This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Accuracy of Noninvasive Pulse Oximeter Measurement of Hemoglobin for Rainbow DCI Sensor (NCT03124901), a na na study. The primary outcome is Accuracy of Sensor.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03124901 |
| Phase | NA |
| Condition | Surgery, Healthy |
| Allocation | Not applicable (single group) |
| Design model | Single Group |
| Masking | None (open label) |
| Enrollment | 147 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Masimo Corporation |
| Results posted | Yes |
2. Clinical Question
Population
The registry describes hospitalized study subjects undergoing surgery and healthy study volunteers. Fluid status was expected to change through fluid loss during surgery for hospitalized subjects or controlled intravenous fluid administration for healthy volunteers. The enrolled population is reported as 147 participants.
Intervention
All subjects were assigned to the test group and received the DCI pulse oximeter sensor. The sensor is a device intended to measure hemoglobin noninvasively.
Comparator
The comparison was between the noninvasive sensor measurement and a hemoglobin value obtained from a blood sample. The record does not describe a separate control group or a randomized comparison.
Primary question
The primary question was how accurately the sensor measured hemoglobin, assessed by comparing sensor readings with blood-sample values and calculating the arithmetic root mean square error. The outcome timeframe was up to 24 hours. The accuracy estimate and its uncertainty are not reported in the registry record.
This is a measurement-accuracy question, not a comparison of treatment effects. It asks whether the sensor’s readings correspond closely to a blood-based measurement under the study conditions. Because all subjects received the sensor and the record does not report a control arm, the design does not support a randomized comparison between sensor use and another intervention. The blood sample serves as the reference measurement for evaluating agreement, rather than as a participant group.
The record gives the study’s rationale in terms of changes in fluid status and hemoglobin concentration. It does not report how measurements were selected or paired for analysis, how many paired readings contributed, or the observed accuracy. Those omissions prevent a reader from judging the size and precision of the sensor-reference differences from the registry record alone.
3. Trial Design
This completed study is registered as a single-group device study of the DCI pulse oximeter sensor. The record lists 147 participants, conditions of surgery and healthy status, and a test group in which all subjects received the sensor. The allocation field is recorded as not applicable, and masking is listed as none. The study is therefore not described as a randomized, masked comparison.
The study description identifies two contexts in which fluid status could change: fluid loss during surgery among hospitalized subjects, and controlled intravenous fluid administration among healthy volunteers. The sensor measurement was to be compared with hemoglobin obtained from a blood sample. The primary outcome is titled “Accuracy of Sensor,” is expressed in g/dL, and has a timeframe of up to 24 hours.
The registry reports the study as completed. Its recorded start date is 2008-09-30, and its primary completion date is 2008-10-27. The sponsor is Masimo Corporation. The phase is recorded as not applicable. These administrative details describe the registered study; they do not establish how closely the device readings matched the reference measurements.
No separate control arm is described. A blood-sample comparator is part of the measurement assessment, but it is not a like-for-like participant group. The record does not report a protocol-level description of the measurement schedule, inclusion or exclusion criteria, participant flow, handling of repeated observations, or the statistical analysis plan. It also does not report a sample-size justification. Such details should not be inferred from the single-group design or the enrollment entry.
Although the record lists one primary outcome measure, it lists no analysis entries. Consequently, the completed status should not be mistaken for a reported primary result. The accuracy estimate, its precision, and the analysis population 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
| Stage | Test Group |
|---|---|
| STARTED | 147 |
| COMPLETED | 147 |
| NOT COMPLETED | 0 |
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.
| Characteristic | Unit | Test Group |
|---|---|---|
| Participants analysed | 147 | |
| Age, Categorical: <=18 years | Participants | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 147 |
| Age, Categorical: >=65 years | Participants | 0 |
| Sex: Female, Male: Female | Participants | 43 |
| Sex: Female, Male: Male | Participants | 104 |
| Race and Ethnicity Not Collected | Participants | — |
| Region of Enrollment: United States | participants | 147 |
6. Primary Endpoint
The registry lists the primary outcome as Accuracy of Sensor, measured over Up to 24 hours.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | g/dL |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is sensor accuracy, assessed by comparing noninvasive hemoglobin measurements from the pulse oximeter with hemoglobin values obtained from blood samples. The registry specifies the arithmetic root mean square error (ARMSE) as the summary measure and gives g/dL as the unit. The measurement timeframe is up to 24 hours.
As a general statistical explanation, a root mean square error summarizes the size of differences between paired measurements by squaring the differences, averaging them, and taking the square root. This gives greater influence to larger discrepancies than a simple average of signed differences would. The resulting measure describes the magnitude of disagreement; on its own it does not indicate whether the sensor tends to read higher or lower than the reference. It also does not show how error varies across the range of hemoglobin values or across participants.
Agreement with a blood-sample measurement is relevant because the study question concerns measurement performance, rather than a clinical outcome or treatment effect. Interpretation would ordinarily depend on the estimated error, the spread and pattern of paired differences, and the precision of the estimate. The registry record does not report the ARMSE value, paired-measurement results, or an uncertainty interval. The endpoint’s definition therefore identifies what was intended to be assessed, but does not establish the sensor’s observed accuracy.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record describes a completed study with an accuracy endpoint, but it does not report the primary outcome analysis. It provides no observed ARMSE value, no measure of uncertainty, and no analysis entries. The record therefore does not show how large the sensor-reference discrepancies were or how precisely accuracy was estimated.
The stated comparison is between a noninvasive sensor measurement and a blood-sample hemoglobin value. For a measurement study, the central statistical question is agreement between paired readings. A single summary such as ARMSE can describe the typical scale of error, but it does not reveal the direction of any systematic bias. Nor does it show whether errors are similar throughout the measurement range or differ by participant or fluid-status context. The registry record does not report these details.
The single-group structure matters. All subjects received the DCI pulse oximeter sensor, and there is no separate randomized control group. This is consistent with comparing a device reading against a reference measurement, but it is not a treatment comparison. The design cannot establish comparative effects against an alternative device or strategy based on the information in the registry record.
The entry for enrollment is 147. That figure does not tell the reader how many observations were available for the accuracy calculation, whether participants contributed repeated measurements, or whether every sensor reading could be paired with a reference value. The number of paired readings and the analysis population are not reported in the registry record. It would be inappropriate to assume that the enrollment count is also the denominator for the primary accuracy analysis.
Completion status is administrative and is not a substitute for a posted result. Likewise, listing a primary endpoint does not establish that the endpoint met a threshold or that a particular degree of accuracy was achieved. No acceptance criterion or clinically meaningful accuracy threshold is reported in the registry record.
Any planning illustrations based on an observed effect must be distinguished from observed trial findings. The registry record does not report the primary effect estimate that would allow readers to assess such an illustration. Planning calculations do not recover missing outcome data and should not be interpreted as evidence that the sensor reached a particular level of accuracy.
The most defensible interpretation is limited: the registered study intended to evaluate noninvasive hemoglobin measurement accuracy against blood-sample values in a single-group setting, with assessment up to 24 hours. The posted registry information does not provide the numerical findings needed to judge the magnitude, precision, or pattern of measurement error.
9. Statistical Methods Explained
The primary outcome is a measurement-agreement endpoint. The registry says that noninvasive hemoglobin measurements from the pulse oximeter would be compared with hemoglobin values from blood samples, with accuracy summarized by arithmetic root mean square error. This is different from testing whether an intervention changes an outcome between randomized groups: the focus is the discrepancy between two measurements of the same underlying quantity.
What does root mean square error describe?
As a general statistical explanation, root mean square error starts with paired differences between a method under evaluation and a reference method. Each difference is squared, the squared differences are averaged, and the square root is taken. The measure is expressed on the same scale as the original measurements. Here, the registered unit is g/dL.
Squaring makes the sign of each difference disappear, so the summary reflects error magnitude rather than whether the sensor tends to over-read or under-read. Squaring also gives relatively more influence to large errors. Thus, a small number of sizeable discrepancies can affect the overall summary. ARMSE alone does not show the distribution of errors, whether error changes with the hemoglobin level, or whether the sensor has a consistent directional bias. Those questions require additional summaries or displays of paired data, none of which are reported in the registry record.
Why does pairing matter?
For a direct accuracy assessment, each sensor reading needs to be matched to a suitable blood-sample reference measurement. Pairing is important because the discrepancy is defined within a measurement pair. The registry record does not report the matching rules, the timing between sensor and blood measurements, the number of valid pairs, or how readings were handled when more than one measurement was available for a participant. Without those details, the calculation cannot be reconstructed or evaluated from the registry record.
Repeated measurements can also affect statistical interpretation. As a general matter, readings from one participant may be related to one another rather than independent. An analysis that treats correlated observations as independent can make uncertainty appear smaller than it is. The record does not say whether the analysis used one measurement per participant, accounted for repeated readings, or applied another method. It also does not report an analysis plan that would settle these questions.
How should uncertainty be assessed?
A point estimate describes the result of a calculation but does not by itself show how precisely it was estimated. As a general statistical explanation, a confidence interval gives a range of values compatible with the data and the method used to calculate it, subject to the interval’s assumptions. For measurement accuracy, uncertainty can depend on the number of usable pairs and on dependence among repeated measurements. The registry record does not report an ARMSE estimate, confidence interval, standard error, or other measure of precision.
A confidence interval would not answer every measurement question. It would quantify uncertainty in a summary, while information about direction and distribution would still be needed to assess systematic bias and individual discrepancies. The record does not report those complementary results either. No uncertainty claim can be made from the outcome definition alone.
Why is a control group not the key feature?
In a measurement study, the reference blood-sample value is the comparator for the sensor reading. It is not a separate control arm of participants. The registry describes a single group, with all subjects receiving the sensor, and records allocation as not applicable. That setup can address agreement with a reference if paired readings are collected and analyzed appropriately. It does not provide a randomized comparison of the sensor against an alternative intervention. The record does not report a like-for-like control group.
The masking field is recorded as none. The registry does not describe whether people taking or interpreting measurements knew the other measurement result, nor does it report procedures to limit measurement or interpretation bias. Those details should not be inferred from the masking field alone.
What can planning calculations establish?
As a general explanation, a sample-size calculation can illustrate how many participants per arm might be required under specified assumptions for a planned comparison. Such calculations depend on choices about the effect to detect, variability, allocation, and statistical method. They are planning exercises, not observed evidence about the accuracy of a device. The registry’s stated planning assumptions describe an illustration based on an observed effect, but the primary outcome estimate is not reported in the registry record. No planning illustration can substitute for the absent accuracy result, and it should not be described as observed power.
The registry record does not report the study’s own sample-size justification, a formal analysis plan, or an interim-analysis plan. It also does not report whether the proposed ARMSE calculation was modified in practice. The outcome definition and design description support an explanation of the intended statistical question, but not a reconstruction of the completed analysis.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists one primary outcome measure, “Accuracy of Sensor,” and lists no analysis entries. A single listed primary measure does not by itself show whether additional outcomes or exploratory analyses were considered, how any multiplicity was handled, or whether the registered outcome was analyzed as planned. Those details are not reported in the registry record.
The record does not report an analysis population for the primary endpoint, the number of paired sensor and blood-sample readings, or the number of participants with usable measurements. Although enrollment is reported as 147, that does not establish the denominator for the accuracy calculation. The handling of missing sensor readings, unavailable blood samples, incomplete pairs, withdrawals, or measurements excluded from analysis is not reported in the registry record.
As a general statistical point, missing paired measurements can affect an accuracy estimate if the unavailable pairs differ systematically from the available ones. The extent and direction of any such effect cannot be assessed without information about missingness and the analysis approach. No imputation method, sensitivity analysis, or missing-data assumption is reported.
The statistical plan, including any adjustment for multiple testing, is not reported in the registry record. Neither an absence of analysis entries nor a single registered primary measure establishes that multiplicity was absent or appropriately addressed. The posted information does not support conclusions about the completeness of outcome reporting or the robustness of the primary analysis.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Test Group | 0 | 147 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Test Group | 0 of 147 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Test Group | 0 of 147 |
The registry reports a test group with 147 subjects at risk for serious adverse events and 0 affected; it also reports 147 at risk for other adverse events and 0 affected. Deaths are recorded as 0. No top serious adverse-event terms are listed.
These entries describe what is recorded in the registry, not proof that no adverse events occurred under all possible definitions or reporting procedures. The record does not explain the collection period or provide further detail about event ascertainment, definitions, or completeness. The serious and other adverse-event entries should be read as registry counts, not as a complete assessment of device safety.
The study’s primary purpose in the record is measurement accuracy, and these safety entries do not answer whether sensor readings agreed closely with blood-sample values. They also do not establish a comparative safety profile: the study is single-group and does not report a separate control group. The registry information supports only the narrow statement that the listed group has zero affected entries for these recorded categories and zero recorded deaths.
12. What the Result Does, and Does Not, Mean
The record establishes that a completed single-group study was registered to assess the accuracy of a DCI pulse oximeter sensor. Its intended comparison was between noninvasive hemoglobin readings and values obtained from blood samples, summarized using arithmetic root mean square error over a timeframe of up to 24 hours. It also describes study contexts involving surgery and healthy volunteers with fluid-status changes.
The record does not establish the magnitude of the sensor’s measurement error. The ARMSE result, paired-reading data, confidence interval, and analysis population are not reported in the registry record. Accordingly, readers cannot determine from this information whether discrepancies were small or large, whether readings tended to be systematically higher or lower, or how consistent accuracy was across observations.
This is not a randomized treatment comparison. The blood sample is the reference for evaluating the sensor measurement, not a separate control group. The record therefore does not support claims about comparative effectiveness or superiority to another device. It also does not report a threshold that would define acceptable measurement accuracy, so the intended endpoint cannot be classified against an acceptance standard from the registry information.
The registered enrollment of 147 does not reveal how many paired measurements entered the analysis. It should not be treated as the number of observations used to calculate accuracy. Missing measurements, repeated readings, and exclusions are not described in the registry record.
The safety entries report zero affected subjects in the listed serious and other adverse-event categories, and zero deaths. Those entries are limited registry data and should not be extended into a broader conclusion about safety. Overall, the record documents the measurement question and the design used to pursue it, but the posted information is insufficient to judge the observed accuracy or its precision.
13. Why This Trial Matters Statistically
This study raises a useful statistical distinction: evaluating the accuracy of a measurement device is not the same as comparing clinical outcomes between treatment groups. The key evidence would be paired measurements from the sensor and a reference method, analyzed to describe the size and pattern of disagreement. A design can be single-group and still address a measurement question, but its interpretation depends on how the pairs are collected and analyzed.
The registry identifies ARMSE as the primary accuracy summary. That choice focuses on the magnitude of measurement discrepancies and gives larger errors greater influence. It does not, by itself, describe directional bias or the range of errors an individual might encounter. A complete interpretation would benefit from the estimate’s uncertainty and additional information about the distribution of paired differences. Those results are not reported in the registry record.
The record also illustrates why enrollment and analysis denominators are not interchangeable. An enrollment entry does not tell readers how many complete sensor-reference pairs contributed to the endpoint, especially when the number of readings and handling of repeated observations are not reported. Those details affect the interpretation of both the point estimate and its precision.
Finally, a completed registration and a clearly stated endpoint do not substitute for posted findings. The record offers enough information to identify the intended measurement comparison, but not enough to evaluate the accuracy achieved. Keeping design description separate from results is essential when interpreting a device study.
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.
Important limitations for interpretation arise from the lack of reported outcome results. The registry record does not provide an ARMSE estimate, confidence interval, or other precision measure. It therefore does not permit a quantitative assessment of the sensor-reference differences.
The record does not report the number of paired readings, timing and matching procedures, whether participants contributed repeated readings, or the treatment of correlated observations. It also does not describe missing pairs, exclusions, withdrawals, or the primary analysis population. These omissions limit assessment of the validity and representativeness of any accuracy estimate.
The study is registered as single-group, with all subjects receiving the sensor, and allocation is not applicable. There is no reported control group for a randomized comparison. The blood-sample value is the reference measurement, but the registry record does not provide enough detail to evaluate the reference procedure or the conditions under which each pair was obtained.
The masking field is listed as none, and further procedures concerning awareness of measurements are not reported. The statistical analysis plan, sample-size justification, and any interim analyses are also not reported in the registry record. The listed enrollment cannot be assumed to equal the number analyzed for the primary outcome.
The safety entries are limited to the recorded counts and do not include detailed event definitions or ascertainment procedures. These limitations do not prove that the study was conducted inadequately; they mean the registry information alone does not allow those aspects to be assessed.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Accuracy of Noninvasive Pulse Oximeter Measurement of Hemoglobin for Rainbow DCI Sensor — NCT03124901.
- Registry record last updated 2017-06-12; status last verified 2017-05.
- Results first posted 2017-06-12.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
The registry describes a completed, single-group device study sponsored by Masimo Corporation. All subjects were assigned to the test group and received the DCI pulse oximeter sensor. The study included surgery and healthy-volunteer contexts involving fluid-status changes.
The primary endpoint was sensor accuracy, assessed by comparing noninvasive hemoglobin readings with blood-sample values and calculating arithmetic root mean square error. The unit is g/dL, and the timeframe is up to 24 hours. Enrollment is recorded as 147; the start date is 2008-09-30, and the primary completion date is 2008-10-27.
The registry record does not report the primary accuracy result, its uncertainty, or an analysis population. It reports 0 affected for serious adverse events, 0 affected for other adverse events, and 0 deaths in the listed test group. These entries do not replace the missing accuracy analysis or establish comparative results.