Trial Between a Computer-Guided Insulin Infusion Protocol Versus a Standard Insulin Infusion Algorithm in Medical ICU: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Trial Between a Computer-Guided Insulin Infusion Protocol Versus a Standard Insulin Infusion Algorithm in Medical ICU (NCT00394524), a randomized na study. The primary outcome is Mean Blood Glucose (BG) in mg/dl Among Glucommander Group Compared to Standard Insulin Infusion.

2. Clinical Question

The registry record describes a comparison of two approaches to continuous intravenous insulin infusion for patients hospitalized in a medical intensive care unit with diabetes or new hyperglycemia. The central question is whether mean blood glucose during insulin infusion differed between the computer-guided Glucommander approach and a standard paper-form insulin infusion algorithm.

The registry record also lists severe hypoglycemia and ICU and hospital length of stay as secondary outcomes. These outcomes address other aspects of the observed clinical course; they do not change the definition of the primary question. The registry record does not report additional eligibility criteria, a detailed glucose-target definition, or a fuller description of how patients entered the outcome analyses.

3. Trial Design

The study is recorded as a randomized, parallel-group trial with no masking. Its brief summary describes it as multicenter, prospective, and open-label. The comparison was between computer-guided continuous insulin infusion using Glucommander and a standard insulin infusion algorithm. The arm descriptions indicate intravenous insulin in both groups and identify Glulisine as the insulin used in both.

The registry lists enrollment as 160. For the primary blood-glucose outcome, the posted groups are Glucommander, with 77 participants, and standard insulin infusion, with 76 participants. These are the group counts attached to that outcome; the registry record does not explain their relationship to the listed enrollment. Randomization details, such as sequence generation and allocation concealment, are not reported in the registry record. The record also does not report whether there were site-specific procedures or stratification factors.

The study status is completed. The registry dates the start to 2006-06 and primary completion to 2008-09. The phase is recorded as NA. The sponsor is Emory University. These administrative details identify the registered study but do not supply information about the conduct or quality of the statistical analysis.

The trial was unmasked, so participants and study personnel were not described as blinded. The registry record does not say whether outcome assessors or analysts were masked. It does not report the original sample-size justification, planned analysis population, treatment of missing observations, or prespecified statistical model. Those omissions matter when interpreting group summaries: the posted means and comparisons describe the results as recorded, but do not establish which analysis procedures were planned in advance.

The primary outcome is a continuous measure, reported as a mean and standard deviation in each group. The secondary outcomes include a participant count for severe hypoglycemia and means for ICU and hospital length of stay. The registry record lists three secondary outcomes and no reported analyses for them. Thus, the record permits a descriptive comparison of the posted summaries, while leaving important details about the full statistical plan and analysis conduct unreported.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Of 160 enrolled, 80 were randomized to each group. Among the 80 subjects in glucommander group, 1 died prior to starting infusion and 2 were excluded because of meeting exclusion criteria for the protocol. Among 80 subject in standard group 1 withdrew consent prior to starting infusion and 3 met exclusion criteria. 153 were analyzed

Overall Study

Reasons for not completing this period:

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 Mean Blood Glucose (BG) in mg/dl Among Glucommander Group Compared to Standard Insulin Infusion, measured over First 10 days of ICU stay.

The primary endpoint is mean blood glucose in mg/dl among patients receiving Glucommander compared with patients receiving standard insulin infusion. The timeframe is the first 10 days of ICU stay. The endpoint description says that daily mean blood glucose concentrations during insulin infusion were measured each day up until 10 days, and that a mean value was calculated. It also states that blood glucose was measured once target levels were achieved after admission.

This endpoint represents glucose concentration as a continuous measurement. A group mean describes the average observed concentration in that group; the standard deviation describes how dispersed individual observations are around that mean. These summaries do not show each patient’s glucose pattern, the timing of individual measurements, or how often measurements were taken for each patient.

As a general statistical point, a continuous mean can summarize the overall level of a measurement, but it may not capture variation over time or unusual values. The registry record does not explain how daily means were combined into the group-level results, how repeated measurements were handled, or whether the analysis accounted for differences in measurement frequency. The endpoint description gives the target and timeframe in broad terms, but a more detailed operational definition of target achievement is not reported in the registry record.

The endpoint is used to compare average blood glucose under the two infusion approaches. The posted means were 103.3 mg/dl with Glucommander and 117.3 mg/dl with standard insulin infusion. This is a statistical description of the registered outcome, not by itself a complete assessment of safety, patient experience, or other clinical outcomes. Those questions require evidence beyond the primary mean.

7. Results Posted in the Registry

• Glucommander: mean 103.30 versus 117.30 on control; difference -14.00 (95% CI -18.20 to -9.80), standardized effect d = -1.06; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Number of Patients With Severe Hypoglycemia Episodes Among the Glucommander Group Compared to Standard Algorithm

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Computer Assisted IV Insulin Infusion: 3.9% (3/77) versus 5.3% (4/76) on control; risk difference -1.4 percentage points (95% CI -8.0 to 5.3), risk ratio 0.74 (0.17–3.20); unadjusted two-proportion z-test p = 0.686.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Computer Assisted IV Insulin Infusion and the control. The interval gives the range of effects that remain compatible with the data.

Mean Length of Intensive Care Unit (ICU) in Days Stay Among Glucommander Group Compared to Standard Insulin Infusion Group

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Glucommander: mean 8.50 versus 13.40 on control; difference -4.90 (95% CI -8.44 to -1.36), standardized effect d = -0.44; z-test p = 0.007.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Glucommander than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Mean Hospital Length of Stay in Days Among the Glucommander Group Compared to Standard Insulin Infusion

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Glucommander: mean 23.90 versus 17.50 on control; difference 6.40 (95% CI -0.37 to 13.17), standardized effect d = 0.30; z-test p = 0.064.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Glucommander and the control. The interval gives the range of effects that remain compatible with the data.

The secondary endpoints add information about severe hypoglycemia and time spent in the ICU and hospital. Their directions are not uniform: the severe-hypoglycemia count and ICU length-of-stay mean are numerically lower in the Glucommander group, while the hospital length-of-stay mean is numerically higher. Their estimates also differ in precision. These outcomes therefore do not form a single consistent pattern that can simply be substituted for the primary blood-glucose comparison.

They relate to the primary result by addressing distinct questions. Mean blood glucose is a continuous measure of glucose concentration during insulin infusion. Severe hypoglycemia is an event count, while ICU and hospital length of stay describe duration of care. A group difference in the primary measure does not automatically determine an event rate or length of stay, and findings on one endpoint do not validate or invalidate the others.

As a general statistical consideration, multiple outcomes create multiple opportunities for apparently notable results. If each is assessed without adjustment, the chance of at least one small p-value can rise, depending on the number and dependence of the tests. The registry record does not report a multiplicity plan, hierarchy among secondary outcomes, or adjusted analyses. The posted secondary p-values are unadjusted, so they carry less weight as confirmatory evidence than results from a prespecified analysis with an appropriate plan for multiple testing.

The record lists three secondary outcomes and no reported analyses. The absence of a reported multiplicity strategy makes it difficult to know whether these outcomes were intended as confirmatory or exploratory. Their results are most appropriately read as additional descriptive comparisons, with attention to their intervals and the limited event counts for severe hypoglycemia. They can inform questions for further study, but the registry record does not establish that any particular secondary difference was a prespecified, adjusted finding.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Number of Patients With Severe Hypoglycemia Episodes Among the Glucommander Group Compared to Standard Algorithm
• Mean Length of Intensive Care Unit (ICU) in Days Stay Among Glucommander Group Compared to Standard Insulin Infusion Group
• Mean Hospital Length of Stay in Days Among the Glucommander Group Compared to Standard Insulin Infusion

9. Statistical Interpretation

10. Statistical Methods Explained

The posted outcomes include means, standard deviations, mean differences, risk differences, intervals, and unadjusted p-values. These quantities answer related but distinct questions. Means describe group averages, differences quantify contrasts between group summaries, intervals communicate uncertainty around estimates, and p-values assess compatibility with a null hypothesis under an analysis model. Their interpretation depends on how the data were collected and analyzed; several relevant details are not reported in the registry record.

What does a mean difference describe?

For a continuous outcome, a mean difference is the mean in one group compared with the mean in the other group. Here, the primary comparison is expressed as the Glucommander mean relative to the standard-insulin-infusion mean. The posted difference is -14.0 mg/dl. Its direction is consistent with the posted group means: 103.3 mg/dl for Glucommander and 117.3 mg/dl for standard insulin infusion. The same type of contrast is reported for ICU and hospital length of stay.

As a general statistical explanation, a mean is informative when the average is an appropriate summary of the observed distribution. It can conceal skew, extreme observations, or distinct patterns among participants. Length of stay can be unevenly distributed, but the registry record supplies means and standard deviations rather than distributional plots or medians. It does not report whether transformations or alternative analyses were considered. Therefore, interpretation should stay with the summaries and comparisons recorded rather than assume that the underlying distributions meet a particular model’s requirements.

What does a standard deviation add?

A standard deviation describes the spread of observations around a group mean. For the primary endpoint, the posted standard deviations are 8.8 mg/dl in the Glucommander group and 16.5 mg/dl in the standard group. For ICU length of stay, they are 7.6 and 13.8 days; for hospital length of stay, they are 26.3 and 15.0 days. These values describe variability within groups, not uncertainty in the difference between group means. The uncertainty of a contrast also depends on the number of observations and on the analysis method.

As a general concept, standard errors quantify the expected variability of an estimate across repeated samples under a statistical model. They are not measures of participant-to-participant spread. The registry comparisons include standard-error values for the continuous outcomes, but the registry record does not report a full analysis specification explaining their role or the procedure used to calculate intervals and p-values.

How should the intervals and p-values be read?

An interval around an estimate represents a range of values that are reasonably compatible with the data and the statistical assumptions used to construct it. Narrower intervals indicate greater precision than wider intervals, in the sense that they leave a smaller range of plausible values. An interval that includes zero for a difference does not rule out no difference under the corresponding method; an interval entirely on one side of zero is inconsistent with zero under that method at its associated confidence level. The confidence level is not stated in the registry record, so no level should be assigned to these intervals.

A p-value is not the probability that the null hypothesis is true, and it is not a measure of the size or practical importance of an effect. The primary comparison has an unadjusted p-value of <0.001; the ICU stay comparison has 0.007; hospital stay has 0.064; and severe hypoglycemia has 0.686. These are reported as unadjusted. The registry record does not report whether they were prespecified, whether a particular hypothesis-testing hierarchy was planned, or whether any adjustments were performed.

What is different about the severe-hypoglycemia comparison?

Severe hypoglycemia is reported as a participant count, so its comparison is expressed as a risk difference rather than a mean difference. The posted counts are 3 participants in the Glucommander group and 4 in the standard group, with outcome group sizes of 77 and 76. The estimated risk difference is -0.014, with an interval from -0.080 to 0.053, and an unadjusted p-value of 0.686. The interval covers values on both sides of zero, indicating substantial uncertainty about the direction and size of the risk difference.

As a general statistical concept, binary outcomes can be sparse when relatively few participants have the event. Sparse outcomes can yield wide intervals and unstable relative measures, so the observed counts alone do not establish that the groups have similar risks. The record also lists relative risk and odds-ratio summaries, but those are not needed to understand the primary endpoint and should not be read as precise when their intervals are wide.

What analysis details remain unknown?

The registry record lists no reported analyses. It does not describe a model for the repeated daily glucose measurements, adjustment for baseline characteristics, handling of missing observations, or a method for accommodating clustering by study center. It also does not report whether participants were analyzed according to randomized assignment or according to the treatment actually received. These are not minor technicalities: analysis choices can affect estimates, uncertainty, and the populations to which results apply.

As a general explanation, repeated measurements from the same participant tend to be related rather than independent. An analysis that treats them as independent can misstate uncertainty; methods that account for within-participant correlation are often considered for repeated data. The registry outcome description indicates daily glucose measurement during the stated timeframe, but the record does not say how such repetition was handled. This general consideration should not be mistaken for evidence that a particular method was or was not used.

11. Confidence Intervals and Precision

The primary mean difference is -14.0 mg/dl, with a reported interval from -18.2 to -9.802 mg/dl. The interval is entirely below zero, indicating that the range of values compatible with the reported comparison points toward a lower mean in the Glucommander group. It also communicates uncertainty about the size of that contrast: the estimate is not an exact, error-free value. The registry record does not state the confidence level or construction method for the interval.

Precision differs across outcomes. The ICU length-of-stay mean difference is -4.9 days, with an interval from -8.437 to -1.363 days. The hospital length-of-stay mean difference is 6.4 days; its reported interval extends from a negative value to 13.17 days, crossing zero. The severe-hypoglycemia risk difference is -0.014, with an interval from -0.080 to 0.053, also spanning zero. Thus, the posted intervals for hospital stay and severe hypoglycemia include values in either direction, while the primary and ICU comparisons have intervals wholly on one side of zero.

As a general statistical explanation, an interval is useful because it shows more than whether an estimate is on one side of zero. It indicates the range of effect sizes that remain compatible with the analysis and its assumptions. A wide interval is less precise and can leave meaningful uncertainty even when the point estimate has a clear direction. Conversely, an interval that does not cross zero does not establish clinical importance or remove concerns about bias, multiplicity, or analysis choices.

The registry record does not report adjusted intervals or a confidence level, and no reported analyses are listed. The precision statements therefore concern the posted comparisons only. They do not account for unreported missing data, repeated-measure structure, or possible differences in the analysis population.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm for the continuous outcomes and the severe-hypoglycemia outcome; they are not total event counts. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect used for planning is a stated fraction of the observed effect. The original trial’s own design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives an illustrative number of participants per arm for a future study to have an 80% chance of detecting the specified effect under those planning assumptions, if that effect and the assumptions are correct. The 90% power column describes the corresponding illustration for a 90% chance. For the primary outcome, planning for 100% of the observed effect gives 14 participants per arm for 80% power and 19 for 90% power. Planning for 75% of the observed effect gives 25 participants per arm for 80% power and 34 for 90% power; planning for 50% gives 56 and 75.

Smaller effects are harder to distinguish from random variation, so the illustrations require more participants per arm as the effect fraction gets smaller. This pattern is visible across the posted planning estimates, including for the secondary outcomes. These are planning illustrations based on the observed effects, not observed power. They should not be used to infer how likely the completed trial was to detect an effect after its results were known.

For the binary outcome, the same per-arm interpretation applies to the listed participant counts. The figures depend strongly on the observed event frequencies and the chosen effect fraction. The registry record does not give the trial’s original target or explain whether its enrollment was selected using a different assumed effect, significance level, or method.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -14.000; pooled standard deviation σ = 13.198.
2. n = 2 × (1.96 + 0.842)² × 13.198² / -14.000² = 14.0, rounded up to 14 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -14                      # observed difference in means
sigma <- 13.1984822526999                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(394524)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 14 (80% power), 19 (90% power)
 75% of observed effect: n per arm = 25 (80% power), 34 (90% power)
 50% of observed effect: n per arm = 56 (80% power), 75 (90% power)
power.t.test: 15 
Simulated power at n = 14 per arm: 0.779 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary endpoint and three secondary endpoints, but does not report a multiplicity adjustment or testing hierarchy. Multiplicity arises when several hypotheses are tested: even when each test is conducted at a conventional threshold, considering many results can increase the chance of at least one apparently strong finding by chance. This is a general statistical issue, not evidence that a particular result in this trial is false. Without a reported plan, the unadjusted p-values should be interpreted as endpoint-specific rather than as controlling the error rate across all outcomes.

The record also lists no reported analyses. It does not state whether the comparisons were prespecified, whether they were adjusted for baseline variables, or whether an intention-to-treat population was used. The record’s enrollment is 160, while the posted primary endpoint groups have outcome counts of 77 and 76; the reason for this difference is not reported in the registry record. No explanation is given for exclusions, missing measurements, withdrawals, or whether all randomized participants contributed to each endpoint.

Missing data can affect both the estimate and its uncertainty. As a general concept, if the availability of an outcome is related to a participant’s health or treatment experience, an analysis limited to observed data may not represent all randomized participants. The record does not report the amount or pattern of missing data, imputation methods, or sensitivity analyses. It is therefore not possible to assess from the registry record whether missingness could have influenced the posted results.

The primary outcome description involves daily measurements during insulin infusion, but the registry record does not specify how repeated measurements were incorporated into a participant-level or group-level analysis. It also does not report handling of differences in measurement schedules or study-center effects. These omissions prevent a complete appraisal of the statistical model and its assumptions. They do not prove that appropriate methods were absent; they mean the methods are not documented in the registry record.

15. 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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: severe hypoglycemia (5.3%).

The registry lists serious adverse events for 3 participants in the Glucommander group, with 77 at risk, and 4 participants in the standard-insulin-infusion group, with 76 at risk. The serious adverse event term listed is severe hypoglycemia at 5.3%. The record also lists other adverse events affecting 42 participants in the Glucommander group and 41 in the standard group. Deaths are listed for 17 and 20 participants, respectively.

These are registry adverse-event counts and should be read with their stated group denominators and event categories. The record does not provide detailed event narratives, timing, severity criteria, attribution to treatment, or an analysis comparing these safety outcomes. The severe-hypoglycemia counts listed as a secondary endpoint are 3 and 4, but the record does not explain how that endpoint definition relates to the adverse-event listing. Counts alone do not establish causation or quantify the uncertainty of a group difference.

As a general statistical point, safety comparisons can be difficult to interpret when events are uncommon or follow-up and ascertainment are not fully described. The registry record does not report those details, so the listed figures are descriptive rather than a complete assessment of comparative safety.

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

The registry results show a lower posted mean blood glucose in the Glucommander group than in the standard-insulin-infusion group during the first 10 days of ICU stay. The mean difference is -14.0 mg/dl, and its reported interval is entirely below zero. This describes the observed primary comparison and its reported precision. It does not by itself establish how clinically important the difference is, because the registry record does not report a threshold for meaningful change.

The secondary findings broaden the statistical picture but do not resolve every question. The ICU length-of-stay comparison has an interval entirely below zero, while the hospital length-of-stay and severe-hypoglycemia intervals span zero. The results are endpoint-specific, and the registry record does not report multiplicity adjustment or a detailed analysis plan. The unadjusted results should therefore not be treated as a set of equally definitive claims.

As a general distinction, statistical evidence about a group mean is not the same as evidence about every individual participant. Group averages can differ even when individual experiences vary widely, and a mean does not describe the full distribution or trajectory of measurements. The posted summary does not show whether the glucose results were consistent across days or across participants. The registry record does not report those details.

The record also does not report enough information to assess the role of missing data, the handling of repeated measures, baseline differences, or the exact analysis population. The study is described as randomized, which is relevant to the design, but randomization alone does not answer questions about execution, attrition, analysis choices, or precision. The listed enrollment and outcome-group counts are both relevant registry figures, but the record does not explain their relationship.

Accordingly, the posted results support a description of the observed mean blood-glucose contrast and the uncertainty captured by its interval. They do not establish a complete account of treatment effects across outcomes, mechanisms, or patient subgroups. Nor do the registry figures alone establish that the intervention caused every observed difference. Interpretation should remain tied to the registered population, outcomes, timeframes, and analysis details that are actually reported.

17. Why This Trial Matters Statistically

This trial is statistically interesting because its primary endpoint is a repeated clinical measurement summarized as a mean, while the secondary outcomes include both a binary event and durations of stay. Those outcome types have different statistical behavior and require different summaries. A mean difference in blood glucose cannot be interpreted in the same way as a count of severe hypoglycemia events or a difference in length of stay.

The primary comparison illustrates why effect size and uncertainty should be read together. The posted difference is -14.0 mg/dl, and the reported interval remains below zero. That gives more information than a p-value alone: it shows the estimated direction and a range of compatible values under the analysis method. Yet the registry record does not report the confidence level, analysis model, or clinical threshold, which limits how far the statistical result can be translated into a broader judgment.

The secondary results also illustrate the importance of multiplicity and precision. Several outcomes are listed, but no multiplicity plan or reported analyses appear in the registry record. The secondary p-values are unadjusted, and some intervals include zero. This is a useful reminder that outcome-by-outcome significance labels can overstate certainty when several endpoints are examined without a documented strategy.

Finally, the primary description includes daily glucose measurements, raising a general methodological question about repeated observations within participants. Correctly representing correlation among repeated measurements can matter for uncertainty estimates. The registry record does not describe how the repeated data were analyzed. This trial therefore highlights both what a registry summary can communicate—group summaries, contrasts, and intervals—and what it may leave unresolved about the analysis behind them.

18. 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.

Several trial-specific limitations of interpretation follow from details absent in the registry record. Although the study is described as randomized, the method for generating the allocation sequence and concealing assignments is not reported in the registry record. The study is open-label, and the record does not say whether outcome assessment or statistical analysis was masked. These gaps make it difficult to evaluate safeguards against selection or measurement bias from the registry information alone.

The primary endpoint is described in terms of daily mean blood glucose during insulin infusion after target levels are achieved, but the record does not specify the exact target, measurement schedule, handling of repeated readings, or method for combining daily measurements. It also does not report baseline glucose values or a baseline-adjusted analysis. Consequently, the posted means cannot establish how the outcome changed from baseline or whether measurement frequency differed by group.

The listed enrollment is 160, while the primary outcome summaries report groups of 77 and 76. The registry record does not explain the relationship between these figures or report participant flow, exclusions, missing data, or endpoint-specific analysis populations. It is therefore not possible to determine from this record whether all randomized participants were included in each analysis or how missing observations may have affected results.

The record lists no reported analyses and does not document a multiplicity plan, model specification, or sample-size justification for the completed trial. Its sample-size illustrations are based on observed effects and are not a substitute for the original design rationale. The registry also gives limited detail on the definitions and ascertainment of adverse events and lengths of stay. These limitations do not demonstrate that the trial procedures were inadequate; they identify information not reported in the registry record and constrain the conclusions available from it.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Trial Between a Computer-Guided Insulin Infusion Protocol Versus a Standard Insulin Infusion Algorithm in Medical ICU — NCT00394524.
• Registry record last updated 2018-10-30; status last verified 2018-10.
• Results first posted 2013-09-30.
• 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.

22. Record Summary

NCT00394524 is a completed, randomized, open-label, parallel-group study described as multicenter and prospective. It compared computer-guided Glucommander continuous intravenous insulin infusion with a standard insulin infusion algorithm among patients hospitalized in a medical ICU with diabetes or new hyperglycemia. The registry lists enrollment as 160 and the phase as NA.

The primary endpoint was mean blood glucose in mg/dl during the first 10 days of ICU stay. The posted group means were 103.3 with Glucommander and 117.3 with standard insulin infusion; the mean difference was -14.0 mg/dl, with an interval from -18.2 to -9.802 and an unadjusted p-value of <0.001. Secondary outcomes covered severe hypoglycemia and ICU and hospital length of stay.

The registry record lists no reported analyses and does not report the original sample-size justification, missing-data procedures, analysis populations, or multiplicity plan. The results therefore describe the posted endpoint summaries and comparisons, with important analysis details remaining unreported in the registry record.