High Flow Therapy for the Treatment of Respiratory Failure in the ED: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for High Flow Therapy for the Treatment of Respiratory Failure in the ED (NCT02236559), a randomized na study. The primary outcome is Treatment Failure Rate.

2. Clinical Question

The trial concerns a treatment decision in the ED, where respiratory support is initiated for respiratory failure. Its stated objective was to evaluate HFT against NIPPV, rather than to compare either intervention with no respiratory support. The registry describes possible advantages of HFT, including ease of use as a first-line intervention and a shorter time to patient stability. These are study aims, not findings established by the information reported.

The hypothesis proposed clinical non-inferiority for HFT in relation to treatment failure, with attention to ventilation indices and intolerance. It also proposed examining hospital disposition and length of stay. A non-inferiority hypothesis asks whether the outcome with a new approach is not unacceptably worse than the comparator under a prespecified margin. That margin and the results needed to assess it are not reported in the registry record. The record therefore states the question and hypothesis, but does not establish whether the comparison met them.

3. Trial Design

This was a randomized, parallel-group trial with no masking reported. The registry lists an enrollment of 204 and identifies the condition as acute respiratory failure. The study title and description place the treatment setting in the ED. The study is marked completed; its recorded start was 2014-09 and its primary completion was 2016-09.

The two interventions were NIPPV and Vapotherm high-flow therapy. For NIPPV, the record describes fitting an oronasal mask with help from a respiratory therapist or another clinician skilled in NIPPV management. Initial pressures were to be at the low end of the suggested range and could be increased as needed to alleviate respiratory distress. For HFT, a clinician skilled in its management applied a Vapotherm adult nasal cannula. Initial flow was described as 35 L/min, with adjustment as needed to address respiratory distress and optimize treatment.

Random allocation is intended to make groups comparable on average, including for factors that may affect outcomes but are not measured. It does not ensure that the groups will be identical, and the registry entries about how the allocation sequence was generated or concealed are not reported in the registry record. With no masking, participants and clinicians were not reported as blinded to treatment assignment. Whether outcome assessors or analysts were masked is not reported in the registry record.

The primary endpoint was treatment failure within 72 hrs, defined by intubation. The registry also lists eight secondary outcomes labelled ventilatory indices and 11 outcome measures overall. The descriptions and comparative results for the secondary measures are not available in the record information. The registry lists no analysis entries. Details of the analysis populations, protocol deviations, and handling of missing outcome data are not reported in the registry record.

The study was sponsored by Vapotherm, Inc. The registry does not report a statistical analysis plan or a sample-size justification. The design information describes the intended comparison; it does not, by itself, tell whether either treatment performed better on the 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.

Before assignment: 24 patients randomized but not enrolled were excluded for meeting exclusion criteria (10), consent not obtained or withdrawn (6), bedside clinician not comfortable with enrollment after randomization (2), \& patient identified to not need NiPPV after initial evaluation, thus failing to meet inclusion criteria (6). One patient met multiple criteria.

Overall Study

Reasons for not completing this period:

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Treatment Failure Rate, measured over Within 72 hrs.

The primary endpoint was treatment failure within 72 hrs, with failure determined by intubation. This is a binary outcome at the participant level: a participant either meets the stated failure definition during the specified period or does not. The endpoint focuses on escalation to invasive ventilation, rather than directly summarizing all changes in respiratory status.

Endpoints need a clear definition so that outcome status can be assessed consistently across randomized groups. A time window also specifies when the event counts for the primary comparison. Here, the registry identifies the window and the event used to define failure. The clinical reasoning for choosing this endpoint is to assess whether the assigned non-invasive support is followed by a need for intubation within that window. The record describes efficacy in relation to treatment failure but does not report additional adjudication procedures or detailed criteria for intubation.

A binary endpoint is often summarized by the number of participants with the event in each group and a comparison between groups. For a randomized study, the comparison is most informative when accompanied by an estimate of uncertainty, such as a confidence interval. The registry information does not report the primary event counts, an effect estimate, a confidence interval, or a test result. The comparative result for this endpoint is therefore not reported in the registry record.

Intubation is only the endpoint specified for the primary treatment-failure definition. It should not be treated as a complete account of ventilation indices, tolerance, hospital disposition, or length of stay. Those topics are mentioned as study aims or secondary outcomes, but their findings are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies randomized allocation and a parallel-group model. In a parallel trial, participants are assigned to one treatment group and outcomes are compared between groups. Randomization is used to reduce systematic differences in prognostic factors at the start of the comparison. In general, it supports a causal interpretation when the allocation process is implemented appropriately and the analysis follows the trial’s intended comparison. Details of sequence generation, allocation concealment, and the analysis population are not reported in the registry record.

What does randomization contribute?

Randomization balances measured and unmeasured characteristics in expectation, but chance imbalances can remain. It also does not prevent problems after assignment, such as missing outcomes, crossovers, or departures from treatment. Those issues require transparent reporting and an analysis strategy. The record does not report participant flow, treatment switching, or how deviations were handled. It is therefore not possible to assess these aspects of the comparison from the registry information.

What does the primary endpoint measure?

The endpoint is a participant-level event: intubation within 72 hrs, used to define treatment failure. For a binary outcome, a common summary is the event proportion in each group, followed by a between-group contrast. The contrast could be expressed on different scales, each answering a somewhat different question. The scale and primary effect measure are not reported in the registry record. Neither are the event counts required to describe a primary comparison.

General statistical principle: estimates should be interpreted together with their uncertainty. A confidence interval gives a range of values compatible with the data and statistical assumptions at a stated confidence level. A narrow interval usually indicates greater precision than a wide interval, but precision is not the same as freedom from bias. No primary confidence interval is reported in the registry record. A p-value, if available, would address compatibility with a specified null hypothesis under the model; it would not measure clinical importance or the probability that a hypothesis is true. No primary p-value is reported in the registry record.

How is non-inferiority assessed?

General statistical principle: non-inferiority trials test whether the intervention is not worse than a comparator by more than a prespecified acceptable margin. The margin must be chosen before examining the results and justified in the context of the outcome. The confidence interval for the treatment contrast is then compared with the margin in the direction that represents worse outcomes. A non-significant superiority test alone does not establish non-inferiority. The registry states a non-inferiority hypothesis, but the margin, the primary contrast, its uncertainty, and a non-inferiority analysis are not reported in the registry record.

The registry’s wording refers to an “at least equivalent” objective and a hypothesis of clinical non-inferiority. Equivalence and non-inferiority are related but distinct statistical objectives. In general, equivalence requires ruling out differences in either direction beyond prespecified limits, whereas non-inferiority focuses on excluding an unacceptable disadvantage. The exact statistical objective and its operational definition are not reported in the registry record, so the wording alone cannot establish which formal decision rule was used.

What does the analysis record show?

The registry lists no analyses. It does not report the statistical model, effect measure, covariate adjustment, significance threshold, confidence-interval method, or method for handling missing data. The record also does not report whether an interim analysis was planned or conducted. These details should not be inferred from the randomized design or from the endpoint definition.

General statistical principle: adjustment can account for prespecified prognostic variables and may improve precision, but an adjusted analysis answers a conditional question and depends on its model. An unadjusted comparison is more directly tied to the randomized group assignment but may be less precise. The record does not identify whether the primary comparison was adjusted or unadjusted. No recalculated primary comparison is available from the reported details, and no sponsor analysis can be contrasted with a recalculation.

Why are multiple outcomes relevant?

The registry lists 11 outcome measures and eight secondary outcomes labelled ventilatory indices. When many outcomes or repeated measurements are tested, the chance of at least one apparently notable result can increase if each is judged separately. General methods for addressing multiplicity include defining a hierarchy, adjusting significance thresholds, or treating secondary findings as exploratory. Which approach, if any, was planned is not reported in the registry record. The outcome labels alone do not reveal the number of statistical tests or their dependence.

The registry also describes planning illustrations based on the observed effect, with equal allocation, a two-sided significance level of 0.05, and a normal approximation. Such illustrations are planning calculations, not observed power and not evidence that the original trial was designed around those assumptions. The stated calculation uses a fraction of the observed effect as the effect to detect. The registry’s sample-size justification and the exact planning results are not reported in the registry record. A planning illustration cannot replace an analysis of the observed outcomes or resolve uncertainty in the treatment estimate.

Why does dispersion matter in general?

Dispersion describes variability in an outcome. For event counts or repeated measurements, variability greater than a model expects can make standard errors too small if the model does not account for it. That can produce confidence intervals that are too narrow and tests that appear more decisive than warranted. This trial’s primary endpoint is described as a participant count, not as a continuous measurement. No dispersion parameter or model-based count analysis is reported in the registry record. General discussion of over-dispersion therefore does not imply that it was present or relevant to the trial’s actual analysis.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures, including the primary treatment-failure endpoint and eight secondary outcomes labelled ventilatory indices. It also identifies study interests in intolerance, hospital disposition, and length of stay. A collection of outcomes can support a broader clinical description, but it creates a statistical interpretation issue if many results are tested and selectively emphasized. General statistical principle: when multiple hypotheses are tested, the chance of at least one apparently unusual result can rise. Prespecifying a primary outcome and a testing strategy helps readers distinguish confirmatory from exploratory findings.

The record does not report a multiplicity strategy, a hierarchy of outcomes, or how secondary analyses were to be interpreted. It does not report whether repeated measurements of ventilatory indices were tested separately or handled in a joint model. These methods are not reported in the registry record, so the outcome list alone cannot establish how much multiplicity was addressed.

Missing-data methods are also not reported in the registry record. The record does not describe missing primary outcomes, withdrawals, exclusions after randomization, or whether all randomized participants were included in the primary analysis. If outcome status is missing, the amount and reasons for missingness can affect how a group comparison should be interpreted. The registry entries available do not permit an assessment of that issue.

The analysis population is not reported in the registry record. In general, an intention-to-treat analysis retains participants in their randomized groups, while a per-protocol analysis applies specified criteria related to adherence and eligibility. Non-inferiority assessments often consider how choices of analysis population may affect interpretation, but no such analysis details are reported. Interim monitoring and any stopping rules are also not reported in the registry record.

11. Safety Results

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

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

Serious events with the highest arm-level rate: death outside study window (2.0%).

The safety entries record serious adverse events in the NIPPV group for 2 affected participants among 100 at risk, and in the high velocity nasal insufflation group for 0 affected participants among 104 at risk. Other adverse events are recorded as 0 affected participants in each group, with the same at-risk counts. These are counts from the registry’s safety record, not calculated rates or a formal comparison.

The registry lists “Death Outside Study Window” as the top serious-event term at 2.0%. The term’s wording matters: it identifies an event outside the study window, not an event within the primary endpoint’s 72 hrs. The registry does not report arm-specific death counts, and the death-count field is not populated. Whether any specific serious event occurred in each treatment group is not given by the top-term entry alone.

Zero recorded events in a group does not establish that the underlying risk is zero. Event counts can be affected by the size of the group, the observation period, event definitions, and reporting completeness. The registry information does not provide confidence intervals, detailed event narratives, or a formal safety analysis. These entries should therefore be read as descriptive safety counts, not as proof of a treatment-related difference or as a complete account of harms.

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

The record is useful for identifying the trial’s intended comparison: Vapotherm high-flow therapy versus NIPPV for respiratory failure in the ED, with intubation within 72 hrs defining the primary treatment-failure endpoint. It also records that the study was randomized, used a parallel-group design, and had no masking reported. These details clarify the question the trial was designed to address.

The record does not report the primary event counts, treatment contrast, confidence interval, or test result. It therefore does not show whether treatment failure was more or less frequent in either group, how large any difference was, or how precise an estimate might be. The stated non-inferiority hypothesis is not a result. Without a reported margin and the relevant estimate and uncertainty, it cannot be judged as met or unmet.

The safety entries give limited descriptive information: serious adverse events were recorded for 2 participants in the NIPPV group and 0 in the high velocity nasal insufflation group. Other adverse events were recorded for 0 participants in each group. These counts do not on their own establish comparative safety, and the serious-event term reported as “Death Outside Study Window” should not be interpreted as a primary-window outcome. Detailed event information and arm-specific death counts are not reported in the registry record.

Nor does the record establish the study’s findings for ventilation indices, intolerance, hospital disposition, or length of stay. Those topics appear as aims or outcome areas, but comparative results are not reported in the registry record. The registry information supports a description of the design, endpoint, and selected safety counts; it does not support a conclusion that either treatment is superior, equivalent, non-inferior, or safer.

In general, statistical evidence needs both an estimate and an account of its uncertainty, alongside transparent methods and outcome reporting. Those elements are not available for the primary comparison in this record. The appropriate interpretation is therefore limited: the trial posed a comparative question, but the reported information does not provide its efficacy answer.

13. Why This Trial Matters Statistically

This trial raises an important statistical distinction between a study hypothesis and evidence that answers it. The registry describes an objective of evaluating whether HFT is at least equivalent to NIPPV and a hypothesis of non-inferiority. Those objectives require a clear decision rule, a prespecified margin, and an estimate with sufficient precision. Naming the objective does not establish that the evidence supports it.

The endpoint also illustrates the importance of matching a statistical result to its clinical definition. Treatment failure is tied to intubation within 72 hrs, while the broader study aims mention ventilation indices, tolerance, disposition, and length of stay. These outcomes represent different aspects of care and should not be treated as interchangeable. A result for one would not automatically answer questions about the others.

Randomization is a strong design feature for comparing assigned groups, but the design alone cannot communicate the result. Readers need outcome counts, a clearly defined contrast, uncertainty estimates, and enough detail to understand who was analyzed and how missing outcomes were handled. Such primary analysis information is not reported in the registry record. The gap between a well-defined research question and an interpretable result is especially visible when a registry record includes design and hypothesis details but no comparative endpoint estimate.

The listed secondary outcomes also make multiplicity relevant. If multiple indices and clinical outcomes are examined, a prespecified approach helps distinguish a primary confirmatory finding from exploratory observations. The record does not report such an approach. This does not show that the trial’s investigators failed to plan one; it means the approach is not reported in the registry record.

Finally, the safety counts demonstrate why descriptive numbers should be kept separate from efficacy conclusions. They provide limited information about recorded adverse events, but without fuller descriptions and uncertainty they cannot settle comparative safety. Statistical interpretation depends not only on which values are reported, but also on what those values measure and what information is absent.

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 information does not provide participant characteristics or detailed eligibility criteria. The population is identified as people with acute respiratory failure in the ED, but the record does not describe how respiratory failure was defined, which patients were eligible, or the clinical features of those enrolled. Generalizability to other settings or populations therefore cannot be assessed from these registry entries.

The primary endpoint is defined as intubation within 72 hrs, but further details about endpoint assessment and decision-making are not reported in the registry record. The record does not describe whether intubation decisions followed standardized criteria, whether outcome assessment was independently adjudicated, or how events outside the stated time window were treated in the efficacy analysis.

The intervention descriptions include adjustable support settings, but the full treatment protocol, adherence, crossover, and co-interventions are not reported in the registry record. Masking is listed as none; whether outcome assessors or data analysts were masked is not reported. Sequence generation and allocation concealment are also not reported in the registry record.

Most importantly, the registry information does not include primary outcome counts or a comparative estimate. It also does not report confidence intervals, a statistical analysis plan, sample-size justification, missing-data handling, or analysis populations. Without these details, neither the primary effect nor its precision can be evaluated. The record lists secondary outcome areas but does not report their comparative findings or a multiplicity strategy.

The safety information is limited to selected counts and a serious-event term. It does not provide full event descriptions, arm-specific death counts, or uncertainty estimates. These limitations restrict interpretation of both benefits and harms; they do not establish that unreported procedures or analyses were absent from the trial itself.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: High Flow Therapy for the Treatment of Respiratory Failure in the ED — NCT02236559.
• Registry record last updated 2019-05-23; status last verified 2019-04.
• Results first posted 2019-05-23.
• 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

NCT02236559 is a completed randomized, parallel-group trial comparing Vapotherm high-flow therapy with noninvasive positive pressure ventilation for acute respiratory failure in the ED. The registry lists enrollment of 204, no masking, a study start of 2014-09, and primary completion in 2016-09. Its primary endpoint was treatment failure within 72 hrs, defined by intubation.

The stated objective included evaluating non-inferiority and potential advantages of high-flow therapy. The record lists 11 outcome measures, including eight secondary outcomes labelled ventilatory indices. Primary comparative results, analysis details, and the non-inferiority margin are not reported in the registry record. Safety entries record serious adverse events for 2 participants in the NIPPV group and 0 in the high velocity nasal insufflation group; other adverse events are recorded as 0 in both groups. These descriptive entries do not establish comparative efficacy or safety.