Desaturation Validation of INVSENSOR00006: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Desaturation Validation of INVSENSOR00006 (NCT03435211), a na phase not stated study. The primary outcome is Oxygen Saturation (SpO2) Accuracy of Sensor.

2. Clinical Question

The study was designed around a controlled change in inhaled oxygen: oxygen concentration was to be reduced gradually until arterial oxygen concentration was approximately 70%, followed by a return to room air. This procedure provides a range of oxygen saturation conditions in which device readings can be compared with reference values. The registry describes the intended procedure, but does not report detailed measurement schedules or participant-level results. Accordingly, the question that can be evaluated from the record is the question the study set out to address, not whether the device met a particular accuracy threshold. No threshold or primary endpoint result is reported in the registry record.

3. Trial Design

This was a completed, single-group device study sponsored by Masimo Corporation. The registry reports enrollment of 52, with all consented subjects enrolled into the test group and receiving INVSENSOR00006. Allocation is listed as NA, and masking is listed as none. The record therefore does not describe a randomized, masked comparison or a separate control group. Arterial blood samples were the measurement reference, not a second treatment arm.

The study is described as taking place in healthy adults during a desaturation procedure. Inhaled oxygen was to be reduced slowly until arterial oxygen concentration was approximately 70%, after which participants were to return to inhaling room air. The primary outcome timeframe is listed as 1-5 hours. These details describe the intended measurement setting; the registry does not report the exact sequence of measurements, how many readings were collected from each participant, or how the procedure was implemented for each participant.

The listed start date is 2018-01-17, and the primary completion date is 2018-02-19. The record status is COMPLETED. Completion status indicates the study record's administrative status; it does not, by itself, establish that a particular accuracy criterion was met. The registry lists one primary outcome measure and zero analyses. It does not provide a numerical primary outcome result or a statistical analysis of the paired device and reference measurements.

Because this is a single-group device validation design, its central comparison is within the measurement process: a noninvasive sensor reading is compared with a laboratory reference value. The study does not estimate a difference between randomized groups, and the registry record gives no basis for interpreting the result as a treatment effect. Details of a protocol, sample-size justification, analysis populations, interim analyses, or a prespecified statistical analysis plan are not reported in the registry record.

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: Of the 52 enrolled participants, 26 met inclusion criteria and proceeded to the study. The remainder of the enrolled participants include screen failures and subjects that did not proceed due to time constraints.

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 Oxygen Saturation (SpO2) Accuracy of Sensor, measured over 1-5 hours.

The primary endpoint is titled “Oxygen Saturation (SpO2) Accuracy of Sensor.” It is intended to assess agreement between oxygen saturation measured noninvasively by the pulse oximeter sensor and arterial oxygen saturation, SaO2, measured from a reference blood sample using a laboratory blood gas analyzer. The planned summary measure is arithmetic root mean square, or ARMS. The endpoint unit is listed as % of oxygen saturated hemoglobin, and its timeframe is 1-5 hours.

In general, an RMS-based accuracy summary combines the size of measurement errors into a single measure: larger discrepancies between device and reference readings contribute more to the summary than smaller discrepancies. It can reflect both systematic offset and variation in readings. That makes it useful for describing overall agreement across measurements, but it does not show the direction of error on its own. Other summaries, such as signed bias or limits of agreement, would address different aspects of the paired readings; whether those were calculated is not reported in the registry record.

The registry explains that ARMS was to be calculated, but does not post the value, the underlying paired readings, or an interval describing uncertainty. It also does not state an acceptance threshold for deciding whether the measured accuracy would meet a specified criterion. Consequently, the endpoint definition identifies what the study intended to measure, but the posted record does not show the endpoint result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The study's statistical question concerns agreement between two measurements of oxygen saturation: noninvasive SpO2 from INVSENSOR00006 and SaO2 from a reference blood sample analyzed by a laboratory blood gas analyzer. This is a measurement-validation problem. It differs from a trial asking whether one assigned treatment changes a clinical outcome relative to another treatment. The registry specifies ARMS as the intended summary, but does not report the analysis details or results.

What does an ARMS summary describe?

In general, a root mean square summary combines measurement discrepancies by squaring them, averaging the squared values, and taking the square root. The squaring step makes larger discrepancies contribute more heavily than smaller ones. The result is expressed on the scale of the measurements. An ARMS value can therefore summarize overall error across paired readings, but it does not reveal whether the sensor tends to read higher or lower than the reference. Nor does it show whether error is concentrated in particular participants or oxygen saturation ranges.

The registry names an arithmetic root mean square calculation, but does not report its precise implementation. The record does not say how readings were paired, whether repeated readings were averaged within participants, how observations from the same participant were handled, or how missing or technically invalid readings were treated. Those details matter because a collection of repeated readings from the same person may not provide the same information as the same number of independent participants. The analysis method cannot be reconstructed from the endpoint description alone.

Why compare against a reference measurement?

A device reading is informative only in relation to a measurement standard or comparator suited to the intended question. Here, the study description identifies arterial blood samples analyzed by a laboratory blood gas analyzer as the reference. Comparing each sensor value with a corresponding reference value can quantify agreement under the conditions in which measurements were taken. The reference method is not a randomized control group; it is the benchmark for evaluating the sensor's measurement performance.

In general, agreement analysis benefits from reporting several complementary features: the average signed difference, the spread of differences, and whether error changes across the measurement range. An overall RMS measure alone can combine these features into one number and conceal distinct patterns. The registry record does not report whether complementary summaries or diagnostic analyses were conducted. It also does not report a target accuracy threshold, so an RMS estimate could not be judged against a stated criterion even if the estimate were posted.

What does uncertainty add?

In general, a confidence interval describes a range of values reasonably compatible with the data and the statistical model used to calculate it. It helps distinguish an estimate that is precise from one that remains uncertain. For a validation study, uncertainty can depend on how many independent participants contributed data, how many paired readings were available, and how strongly repeated readings within a participant were related. The registry reports enrollment of 52 but does not report the number of analyzed participants or paired measurements. No confidence interval or other uncertainty estimate for ARMS is reported in the registry record.

Enrollment alone is not enough to establish the precision of an endpoint estimate. If each participant contributes repeated observations, those observations may be correlated. Treating correlated measurements as independent can make uncertainty appear smaller than it is. In general, statistical methods should account for the way observations are clustered within participants. Whether the study analysis accounted for repeated measurements is not reported in the registry record.

What is over-dispersion, and is it relevant?

In general, over-dispersion means that the observed variability in data is greater than a model assumes. It is a concern in some count or categorical analyses, where an overly restrictive variance model can understate uncertainty. The listed primary endpoint is a measurement-accuracy summary, not a count outcome, and the record does not report a model for it. No over-dispersion assessment is reported in the registry record. The issue cannot be evaluated from the posted information.

How should planning illustrations be read?

The study record does not report its own sample-size justification. A planning illustration described for the associated calculations uses a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with an effect to detect stated as a fraction of the observed effect. These are planning assumptions, not a description of this study's single-group validation design. They do not establish that the trial used a two-arm design, nor do they establish observed power. Sample-size illustrations are hypothetical planning quantities, not observed participant counts or evidence about the precision of this endpoint.

In general, statistical power is a design property calculated under assumptions about the effect, variability, significance level, and analysis. It is not a measure of how likely the observed result is to be true. A calculation based on an observed effect is especially unsuitable as a retrospective claim of observed power. The registry record does not report the assumptions used to justify enrollment of 52, and no conclusion about the adequacy of that enrollment for the primary endpoint can be drawn from the record.

Finally, no inferential analysis is posted: the record lists zero analyses. It therefore offers no reported test statistic, p-value, effect estimate, or uncertainty interval for the primary endpoint. General explanations of these methods cannot fill in the missing analysis result. The appropriate statistical reading is that the endpoint and intended comparison are described, while the numerical findings and the details required to evaluate their reliability are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and zero analyses. It does not report a multiplicity adjustment, a hierarchy of outcomes, or a prespecified plan for controlling the chance of false-positive findings across multiple statistical tests. Since no inferential results are posted, there is no reported set of tests from which to assess how multiplicity was handled. The absence of a reported adjustment does not show that none was used; the approach is not reported in the registry record.

The record does not describe how missing measurements were handled. It does not say whether all enrolled participants contributed reference and device readings, how incomplete measurement pairs were treated, or whether any data were excluded because of measurement quality. In general, missingness can affect a measurement-accuracy summary if the unavailable readings differ systematically from the available readings. The extent and handling of missing data cannot be assessed from the registry entry.

Analysis populations are also not described. The record gives enrollment of 52 and says all consented subjects were enrolled into the test group, but does not state how many participants were included in the primary endpoint analysis or whether any participants were excluded. It does not define a per-protocol, evaluable, or other analysis set. These details are important for understanding which measurements contribute to an endpoint estimate.

A statistical analysis plan, interim analyses, and a sample-size justification are not reported in the registry record. The record also does not identify adjustments for participant-level clustering, covariates, or repeated readings. These omissions limit assessment of the analysis process; they do not, on their own, show that the underlying study lacked such procedures.

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 entry for the INVSENSOR00006 group reports zero serious adverse events among 26 at risk and zero other adverse events among 26 at risk. The record lists zero deaths. These are registry counts for the listed categories and denominators. The record does not explain why the at-risk denominators are 26 when total enrollment is 52, so the safety denominators should be reported as listed rather than treated as interchangeable with total enrollment.

In general, an observed count of zero means that no events in that category were recorded for the stated denominator and reporting period. It does not prove that the underlying event risk is zero. Interpretation depends on how events were collected, what qualified as an event, the observation period, and how complete reporting was. Those details are not reported in the registry record. The entry does not identify serious adverse-event terms, and no top serious terms are listed.

These safety counts should not be used to infer sensor accuracy, and they do not establish an absence of all possible harms. They describe only the adverse-event information posted for this study.

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

The record supports a limited description of the study: it was a completed, single-group evaluation of INVSENSOR00006 in healthy adults, with sensor SpO2 intended to be compared against SaO2 from laboratory blood samples. The planned endpoint was ARMS, and the described procedure included reducing inhaled oxygen until arterial oxygen concentration was approximately 70%, then returning participants to room air.

The record does not report the ARMS result. It also does not provide paired readings, a confidence interval, or a stated accuracy threshold. Therefore, it does not show how close the sensor readings were to the reference measurements or how precisely that agreement was estimated. It is not possible to conclude from the registry entry that the sensor did or did not meet an accuracy standard.

The design also does not answer a comparative treatment question. There is no randomized comparator arm; the blood gas analyzer provides a measurement reference. The study's results, if fully reported elsewhere, would concern device measurement agreement under the study conditions, not the effect of a treatment on a clinical outcome. Generalizing beyond the described healthy-adult desaturation procedure would require evidence not present in the registry record.

The safety entry reports zero serious adverse events and zero other adverse events for the listed denominators. Such counts summarize the categories recorded, but do not establish zero risk or resolve the missing measurement-performance result. The most defensible conclusion is that the registry describes the intended validation question and study setup, while leaving the primary numerical result and key analysis details unreported.

13. Why This Trial Matters Statistically

Device validation studies raise a distinct statistical question: whether a measurement agrees sufficiently with a reference method for the intended conditions. This study's registry record makes that question explicit by identifying sensor SpO2, reference SaO2, and ARMS as the planned accuracy measure. Separating measurement agreement from treatment-effect questions helps prevent misinterpretation of a single-group design.

The desaturation procedure is also relevant to the statistical target. Device accuracy can vary across measurement conditions, so a study that deliberately changes inhaled oxygen is designed to assess readings across a changing oxygenation context rather than only under an unchanging condition. The registry describes the intended reduction to approximately 70% arterial oxygen concentration and subsequent return to room air. It does not report results by measurement range, so the extent to which performance varied across conditions cannot be assessed.

The record is a useful reminder that an endpoint definition is not an endpoint result. Naming ARMS does not provide the estimate, its uncertainty, or evidence that a prespecified criterion was met. For statistical interpretation, readers need the numerical result and enough detail about participants, paired readings, missing data, and analysis methods to understand what the summary represents.

Finally, completion status and enrollment are not substitutes for analysis. A completed study with enrollment of 52 may have a defined objective and procedure, yet the posted registry information can still be insufficient to assess the primary question. Transparent reporting of estimates and uncertainty is central to evaluating measurement studies, especially when the aim is to quantify agreement with a reference method.

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 important limitation of the registry record is the absence of the primary endpoint result. Although the endpoint is defined as an ARMS comparison of sensor SpO2 with reference SaO2, the record does not report the ARMS estimate, uncertainty, or underlying paired measurements. The study's stated accuracy question therefore cannot be answered numerically from this entry.

The record also omits details needed to assess the measurement analysis. It does not report the number of paired readings, the number of participants contributing to the endpoint analysis, handling of repeated readings, or procedures for missing or invalid measurements. It does not say whether accuracy was examined across different oxygen saturation levels or whether complementary measures of bias and spread were calculated.

The registry does not report an acceptance threshold or a protocol-based sample-size justification. It also does not provide a statistical analysis plan, analysis-population definitions, or information about interim analyses. These omissions prevent readers from determining how the endpoint was prespecified and how the analysis was intended to support a decision about accuracy.

The population and setting described are specific: healthy adults undergoing a controlled desaturation procedure. The registry does not provide participant characteristics or results from other clinical settings. Consequently, the record alone does not establish how the sensor would perform in populations or conditions not described. This is a limit on what can be inferred from the posted information, not evidence of a particular performance difference.

Finally, the adverse-event denominators are listed as 26 for serious events and 26 for other events, while enrollment is 52. The registry does not explain this presentation. The safety figures should therefore be interpreted using their stated denominators, without assuming how they relate to enrollment or to the primary endpoint analysis population.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Desaturation Validation of INVSENSOR00006 — NCT03435211.
• Registry record last updated 2024-09-19; status last verified 2024-08.
• Results first posted 2021-03-30.
• 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

NCT03435211 is a completed, single-group device study sponsored by Masimo Corporation, with enrollment of 52. The study evaluated INVSENSOR00006 in healthy adults and planned to compare noninvasive SpO2 readings with SaO2 values from laboratory blood gas analyzer reference samples during induced desaturation. The primary endpoint was an ARMS accuracy measure with a timeframe of 1-5 hours. The study description says oxygen was to be reduced until arterial oxygen concentration was approximately 70%, followed by a return to room air.

The record lists allocation as NA and masking as NONE. It reports one primary outcome measure and zero analyses. No numerical ARMS result, uncertainty estimate, analysis details, or accuracy threshold is reported in the registry record. The adverse-event entry lists zero serious adverse events and zero other adverse events for the stated denominators. The posted record describes the validation objective and procedure but does not provide the primary results needed to judge sensor accuracy.