects.

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 registry record offers limited information for reconstructing the trial's statistical methods. It does not report the randomization sequence, allocation concealment, baseline characteristics, or the statistical analysis plan. Although allocation is described as randomized and masking as none, details about implementation and outcome assessment are not reported in the registry record. The reason for termination is also not reported.

The primary outcome is defined, but numerical observations and a group comparison are absent. The record lists zero analyses, so it is not possible to assess the direction or size of an effect, its precision, or whether a statistical test was conducted. The note about missing 48-hour data in the control arm identifies an incomplete measurement point, but does not explain why observations were absent or how they were handled.

Enrollment was 11, and the listed at-risk counts for adverse events are five in the AirFLO2 group and six in the standard-of-care group. Such a small study provides limited information about uncommon outcomes, and the record does not give follow-up duration or detailed event definitions. Zero recorded events should not be interpreted as proof of no risk.

The registry lists nine secondary outcomes but does not report their results or analyses. It also does not report the planned analysis population, a missing-data strategy, multiplicity handling, or a trial-specific sample-size justification. The planning assumptions described in the record do not substitute for the study's own justification or for observed outcome estimates. These limitations restrict the record to describing the design and intended measures rather than supporting a statistical assessment of comparative outcomes.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: AirFLO2 Treatment for Hypoxia and/or Tachypnea in Patients With COVID-19 — NCT04649775.
• Registry record last updated 2022-11-25; status last verified 2022-04.
• Results first posted 2022-11-25.
• 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.

18. Record Summary

NCT04649775 is a Duke University-sponsored, randomized, unblinded, parallel-group trial of AirFLO2 versus standard of care for patients admitted to Duke Hospital with COVID-19 and tachypnea and/or hypoxia. The registry lists enrollment of 11, study phase NA, a start date of 2021-04-09, a primary completion date of 2021-11-08, and terminated status.

The primary outcome is mean oxygenation measured by the P:F ratio at baseline, 24 hours, and 48 hours. The record notes that 48-hour data were not collected for two participants in the control arm. Nine secondary outcomes are listed, while the registry lists zero analyses. No numerical primary or secondary outcome comparisons are reported. The reason for termination, analysis methods, and missing-data handling are not reported in the registry record.