Research Study to Look at Fast-acting Insulin Aspart With the Insulin Pump System 'iLet™' in Adults With Type 1 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Research Study to Look at Fast-acting Insulin Aspart With the Insulin Pump System 'iLet™' in Adults With Type 1 Diabetes (NCT03816761), a randomized phase 2 study. The primary outcome is Time in Low Interstitial Glucose (Defined as Below 54 mg/dL [3 mmol/L]) From Initiation of Treatment (Day 1) to End of Treatment (Day 7) (Percentage).

2. Clinical Question

This is an exploratory study of a particular insulin and device combination, not a broad comparison of diabetes treatments. The registry describes three cohorts. Participants in cohort 1 had t50 and t65 periods; cohort 2 had t40 and t65 periods; cohort 3 had t30 and t65 periods. These are within-cohort setting comparisons in a crossover framework, rather than one common non-default setting tested across every participant.

3. Trial Design

The registry describes a randomized, double-masked, phase 2 crossover trial sponsored by Novo Nordisk A/S. It is marked completed. The recorded start date is 2019-02-25, and the primary completion date is 2019-05-31. Enrollment was 24.

Participants used fast-acting insulin aspart with the iLet™ and a continuous glucose monitoring device. The treatment settings differed in tmax: the default was t65 = 65 minutes, while non-default settings were t50 = 50 minutes, t40 = 40 minutes, or t30 = 30 minutes. The arm descriptions specify three cohorts with two treatment periods in each cohort. In a crossover design, the same participant receives the compared settings in different periods, and the order is assigned by chance according to the registry description. This design allows observations under different settings to be made within a participant, although the posted summaries alone do not show the paired differences.

The brief summary says each setting was used for 7 days and that the study lasted about 5 to 9 weeks. It also describes 4 visits and 1 phone contact. The registry identifies the study as double-masked, but which people were masked and how masking was maintained are not reported in the registry record. Details of the randomization procedure, treatment-period sequence, and any washout between periods are also not reported in the registry record.

The arm-level endpoint results are organized by cohort and setting. The reported counts are 8 for each setting in cohorts 1 and 2, 8 for the non-default setting in cohort 3, and 7 for its t65 setting. Because the trial is a crossover, these arm labels represent treatment periods and should not be read as separate parallel groups of participants. A formal model accounting for the crossover structure 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.

First Treatment Period (7 Days)

Reasons for not completing this period:

Second Treatment Period (7 Days)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set: included all randomised subjects receiving treatment.Number of participants analysed=participants with available data.

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 Time in Low Interstitial Glucose (Defined as Below 54 mg/dL [3 mmol/L]) From Initiation of Treatment (Day 1) to End of Treatment (Day 7) (Percentage), measured over Day 1 to day 7.

The primary endpoint is time in low interstitial glucose, defined as glucose below 54 mg/dL [3 mmol/L], from initiation of treatment on day 1 to the end of treatment on day 7. Interstitial glucose is glucose measured in fluid between cells. The endpoint expresses the time below the specified threshold as a percentage of available interstitial glucose values.

This measure condenses repeated glucose readings over a treatment period into an interpretable summary of how often readings fell below the chosen threshold. In general, continuous glucose measurements can describe exposure over time rather than relying on a single measurement. The percentage scale also expresses the endpoint relative to the available readings for a period. These are general features of the measure; they do not establish that the available readings were complete or equally distributed across participants.

The outcome is reported as a mean with a standard deviation. The mean describes the central value across the participants with available data in each listed group, while the standard deviation describes spread around that mean. The registry defines the endpoint and analysis population, but does not report participant-level readings, confidence intervals, or a model-based comparison between settings. The posted group summaries therefore describe the observed values without establishing the precision of a treatment-setting contrast.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean summarize?

The primary measure is continuous: each participant’s result is a percentage of available interstitial glucose values below the defined threshold during the treatment period. A mean combines the participant-level results within a listed cohort and setting into one central summary. It is useful for describing the average observed endpoint, but it does not show the distribution of individual results or whether a small number of values influence the average.

The standard deviation is a descriptive measure of how dispersed the observations are around the mean. A larger standard deviation indicates greater spread in the observed results, while a smaller one indicates more clustering around the mean. It is not the standard error, a confidence interval, or a direct measure of the uncertainty in a comparison. In particular, an SD for each setting separately does not supply the within-person covariance needed to estimate uncertainty for a crossover contrast.

Why does the crossover structure matter?

In general, crossover designs compare treatments received by the same participant in different periods. A participant’s own results can serve as a reference, which can account for stable differences between people. However, a suitable analysis must preserve the pairing and consider the sequence and period structure. The registry identifies random assignment of setting order and a crossover model, but the posted primary results are group summaries. The statistical model, adjustment terms, and handling of period or sequence effects are not reported in the registry record.

These results should not be treated as if the treatment periods came from independent parallel groups. Nor can the paired effect be obtained by simply comparing the displayed means: the individual paired values and their relationship are not given. The trial’s design and the form of the posted summaries answer different questions. The design describes how settings were assigned; the summaries describe the observed mean and spread by cohort and setting.

What can be inferred from the posted summaries?

The means describe the observed endpoint values in each listed group. The standard deviations describe observed variability. Neither supplies a confidence interval for a setting effect. A confidence interval, in general, gives a range of effect estimates compatible with the data under a specified statistical procedure and conveys precision as well as direction. No such interval is reported for the primary comparison in the registry record.

A p-value, in general, measures how incompatible the observed data are with a stated null model, under the assumptions of the test. It does not measure the probability that a treatment setting is effective, and it does not describe the size or clinical importance of an effect. No p-value or inferential analysis is posted for the primary endpoint. The means should therefore be reported as descriptive results, without a significance label.

For a continuous endpoint, an analysis might estimate a mean difference between settings and its uncertainty. In a crossover study, the analysis would need to use paired observations and account for the study periods as appropriate. The registry record does not state that such an estimate was made, and the available group summaries do not permit its reconstruction. A recalculated unadjusted comparison is not presented in the record.

Why do sample-size illustrations need careful reading?

In general, planning calculations use assumptions about the effect size to be detected, the significance level, allocation, and statistical model. The registry description identifies planning illustrations based on the observed effect and states a two-sided significance level of 0.05, equal allocation, and a normal approximation. It also describes the target effect as a stated fraction of the observed effect. These are planning assumptions, not results from the trial’s own design process.

Such illustrations do not show the trial’s observed power. Observed power is not a substitute for an effect estimate and confidence interval, and a planning calculation based on an observed effect does not establish that the trial was designed to detect a particular effect. The sample-size justification, intended target effect, and the trial’s own design calculations are not reported in the registry record. The number of participants in the study should not be interpreted as proof that a comparison was adequately powered.

What does over-dispersion mean in general?

Over-dispersion is a general concern for count outcomes when the observed variation is greater than a model’s assumptions allow. It can make standard errors too small if the extra variability is ignored. This trial lists secondary outcomes based on hypoglycaemic episode counts, but the primary endpoint is a continuous percentage measure. The registry record does not report a count model or an over-dispersion assessment. The concept is relevant to choosing methods for count outcomes generally, not evidence that over-dispersion occurred in these data.

Overall, the statistical information posted for the primary endpoint supports a descriptive reading of cohort-specific means and SDs. It does not report a treatment contrast, its uncertainty, or a hypothesis test. Those omissions prevent a statistical conclusion about the comparative effect of the different settings from the registry results alone.

10. Confidence Intervals for Each Arm

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

R code and output

arm <- c("t50 Faster Aspart Cohort 1", "t65 Faster Aspart Cohort 1", "t40 Faster Aspart Cohort 2", "t65 Faster Aspart Cohort 2", "t30 Faster Aspart Cohort 3", "t65 Faster Aspart Cohort 3")
n   <- c(8, 8, 8, 8, 8, 7)
m   <- c(0.98, 0.89, 0.69, 0.5, 0.61, 0.37)
sp  <- c(0.72, 1.16, 0.5, 0.79, 0.56, 0.41)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  t50 Faster Aspart Cohort 1  8  0.980  0.720  0.378  1.582
CARM  t65 Faster Aspart Cohort 1  8  0.890  1.160  -0.080  1.860
CARM  t40 Faster Aspart Cohort 2  8  0.690  0.500  0.272  1.108
CARM  t65 Faster Aspart Cohort 2  8  0.500  0.790  -0.160  1.160
CARM  t30 Faster Aspart Cohort 3  8  0.610  0.560  0.142  1.078
CARM  t65 Faster Aspart Cohort 3  7  0.370  0.410  -0.009  0.749

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 15 outcome measures and 13 posted secondary outcomes. The secondary outcomes cover severe, self-manageable, overall, daytime, and nocturnal hypoglycaemic episodes, with some classifications specified by the American Diabetes Association and others by Novo Nordisk. The registry record lists zero analyses. It does not report which outcomes were included in a statistical testing hierarchy, whether any endpoint was designated for formal hypothesis testing beyond the primary outcome description, or how multiplicity was addressed.

Multiplicity is a general issue when several outcomes or comparisons are examined: the chance of at least one apparently unusual result can increase as the number of tests grows. Methods such as a prespecified hierarchy or adjustment can control that issue, but no such plan is reported in the registry record. Since inferential results are not posted, it is not possible to determine how multiple outcomes were handled in analysis or interpretation.

The primary analysis population is described as the full analysis set, including all randomized subjects who received treatment. The number analysed is defined as participants with available data. This identifies the general population intended for the endpoint summary, but it does not specify the missing-data method. The registry record does not report the amount or pattern of missing glucose readings, reasons for missingness, imputation, sensitivity analyses, or rules for excluding incomplete periods.

In general, missing data can affect a percentage based on available readings if the missing readings differ systematically from observed readings. The direction and size of any effect cannot be inferred without information about the missingness pattern and assumptions. The endpoint counts therefore need to be read as counts with available data, as described in the record, rather than as a guarantee of complete observations for every randomized participant.

12. 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 5% reporting threshold.

Most frequent non-serious adverse event terms

The adverse-event record lists zero participants with serious adverse events in each of the six cohort-setting groups, and zero deaths in each group. It also lists other adverse events affecting 2 participants in the t50 cohort 1 group, 1 in t65 cohort 1, 1 in t40 cohort 2, 0 in t65 cohort 2, 0 in t30 cohort 3, and 1 in t65 cohort 3. The group counts at risk are 8 for the first five entries and 7 for t65 cohort 3.

These are descriptive counts for the treatment groups as recorded. They do not establish that no serious event could occur, nor do they provide a comparative estimate of safety. The record does not supply event-level details or serious adverse-event terms; the serious-event term list is empty. It also does not describe whether events were attributed to a particular setting or how safety was compared statistically. In a crossover study, participants may contribute experience in more than one treatment period, so group counts should not be added or treated as separate parallel populations.

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

The posted primary endpoint results describe the mean percentage of available interstitial glucose values below 54 mg/dL [3 mmol/L] during day 1 to day 7 for each cohort and setting. They show low mean percentages in the listed groups, with the reported standard deviations describing variation among observed results. The record also reports adverse-event counts by setting and cohort, including zero serious adverse events and zero deaths in each listed group.

These observations do not establish that one insulin-delivery setting is safer or more effective than another. There is no reported statistical test, confidence interval, or model-based treatment contrast for the primary endpoint. The group means are not a substitute for a paired crossover analysis, and their differences should not be interpreted as statistically reliable effects. The registry record does not report the clinical importance threshold for a difference in this outcome.

The results also do not establish that the same findings would occur with other devices, insulin preparations, populations, or treatment durations. The trial concerns adults with type 1 diabetes using fast-acting insulin aspart with the iLet™ under specified settings, and the reported treatment periods lasted 7 days. Broader generalization is not justified by these cohort summaries alone.

Absence of reported serious adverse events in these groups is not proof of an absence of risk. The record gives a limited set of safety counts and does not provide a detailed event narrative or comparative safety analysis. Similarly, the exploratory design and descriptive endpoint summaries cannot resolve longer-term effects or answer questions not included in the posted outcomes.

In general, evidence about a treatment contrast depends on the estimated effect, its uncertainty, study design, and outcome relevance. Here, the means and SDs describe the observed data, while the uncertainty around the within-person comparisons remains unreported. The most defensible reading is therefore descriptive: the registry records cohort-specific endpoint values and safety counts, but does not establish a comparative conclusion about the settings.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of why design and reporting need to be read together. The study used a randomized crossover design, which can support within-person comparisons when the analysis uses the paired data and accounts for treatment periods. The posted primary outcome, however, is reported as separate means and standard deviations by cohort and setting. Those summaries do not expose the paired differences that are central to a crossover comparison.

The study also separates participants into cohorts with different non-default settings: t50 in cohort 1, t40 in cohort 2, and t30 in cohort 3, each alongside t65. That structure means the results are not a single direct comparison of one alternate setting against the default across all cohorts. Cohort-specific results are essential to avoid presenting the setting summaries as if they came from one uniform comparison.

The endpoint illustrates another reporting distinction. A percentage of available glucose readings has an intuitive interpretation as time below a threshold, but it depends on the availability of readings and on how repeated observations are summarized. The registry defines the measure and analysis population, yet does not report missing-data handling or the model for the crossover comparison. A clear definition of the outcome does not, by itself, establish the precision or validity of a treatment contrast.

Finally, the registry’s descriptive summaries should not be confused with inferential results or planning evidence. Without a reported contrast and its uncertainty, readers cannot assess the precision of the comparison. Planning illustrations based on an observed effect are not observed power and do not replace the trial’s design rationale, which is not reported in the registry record. These distinctions matter when translating compact registry summaries into appropriately cautious statistical conclusions.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary outcome is available as cohort-specific means and standard deviations, but no inferential analysis is posted. Confidence intervals, p-values, and model-based setting contrasts are not reported in the registry record. Because the design is crossover, the missing paired summaries and model details are especially limiting: the separate group summaries do not show how each participant’s result changed between settings.

The settings were not identical across cohorts apart from the t65 reference setting. Cohort membership and non-default tmax setting vary together, so the results cannot be read as a randomized comparison of all non-default settings against one common alternative. The registry does not report details of the randomization sequence, treatment-period effects, or any washout procedure. These design and analysis details limit interpretation of possible carryover or period influences.

The endpoint is based on available interstitial glucose values, but missing readings and their handling are not described. The listed endpoint counts include 7 for one group and 8 for each other group; the record does not explain the reason for the difference. The analysis population is defined as randomized participants receiving treatment with available data, but the impact of missingness cannot be assessed from the posted information.

Multiplicity procedures, the sample-size justification, the statistical model, and any sensitivity analyses are not reported in the registry record. The safety information is also limited to counts by group, without event descriptions or a comparative analysis. These gaps mean that the available results are best treated as descriptive evidence from an exploratory study, rather than a definitive estimate of the comparative effects of the settings.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Research Study to Look at Fast-acting Insulin Aspart With the Insulin Pump System 'iLet™' in Adults With Type 1 Diabetes — NCT03816761.
• Registry record last updated 2020-06-11; status last verified 2020-05.
• Results first posted 2020-06-11.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT03816761 is a completed, randomized, double-masked, phase 2 crossover study sponsored by Novo Nordisk A/S. It enrolled 24 adults with type 1 diabetes and evaluated fast-acting insulin aspart with the iLet™ closed-loop delivery system under different tmax settings. The primary endpoint was the percentage of available interstitial glucose values below 54 mg/dL [3 mmol/L] from day 1 to day 7.

Results are reported by three cohorts, each pairing a non-default setting with t65. The registry gives means, standard deviations, and endpoint group counts, but no inferential analysis or confidence intervals. It also reports zero serious adverse events and zero deaths in each listed group, alongside counts for other adverse events. The registry record does not report the statistical model, missing-data procedures, multiplicity approach, or sample-size justification. The posted primary results therefore support descriptive interpretation, not a firm comparative conclusion.