Feasibility Study to Evaluate Safety and Device Performance of the Hospital Glucose Management System (HGMS): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Feasibility Study to Evaluate Safety and Device Performance of the Hospital Glucose Management System (HGMS) (NCT01763567), a registered phase not stated study. The primary outcome is Device Performance: Accuracy of HGMS.

2. Clinical Question

The central question is one of device performance rather than a randomized test of clinical benefit. The registered primary outcome is accuracy, expressed as MARD in percent difference, with the mean and standard deviation specified as the summary measures. The endpoint timeframe is up to 72 hours. The registry record describes the purpose as assessing safety and device performance; it does not report a treatment-effect question or a clinical outcome comparison.

The iSTAT measurement serves as the reference in the stated accuracy calculation. That use does not make the study a randomized comparison of HGMS against a separate treatment or care strategy. Details of the clinical setting, measurement schedule, and how participants were selected are not reported in the registry record. Those omissions limit how narrowly the question can be framed beyond the registered population, device, reference measurements, and endpoint.

3. Trial Design

NCT01763567 is titled a feasibility study evaluating the safety and device performance of the Hospital Glucose Management System. Its official title describes a multi-phased, multi-center study in critically ill adult patients. The registry record lists hyperglycemia as the condition, a total enrollment of 19, and a completed status. The study start is recorded as 2012-03, with primary completion in 2012-08.

The stated purpose was to assess safety and device performance for up to 72 hours. The primary outcome concerns HGMS accuracy, measured by MARD using sensor glucose and iSTAT glucose values. The registry record lists five outcome measures and five analyses, but does not give the numerical primary endpoint result or the values for the other analyses in the information available for this account.

The record explicitly identifies this as not a randomized comparison and states that allocation is not stated. One safety group is named Observation Arm. The arm information does not establish a randomized treatment-control structure, and the record warns that arms are not a like-for-like control comparison. No comparative treatment interpretation should be inferred from the arm label or the accuracy reference measurement.

Further operational details are not reported in the registry record. These include the number of centers, the phases and their sequence, participant allocation procedures, and the timing or frequency of measurements. The record also does not describe a comparator arm suitable for causal comparison. Accordingly, the design is best characterized from the stated feasibility purpose, registered population, device-performance endpoint, enrollment, and non-randomized-comparison status, without filling in unreported protocol details.

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

Analysis population for the primary outcome, as defined in the registry: To assess safety and device performance for the Hospital Glucose Managment system

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 Device Performance: Accuracy of HGMS, measured over up to 72 hours.

The primary endpoint is titled Device Performance: Accuracy of HGMS. It is a continuous measure summarized as a mean with a standard deviation, using the unit percent difference and a timeframe of up to 72 hours. The registry record defines MARD using the absolute difference of the relative difference between sensor glucose values and iSTAT glucose values, with iSTAT values in the denominator. The description identifies iSTAT as the comparison measurement for the calculation.

In general, an absolute relative difference expresses the size of disagreement without allowing positive and negative deviations to cancel each other. Averaging those absolute relative differences gives an overall measure of the discrepancy between the sensor and reference measurements. A lower value would represent less disagreement under this definition, but the registry record does not report the observed MARD, so the accuracy level in this study cannot be characterized from the record.

The mean communicates the average discrepancy across the observations included in the analysis. The standard deviation describes how much individual measured discrepancies vary around that mean; it is not a confidence interval and does not, by itself, show the precision of the estimated mean. The number of observations contributing to the endpoint and the exact analysis population are not reported in the registry record. Thus the endpoint definition is clear at a high level, while its numerical result and the detail needed to assess its precision are unavailable.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary measure represent?

The registered primary measure is MARD, a continuous summary of the absolute relative differences between HGMS sensor glucose values and iSTAT glucose values. The registry record describes the calculation as the absolute difference of the relative difference, with sensor glucose values compared against iSTAT values and the iSTAT value used as the denominator. The unit is percent difference. Because the absolute value is used, deviations on either side of the reference contribute as discrepancy rather than cancelling one another.

In general, a continuous endpoint can take many values across its measurement scale. Reporting a mean gives a single summary of the observed values, while the standard deviation describes their spread. The mean and standard deviation answer different questions: the mean concerns the average discrepancy, and the standard deviation concerns variation among the measured discrepancies. Neither statistic is reported numerically for this endpoint in the registry record.

How should the mean and standard deviation be read?

In general, the mean is useful for describing the central level of a continuous outcome, but it can conceal variation among observations. A standard deviation helps show whether measurements cluster relatively closely around the mean or are more dispersed. It is not an estimate of uncertainty in the mean. A confidence interval is a different kind of summary: in general, it describes uncertainty around an estimated quantity under the assumptions of a specified statistical procedure. A confidence interval for MARD is not reported in the registry record.

For repeated glucose measurements, interpretation can also depend on how observations are distributed across participants and time. If some participants contribute more measurements than others, a pooled mean could give more weight to those participants unless the analysis accounts for that structure. This is a general consideration, not a claim about the analysis used in this study. The measurement schedule, number of paired readings, and treatment of repeated observations are not reported in the registry record.

What can be learned from the analysis description?

The record states that the primary parameter is a mean and that dispersion is summarized by a standard deviation. It does not report the estimate, standard deviation, confidence interval, test statistic, or p-value for the primary outcome. Without those results, it is not possible to describe the observed accuracy, compare it with a threshold, or quantify how precisely the mean was estimated. No accuracy threshold or decision criterion is reported in the registry record.

The record lists five outcome measures and five analyses. A count of measures and analyses does not show which analysis corresponds to each outcome, whether analyses were confirmatory or exploratory, or whether adjustments were made for multiple testing. Those details are not reported in the registry record. In general, when several outcomes or analyses are considered, the interpretation of isolated statistical tests may depend on a prespecified analysis plan and on how multiplicity is handled. No such plan is described in the record available for this account.

How should planning calculations be understood?

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. They define the effect to detect as a stated fraction of the observed effect. The assumptions also distinguish time-to-event calculations, which use a total number of events under the Schoenfeld method, from other calculations, which are expressed as participants per arm. These are planning conventions, not descriptions of the trial's own design or evidence of its achieved statistical power.

The illustrations are described as based on the observed effect, but the observed primary endpoint value is not reported in the registry record. Numerical planning estimates cannot therefore be interpreted from the registry information. Such estimates would be hypothetical planning quantities, not observed enrollment outcomes and not observed power. The equal-allocation assumption also does not establish that this study actually assigned participants equally between two arms; the record states that allocation is not stated and that the study is not a randomized comparison.

Why does over-dispersion matter in general?

In general, over-dispersion means that observed variability is greater than a statistical model expects. If variability is underestimated, standard errors may be too small and uncertainty may appear narrower than it is. For repeated measurements, clustering within participants can also make observations less independent than a simple analysis assumes. These are general statistical issues for continuous, repeated measurements. The registry record does not report a model, a dispersion assessment, or whether within-participant dependence was addressed, so no conclusion about the analysis in this study can be drawn on those points.

For a feasibility study, the endpoint definition and the completeness of measurement data are important alongside the point estimate. In general, missing observations can affect both the number of usable paired readings and the composition of the analyzed sample. The record does not describe missing-data handling or the analysis population for MARD. Therefore, even if a summary estimate were available elsewhere, the registry information would not establish how missing measurements or repeated observations were handled.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Oberservation Arm")
n   <- c(19)
m   <- c(12.8)
sp  <- c(9.5)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Oberservation Arm  19  12.800  9.500  8.221  17.379

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists five outcome measures and five analyses. It does not report how those measures and analyses were organized, which were primary beyond the stated device-accuracy endpoint, or whether any outcomes were designated as secondary or exploratory. A count alone does not establish the number of statistical tests performed or whether results were selected for emphasis.

In general, considering multiple outcomes can increase the chance that at least one result appears notable by chance if each is assessed separately. Statistical approaches can address this concern, but the choice depends on the analysis plan and the purpose of the study. The multiplicity strategy, any adjustment, and any hierarchy of outcomes are not reported in the registry record. The numerical test results are also not reported, so the effect of any multiplicity approach cannot be evaluated.

The record does not describe how missing data were handled. It does not report the number of participants with evaluable MARD data, the number of missing sensor or iSTAT measurements, reasons for missingness, or any imputation or sensitivity analysis. In general, missingness can affect the composition of an analyzed dataset and the precision of an estimate; its implications depend on why data are missing and how the analysis handles them. No assumptions about those matters can be made from the record.

The analysis population is not described. It is not reported whether all enrolled participants, only participants with measurements, or another defined group contributed to each analysis. The enrollment of 19 should therefore not be treated as the endpoint denominator without supporting information. No protocol-specific statistical analysis plan, interim-analysis information, or sample-size justification is reported in the registry record.

12. 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 record lists an Observation Arm with 19 participants at risk for serious adverse events and 0 affected. It also lists 19 participants at risk for other adverse events and 0 affected. No top serious adverse-event terms are listed. The death count is not reported in the registry record.

These figures describe the recorded safety counts and denominators; they do not supply event-level detail. The registry record does not report the safety observation window, event definitions, severity grading, relatedness assessment, or the process used to collect events. It is also not reported whether the at-risk denominators remained constant throughout follow-up. Without those details, the counts cannot establish the completeness or intensity of safety surveillance.

In general, a zero count means that no events of the specified type are recorded for the stated group in that registry entry. It is not proof that the risk is absent, and it does not provide a precise estimate of how frequently an event might occur in a broader population. Here, the group is an Observation Arm, not a randomized comparator, and the record does not describe comparative safety analyses. The figures should therefore be read as descriptive registry counts rather than evidence of a comparative safety effect.

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

The registry record documents a completed feasibility study of HGMS in critically ill adult patients, with an intended assessment of safety and device performance for up to 72 hours. It identifies MARD, calculated from sensor glucose and iSTAT glucose values, as the primary accuracy endpoint. It also reports an enrollment of 19 and records zero affected participants for serious and other adverse events in the Observation Arm. These details describe the study's stated purpose, endpoint definition, enrollment, and recorded safety counts.

The record does not report the primary MARD result, its standard deviation, or an uncertainty interval. It therefore does not show the magnitude or precision of the observed sensor-reference discrepancy. It also does not report the numerical findings for the other outcome measures and analyses. Without those results, the posted information cannot establish how accurately HGMS performed in the study or how consistent the measurements were.

This is not a randomized comparison, and allocation is not stated. The registry record explicitly cautions that arms are not a like-for-like control comparison. The iSTAT measurements are used in the endpoint calculation and should not be mistaken for a randomized comparison group. The record does not support a causal claim about differences between treatments or care strategies.

In general, a feasibility study can describe whether a device was evaluated in a particular setting and can identify issues relevant to planning further evaluation. The record does not report a decision threshold for acceptable accuracy, a sample-size justification, or details sufficient to assess completeness of measurement data. Consequently, it cannot show whether a prespecified performance target was met or how broadly its findings might apply beyond the enrolled participants.

The safety counts are descriptive and have limited context in the record. Zero recorded events does not establish an absence of risk, and the unreported death count and safety-collection details leave important questions unanswered. The most defensible reading is that the registry documents a feasibility evaluation and its intended measures, while the outcome values and analytical details required for a fuller statistical interpretation are not reported in the registry record.

14. Why This Trial Matters Statistically

This study illustrates why a clearly named endpoint is not enough to interpret device performance. The registry record identifies MARD, defines its relationship to sensor and iSTAT glucose values, specifies percent difference, and states that the mean and standard deviation are the intended summaries. Those details tell readers what quantity was meant to be assessed. But without the numerical estimate, the endpoint definition cannot answer how large the observed discrepancy was or how much it varied.

The distinction between measurement reference and treatment comparator is also important. iSTAT values enter the accuracy calculation, while the record says this was not a randomized comparison and allocation is not stated. Treating the reference measurement as a control arm would confuse an agreement question with a causal comparison. The record's warning that arms are not like-for-like reinforces the need to keep those questions separate.

The study also highlights the importance of reporting the analysis denominator and repeated-measure structure. The record gives total enrollment as 19, but does not report how many participants or paired measurements contributed to the primary endpoint. In general, the number and distribution of observations can affect both the meaning of a pooled average and its precision. Reporting those details would make the endpoint easier to interpret without changing the underlying study question.

Finally, registry safety counts require context. The listed zeros are useful descriptive information, but the record does not state the event-collection window or provide event definitions and relatedness details. Statistical interpretation depends not only on the count, but also on how events were sought and classified. This feasibility record therefore provides a useful example of how endpoint definition, design status, denominators, uncertainty, and reporting completeness all shape what can be concluded from clinical data.

15. 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 information is that the numerical primary endpoint result is not reported in the registry record. The MARD mean and standard deviation are specified, but no values or interval estimates are given. The record also does not give numerical results for the other listed outcome measures and analyses. This prevents assessment of observed accuracy, variability, and statistical precision.

Enrollment is reported as 19, but the endpoint analysis denominator is not reported in the registry record. The number of paired glucose measurements, measurement schedule, and distribution of readings across participants are also not described. These omissions matter for a repeated-measure accuracy endpoint because a participant with many observations could contribute differently from one with fewer observations, depending on the analysis method. The method used is not reported.

The record describes a multi-phased, multi-center study, but the number of centers, phase structure, and phase-specific enrollment are not reported. Eligibility criteria and participant characteristics beyond critically ill adults and the listed condition are not reported in the registry record. This limits assessment of how representative the enrolled group may have been.

The record says the study was not a randomized comparison, allocation is not stated, and arms are not a like-for-like control comparison. It therefore does not provide a basis for causal comparison between arms. The statistical analysis plan, sample-size justification, multiplicity handling, missing-data methods, and analysis populations are not reported in the registry record. Safety-event definitions, ascertainment procedures, and observation windows are also not described. Together, these gaps limit independent evaluation of the endpoint estimates, data completeness, and safety counts.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Feasibility Study to Evaluate Safety and Device Performance of the Hospital Glucose Management System (HGMS) — NCT01763567.
• Registry record last updated 2015-05-07; status last verified 2015-05.
• Results first posted 2015-05-07.
• 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.

19. Record Summary

NCT01763567 is a completed feasibility study titled Feasibility Study to Evaluate Safety and Device Performance of the Hospital Glucose Management System (HGMS). The official title describes a multi-phased, multi-center evaluation of the Medtronic Hospital Glucose Management System in critically ill adult patients. Hyperglycemia is the listed condition. The record reports enrollment of 19, a start of 2012-03, and primary completion in 2012-08.

The stated purpose was to assess safety and device performance for up to 72 hours. The primary endpoint is HGMS accuracy, measured by MARD from sensor and iSTAT glucose values and summarized as a mean with standard deviation. The record lists five outcome measures and five analyses but does not report the numerical primary endpoint result or other outcome values in the information available for this account.

The study is identified as not a randomized comparison; allocation is not stated, and arms are not described as a like-for-like control comparison. The Observation Arm safety entry lists 0 serious adverse events and 0 other adverse events among 19 participants at risk. Deaths are not reported in the registry record.