A Trial Investigating the Efficacy and Safety of Insulin Detemir Versus Insulin NPH in Combination With Metformin and Diet/Exercise in Children and Adolescents With Type 2 Diabetes Insufficiently Controlled on Metformin With or Without Other Oral Antidiabetic Drug(s) With or Without Basal Insulin: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Trial Investigating the Efficacy and Safety of Insulin Detemir Versus Insulin NPH in Combination With Metformin and Diet/Exercise in Children and Adolescents With Type 2 Diabetes Insufficiently Controlled on Metformin With or Without Other Oral Antidiabetic Drug(s) With or Without Basal Insulin (NCT02131272), a randomized phase 3 study. The primary outcome is Change in HbA1c (Glycosylated Haemoglobin).

2. Clinical Question

3. Trial Design

This was a phase 3, randomized, open-label, two-arm, parallel-group, multicentre trial. It was conducted globally and compared insulin detemir with insulin NPH, with metformin and diet/exercise in both groups. The stated treatment period for the primary endpoint was from baseline, or week 0, to week 26. The arm descriptions state that current oral antidiabetic drugs were continued unchanged.

The registry record lists enrollment of 42 and status as terminated. The recorded start date is 2014-06-11, and primary completion is listed as 2016-06-14. The reason for termination and the timing or circumstances of the termination are not reported in the registry record. These entries describe trial administration; they do not by themselves explain why the study stopped or establish how termination affected the analysis.

The primary analysis was described as a full analysis set including all randomized subjects. The posted arm summaries list 20 participants for insulin detemir and 22 for insulin NPH. For some secondary outcomes, the posted analysis denominators differ from these primary arm counts. The registry record does not explain those differences.

Open-label allocation can matter when outcomes involve judgment or participant reporting, because knowledge of assigned treatment may influence behavior or assessment. HbA1c is a laboratory-based outcome, but the record does not describe laboratory procedures, outcome-assessor masking, or other safeguards. The method used to generate the random sequence, conceal allocation before assignment, or handle protocol deviations is not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: Subjects continued treatment with metformin on their pre-study dose(s) throughout the trial.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set included all the randomised subjects.

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 Change in HbA1c (Glycosylated Haemoglobin), measured over week 0, week 26.

The primary endpoint was estimated mean change in HbA1c, or glycosylated haemoglobin, from baseline at week 0 to week 26. It was expressed as percentage of glycosylated haemoglobin. Change measures how the outcome moved over the stated treatment period rather than reporting only a single end-of-treatment value. The trial description identifies glycaemic control as its focus, and HbA1c was selected as the primary measure for that question.

The registry record reports least-squares mean change for each arm, along with standard errors. Least-squares means are model-based estimates; they are not necessarily simple arithmetic averages. The registry record names the primary analysis population as the full analysis set, including all randomized subjects, but does not report details of the model used to estimate the arm means in the primary outcome summary.

Change scores can summarize within-group movement, but the comparative question is addressed by the difference between groups. A decrease in either group alone does not establish which group had the more favorable change relative to the other. The estimated difference and its uncertainty are therefore important alongside the arm-specific estimates. The endpoint is a measure of a glycaemic outcome; it does not, by itself, answer every question about safety, other outcomes, or longer-term effects.

7. Results Posted in the Registry

• Insulin NPH + Metformin + Diet/Exercise: mean -0.81 versus -0.64 on control; difference -0.17 (95% CI -1.04 to 0.70), standardized effect d = -0.12; z-test p = 0.703.

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.

Change in Body Weight Standard Deviation Score (SDS)

• 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:
  • Insulin NPH + Metformin + Diet/Exercise: mean 0.10 versus 0.01 on control; difference 0.09 (95% CI -0.01 to 0.19), standardized effect d = 0.55; z-test p = 0.078.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Insulin NPH + Metformin + Diet/Exercise and the control. The interval gives the range of effects that remain compatible with the data.

Proportion of Subjects Achieving HbA1c Below 7.0%, Who Have Not Experienced Any Treatment Emergent Severe Hypoglycaemic Episodes Within the Last 14 Weeks of Treatment.

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Insulin NPH + Metformin + Diet/Exercise: 33.3% (7/21) versus 25.0% (5/20) on control; risk difference 8.3 percentage points (95% CI -19.4 to 36.0), risk ratio 1.33 (0.51–3.52); unadjusted two-proportion z-test p = 0.558.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Insulin NPH + Metformin + Diet/Exercise and the control. The interval gives the range of effects that remain compatible with the data.

Proportion of Subjects Achieving HbA1c Below 7.5%, Who Have Not Experienced Any Treatment Emergent Severe Hypoglycaemic Episodes Within the Last 14 Weeks of Treatment

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Insulin NPH + Metformin + Diet/Exercise: 38.1% (8/21) versus 30.0% (6/20) on control; risk difference 8.1 percentage points (95% CI -20.8 to 37.0), risk ratio 1.27 (0.54–3.01); unadjusted two-proportion z-test p = 0.585.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Insulin NPH + Metformin + Diet/Exercise and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary results span different kinds of outcomes: change in body weight standard deviation score, achievement of HbA1c thresholds together with no severe hypoglycaemic episodes during a specified period, hypoglycaemic episode counts, and adverse events. The detailed numeric summaries in the record cover only some of the listed secondary outcomes. Results for the other listed outcomes are not reported in the registry record.

Across the detailed secondary comparisons, the point estimates do not all represent the same quantity or clinical question. The body weight outcome is continuous, while the threshold outcomes are proportions. Their estimates and uncertainty cannot be directly combined into a single overall measure. The primary outcome is change in HbA1c; secondary outcomes provide additional descriptions but do not change what the primary analysis estimated. Taken together, they do not resolve the uncertainty in the primary comparison.

These outcomes should be treated as exploratory evidence unless an analysis plan establishes otherwise. The registry record does not report a multiplicity adjustment or a hierarchy specifying how secondary outcomes were to be interpreted. When many outcomes are tested, some may appear notable by chance even if there is no underlying difference. This makes isolated unadjusted p-values less persuasive than results from a prespecified primary analysis, especially when several outcomes and definitions are considered.

Absence of a reported numeric result for a listed secondary outcome is not evidence that the outcome had no events or no difference. It means the outcome's results are not reported in the registry record. The same caution applies to endpoints with different denominators: the record does not explain why the analyzed counts differ for some outcomes. Secondary findings can add context, but they should not be used to claim consistency or inconsistency beyond what the posted results support.

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

• Change in Body Weight Standard Deviation Score (SDS)
• Proportion of Subjects Achieving HbA1c Below 7.0%, Who Have Not Experienced Any Treatment Emergent Severe Hypoglycaemic Episodes Within the Last 14 Weeks of Treatment.
• Proportion of Subjects Achieving HbA1c Below 7.5%, Who Have Not Experienced Any Treatment Emergent Severe Hypoglycaemic Episodes Within the Last 14 Weeks of Treatment
• Total Number of Treatment Emergent Nocturnal (23:00-06:59) Severe or Blood Glucose (BG) Confirmed Symptomatic Hypoglycaemic Episodes
• Total Number of Treatment Emergent Severe or BG Confirmed Symptomatic Hypoglycaemic Episodes
• Incidence of Adverse Events (AEs)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: it is a numerical measure of change in HbA1c. The registry reports least-squares mean estimates and standard errors. A least-squares mean is an estimate produced by a statistical model, often adjusted for terms included in that model. The model terms and estimation details for the primary outcome are not reported in the registry record. A standard error describes the estimated sampling uncertainty of an estimate; it is not the same quantity as the spread of individual participant values.

What does the reported comparison estimate?

The sponsor's reported comparison is a least-squares mean difference from a mixed models analysis. The record gives the estimate as 0.17, its 95% confidence interval from -0.74 to 1.09, and p=0.307. The unadjusted recalculation instead compares the posted arm summaries and gives a mean difference of -0.17 for NPH relative to detemir. Its calculation uses an approximate pooled standard deviation derived from the standard errors and sample sizes reported for the arm summaries. It is a summary-statistic recalculation, not a reproduction of the sponsor's mixed model.

These signs reflect different contrast directions: reversing which arm is subtracted from which changes the sign of a mean difference, without changing the underlying comparison. The estimates also differ in analysis method. As the discrepancy flag is false, the registry record does not mark a conflict requiring a special precedence statement. The sponsor analysis remains the analysis reported by the sponsor, while the recalculation is best read as a separate, unadjusted statistical check.

What does a confidence interval convey?

A confidence interval gives a range of estimates compatible with the data and the method used to calculate it, under the assumptions of that method. It communicates uncertainty around an estimate more directly than a p-value alone. An interval that includes no difference does not establish that there is no difference; it indicates that the data and method do not isolate a nonzero difference at the corresponding confidence level. The interval's width also matters: a broad interval leaves more uncertainty about the size and direction of the comparative effect.

Why distinguish model-based and unadjusted results?

A mixed model can use the structure of repeated or otherwise related observations and may account for specified sources of variation. The registry names a mixed models analysis for the reported comparison but does not describe its exact specification, covariance structure, included terms, or missing-data assumptions. The recalculation uses only the posted summary values and cannot account for model features that are not encoded in those summaries. Consequently, it is useful as a transparent approximation but does not replace the sponsor's analysis.

Unadjusted comparisons are also sensitive to the limited information available in aggregated summaries. They cannot examine participant-level distributions, baseline covariates, protocol deviations, or patterns of missing outcomes. The record does not report enough detail to assess these features. No claim about a particular model assumption, adjustment, or missing-data procedure should be inferred where those details are absent.

How should p-values be read?

A p-value measures how incompatible the observed data are with a specified null model, under the assumptions of the analysis. It does not measure the probability that the null hypothesis is true, nor does it quantify the clinical importance of an effect. The recalculated unadjusted p-value and the sponsor's mixed-model p-value arise from different analyses and should be kept distinct. The registry's non-inferiority or equivalence label does not supply the margin or decision rule needed to interpret that framework fully.

The primary endpoint is not the only outcome in the record. Secondary measures include body weight standard deviation score, glycaemic thresholds combined with absence of severe hypoglycaemic episodes, hypoglycaemic episode counts, and adverse events. Different outcome types require different analytical approaches. Continuous outcomes may be summarized with means; binary outcomes concern whether an event or criterion was met; event counts can require methods that account for their distribution. The record does not report full analysis details for all these outcomes. For any secondary comparison, the reported estimate, uncertainty, outcome definition, and analysis method should be considered together rather than treating a single p-value as a stand-alone conclusion.

11. Confidence Intervals and Precision

Precision concerns how narrowly the data estimate a comparative effect. The sponsor's primary result is a least-squares mean difference of 0.17, with a 95% confidence interval from -0.74 to 1.09. The interval spans zero and includes estimates on either side of zero. It therefore does not isolate a clear direction of difference. Its width indicates substantial uncertainty about the size of the comparison under the reported analysis.

The unadjusted recalculation gives a mean difference of -0.17 for NPH relative to detemir, with a 95% confidence interval from -1.043 to 0.703. This interval also spans zero and includes values on both sides. It is based on an approximate standard deviation derived from the posted summary statistics and is not the sponsor's model-based interval. The contrast direction is reversed relative to the sponsor estimate, so the signs should not be compared without noting the orientation.

A confidence interval is not a range guaranteed to contain the true effect, and its endpoints are not equally likely values in a simple probability statement. It is a measure of uncertainty under the analysis method and its assumptions. Broad intervals can be compatible with a range of effects; they do not demonstrate that all values in the interval are equally plausible or clinically meaningful. The registry record does not provide a threshold for interpreting the size of a clinically important HbA1c difference.

The small arm counts reported for the primary outcome contribute to the limited precision, although the record does not provide a formal precision target or original sample-size justification. The status of the trial as terminated adds context, but its reason and impact on precision are not reported in the registry record. The results should therefore be read as estimates with considerable uncertainty, not as evidence that the regimens are identical.

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 illustrations based on the observed effect, not the original design assumptions for this trial. For the continuous primary comparison, the listed figures are participants per arm. The stated calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers planning for 100%, 75%, or 50% of the observed effect. The trial's own original sample-size assumptions and justification are not reported in the registry record.

The 80% power column describes the number of participants per arm that the planning calculation estimates would be needed for an 80% chance of detecting the specified effect under its assumptions. The 90% power column gives the corresponding planning figure for a higher chance of detection. These are design illustrations, not evidence about the chance that this completed trial detected an effect. This is not observed power.

The rows that plan for 75% or 50% of the observed effect consider smaller effects than the row planning for 100% of the observed effect. Smaller effects are harder to distinguish from random variation, so the planning figures rise as the effect to detect becomes smaller. The same general relationship applies when the target power is higher: a design seeking greater probability of detection needs more participants under the stated assumptions.

The calculation is anchored to an effect estimated from this trial, which itself is uncertain. It should not be read as a recommendation for a future study or as a reconstruction of the sponsor's design process. The assumptions do not include details that are not listed, such as missing outcomes or the sponsor's analysis model. The sample-size curve should be interpreted as a visual presentation of these planning relationships, not as additional observed trial evidence.

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 = -0.170; pooled standard deviation σ = 1.443.
2. n = 2 × (1.96 + 0.842)² × 1.443² / -0.170² = 1,131.3, rounded up to 1,132 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 <- -0.17                      # observed difference in means
sigma <- 1.44317531852509                     # 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(2131272)
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 = 1,132 (80% power), 1,515 (90% power)
 75% of observed effect: n per arm = 2,012 (80% power), 2,693 (90% power)
 50% of observed effect: n per arm = 4,526 (80% power), 6,058 (90% power)
power.t.test: 1132.3 
Simulated power at n = 1,132 per arm: 0.802 (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. Analyses Reported by the Sponsor

The sponsor's reported primary analysis is a mixed models analysis of least-squares mean difference. It gives an estimate of 0.17, a 95% confidence interval from -0.74 to 1.09, and p=0.307. The unadjusted recalculation from the posted arm summaries gives -0.17 for NPH relative to detemir, a 95% confidence interval from -1.043 to 0.703, and an unadjusted p-value of 0.703.

The contrast signs point in opposite directions because the reported analysis and recalculation use reversed arm order. The methods also differ: one is the sponsor's mixed models analysis, and the other is a summary-statistic recalculation. The sponsor's model details are not reported in the registry record, so the recalculation cannot be expected to reproduce all aspects of that model. The discrepancy flag is false.

The registry labels the sponsor result as non-inferiority or equivalence, but does not report the margin or decision criterion. The estimate, interval, and p-value can be described as posted; the missing criterion prevents a complete evaluation of the non-inferiority or equivalence claim from this record alone. The unadjusted recalculation is supplementary and should not be substituted for the sponsor's stated analysis.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when an analysis examines multiple outcomes, comparisons, time points, or subgroups. As the number of statistical tests grows, the chance of at least one apparently unusual result can rise even when the underlying effects are absent. A prespecified testing hierarchy or adjustment can help control the interpretation of multiple tests. The registry lists several secondary outcomes, but a multiplicity strategy or hierarchy is not reported in the registry record.

The primary endpoint is identified, but the detailed protocol and statistical analysis plan are not reported in the information available in the registry record. The record does not state whether interim analyses were planned or conducted, whether stopping rules were used, or how deviations from a planned analysis were handled. These matters should not be inferred from the terminated status.

The primary population is described as the full analysis set, including all randomized subjects. The record does not explain how that population was operationalized, how protocol deviations were treated, or whether separate per-protocol analyses were performed. It also does not describe the amount or pattern of missing outcome data, the method for handling missingness, or any sensitivity analyses. Those missing details limit assessment of whether the analysis population and method could affect the estimate.

Some secondary denominators differ from the primary arm counts. The registry record does not explain whether those differences reflect missing measurements, eligibility for a particular outcome, or another reason. The cause should not be guessed. Results based on different analysis sets may have different precision and may represent different subsets of the randomized population.

The record also identifies the reported analysis as non-inferiority or equivalence without stating the margin or decision rule. The margin is central to that framework; without it, the statistical claim cannot be fully assessed. These omissions do not invalidate the posted estimates, but they constrain what can be concluded from the registry record.

16. 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: migraine (4.5%).

The registry reports serious adverse events for the two treatment groups. In the insulin detemir group, 0 of 20 participants were reported as affected by a serious adverse event; in the insulin NPH group, 1 of 22 was reported as affected. These figures describe the posted serious-event counts and denominators. They do not establish that the risk is the same between groups or that no other safety concerns occurred.

The record lists migraine at 4.5% among the top serious adverse event terms. It does not provide enough context in this listing to interpret that percentage as a comparative rate or to assign it to a particular treatment group. Non-serious adverse events are reported for 8 of 20 participants in the detemir group and 9 of 22 in the NPH group. These are descriptive counts; the record does not provide an inferential comparison for them in the reported safety summary.

Deaths are not reported in the registry record. Safety summaries depend on outcome definitions, observation periods, coding, and completeness of reporting. The posted counts should be read as a summary of recorded events, not as a comprehensive assessment of comparative safety or as a basis for treatment decisions.

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

The results describe a randomized comparison of insulin detemir and insulin NPH, each used with metformin and diet/exercise, in children and adolescents with type 2 diabetes. The primary endpoint was change in HbA1c through week 26. The registry reports decreases in both arms and a sponsor mixed-model comparison with a confidence interval that includes zero. The unadjusted recalculation also has an interval spanning zero, but it uses a different method and contrast orientation.

These findings do not establish that one regimen is superior, and they do not establish that the regimens are equivalent. A result that does not demonstrate a clear difference is not proof of no difference. The interval remains compatible with effects in either direction, and the registry does not provide the non-inferiority margin or equivalence criterion needed to assess the label attached to the sponsor's analysis.

The results also do not answer every question about safety or other outcomes. Serious adverse event counts are descriptive, and several listed secondary outcomes do not have numeric results reported in the registry record. The trial's status is terminated, but the reason and its consequences for the analysis are not reported.

Interpretation is limited by the open-label design, the small posted primary arm counts, incomplete reporting of methods, and uncertainty around the primary estimate. The registry record does not report a threshold for what would constitute a clinically important HbA1c difference, so statistical estimates cannot be translated into a judgment about clinical importance from this information alone. These results are a record of the reported trial findings, not individualized medical advice or a basis for choosing a treatment.

18. Why This Trial Matters Statistically

This trial illustrates why an estimate and its uncertainty are more informative than a binary reading of statistical significance. The primary comparison is accompanied by a point estimate, confidence interval, and p-value, but the interval spans effects in either direction. That uncertainty remains important even though the registry labels the analysis as non-inferiority or equivalence. Without the decision margin, the label alone is not enough to assess whether a prespecified criterion was met.

It also illustrates the distinction between a sponsor's model-based result and a recalculation from summary statistics. The recalculation is transparent about its unadjusted nature, but it cannot reproduce unreported model details. Reversing the direction of a contrast changes the sign, so interpreting an estimate requires identifying which arm is subtracted from which. Comparisons should account for both the estimand and the method, not just the displayed number.

The trial's termination and the absence of reported details about its reason, missing data, and original sample-size plan are important context. They show why a registry record can support a careful description of results while leaving meaningful questions unanswered. Planning illustrations based on an observed effect are not a substitute for a documented design justification, and secondary outcomes do not erase uncertainty in the primary analysis.

For clinical researchers, the central statistical lesson is to distinguish what was estimated from what remains unknown. The record gives useful endpoint definitions and outcome summaries, but it does not provide every detail needed to evaluate model assumptions, missingness, multiplicity, or non-inferiority criteria.

19. 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 identifies the trial as terminated but does not report why it was terminated or how termination affected recruitment, follow-up, or analysis. The trial is described as open label, and no details about masking of outcome assessors or laboratory personnel are reported. The randomization sequence and allocation concealment methods are also not reported in the registry record.

The primary outcome summary reports least-squares means and standard errors, but the model specification is not reported. The sponsor's result is identified as a mixed models analysis, yet the record does not give the model terms, covariance structure, missing-data assumptions, or sensitivity analyses. The unadjusted recalculation is based on summary statistics and an approximate standard deviation; it cannot address individual-level data or reproduce unreported model features.

The non-inferiority or equivalence label lacks a reported margin or decision rule. The original sample-size justification and design assumptions are not reported, so the planning illustrations based on observed effects should not be mistaken for the trial's planned sample-size calculation. A prespecified multiplicity approach is also not reported despite the listing of multiple secondary outcomes.

Some secondary outcome denominators differ from the primary arm counts, and the registry does not explain why. Numeric results for several listed secondary outcomes are not reported in the registry record. The reason for these gaps and the amount or pattern of missing data cannot be inferred. Safety reporting provides counts for serious and other adverse events, but deaths are not reported in the registry record. These limitations constrain interpretation of the reported estimates and the completeness of comparisons.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Trial Investigating the Efficacy and Safety of Insulin Detemir Versus Insulin NPH in Combination With Metformin and Diet/Exercise in Children and Adolescents With Type 2 Diabetes Insufficiently Controlled on Metformin With or Without Other Oral Antidiabetic Drug(s) With or Without Basal Insulin — NCT02131272.
• Registry record last updated 2018-09-10; status last verified 2018-08.
• Results first posted 2017-09-01.
• 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.

23. Record Summary

The registry record describes a phase 3, randomized, open-label, two-arm, parallel-group, multicentre trial of insulin detemir versus insulin NPH, each with metformin and diet/exercise, in children and adolescents with type 2 diabetes. It lists enrollment of 42, status as terminated, a start date of 2014-06-11, and primary completion on 2016-06-14.

The primary outcome was estimated mean change in HbA1c from week 0 to week 26. The sponsor's mixed models analysis reports a least-squares mean difference of 0.17 with a 95% confidence interval from -0.74 to 1.09 and p=0.307. The unadjusted recalculation gives a difference of -0.17 for NPH relative to detemir, with a 95% confidence interval from -1.043 to 0.703 and an unadjusted p-value of 0.703.

The intervals include zero, and the record does not report the non-inferiority or equivalence margin. The reason for termination, original sample-size assumptions, multiplicity plan, and missing-data methods are not reported in the registry record.