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 Continuous Glucose Monitoring With a Subcutaneous Sensor During Critical Illness and Surgery (NCT00694473), a na na study. The primary outcome is Sensitivity and Specificity of Navigator Threshold Alarms for Hypoglycemia (<60 mg/dl)..
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT00694473 |
| Phase | NA |
| Condition | Critical Illness |
| Allocation | Not applicable (single group) |
| Design model | Single Group |
| Masking | None (open label) |
| Enrollment | 76 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Massachusetts General Hospital |
| Results posted | Yes |
2. Clinical Question
Population
Surgical and ICU patients with critical illness were the population of interest. The registry describes continuous glucose monitoring in these settings, but does not report further eligibility details or a more specific analysis population.
Intervention
Participants received continuous monitoring with the Freestyle Navigator, a subcutaneous sensor system. The arm description specifies monitoring for 72 hours or until discharge from the ICU.
Comparator
There was no separate control arm. For the intended assessment of sensor performance, Navigator readings were to be compared with reference blood glucose values. That measurement comparison is not a randomized treatment comparison.
Primary question
The primary outcome concerned sensitivity and specificity of Navigator threshold alarms for hypoglycemia, defined as blood glucose <60 mg/dl, over 72 hours. The registry says this measure was not calculated: there was not enough hypoglycemia to calculate sensitivity, and blood glucose check times were discordant for calculating specificity.
The study’s broader stated hypothesis was that interstitial fluid glucose measurements from the Navigator would correlate with blood glucose values with a clinically useful degree of accuracy. This is a device-measurement question: it asks how sensor readings and alarms correspond with reference blood glucose measurements during critical illness and surgery. It does not test whether a treatment improves a clinical outcome. The registry lists accuracy and alarm-related outcomes among the secondary measures, but numerical results for those measures are not reported in the registry record.
3. Trial Design
The registry identifies NCT00694473 as a completed, single-group study sponsored by Massachusetts General Hospital. Its listed enrollment is 76, and the condition is critical illness. The study began in 2008-06 and reached primary completion in 2010-12. The phase and allocation are listed as NA, and masking is listed as NONE.
The only intervention arm is Freestyle Navigator, a device intervention. Its description says participants were monitored continuously for 72 hours or until discharge from the ICU. The registry does not describe a parallel control group, randomized assignment, or a like-for-like control comparison. The intended comparison of sensor readings with reference blood glucose values is a comparison between measurement methods, not between assigned treatment groups.
The registry lists one primary outcome, concerning sensitivity and specificity of threshold alarms for hypoglycemia (<60 mg/dl), with a 72-hour timeframe. It also lists seven secondary outcome measures, including alarm performance at glucose thresholds, accuracy against reference blood glucose values, readings across blood sugar ranges, and accuracy by category. Numerical results for these outcomes are not reported in the registry record. The primary measure is explicitly described as not calculated.
The registry does not report a sample-size justification, a detailed statistical analysis plan, or specific procedures for participant selection, sensor placement, or reference-measurement timing. The enrollment figure is a registry description of the study, not evidence by itself of the number contributing data to each outcome. The record does not state how many participants had usable data for the primary or secondary measures.
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 | Freestyle Navigator |
|---|---|
| STARTED | 76 |
| COMPLETED | 69 |
| NOT COMPLETED | 7 |
Analysis population for the primary outcome, as defined in the registry: This measure was not calculated as we did not have sufficient data to calculate sensitivity (not enough hypoglycemia) and discordant BG check times to calculate specificity.
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: 76 were enrolled, analysis was completed for 69
| Characteristic | Unit | Freestyle Navigator |
|---|---|---|
| Participants analysed | 69 | |
| Age, Categorical: <=18 years | Participants | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 45 |
| Age, Categorical: >=65 years | Participants | 24 |
| Age, Continuous | years | 59 (SD 15.2) |
| Sex: Female, Male: Female | Participants | 31 |
| Sex: Female, Male: Male | Participants | 38 |
| Region of Enrollment: United States | Participants | 69 |
6. Primary Endpoint
The registry lists the primary outcome as Sensitivity and Specificity of Navigator Threshold Alarms for Hypoglycemia (<60 mg/dl)., measured over 72 hours.
| Element | Registry entry |
|---|---|
| Result type | N/A |
| Measure of spread | not applicable |
| Unit | not stated |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint was intended to assess sensitivity and specificity of Navigator threshold alarms for hypoglycemia, defined as reference blood glucose <60 mg/dl, during the 72-hour timeframe. Sensitivity describes how often an alarm identifies a low-glucose event when the reference measurement meets the low-glucose definition. Specificity describes how often an alarm does not identify such an event when the reference measurement does not meet that definition. These are general definitions of diagnostic-performance measures.
Such measures are used to characterize how an alarm corresponds to a prespecified reference definition. They address different kinds of classification error, so reporting both can give a more complete account of alarm behavior than either measure alone. They do not, by themselves, establish that use of the device changes clinical outcomes.
In this study, the registry states that the primary measure was not calculated. It gives two reasons: insufficient hypoglycemia to calculate sensitivity and discordant blood glucose check times for calculating specificity. Consequently, the registry record contains no primary-endpoint sensitivity or specificity estimate, and it does not report confidence intervals for those measures. The endpoint’s stated purpose can be explained, but its observed performance cannot be quantified from the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The study addresses whether continuous interstitial-fluid glucose measurements and Navigator threshold alarms correspond to reference blood glucose values in surgical and ICU patients. Its registry hypothesis describes a clinically useful degree of accuracy as the goal, but the record does not establish that this goal was met. The primary measure is explicitly marked as not calculated, so there is no reported sensitivity or specificity estimate for hypoglycemia alarms.
The reason given for not calculating sensitivity is insufficient hypoglycemia. The reason given for not calculating specificity is discordant blood glucose check times. These are distinct data limitations: the first concerns whether enough low-glucose events were available for the event-detection measure; the second concerns the timing alignment needed for evaluating non-event classifications. The registry does not provide event counts, paired-measurement counts, or another quantitative account of these limitations.
The registry lists secondary outcomes related to hyperglycemia alarms, accuracy against reference blood glucose, projected low and high blood sugar alarms, ranges of readings, and accuracy categories. However, it does not report numerical findings for those outcomes. The record’s list of intended outcome measures should not be read as evidence that each measure was analyzed or that any particular performance level was observed. It lists zero analyses, and no numerical secondary results are included in the registry record.
Because the study is single-group, it does not offer a randomized comparison of the device with a separate control arm. Comparing sensor values with reference blood glucose values can address agreement or alarm classification, but it is not a treatment-effect comparison. The posted information therefore cannot establish that using the device causes a change in patient outcomes, nor can it show comparative effectiveness against another monitoring strategy.
The enrollment listed in the registry is 76, while the safety record lists 69 participants at risk in the Freestyle Navigator group. The registry does not explain that difference, and it should not be assumed that all enrolled participants contributed to every endpoint. Likewise, the reported safety figures do not resolve the missing primary accuracy estimates. They describe adverse-event reporting, a separate topic from sensor performance.
Any sample-size figures used for planning are illustrations rather than observed power. The listed planning assumptions use a two-arm design with equal allocation, whereas the registered study is single-group and does not have a randomized control comparison. Those assumptions do not describe the actual design or supply evidence about the precision of the uncalculated primary measure. Overall, the registry record documents the study’s aim, design, and reason the primary measure was not calculated, but leaves the magnitude and precision of sensor accuracy unreported.
9. Statistical Methods Explained
Statistical interpretation depends on separating a study’s question, its measurement plan, and the results that were actually calculated. This registry describes a device-measurement study whose stated aim was to compare continuous interstitial-fluid glucose readings with blood glucose values. The primary outcome was framed as alarm sensitivity and specificity for a low-glucose threshold. The registry states that this outcome was not calculated, so statistical definitions can clarify what the endpoint was intended to measure, but cannot supply an estimate of performance.
What sensitivity and specificity mean
General explanation: sensitivity is the proportion of reference-defined events that an alarm detects. In this study’s primary outcome, the reference event was hypoglycemia defined as blood glucose <60 mg/dl. A sensitivity estimate would therefore require enough reference-defined low-glucose events and corresponding alarm information. The registry explicitly says there was not enough hypoglycemia to calculate sensitivity.
General explanation: specificity is the proportion of reference observations that do not meet the event definition and are also not identified by the alarm. For a measurement comparison, the alarm and reference value need to correspond to the same relevant time point or a clearly defined time window. The registry says discordant blood glucose check times prevented calculation of specificity. It does not describe the timing rules or the amount of usable paired data.
Why event frequency and timing matter
General explanation: an event-based performance measure relies on data that can be classified against a reference definition. If the event of interest is too sparse, the estimate may not be calculable or may have limited precision. This is particularly important for sensitivity, which depends on observations meeting the event definition. Specificity also depends on correctly pairing the test and reference observations. If measurements are not aligned in time, an apparent disagreement can reflect timing rather than the device’s classification performance. The registry gives these data limitations but does not report counts or estimates that would allow their magnitude to be assessed.
General explanation: a confidence interval describes uncertainty around an estimate under the statistical method used. It is not a guarantee about the true value, and its width depends on the amount and structure of the data as well as the method. Since the registry does not report a calculated primary estimate, it also provides no primary confidence interval to interpret. No precision claim about the primary alarm measure can be drawn from the record.
Measurement comparison is not a randomized comparison
General explanation: agreement, correlation, and diagnostic classification answer different questions. Correlation describes whether values tend to vary together; it does not necessarily show that measurements are close enough to be interchangeable. Diagnostic sensitivity and specificity assess classification relative to a specified reference threshold. The study’s hypothesis refers to correlation with blood glucose values, while its primary outcome concerns alarm classification at a low-glucose threshold. These are related but not identical statistical questions.
The registry identifies a single group receiving continuous monitoring and does not describe a separate control arm. The Navigator-to-reference comparison is therefore a comparison of measurements, not a randomized estimate of a treatment contrast. A control-group method would not be justified by this design as described. The registry record does not report a detailed analysis plan specifying how correlation, agreement, repeated readings, or within-participant dependence were handled.
Repeated measurements and data structure
General explanation: continuous monitoring can generate multiple readings from the same participant. Such readings are not necessarily independent, because observations from one participant may be more alike than observations from different participants. Methods that treat all readings as independent can give an inaccurate impression of precision when the data are clustered or repeated. The registry identifies continuous monitoring but does not report the number of readings, the statistical approach to repeated measurements, or whether participant-level clustering was accounted for. No specific method should be presumed.
General explanation: correlation and agreement can also be affected by the distribution of measured values. A narrow range may make a relationship appear weaker even when paired values are close, while a broad range can yield a strong correlation despite meaningful differences between measurement methods. The registry provides no numerical distributions or analysis details for evaluating these issues. It is therefore not possible to reconstruct a correlation or agreement result from the record.
Planning illustrations are not observed power
The planning assumptions described for sample-size illustrations use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect is framed as a stated fraction of an observed effect. For time-to-event illustrations, the specified quantity is total events; other illustrations are framed as participants per arm. These are planning assumptions, not the study’s reported design. The actual registry describes a single-group study, and the record does not report that the study was designed as a two-arm randomized comparison.
General explanation: a sample-size calculation is conditional on assumptions about the outcome, effect size, variability, allocation, and analysis method. It is not a measure of what the study actually observed. “Observed power” is not a substitute for reporting an estimate and its uncertainty; an uncalculated endpoint remains uncalculated regardless of a planning illustration. The registry does not report its own sample-size justification, so the assumptions used for a planning illustration should not be attributed to the original protocol.
Finally, no recalculated result is reported for the primary endpoint. A recalculation from summary statistics would still require posted statistics describing the endpoint, and the registry says it was not calculated. The statistically faithful reading is therefore to describe the intended measures and the stated reasons for noncalculation, while leaving performance and precision unresolved.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists a primary measure and seven secondary outcome measures. The secondary measures span alarm thresholds, comparisons with reference blood glucose values, blood sugar ranges, and accuracy categories. When a study examines multiple outcomes, there is a general statistical question about how to interpret the chance of apparently notable findings across those outcomes. The registry record does not report a multiplicity adjustment, a hierarchy among secondary outcomes, or a prespecified approach for interpreting multiple tests. It would be inappropriate to assume that an adjustment was or was not used beyond what the record states.
The primary hypoglycemia-alarm measure was not calculated because the registry says there was insufficient hypoglycemia for sensitivity and discordant blood glucose check times for specificity. The record does not describe a general missing-data method, how unavailable readings were handled, or whether incomplete monitoring affected particular outcome calculations. These details are not reported in the registry record.
The registry lists enrollment of 76, but does not state the analysis population for each outcome or how many participants contributed usable data to each analysis. The adverse-event group lists 69 participants at risk; that figure should not be treated as the denominator for accuracy outcomes. The record does not explain the relationship between the enrollment and safety denominators.
The registry lists zero analyses and does not include numerical results for the secondary outcomes. Consequently, the record does not allow assessment of whether all planned outcomes were analyzed, whether outcome availability varied, or how any multiplicity considerations affected interpretation. The registry also does not report interim analyses or stopping rules.
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 |
|---|---|---|---|
| Freestyle Navigator | 0 | 69 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Freestyle Navigator | 1 of 69 |
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 |
|---|---|
| Freestyle Navigator | 5 of 69 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Freestyle Navigator |
|---|---|---|
| Bleeding at Insertion Site | Injury, poisoning and procedural complications | 4 of 69 |
| Culture proven bacteremia | Blood and lymphatic system disorders | 1 of 69 |
The safety record lists a Freestyle Navigator group with 69 participants at risk. It reports 0 participants affected by serious adverse events, 5 affected by other adverse events, and 1 death. These are registry-reported safety figures and should be read in their listed categories rather than combined with the device-accuracy outcomes.
The record does not provide serious adverse-event terms; the list of top serious terms is empty. It also does not report details in this summary about the nature, timing, severity beyond the serious-versus-other classification, or relatedness of the other adverse events or the death. The figures alone do not establish whether an event was caused by the device or by the circumstances of critical illness and surgery. No causal conclusion should be drawn from these counts.
The safety denominator is 69, while the listed enrollment is 76. The registry record does not explain the difference. The safety figures should not be assumed to describe the same participants who contributed to a particular glucose-monitoring analysis, nor should the safety denominator be substituted for an outcome-specific analysis population.
12. What the Result Does, and Does Not, Mean
The registry establishes that the study was intended to evaluate continuous glucose monitoring with the Freestyle Navigator in surgical and ICU patients, with blood glucose values serving as a reference for assessing sensor readings and alarms. It also establishes that the primary hypoglycemia-alarm sensitivity and specificity measure was not calculated. Insufficient hypoglycemia prevented calculation of sensitivity, and discordant blood glucose check times prevented calculation of specificity.
That means the posted record does not provide a numerical answer to the primary alarm-performance question. It does not report an estimate of sensitivity or specificity, or a confidence interval indicating the precision of either measure. The listed secondary outcomes likewise have no numerical results in the record. The study’s stated hypothesis about clinically useful accuracy is an aim, not a reported finding.
The record also does not establish that the device did or did not improve a clinical outcome. The design is single-group, without a separate control arm, and the stated outcomes concern sensor measurements and alarm performance. A comparison between readings and a reference measurement is not a randomized comparison of care strategies. The record cannot support a causal claim about patient outcomes.
The safety figures are a separate part of the record: 0 serious adverse events among 69 at risk, 5 affected by other adverse events, and 1 death are listed. These figures do not determine sensor accuracy and, without event details or relatedness information, do not establish causation.
In short, the registry record documents the study’s aim, intervention, and reason the primary measure was not calculated. It leaves the degree of sensor accuracy, the precision of any accuracy estimates, and the numerical results for the listed secondary outcomes unreported. Those questions remain unresolved by the posted results.
13. Why This Trial Matters Statistically
This study illustrates a central issue in diagnostic and device evaluation: a clear measurement question does not guarantee that the planned performance statistic can be estimated from the collected data. Sensitivity depends on having reference-defined events to evaluate. The registry says hypoglycemia was not frequent enough for the primary sensitivity calculation. Specificity requires appropriately aligned test and reference observations; the registry identifies discordant blood glucose check times as the obstacle to that calculation.
The record also distinguishes a measurement study from a treatment comparison. The Navigator readings were intended to be evaluated against reference blood glucose values, while the design is listed as single-group. That can frame questions about correspondence and alarm classification, but it does not create a randomized control comparison or support a causal estimate of clinical benefit.
Statistically, the outcome labels alone are not results. The registry lists several secondary measures, but numerical findings are absent, and the primary measure is marked as not calculated. Careful interpretation therefore requires separating planned endpoints from estimates that were actually reported. It also requires resisting the temptation to infer performance from the study’s stated hypothesis or from the existence of a continuous monitoring intervention.
The study further highlights why analysis details matter when measurements are repeated and reference checks occur at particular times. General statistical methods for repeated observations and time alignment can affect estimates, but the registry does not report the relevant analytic procedures. Without those details and numerical results, the posted record supports a description of the question and limitations, rather than a quantified conclusion about accuracy.
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 principal limitation for interpreting the primary outcome is explicit: sensitivity and specificity were not calculated. The registry attributes this to insufficient hypoglycemia for sensitivity and discordant blood glucose check times for specificity. Thus, the outcome that most directly addresses low-glucose alarm performance has no reported estimate or precision measure.
The record does not provide numerical results for the listed secondary outcomes. Although those measures include alarm performance and accuracy against reference blood glucose values, their presence in the outcome list does not establish that results were available. The registry’s analysis count is zero, and the record does not supply participant-level or measurement-level analysis counts.
The single-group design has no separate control arm, so the study is not a randomized comparison of the device with another monitoring approach. The measurement comparison with reference blood glucose is distinct from a causal comparison of treatment strategies. The registry also does not report a detailed statistical analysis plan, a sample-size justification, or methods for handling repeated readings and missing observations.
Enrollment is listed as 76, whereas the safety group lists 69 participants at risk. The record does not explain this difference or identify the population contributing to each outcome. Safety details are limited: serious-event terms are not listed, and the record does not give event timing or relatedness in the information summarized. These limitations restrict interpretation of both performance and safety, without implying that unreported findings were favorable or unfavorable.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Continuous Glucose Monitoring With a Subcutaneous Sensor During Critical Illness and Surgery — NCT00694473.
- Registry record last updated 2018-03-08; status last verified 2018-02.
- Results first posted 2017-10-06.
- 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
NCT00694473 is a completed, single-group device study sponsored by Massachusetts General Hospital. The registry lists enrollment of 76, the condition critical illness, a start date of 2008-06, and primary completion in 2010-12. Participants received continuous monitoring with the Freestyle Navigator for 72 hours or until discharge from the ICU.
The primary outcome was sensitivity and specificity of threshold alarms for hypoglycemia (<60 mg/dl) over 72 hours. The registry says it was not calculated because there was insufficient hypoglycemia for sensitivity and discordant blood glucose check times for specificity. Numerical results for the listed secondary outcomes are not reported in the registry record.
The safety record lists 69 participants at risk, 0 affected by serious adverse events, 5 affected by other adverse events, and 1 death. The record does not explain the difference between enrollment and the safety denominator or report further serious-event terms.