Comparison of CPAP Machines With Reusable vs Disposable Circuits: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparison of CPAP Machines With Reusable vs Disposable Circuits (NCT03121612), a randomized na study. The primary outcome is FiO2.

2. Clinical Question

The study also planned to assess physiological measurements and monitor side effects and adverse events. Its description identifies the central statistical question as a comparison between two device categories, rather than an evaluation against no CPAP. The record does not state a formal equivalence or non-inferiority analysis, nor does it explain how the stated threshold of 0.63 standard deviations would be used in a decision rule.

3. Trial Design

The registry describes a randomized, parallel-group trial with no masking. It lists the study as terminated and gives the enrollment as 51. The record lists the start date as 2017-08-01 and the primary completion date as 2018-08-15. A reason for termination is not reported in the registry record.

The two groups were defined by the CPAP machine and circuit type. The Dolphin machine is described as having reusable circuits, an oxygen intake, an on-board air compressor, and an integrated blender and humidifier. It delivered heated, humidified blended gases through a Fisher-Paykel nasal mask. Although the machine had a built-in pulse oximeter, that device was not used in the study. The comparator was a Fisher-Paykel CPAP machine with disposable circuits, separate oxygen and medical-air intakes, and blender equipment; it also delivered heated, humidified blended gases through a Fisher-Paykel nasal mask.

The primary endpoint was FiO2, expressed as a change from baseline and measured at several time points through 72 hours after treatment commenced. The registry specifies median and inter-quartile range as the summary format. It also lists secondary measurements including SpO2, respiratory rate, arterial pH, PaO2 and PaCO2, as well as outcomes related to CPAP failure, intubation, mechanical ventilation and surfactant use.

The registry does not report the randomization method, allocation concealment, eligibility criteria, recruitment setting, or the reason for choosing the planned enrollment. It also does not describe masking of outcome assessors. Those details should not be inferred from the randomized, parallel-group label or from the trial’s termination status.

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: There were no wash-out or run-in procedures. No enrolled participant was excluded from the study before randomisation.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Number missing at each time point: 6 hrs: 1 Dolphin (failed), 0 FP 12 hrs: 1 Dolphin (failed), 1 FP (recovered) 24 hrs: 6 Dolphin (2 failed + 4 recovered), 4 FP (all recovered) 48 hrs: 16 Dolphin (3 failed + 13 recovered), 11 FP (1 failed + 10 recovered) 72 hrs: 21 Dolphin (3 failed + 18 recovered), 21 FP (1 failed + 20 recovered)

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 FiO2, measured over 6, 12, 24, 48 and 72 hours after treatment commencement.

The primary endpoint was fraction of inspired oxygen, or FiO2, measured as change from baseline at 6, 12, 24, 48 and 72 hours after treatment commenced. FiO2 represents the oxygen concentration in the gas mixture delivered by the CPAP machine. The registry describes the outcome as the value shown on the machines and specifies that results were to be summarized using the median and inter-quartile range.

Tracking FiO2 over time can describe how much oxygen is being delivered while infants receive CPAP. It is a physiological measurement rather than a direct measure of survival or respiratory support failure. The record also lists SpO2 and other physiological and clinical outcomes, so FiO2 was one part of a broader assessment.

The registry flags an important interpretive problem: the intended SpO2 target was 90-95%, and FiO2 should be reduced when SpO2 exceeds that target. The majority of readings were at SpO2 > 95%. For those readings, the recorded FiO2 therefore reflects oxygen delivered, not the amount required to achieve the target range; the record says the delivered amount was excessive by an unknown amount. This contamination limits the clinical meaning of the FiO2 measurements and could complicate comparisons between machines.

Comparative medians, inter-quartile ranges, and changes from baseline are not reported in the registry record. The direction, size and precision of any between-group difference therefore cannot be described from the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Randomization is intended, in general, to make treatment groups comparable at the start of a study by assigning participants through a chance process. It can reduce systematic differences in measured and unmeasured characteristics, but it does not guarantee balance in a particular trial. The registry identifies this trial as randomized, but the random sequence method and allocation concealment are not reported in the registry record. The parallel-group design means the record describes separate groups receiving the two machine types, rather than a crossover comparison in which each participant receives both.

What does a median and inter-quartile range describe?

In general, the median is the middle of an ordered set of observations. The inter-quartile range describes the spread between the lower and upper quartiles. These summaries can be useful when measurements are skewed or include outlying values, because they do not rely on the mean and standard deviation in the same way. The registry specifies median and inter-quartile range for FiO2, but the actual summaries by group are not reported in the registry record. Without those values, neither the central tendency nor the spread of FiO2 can be compared.

Why does repeated measurement matter?

FiO2 was measured at 6, 12, 24, 48 and 72 hours after treatment commenced. In general, repeated observations on the same participant are related to one another, so they should not automatically be treated as independent measurements. A statistical model may account for that within-participant relationship and may estimate changes over time or a group-by-time pattern. The registry record does not report whether such a model was used, which time point was designated for a formal comparison, or how the multiple time points were handled in the primary analysis.

Repeated measurements also raise a question about how to interpret an overall result. A group difference at one time point can differ from the pattern across the full period. A single summary may conceal changes in the direction or size of a difference over time. The record does not report the group-level time profiles or a model-based estimate, so no such pattern can be assessed.

What is the role of a confidence interval?

In general, a confidence interval gives a range of values compatible with the data and the statistical model used to calculate it. It communicates precision as well as the estimated direction and size of an effect. A narrow interval indicates greater precision than a wide interval, but its interpretation depends on the design and analysis. No confidence interval for the primary comparison is reported in the registry record. There is therefore no basis for describing how precisely the trial estimated a FiO2 difference.

What does the stated threshold mean?

The registry states a null hypothesis that infants treated on the two machine categories would not differ in key physiological parameters by more than 0.63 standard deviations. In general, a difference expressed in standard-deviation units places a result on a scale that reflects outcome variability. Such a threshold can be part of a prespecified hypothesis, but its role depends on the statistical framework and analysis plan. The registry does not report the framework, the calculations behind the threshold, or a test result. It should not be interpreted as proof that differences were within that boundary.

How should planning assumptions be read?

The record describes a planning illustration based on an observed effect, using a two-sided significance level of 0.05, equal allocation between groups and a normal approximation. The effect to detect is described as a stated fraction of the observed effect. For time-to-event tables, the planning approach concerns the total number of events; other planning outputs concern participants per arm. These are planning assumptions, not evidence of the study’s achieved power. The record does not report the numerical planning outputs or a sample-size justification used by the trial itself. In general, a sample-size illustration is conditional on its assumptions and should not be confused with the information actually observed in a trial.

Why can over-dispersion matter?

In general, over-dispersion occurs when observed variability is greater than a model assumes. If a model underestimates variability, standard errors may be too small and confidence intervals too narrow. The record does not identify an over-dispersion analysis for this trial, and there is no basis to state that over-dispersion affected its results. More broadly, uncertainty estimates depend on whether the chosen statistical model fits the outcome and study structure.

The registry lists 19 outcome measures and 12 analyses. The listing does not establish which analyses were prespecified, which results were included in any report, or whether adjustments for multiple comparisons were made. Those details are not reported in the registry record. Statistical interpretation should therefore distinguish the stated design and measurement plan from analysis choices and results that the record does not describe.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 19 outcome measures and 12 analyses. It identifies multiple primary FiO2 time points and also lists secondary physiological and clinical outcomes. When several outcomes or time points are examined, the chance of finding an apparently notable result by chance can increase. In general, multiplicity procedures or a clearly prespecified hierarchy can help frame such comparisons. Whether this trial used an adjustment, a hierarchy, or another approach is not reported in the registry record.

Missing primary-outcome observations are described at several time points, with reasons that include treatment failure and recovery. At 6 hours, the record lists 1 missing Dolphin observation due to failure and 0 missing Fisher-Paykel observations. At 12 hours, it lists 1 Dolphin observation due to failure and 1 Fisher-Paykel observation due to recovery. At 24 hours, it lists 6 Dolphin observations, including 2 due to failure and 4 due to recovery, and 4 Fisher-Paykel observations, all due to recovery. At 48 hours, it lists 16 Dolphin observations, including 3 due to failure and 13 due to recovery, and 11 Fisher-Paykel observations, including 1 due to failure and 10 due to recovery.

In general, missingness can affect interpretation if the probability of an observation being absent is related to the outcome or differs between groups. Missingness due to failure or recovery may be especially relevant to how a repeated physiological measurement is summarized, but the registry does not report the method used to handle missing observations. The available record also does not report an analysis population definition, imputation method, sensitivity analysis, or participant flow by analysis. These details should not be inferred from the missing-observation counts.

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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: death (3.8%).

The registry lists serious adverse events affecting 1 participant in the Dolphin CPAP group, with 26 at risk, and 0 participants in the Fisher-Paykel CPAP group, with 25 at risk. It lists other adverse events affecting 4 participants in the Dolphin group and 2 in the Fisher-Paykel group. Deaths are listed as 1 and 0, respectively. The serious adverse-event terms include Death, listed at 3.8%.

These figures describe the safety information recorded for the two groups. They do not, by themselves, show whether an event was caused by a device or whether the groups differed reliably in safety. The registry does not report event definitions, the relationship of events to the machines, event timing, statistical comparisons, or uncertainty estimates. A count of participants affected is also not necessarily the same as a count of all event episodes; the record does not clarify that distinction.

The at-risk counts are the denominators stated for the adverse-event groups. They should not be treated as proof of the number analyzed for every other outcome. In general, safety counts are most informative when read alongside definitions, follow-up duration and completeness, and an account of causality assessment. Those details are not reported in the registry record, so the figures should be interpreted descriptively rather than as a causal comparison.

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

The registry record establishes that the trial was designed as a randomized, parallel comparison of CPAP machines with reusable and disposable circuits in newborns with respiratory distress syndrome. It identifies FiO2 change from baseline, measured at several time points, as the primary outcome and lists additional physiological and clinical outcomes. It also records adverse-event counts and notes that the study was terminated.

What the record does not establish is the result of the primary comparison. Group-specific FiO2 summaries, an estimated difference, a confidence interval and a p-value are not reported in the registry record. No conclusion can therefore be drawn from this record about whether the measured FiO2 differed between the machine groups, how large any difference was, or how precisely it was estimated.

The stated null hypothesis refers to a threshold of 0.63 standard deviations, but the record does not describe a statistical framework or result that would allow that threshold to be evaluated. The trial cannot be interpreted as having established equivalence, non-inferiority, or a lack of difference on the basis of the stated hypothesis alone.

The FiO2 outcome also has a documented interpretive limitation: the majority of readings were at SpO2 > 95%, above the stated target of 90-95%. In those circumstances the recorded FiO2 reflects oxygen delivered rather than the amount required to meet the target, and the amount of excess is unknown. This should be kept in mind even when considering any future reported comparison.

The adverse-event counts are descriptive, and the registry does not establish causation or provide comparative uncertainty. The termination status does not reveal why the study ended, and that reason is not reported in the registry record. In general, absence of reported results is not evidence that the groups were the same or that one was better. The defensible reading is that the registry describes a planned comparison and some safety counts, while leaving the main comparative findings and important analysis details unreported.

13. Why This Trial Matters Statistically

This trial raises several statistical issues that are common in clinical device comparisons. Randomization defines the intended basis for comparing groups, but interpretation also depends on how allocation was implemented, how outcomes were measured, and which participants contributed to each analysis. The registry confirms randomization and a parallel-group structure, while leaving several implementation details and the analysis population unreported.

The primary endpoint illustrates the distinction between a measurement and the clinical quantity it is intended to represent. FiO2 is recorded from the machines, but the registry notes that readings were often associated with SpO2 above the stated target. In those circumstances, delivered oxygen does not identify the oxygen amount required to reach the target, and the excess is unknown. This makes interpretation dependent on context, not just on the numerical value of FiO2.

The repeated time points create another important issue. A comparison over time may depend on the treatment of within-participant correlation, missing observations and the choice of summary or model. The record names several measurement times but does not report the statistical approach. This leaves uncertainty about how a single primary conclusion would be drawn from a time-varying outcome.

The registry also presents a stated threshold in standard-deviation units and a planning framework based on an observed effect. These ideas are not interchangeable: a planning illustration describes assumptions for a hypothetical design, while an analysis result requires observed data and a specified model. The record does not report a numerical planning output or the primary comparative estimate, so neither should be treated as evidence of observed precision or power.

Finally, the study has multiple listed outcomes and analyses, and a nontrivial description of missing measurements. Clear reporting of prespecification, multiplicity handling, missing-data methods and uncertainty would be needed to judge the strength of any eventual result. The registry record makes these reporting questions visible without answering them.

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 most consequential limitation of the registry record for statistical interpretation is that it does not report the primary comparative results. Group-specific FiO2 summaries, a treatment-effect estimate, uncertainty intervals and a p-value are absent. Consequently, the record cannot establish the direction, size or precision of a difference between machines.

The primary outcome is explicitly described as contaminated. The majority of readings were at SpO2 > 95%, while the stated target was 90-95%. For these readings, the FiO2 value indicates oxygen delivered rather than oxygen required to achieve the target, with the excess amount unknown. This complicates the meaning of the endpoint and potentially the interpretation of changes from baseline.

FiO2 was measured repeatedly, and the registry identifies multiple time points without reporting how they were analyzed. It does not describe a model for within-participant correlation, the handling of time-point comparisons, or a prespecified summary across time. Missing observations are also noted, with failure and recovery among the stated reasons, but the missing-data method and analysis populations are not reported.

The record lists multiple outcomes and analyses but does not report whether multiplicity was addressed. It does not provide details of randomization implementation, allocation concealment, eligibility criteria, outcome-assessor masking, or the reason for termination. The enrollment and adverse-event at-risk counts should not be assumed to define the analysis population for every outcome.

Finally, the safety information is presented as counts without reported event definitions, attribution, comparative uncertainty or an explanation of follow-up completeness. These limitations restrict both efficacy and safety interpretation. They reflect what is not reported in the registry record and should not be filled in by assumptions about the protocol or analysis.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Comparison of CPAP Machines With Reusable vs Disposable Circuits — NCT03121612.
• Registry record last updated 2019-11-20; status last verified 2019-11.
• Results first posted 2019-11-04.
• 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

The registry describes a randomized, unmasked, parallel-group trial comparing Dolphin CPAP with reusable circuits against Fisher-Paykel CPAP with disposable circuits in newborns diagnosed with respiratory distress syndrome. Enrollment is listed as 51, and the study status is terminated. The primary outcome was change from baseline in FiO2 at 6, 12, 24, 48 and 72 hours, summarized using median and inter-quartile range.

The registry flags contamination of the FiO2 endpoint because the majority of readings were at SpO2 > 95%, above the stated target of 90-95%. It also lists secondary physiological and clinical outcomes and descriptive adverse-event counts. Primary comparative results, statistical analysis details, multiplicity procedures, and the reason for termination are not reported in the registry record. The recorded information therefore describes the study question and some design and safety features, but does not establish the comparative primary outcome.