Validation of Brain Oxygenation Monitor: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Validation of Brain Oxygenation Monitor (NCT00815490), a na na study. The primary outcome is Accuracy of Sensor.

2. Clinical Question

The study’s stated aim was to determine the accuracy of a non-invasive cerebral oxygen saturation monitor. The sensor sits on the forehead and emits and detects light; the monitor uses the absorbed light to display a saturation value. Blood measurements taken at the same time provide the reference against which those displayed values can be assessed.

To create a range of oxygen levels for evaluation, inspired oxygen was reduced in a controlled manner. The registry describes this as equivalent to ascending to an altitude of 16,000 feet. This is a measurement-validation question in healthy volunteers, rather than a treatment comparison in patients. It asks whether sensor readings track the blood measurements under the study conditions; the registry record does not report the numerical findings needed to quantify how closely they tracked.

3. Trial Design

NCT00815490, titled “Validation of Brain Oxygenation Monitor,” is a completed study sponsored by Duke University. The registry gives a start date of 2008-06 and a primary completion date of 2008-11. The official title describes calibration and validation of the Nonin non-invasive cerebral oxygen saturation oximeter and cerebral sensor in human volunteers. The listed phase and allocation are NA.

The registry classifies the model as single-group and masking as none. The study involved healthy volunteers undergoing oxygen desaturation, with the aim of evaluating sensor accuracy over a stated range of oxygen saturations of 70 - 100%. It is not a randomized treatment comparison, and the registry explicitly cautions that its arms are not a like-for-like control comparison. The labels Calibration and Validation appear in the adverse-event group listing; they do not establish that one group was a control for the other.

For the measurement comparison, the forehead sensor’s displayed saturation was paired conceptually with blood saturation measured at the same time. The registry describes arterial blood samples from the wrist and venous blood samples from the neck. Inspired oxygen was reduced in a controlled manner to produce desaturation. The record says the study would collect individual-participant data over a 1 hour timeframe and collect cohort data over a 6 month period.

The registry lists enrollment of 18. Its primary outcome description says that multiple data points from each individual were pooled and presented as group data. This detail matters: the reported endpoint was intended to describe pooled readings, rather than a separate summary for each volunteer. Numerical results for the primary measure and details of how the pooled observations were analyzed are not reported in the registry record. The record lists 1 outcome measure and 0 analyses.

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: Healthy volunteers recruited with specific exclusion of anemia \& hemoglobinopathy.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Multiple data points from each individual were pooled and presented as group data.

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 Accuracy of Sensor, measured over Data collected from individual participants over 1 hour timeframe. Data from cohort of subjects collected over 6 month period..

The primary outcome is titled “Accuracy of Sensor.” Its unit is percentage saturation, its parameter is mean, and its dispersion is described as full range. The assessment concerns how the forehead sensor’s saturation readings correspond to measured blood saturation. The registry describes displaying the sensor value on the vertical axis and blood saturation on the horizontal axis of a scatterplot, with a line of identity representing exact agreement.

Distance from that identity line provides a way to describe disagreement between the sensor and blood measurements. The named summary measure is Arithmetic Root Mean Square, or ARMS. In general, an RMS-type measure combines deviations from a reference into a summary that reflects their typical magnitude; for this endpoint, smaller ARMS is described in the registry as meaning that observed points lie closer to the identity line. The measure is intended to summarize agreement across readings, not to test a treatment effect.

That type of accuracy measure is useful because a sensor may be evaluated across a range of saturation values, rather than at only one value. The registry describes controlled desaturation for this purpose and states a range of 70 - 100%. However, the numerical ARMS result, the plotted values, and a confidence interval for the measure are not reported in the registry record. The endpoint’s definition therefore explains what the study intended to quantify, but does not establish the magnitude or precision of the observed agreement.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry’s primary endpoint is a measurement-agreement assessment, not a comparison of clinical outcomes. It describes paired values: saturation displayed by the forehead sensor and saturation measured in blood. A scatterplot places sensor saturation on the y-axis and measured blood saturation on the x-axis. If each reading matched its reference exactly, the points would lie on the line of identity. Departures from that line represent disagreement between the two measurements.

What does ARMS describe?

ARMS means Arithmetic Root Mean Square. In general, a root-mean-square summary squares deviations, averages them, and takes the square root, producing a measure of their overall magnitude. The registry identifies ARMS as the summary of dispersion around the line of identity and explains that a smaller value corresponds to points lying closer to that line. This is an accuracy-oriented summary: it describes the size of differences from a reference, rather than whether a treatment changed an outcome.

ARMS has an intuitive advantage over looking at individual deviations alone: it condenses a collection of discrepancies into a summary statistic. It does not, by itself, show the full shape of those discrepancies. For example, a summary does not tell a reader whether disagreement is similar across the tested saturation range or whether it is larger in some part of the range. The registry mentions a scatterplot and a full-range dispersion description, but the actual plotted values and numerical ARMS result are not reported in the registry record.

Why do repeated measurements matter?

The endpoint description says that multiple data points from each individual were pooled and presented as group data. In general, repeated readings from the same person may be related to one another. If such readings are treated as independent when they are not, a statistical analysis can overstate how much independent information the data contain and make uncertainty appear smaller than it should. Appropriate analysis may need to account for the clustering of observations within participants. The registry record does not report whether or how that dependence was handled, nor does it state the number of measurement pairs contributing to the pooled result.

Pooling can also hide variation between people. A group-level estimate describes the combined data as analyzed, but it does not necessarily communicate how much sensor agreement varied from volunteer to volunteer. Participant-level estimates or an analysis that explicitly models participant variation could answer related questions, but such results or methods are not reported in the registry record. No inference about those unreported analyses can be made from the endpoint description alone.

What does the study design permit?

The registry classifies the study as single-group, with allocation listed as NA and masking as none. The aim is to compare two ways of measuring oxygen saturation during a controlled procedure, not to compare the effects of assigned treatments. The simultaneous blood samples serve as measurement references in the accuracy question; they are not a parallel randomized control group. The labels Calibration and Validation in the adverse-event listings do not alter that design description.

In general, comparing a device reading with a reference measurement requires attention to whether both measurements refer to the same time and underlying condition. The brief summary says that blood samples were taken simultaneously with the monitor readings. It identifies arterial samples from the wrist and venous samples from the neck. The registry does not report further details of the sampling protocol, calibration procedures, or how the two blood sources were incorporated into the primary analysis.

What can confidence intervals and tests add?

A point estimate describes the observed magnitude of a statistic; a confidence interval communicates uncertainty around that estimate under a specified statistical method. A hypothesis test instead evaluates data relative to a defined null hypothesis and test procedure. These tools answer different questions. An accuracy estimate and its uncertainty would help characterize how precisely agreement was measured. A p-value alone would not explain the size or practical pattern of disagreement.

The registry lists 0 analyses and does not report an ARMS value, confidence interval, p-value, or formal testing procedure for the primary endpoint. It is therefore not possible to describe the inferential method actually used or the precision of its result from the registry record. General explanations of agreement statistics should not be mistaken for a claim that a particular statistical procedure was applied in this study.

How should planning calculations be read?

The record gives assumptions for planning illustrations based on the observed effect: a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It says that the effect to detect is a stated fraction of the observed effect, and distinguishes participants per arm from total events for time-to-event calculations. These are planning assumptions, not a description of the study’s own design. The study is listed as single-group, and no numerical observed effect or planning outputs are reported in the registry record.

In general, a planning calculation estimates the sample size needed under its chosen assumptions. It is not observed power, and it does not show that a study was adequately designed or that an endpoint result is precise. Illustrations that scale a detectable effect relative to an observed effect also depend on the reliability of that observed estimate. Without the numerical observed effect and planning outputs, the assumptions cannot be used to draw a quantitative conclusion about future sample size.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 1 outcome measure and 0 analyses. The listed primary measure is Accuracy of Sensor, described using ARMS and pooled group data. The record does not report a statistical analysis plan, a hierarchy of outcomes, or methods for adjusting for multiple comparisons. It therefore is not possible to determine whether multiplicity was considered in an analysis beyond what the registry describes.

Missing-data handling is not reported in the registry record. The record does not describe whether any sensor readings or blood samples were missing, how incomplete participant data would be handled, or whether exclusions were made. No imputation approach or sensitivity analysis is reported. The absence of those details does not establish that missing data occurred; it means the registry does not explain how that issue was assessed or addressed.

The endpoint description says multiple data points from each individual were pooled and presented as group data. This identifies the stated presentation population as pooled measurements, but does not specify a participant-level inclusion rule or the analysis population used for any numerical estimate. The record does not report whether all enrolled volunteers contributed to the primary analysis or how repeated measurements within a volunteer were treated. It also does not describe a per-protocol or other analysis population.

These omissions limit assessment of how the reported estimate would have been selected and calculated. In particular, readers cannot evaluate whether the number of observations contributing to an estimate matched the number planned, whether missingness could have affected the result, or whether additional analyses were considered. No numerical endpoint result is reported in the registry record, so these questions cannot be resolved from the posted information.

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.

The registry’s adverse-event listing separates groups labeled Calibration and Validation. For each listed group, the number at risk for serious adverse events is 9 and the number affected is 0. The number at risk for other adverse events is also 9 and the number affected is 0. Deaths are listed as null, and no serious-event terms are listed.

These entries are registry counts, not a comparative safety analysis. The study is classified as single-group, allocation is NA, and the registry cautions that the arms are not a like-for-like control comparison. The group labels should therefore not be read as evidence of randomized treatment groups or used to infer a difference between interventions.

A count of zero affected participants in the listed categories does not establish that an event is impossible or that the device or procedure is risk-free. The record does not report further detail about event definitions, collection methods, severity beyond the serious-event category, timing, or follow-up for adverse events. The null entry for deaths does not provide a numerical death count; death information is not reported in the registry record. These figures describe what the registry lists, with those limitations.

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

The registry supports a limited, specific description of the study: healthy volunteers underwent controlled oxygen desaturation while a forehead cerebral oxygen sensor was evaluated against blood saturation measurements. The stated aim was to assess accuracy over a saturation range of 70 - 100%. The primary endpoint was designed around agreement with a line of identity, summarized by ARMS. These details explain the measurement question and the intended form of its answer.

The record does not report the numerical ARMS result, confidence interval, paired readings, or primary analysis output. It therefore does not allow readers to quantify how closely the sensor readings matched the blood measurements, how uncertain an estimate was, or whether agreement varied across the tested range. The endpoint description says that smaller ARMS corresponds to data points closer to the identity line, but that general interpretation cannot substitute for an observed value.

The result, even if fully reported, would address the measurement conditions and population studied. The record describes healthy volunteers and induced desaturation, not a patient treatment study. It cannot establish performance in populations or circumstances not described in the registry. Nor does the study design test whether using the monitor changes clinical decisions or patient outcomes.

The posted safety entries list no affected participants for the serious or other adverse-event categories in the Calibration and Validation groups, while deaths are listed as null. Those entries should not be interpreted as proof that no risk exists, nor as a comparative safety finding. Taken together, the record supplies the rationale and measurement framework for validation but does not supply the primary numerical result needed to judge the observed accuracy.

13. Why This Trial Matters Statistically

Device validation is a statistical problem of measurement agreement. A monitor’s displayed value is useful to evaluate against a reference measurement, and the comparison must consider how far readings depart from the reference across the intended measurement range. This study’s stated use of simultaneous sensor and blood measurements, a line of identity, and ARMS makes that aim explicit.

The endpoint also highlights why data structure matters. Multiple observations from each individual were pooled into group data. In general, a collection of readings is not necessarily equivalent to a collection of independent participants: repeated observations from the same person may share characteristics. Describing how those observations are handled is important for interpreting uncertainty and the weight of the resulting evidence. The registry record does not report that analysis detail or the numerical endpoint result.

The design distinction matters as well. Calibration and Validation are labels in the adverse-event listing, not evidence of a randomized control comparison. The record describes a single-group study with no masking and controlled oxygen desaturation. The statistical question is measurement accuracy under those conditions, not whether one treatment is better than another.

Finally, the difference between a study result and a planning illustration is important. Sample-size planning depends on assumptions about the effect to detect and the analysis. The registry’s stated planning assumptions are not the trial’s design, and illustrations based on an observed effect are not observed power. Since numerical accuracy results and planning outputs are not reported in the registry record, the study’s statistical value lies chiefly in its defined measurement question and described design, rather than a quantifiable posted estimate.

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 in the registry record is the absence of the numerical primary outcome. Although the endpoint description explains ARMS and how it relates to the line of identity, the observed ARMS value and its uncertainty are not reported. This prevents an assessment of the magnitude and precision of agreement from the posted results.

The study involved healthy volunteers and controlled reduction of inspired oxygen. The registry does not describe patient enrollment. Findings from this setting, even if numerical results were available, would not by themselves establish how the monitor performs in clinical populations or under conditions not represented in the study.

The endpoint uses multiple data points per individual, pooled as group data. The number of measurement pairs, participant-level variation, and method for accounting for within-person dependence are not reported in the registry record. Without those details, readers cannot assess the effective information in the pooled data or whether uncertainty properly reflects the repeated-measure structure.

The study is listed as single-group, with allocation NA and masking none. The Calibration and Validation labels do not define a randomized comparison, and the registry cautions that the arms are not a like-for-like control comparison. The record also does not report details of missing-data handling, the full analysis procedure, or a confidence interval for the primary endpoint.

Finally, the stated saturation range and simultaneous blood sampling describe the intended measurement conditions, but the registry does not report how arterial and venous reference values were incorporated into the primary analysis. Together, the design and reporting limitations constrain what can be concluded: the record describes a validation approach, but does not provide a numerical basis for judging its accuracy.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Validation of Brain Oxygenation Monitor — NCT00815490.
• Registry record last updated 2013-04-16; status last verified 2013-03.
• Results first posted 2013-04-16.
• 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

NCT00815490 is a completed, single-group study sponsored by Duke University, with enrollment of 18 and healthy volunteers as the study population. The registry lists a start date of 2008-06 and a primary completion date of 2008-11. Masking is listed as none, allocation as NA, 1 outcome measure, and 0 analyses.

The study evaluated a forehead cerebral oxygen saturation sensor during controlled oxygen desaturation, comparing displayed readings with simultaneously measured blood saturation. Its primary endpoint, Accuracy of Sensor, was described using a scatterplot around the line of identity and ARMS. The record says that multiple data points per individual were pooled as group data.

The numerical primary result and its uncertainty are not reported in the registry record. The adverse-event listings report 0 affected participants for serious and other events in each listed group; deaths are null. The record does not describe a randomized comparison or report sufficient numerical endpoint results to quantify sensor agreement.