SpO2 System Accuracy Testing With Different Sensors: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for SpO2 System Accuracy Testing With Different Sensors (NCT01613222), a na na study. The primary outcome is Accuracy Root Mean Square (ARMS).

2. Clinical Question

3. Trial Design

This completed study, NCT01613222, is titled “SpO2 System Accuracy Testing With Different Sensors.” GE Healthcare is listed as the sponsor. The record gives a start date of 2012-06-19 and a primary completion date of 2013-02-15. It lists an enrollment of 63 and classifies the study as “other,” with phase recorded as NA.

The design is listed as a single-group model, with allocation recorded as NA and masking as none. It is therefore not a randomized comparison of treatment arms. The listed configurations describe combinations of sensors with monitors or other equipment; they should not be read as independent parallel groups with a conventional control arm. The registry record specifically indicates that the arms are not a like-for-like control comparison.

The study description concerns accuracy testing using paired sensor and reference readings. Multiple data pairs were collected from each participant, although the record says that not every sensor, monitor, co-oximeter, and module configuration was used with each participant. The primary endpoint timeframe is 60 minutes. Details about the timing and order of measurements, assignment of configurations, participant flow, and handling of repeated measurements are not reported in the registry record.

The registry lists one primary outcome measure and zero analysis entries. That does not establish that no analysis was conducted; it means that analysis results are not posted as registry analysis entries. The protocol’s statistical analysis plan, sample-size justification, and any interim analyses 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The data analyzed was SpO2 data pairs using different U-Trusginal SpO2 sensors, patient monitors, co-oximeters, and various modules. Multiple data pairs were collected from each study participant, but not all sensors, patient monitors, co-oximeters and module configurations were used with each participant.

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.

Population: 63 subjects were enrolled. Data from 51 subjects was used for analysis.

6. Primary Endpoint

The registry lists the primary outcome as Accuracy Root Mean Square (ARMS), measured over 60 minutes.

The primary endpoint is Accuracy Root Mean Square (ARMS), described as a measure of accuracy of the test SpO2 device relative to a reference SpO2 device. The registry describes plotting the forehead sensor saturation against measured blood saturation and comparing the paired points with a line of identity. The line represents perfect agreement: each sensor reading would match its paired reference reading.

As a general statistical concept, a root mean square summarizes the magnitude of deviations by squaring them before averaging and taking the square root. This makes larger deviations contribute more to the summary than small deviations. In this context, a smaller ARMS indicates that the paired observations lie closer to the line of identity, as the registry description states. The measure can therefore characterize agreement across the observed paired readings rather than simply report whether readings tend to be above or below the reference.

The outcome is specified as a mean with full range as its dispersion, over 60 minutes. The registry cites ISO 80601-2-61, 2011, Annex EE and the FDA Draft Guidance for Pulse Oximeters dated July 19, 2007, as the stated basis for the analysis approach. A numerical ARMS estimate, its full range, and any threshold used to interpret the result are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does ARMS describe?

ARMS is a summary of the distance between paired sensor readings and reference readings. The registry’s endpoint description uses a scatterplot with the forehead sensor saturation on one axis and measured blood saturation on the other. The line of identity marks exact agreement. Points close to that line indicate readings close to their paired references; points farther away indicate greater disagreement. The registry states that the dispersion around this line is summarized using an arithmetic root mean square calculation.

As a general statistical explanation, a root-mean-square measure combines deviations into a single scale by squaring deviations, averaging those squared values, and taking a square root. Because deviations are squared, larger discrepancies have a greater influence on the summary than smaller ones. The resulting metric reflects the magnitude of deviations, not only their direction. It does not by itself show whether readings tend to be higher or lower than the reference; that pattern would require additional information about signed differences or the displayed paired data.

Why use a reference comparison?

For a measurement-accuracy question, a reference method allows each test reading to be assessed against a corresponding reference reading. The co-oximetry measurement is identified as the gold reference in the registry description. Agreement is distinct from association: two measures can move together while still differing systematically. Comparing paired observations with the line of identity focuses on whether they are numerically close, not merely whether they vary in a similar direction. This is a general explanation of measurement agreement, not an additional result from this study.

The direction of the ARMS scale is also important. The registry says that smaller ARMS values correspond to points closer to the identity line and therefore closer agreement. But without the numerical ARMS result, no conclusion about the magnitude of agreement can be drawn. A metric’s definition tells readers how to interpret a reported value; it cannot substitute for that value.

What does the reported dispersion mean?

The endpoint is identified as a mean, with full range as the dispersion. In general, a range describes the spread from the smallest to the largest observed value, while a mean gives a summary of central tendency. These summaries answer different questions: the mean describes an overall level, and the range indicates how far the observed values extend. Neither is reported numerically in the registry record. A range alone would not show the shape of the distribution, and a mean alone would not describe variation among readings or participants.

The record names ISO 80601-2-61, 2011, Annex EE, and the FDA Draft Guidance for Pulse Oximeters, dated July 19, 2007, in describing the analysis approach. The registry does not report enough detail to determine the specific implementation, the exact set of observations included, or how any acceptance criterion was applied. No numerical endpoint result is posted.

Why do repeated readings matter?

The registry says that multiple data pairs were collected from each participant. As a general statistical matter, measurements from the same participant may be more alike than measurements from different participants. Treating every paired reading as independent when readings are correlated can make uncertainty appear smaller than it is. Statistical methods can account for within-participant dependence, but the specific approach used for these repeated measurements is not reported in the registry record.

Not every equipment configuration was used with every participant. This means the amount and composition of data may differ across configurations. Configuration-specific estimates would be most informative if accompanied by the number of contributing participants and pairs, along with an account of how repeated observations were handled. Those details and numerical comparisons are not reported in the registry record. The configuration labels should therefore be read as testing setups, not as evidence that every setup was tested in an identical participant group.

What can be said about uncertainty?

A confidence interval is a general statistical way to express the precision of an estimate: it gives a range of values compatible with the data under the stated method and assumptions. The registry record does not report a confidence interval or another uncertainty estimate for ARMS. Without one, readers cannot assess how precisely the endpoint was estimated. The record also contains zero analysis entries, so there are no posted registry analysis results to interpret. This does not establish whether an analysis was performed outside the registry.

Because the numerical outcome and its uncertainty are missing, there is no basis to calculate or infer an effect, a confidence interval, or a significance test. The endpoint definition, reference method, and cited standards clarify the intended measurement framework, while leaving the empirical result and its precision unreported.

10. Multiplicity, Missing Data and Analysis Populations

The registry identifies one primary outcome measure, Accuracy Root Mean Square (ARMS), and lists zero analysis entries. It does not report a statistical analysis plan, so the prespecified handling of multiple comparisons or multiple sensor and monitor configurations is not reported in the registry record. It is also not reported whether any additional exploratory analyses were planned or performed.

Multiplicity is a general concern when many outcomes or comparisons are examined: the more tests considered, the greater the chance of finding an apparently notable result by chance if results are interpreted without accounting for the testing strategy. The listed configurations do not by themselves establish how many formal comparisons were made. No adjustment method or comparison hierarchy is reported, so the registry record does not support a judgment about whether multiplicity was addressed.

Rules for missing or unusable readings, participant-level exclusions, and incomplete configuration data are not reported in the registry record. The number of data pairs available for each configuration and the handling of participants who did not contribute every configuration are also not reported. These details are relevant to understanding which observations entered each endpoint summary.

The analysis populations, including any distinction between enrolled participants and participants with evaluable measurements, are not reported in the registry record. Interim analyses and stopping rules are not reported either. These omissions do not show that such procedures were absent; they mean their use cannot be established from the registry information.

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 registry lists safety entries for sensor and equipment configurations. For each listed entry, it records 63 participants at risk, zero serious adverse events, zero other adverse events, and zero deaths. These are the posted counts for the entries; they do not establish that no events occurred outside the reporting captured in the registry record.

The same at-risk count appears across multiple configuration entries. Because the study is a single-group device assessment and participants may contribute measurements under different setups, these entries should not be treated as distinct sets of participants or added together to create a larger unique-participant total. The record does not report how exposure time or safety follow-up was defined for each setup.

Zero recorded events is a count, not proof that the underlying risk is zero. The registry record does not provide event-rate estimates, uncertainty intervals, detailed event narratives, or a description of safety ascertainment. The serious adverse event figures therefore describe the posted registry counts only and should be interpreted within those reporting limits.

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

The study addresses a measurement question: how closely SpO2 sensor readings correspond to paired co-oximetry reference readings across different equipment configurations. The registry identifies ARMS as the primary accuracy measure and explains that smaller values represent closer proximity to the line of identity. This makes the intended interpretation of the metric clear in principle.

However, the numerical ARMS result is not reported in the registry record. The record therefore does not allow readers to quantify the observed disagreement, assess its variability, or determine how precisely it was estimated. The arm description states that specified accuracy is met, but no numerical endpoint estimate, accuracy threshold, or configuration-specific result accompanies that statement. It should not be treated as a substitute for the missing results.

The design is not a randomized treatment comparison. It evaluates device measurement performance against a reference and includes multiple configurations, with multiple pairs collected from participants and not every configuration used for every participant. The record does not provide the detailed analysis needed to compare these configurations or determine whether differences between them are meaningful.

More generally, an accuracy summary does not automatically establish performance in every setting or for every participant. The registry record does not provide enough information about participant characteristics, measurement conditions, or configuration-specific results to define the scope of any generalization. It also does not provide the statistical plan, missing-data rules, or uncertainty estimates needed for a fuller assessment.

Accordingly, the posted information establishes the study’s intended question, endpoint, reference approach, enrollment, and design features. It does not establish a numerical level of accuracy or the precision of that level. The absence of a reported endpoint result should not be interpreted as evidence either for or against the measurement performance of the tested devices.

13. Why This Trial Matters Statistically

Device-accuracy studies rely on a different statistical question from trials that compare clinical outcomes between treatment groups. Here, the focus is agreement between a test measurement and a reference measurement. The line-of-identity description and ARMS endpoint make that distinction explicit: the aim is to summarize how far paired readings deviate from exact agreement.

The study also illustrates why the measurement structure matters. Multiple pairs were collected from each participant, while not every participant used every sensor and equipment configuration. These features affect how observations should be summarized and how uncertainty should be estimated. As a general statistical issue, repeated measurements from one participant may be correlated, and varying configuration coverage can complicate direct comparisons. The registry does not report how these issues were handled in analysis.

The record names recognized methodological references, including ISO 80601-2-61, 2011, Annex EE, and the FDA Draft Guidance for Pulse Oximeters dated July 19, 2007. Such references give context for the stated measurement approach, but they do not replace the need for numerical results, uncertainty estimates, and transparent reporting of the analysis population.

For statistical interpretation, the key lesson is to separate a clear endpoint definition from evidence about the endpoint’s observed value. In this case, the measure is described, but its result is not reported in the registry record. Without that result, readers cannot evaluate the magnitude or precision of the observed agreement, regardless of how clearly the intended method is stated.

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 for interpreting the posted results is that the numerical primary endpoint is not reported in the registry record. No ARMS estimate, numerical range, confidence interval, or configuration-specific endpoint value is available. The stated conclusion in the arm description cannot be independently evaluated from those missing results.

The study includes multiple data pairs per participant, and the record says that not all configurations were used by each participant. The number of observations and participants contributing to each setup is not reported. Nor is the method for handling within-participant dependence described. These gaps limit assessment of how the estimates were constructed and how precise they might be.

The registry does not report participant characteristics, detailed measurement procedures, criteria for including or excluding readings, handling of missing measurements, or analysis populations. It does not give a statistical analysis plan or a sample-size justification. The basis and numerical value of the specified accuracy criterion are also not reported in the registry record.

The single-group, nonrandomized structure is appropriate to distinguish from a conventional randomized control comparison, but it does not support interpreting the configuration labels as randomized arms. The record does not give enough detail to determine whether the participant groups contributing to different configurations were comparable or whether direct statistical comparisons were intended.

Finally, safety entries show zero recorded events in the listed categories, but the safety ascertainment period and level of detail are not reported. These limitations concern what can be concluded from the registry record; they do not establish that procedures or analyses were absent outside it.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: SpO2 System Accuracy Testing With Different Sensors — NCT01613222.
• Registry record last updated 2019-02-05; status last verified 2019-01.
• Results first posted 2019-02-05.
• 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

NCT01613222 is a completed, single-group study sponsored by GE Healthcare, with enrollment of 63 and a primary endpoint of Accuracy Root Mean Square (ARMS) over 60 minutes. It tested SpO2 sensor configurations against co-oximetry reference measurements, with multiple paired readings collected from participants and not every configuration used with each participant.

The registry describes the ARMS approach and cites ISO 80601-2-61, 2011, Annex EE, and FDA Draft Guidance for Pulse Oximeters dated July 19, 2007. It lists one primary outcome measure and zero analysis entries. The numerical primary result, uncertainty estimates, detailed statistical plan, missing-data rules, and analysis populations are not reported in the registry record. Safety entries list zero serious adverse events, zero other adverse events, and zero deaths for each listed configuration.