Veterans Inpatient Insulin Study and Transition to Outpatient Therapy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Veterans Inpatient Insulin Study and Transition to Outpatient Therapy (NCT00821795), a randomized phase 4 study. The primary outcome is For Transition From Inpatient to Outpatient Care, Evaluate Glycemic Control in Subjects Randomized to Receive 70/30 NPH/Regular Insulin or Aspart Analog 70/30 Mix Bid- Main Outcome Will be HbA1c During Transition Outpatient Therapy Phase..

2. Clinical Question

The registry’s brief summary frames the question as whether twice-daily aspart protamine 70/30 is as good as twice-daily NPH/Regular 70/30 after discharge. It also identifies safety and treatment during the move from inpatient therapy to outpatient care as research interests. The posted primary outcome, however, is specifically a continuous measure: mean hemoglobin A1c, expressed as % hemoglobin A1c, during the transition outpatient therapy phase.

The comparison is between assigned insulin regimens, not between insulin use and no insulin. The record describes random allocation and a parallel-group design, with no masking. The results therefore address the recorded group means under this study’s circumstances. The registry record does not report a formal equivalence or noninferiority margin, so the stated aim of being “as good as” should not be treated as proof that a formal equivalence criterion was met.

3. Trial Design

This was a randomized, open-label, parallel-group phase 4 study of insulin transition therapy. Its registry status is completed, and the listed enrollment is 120. The registry describes volunteers with diabetes and an illness requiring hospitalization. The arm descriptions specify transition insulin therapy for outpatient glycemic control following hospitalization for a non-critical acute illness.

The two assigned regimens were NPH/Regular 70/30 mix and an aspart insulin analog 70/30 mix. Masking is listed as none, meaning the record does not describe a blinded comparison. The primary endpoint concerned mean hemoglobin A1c during the outpatient transition phase, with a timeframe of 16 weeks. The primary outcome summary lists 44 participants for the NPH/Regular group and 60 for the aspart group. Those outcome denominators are distinct from the safety denominators recorded elsewhere; the registry record does not explain the difference.

The record lists four secondary outcomes: seven-point blood glucose profiles, patient-reported outcomes using the Diabetes Symptom Checklist to evaluate perception of diabetes control, hypoglycemia, and twice-daily blood glucose monitoring using a telehealth transmitter. The registry identifies five outcome measures overall but lists no analyses in its analysis count. This does not establish why outcome results or analyses may be absent; the reason is not reported in the registry record.

The record does not report the original sample-size justification, the planned statistical analysis, or the handling of interim analyses. The planning illustrations associated with the observed result are separate from the trial’s original design assumptions. These omissions limit what can be inferred about how the investigators intended to analyze the trial before results were available.

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

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 For Transition From Inpatient to Outpatient Care, Evaluate Glycemic Control in Subjects Randomized to Receive 70/30 NPH/Regular Insulin or Aspart Analog 70/30 Mix Bid- Main Outcome Will be HbA1c During Transition Outpatient Therapy Phase., measured over 16 weeks.

The primary endpoint is hemoglobin A1c, reported as a mean in % hemoglobin A1c during the outpatient transition therapy phase. The registry gives a timeframe of 16 weeks. The posted group summaries report the mean and standard deviation for each arm.

Hemoglobin A1c is a continuous measure used to describe blood glucose control over a period of time, rather than a single glucose reading. In general, continuous measures can be summarized with a group mean to describe the typical recorded value and a standard deviation to describe spread among individual observations. Those summaries answer different questions: the mean describes the group’s central value, while the standard deviation describes variation within the group.

The registry’s primary question focuses on the group average after transition to outpatient therapy. A mean comparison does not describe every participant’s result, nor does it by itself show the distribution of outcomes or changes from baseline. The record does not report a baseline-adjusted endpoint, a change-from-baseline definition, or a clinically meaningful threshold for interpreting the observed difference. Those details are therefore not part of the posted endpoint description.

7. Results Posted in the Registry

• Aspart Insulin Analog Biphasic Mix: mean 8.35 versus 8.32 on control; difference 0.03 (95% CI -0.63 to 0.69), standardized effect d = 0.02; z-test p = 0.929.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous, so its central summary is a mean rather than a count or proportion. The record also reports a standard deviation for each arm. In general, the standard deviation describes how dispersed individual observations are around their group mean; it is not the uncertainty in the mean difference itself. The uncertainty in the between-group comparison is represented by the confidence interval.

What does the mean difference represent?

The reported difference is 0.03, calculated in the direction of the aspart group minus the NPH/Regular group. The group means are 8.35 and 8.32, respectively. A positive difference in this direction means that the aspart group’s posted mean is higher; a negative difference would mean it is lower. The estimate is small relative to the reported spread, but its meaning should be judged alongside its confidence interval and the study’s clinical context. The registry does not report a threshold for a clinically important difference.

What does the p-value test?

The unadjusted p-value is 0.929. In general, a p-value describes how compatible the observed data are with a specified null hypothesis, under the assumptions of the statistical test. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. A large p-value does not demonstrate equivalence. It indicates that the analysis did not produce strong statistical evidence against its null hypothesis; it does not rule out differences that the study may have estimated imprecisely.

The registry labels this p-value unadjusted. The record does not report the full test specification, whether the analysis adjusted for baseline values or other covariates, or a prespecified primary analysis model. Consequently, the p-value should be interpreted as the posted unadjusted comparison, not as a complete account of every possible analysis.

Why does the confidence interval matter?

A confidence interval expresses uncertainty around an estimate. In general, a narrower interval indicates greater precision, while a wider interval leaves more possible values compatible with the data and analysis. The interval for this mean difference extends below and above zero. That means it includes no difference and also allows for differences in either direction. Its span cautions against interpreting the point estimate alone as an exact measure of the group contrast.

When a confidence interval includes zero, that is consistent with the unadjusted p-value not indicating a statistically detectable difference under a conventional two-sided test. The interval is more informative than a binary significant/not-significant label because it shows the range of uncertainty. However, the registry does not provide a clinical threshold against which to judge whether values within that interval would be important.

What does variation within groups contribute?

The reported standard deviations are 1.5 for NPH/Regular and 1.95 for aspart. In general, greater within-group variation makes it harder to estimate a mean precisely when other features are similar. Here the two arms also have different outcome sample sizes, 44 and 60, so each group mean has its own sampling uncertainty. The record’s comparison presents an interval for the difference, which incorporates uncertainty in both group means under the analysis used for that comparison.

Over-dispersion is a concept most often discussed for count outcomes, where variability exceeds what a simple count model expects. It is not the central issue for this continuous endpoint. For continuous measurements, the relevant practical point is that individual values vary around each group mean, and that variation contributes to uncertainty in the estimated contrast. No special modeling approach for over-dispersion is reported in the registry record.

What does “as good as” require statistically?

In general, a noninferiority analysis asks whether a new treatment is not worse than a comparator by more than a prespecified acceptable margin. An equivalence analysis asks whether the difference falls within prespecified bounds in both directions. Both require a margin and an analysis strategy defined for that purpose. The registry brief summary uses the phrase “as good as,” but the registry record does not report a margin or a formal equivalence or noninferiority analysis. The ordinary difference estimate, p-value, and confidence interval cannot substitute for those missing design features.

The record describes randomized allocation, a parallel model, and no masking. Randomization is intended in general to balance characteristics between groups on average, but the registry does not report baseline characteristics or how allocation was generated or concealed. With no masking, participants and study personnel were not described as blinded. The record does not say whether outcome assessment or analysis was masked.

Finally, the posted comparison should be distinguished from a full analysis plan. The registry does not report a sample-size justification, interim-analysis plan, analysis population, missing-data procedure, or multiplicity adjustment. Those details are not inferable from the numerical result alone.

10. Confidence Intervals and Precision

The estimated mean difference is 0.03, with the contrast defined as aspart minus NPH/Regular. The confidence interval extends on both sides of zero. This indicates that the posted estimate is not precise enough to exclude no difference, and it also leaves uncertainty about the direction and size of the underlying group contrast. The registry reports an unadjusted p-value of 0.929, which is consistent with the interval’s inclusion of zero.

Precision is different from whether an estimate is statistically significant. The point estimate gives one value for the observed contrast, while the confidence interval communicates uncertainty around that value. A reader should not treat the estimate as proof of an exact difference, nor interpret the interval as demonstrating equivalence. The record does not report an equivalence margin or a clinical threshold that would define what range of differences is acceptable or important.

The group standard deviations, 1.5 and 1.95, describe variation in individual hemoglobin A1c observations. The outcome sample sizes are 44 for NPH/Regular and 60 for aspart. These features contribute to the precision of the group means and their comparison. The registry does not report a prespecified precision target or the original sample-size justification, so it is not possible to determine whether the trial was designed to estimate the difference within a particular margin.

The confidence interval is a useful expression of uncertainty for the reported analysis, but its interpretation depends on the statistical model and analysis population. Those details are not fully reported in the registry record. The interval should therefore be read as uncertainty around the posted comparison, not as a guarantee that the true difference lies within a clinically acceptable range.

11. 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 sample-size illustrations are planning calculations based on the observed effect, not the original trial design. They use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning row uses the stated fraction of the observed effect as the effect to detect. These are participants per arm, not total participants or total events; the registry’s planning note distinguishes these continuous-outcome illustrations from time-to-event displays, which use total events.

The 80% power column describes the number of participants per arm that the planning calculation associates with an 80% chance of detecting the chosen effect under its assumptions, if that effect is the true effect and the model assumptions hold. The 90% column uses a higher target chance of detecting that effect, so its participant requirement is larger. For example, planning for 100% of the observed effect gives 54,908 participants per arm for 80% power and 73,506 participants per arm for 90% power. Planning for 75% of the observed effect gives 97,614 and 130,677 participants per arm, respectively. Planning for 50% of the observed effect gives 219,631 and 294,023 participants per arm.

Smaller effects are harder to distinguish from random variation, so the illustrations require more participants per arm when the target is a smaller fraction of the observed effect. These calculations are planning illustrations, not observed power. They do not say what power the completed trial had, and they do not establish that the original study was planned using these assumptions. The original design assumptions are not reported in the registry record.

12. 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 = 0.030; pooled standard deviation σ = 1.774.
2. n = 2 × (1.96 + 0.842)² × 1.774² / 0.030² = 54,907.5, rounded up to 54,908 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 <- 0.0299999999999994                      # observed difference in means
sigma <- 1.77426455352136                     # 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(821795)
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 = 54,908 (80% power), 73,506 (90% power)
 75% of observed effect: n per arm = 97,614 (80% power), 130,677 (90% power)
 50% of observed effect: n per arm = 219,631 (80% power), 294,023 (90% power)
power.t.test: 54908.5 
Simulated power at n = 54,908 per arm: 0.801 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists secondary outcomes involving seven-point blood glucose profiles, patient-reported diabetes symptoms and perceived control, hypoglycemia, and twice-daily glucose monitoring using a telehealth transmitter. It identifies five outcome measures overall, but the analysis count is zero. The record does not report results for those secondary outcomes or explain why analyses are not listed. Their absence from the posted analysis information should not be interpreted as evidence that no measurements were collected.

When a study evaluates multiple outcomes or performs multiple statistical tests, the chance of at least one apparently unusual result can increase. In general, multiplicity procedures manage that issue by defining which comparisons are primary and how statistical evidence is interpreted across tests. The registry record does not report a multiplicity plan, an adjustment method, or a hierarchy of secondary analyses. The reported primary p-value is explicitly unadjusted, so it should not be read as accounting for other possible outcome tests.

The registry record also does not report how missing primary or secondary outcome measurements were handled, whether any imputation was used, or whether the analysis included all randomized participants. The analysis population is not reported in the registry record. The primary outcome summaries list 44 participants in the NPH/Regular group and 60 in the aspart group, while the safety record uses different denominators. The reason for these differences is not reported.

Other details relevant to interpretation are also absent: the original sample-size justification, interim analyses, the statistical analysis plan, and the approach to protocol deviations are not reported in the registry record. These gaps do not invalidate the numerical summaries, but they limit assessment of how completely the posted analyses reflect the intended trial methods.

14. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cardiovascular disease (14.0%), cerebral vascular disease (2.0%), gastrointestinal (2.0%), genitourinary (2.0%), hypoglycemia (2.0%).

The registry reports serious adverse event counts using groups of 50 at risk for NPH/Regular and 64 at risk for aspart. It lists 15 affected participants in the NPH/Regular group and 10 in the aspart group. For other adverse events, the at-risk denominators are 50 and 64, with 6 and 9 affected participants, respectively. These are descriptive figures; the record does not give a statistical comparison or an adjusted safety estimate.

The listed top serious-event terms include cardiovascular disease at 14.0%, and cerebral vascular disease, gastrointestinal, genitourinary, hypoglycemia, and infection at 2.0% each. The registry record does not provide enough context in this listing to interpret those percentages as comparative risks between assigned treatments. The death fields are not populated, so death counts are not reported in the registry record.

Safety denominators differ from the primary outcome denominators. The reason is not reported, so the adverse-event figures should not be directly reconciled with the hemoglobin A1c analysis groups. The registry does not report event timing, severity grading beyond the serious/other distinction, or a formal statistical analysis of these safety data.

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

The posted primary result shows similar observed group means at 16 weeks: 8.32 for NPH/Regular 70/30 and 8.35 for the aspart insulin analog biphasic mix. The mean difference was 0.03, with an unadjusted p-value of 0.929. The confidence interval extends across zero, so the comparison does not establish a statistically detectable difference in this analysis.

That finding does not prove that the regimens are equivalent or that any possible difference is unimportant. A conclusion of equivalence or noninferiority requires a prespecified margin and an analysis designed to address it. Neither is reported in the registry record. The result also does not establish that the treatments have identical effects for every participant: group means can conceal individual variation, and the reported standard deviations show that observations varied within both groups.

The result is specific to the posted endpoint, population description, treatment transition, and timeframe. It concerns mean hemoglobin A1c during outpatient transition therapy at 16 weeks. It does not, by itself, answer every question about daily glucose profiles, symptoms, hypoglycemia, telehealth monitoring, or safety. Those outcomes are listed, but their analyses are not reported in the registry record.

The study is recorded as randomized and open-label, but the registry does not report enough detail to assess allocation procedures, masking of outcome assessment, missing-data handling, or the analysis population. The primary outcome summaries and safety summaries also use different denominators, with no explanation recorded. These limitations mean the posted result should be interpreted as an estimate with uncertainty, rather than as a definitive statement about comparative treatment performance across all outcomes or settings.

16. Why This Trial Matters Statistically

This trial illustrates why a point estimate, a p-value, and a confidence interval answer different statistical questions. The mean difference of 0.03 describes the observed contrast. The unadjusted p-value of 0.929 describes the evidence against a null hypothesis under the analysis used. The confidence interval communicates uncertainty and extends across zero. Together, these quantities do not support a claim of proven equivalence; they indicate that the observed comparison does not resolve the range of possible differences.

The registry’s wording of an “as good as” research aim is statistically important. Similar observed means and a large p-value are not enough to establish noninferiority or equivalence. Those claims require a prespecified margin and a suitable analysis, neither of which is reported. This distinction is useful when reading clinical trial results because absence of evidence for a difference and evidence that differences are acceptably small are not the same conclusion.

The study also highlights the need to distinguish outcome denominators from safety denominators. The hemoglobin A1c summaries list 44 and 60 participants, whereas the adverse-event record lists different at-risk groups. Without participant flow and analysis-population details, those records cannot be reconciled. Clear reporting of who was included in each outcome analysis is important for evaluating both precision and generalizability.

Finally, planning calculations derived from an observed effect can illustrate how sample requirements change with the target effect and desired power. They do not reconstruct the original design or describe the completed trial’s power. The registry does not report the original sample-size justification. Keeping the planning exercise separate from the observed trial result prevents a post hoc illustration from being mistaken for a prospective design decision.

17. 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 registry record does not report the original sample-size rationale, a protocol-defined equivalence or noninferiority margin, or the complete planned statistical analysis. Although the brief summary frames the research question as whether one regimen is “as good as” the other, the posted result is an ordinary mean comparison and does not establish a formal equivalence conclusion.

Important analysis details are missing from the registry record, including the analysis population, how missing outcome data were handled, whether the primary result was adjusted for baseline values, and whether any interim analyses were planned or conducted. The record also does not describe how allocation was generated or concealed, or whether outcome assessment was masked. The study is listed as having no masking.

The primary outcome summaries list 44 participants in the NPH/Regular arm and 60 in the aspart arm, while adverse-event reporting uses at-risk denominators of 50 and 64. The reason for those differences is not reported. The record does not provide participant flow information that would explain which participants contributed to each result.

The registry lists secondary outcomes but does not report their analyses. It also does not report a multiplicity strategy, clinical thresholds for interpreting hemoglobin A1c differences, or detailed safety analysis methods. These omissions restrict interpretation of the breadth, precision, and clinical meaning of the posted comparison. The record describes transition therapy after hospitalization for non-critical acute illness; it does not establish how results apply outside that described context.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Veterans Inpatient Insulin Study and Transition to Outpatient Therapy — NCT00821795.
• Registry record last updated 2020-03-04; status last verified 2020-02.
• Results first posted 2020-03-04.
• 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.

21. Record Summary

The registry describes a completed, randomized, open-label, parallel-group phase 4 study of NPH/Regular 70/30 mix versus an aspart insulin analog 70/30 mix for transition from inpatient to outpatient therapy. The listed enrollment is 120, and the condition is type 2 diabetes mellitus.

The primary endpoint is mean hemoglobin A1c during the outpatient transition phase at 16 weeks. The posted group means are 8.32 and 8.35, and the mean difference is 0.03, with an unadjusted p-value of 0.929. The confidence interval includes values on both sides of zero. The registry does not report a formal equivalence or noninferiority margin.

Secondary outcomes include glucose profiles, patient-reported diabetes symptoms, hypoglycemia, and twice-daily glucose monitoring. Their analyses, the original design assumptions, the analysis population, and missing-data methods are not reported in the registry record.