(R33 Phase) DELISH Study: Diabetes Education to Lower Insulin, Sugars, and Hunger: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for (R33 Phase) DELISH Study: Diabetes Education to Lower Insulin, Sugars, and Hunger (NCT03837405), a randomized na study. The primary outcome is Diet Adherence Between Intervention Arms as Measured by % of Ketones >= 0.5 mmol/L.

2. Clinical Question

3. Trial Design

The DELISH study is registered as a randomized, single-masked, parallel-group trial in people with type 2 diabetes. The registry lists the phase as NA and the status as completed. Enrollment is listed as 125, while the brief summary describes a plan to randomize 120 people in a 1:1 ratio. These are different registry descriptions: the record does not explain their relationship.

Both groups were assigned diet education focused on following a carbohydrate-restricted diet, along with basic behavioral strategies. The registry describes weekly, in-person group sessions for 3 months. It characterizes the diet as approximately 10% of calories from carbohydrate, typically 50 grams per day or fewer, not including fiber. Participants assigned to the education-plus-mindfulness group were also to receive mindfulness components using the Eat Right Now platform. Further details of those components are not reported in the registry record.

The primary endpoint uses ketone measurements beginning at week 5 and continuing to month 12. Measurements were taken daily over 7 days at five study timepoints: 5 weeks, 3 months, 6 months, 9 months, and 12 months. The record does not specify who was masked, how the random sequence was generated, how allocation was concealed, or the trial's original sample-size justification. The posted primary arm summaries include 62 participants in the education group and 60 in the education-plus-mindfulness group with endpoint data.

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.

Only the first 6 of 10 groups are shown.

Period 1: 3-month Main Intervention

Period 2: 9-month Maintenance

Analysis population for the primary outcome, as defined in the registry: All trial participants with ketone measures. Ketone measures started at week 5 of the trial and ended at month 12.

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 Diet Adherence Between Intervention Arms as Measured by % of Ketones >= 0.5 mmol/L, measured over Measures from ketone measure initiation (week 5) to 12 months. Ketones were measured daily over 7 days at 5 study timepoints: 5 weeks, 3, 6, 9, and 12 months..

The primary endpoint is the mean percentage of a participant's ketone measurements that were at least 0.5 mmol/L, compared between the two randomized groups. The registry identifies this as a measure of diet adherence. It describes measurements starting at week 5 and continuing through month 12, with daily measurement over 7 days at five timepoints. The endpoint therefore reflects repeated measurements across the study period, rather than one measurement taken at a single visit.

As a general statistical matter, a repeated-measure summary can capture how often a measured value meets a prespecified threshold across the observation periods. That makes the endpoint an operational way to represent adherence over time, rather than a direct measure of every aspect of diet or behavior. The posted summaries report the mean and standard deviation of the percentage for each group. The registry does not describe how participant-level percentages were combined across timepoints, how missing daily measurements were handled, or whether the primary comparison adjusted for any other variables.

The result should be interpreted as a comparison of the posted adherence measure, not as a complete account of dietary behavior. The endpoint description does not establish that ketone readings were available for every randomized participant. The registry defines the endpoint population as trial participants with ketone measures.

7. Results Posted in the Registry

• Diet Education + Mindfulness: mean 70.40 versus 72.20 on control; difference -1.80 (95% CI -9.72 to 6.12), standardized effect d = -0.08; z-test p = 0.656.

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.

Diet Adherence Between Intervention Arms as Measured by Mean Grams of Non-fiber Carbohydrate Consumed Per Day From 24- Hour Diet Recall

• 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:
  • Diet Education + Mindfulness: mean 55.20 versus 61.40 on control; difference -6.20 (95% CI -18.40 to 6.00), standardized effect d = -0.18; z-test p = 0.319.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Diet Education + Mindfulness and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists six secondary outcome measures spanning emotion-related eating, glycemic control, craving-related eating, dietary carbohydrate intake, stress-related eating, and weight change. The detailed secondary result includes mean non-fiber carbohydrate intake from diet recalls between 3 and 12 months. Its posted group means are 61.4 grams per day for diet education and 55.2 grams per day for diet education plus mindfulness; the mean difference is -6.2 grams per day, with a 95% confidence interval from -18.4 to 6.001 and an unadjusted p-value of 0.319.

That estimate points toward lower reported carbohydrate intake in the mindfulness group, but its interval includes no difference and also allows for estimates in the opposite direction. In that respect, it does not provide a clear contrast that changes the interpretation of the primary result. Both posted comparisons leave uncertainty about the direction and magnitude of any between-group difference. The registry does not provide detailed results for every secondary measure, so the pattern across all secondary endpoints cannot be characterized from the record.

Secondary results generally carry less weight than a prespecified primary comparison because they may involve multiple outcomes, timepoints, or analytic choices. As a general statistical matter, conducting many tests increases the chance that at least one result appears notable by chance if the tests are interpreted without accounting for multiplicity. The registry does not report a multiplicity adjustment or a hierarchy for testing secondary outcomes. The detailed secondary p-value is unadjusted, and should not be treated as confirmatory evidence in isolation.

The registry lists no reported analyses in its analysis inventory, while also posting outcome summaries. That limits clarity about which secondary comparisons were prespecified, which were actually tested, and how missing data were handled. Secondary findings are best treated as descriptive or exploratory unless a complete analysis plan and results provide stronger context.

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

• Emotion-related Eating
• Glycemic Control--HbA1c
• Frequency of Eating in Response to Cravings
• Diet Adherence Between Intervention Arms as Measured by Mean Grams of Non-fiber Carbohydrate Consumed Per Day From 24- Hour Diet Recall
• Salzburg Stress Eating Scale
• Weight Change

9. Statistical Interpretation

10. Statistical Methods Explained

What does a mean difference describe?

A mean is a measure of the center of a set of observations. For this endpoint, the group mean is the average percentage of ketone measurements meeting the threshold. A mean difference places the two group averages on the same scale. The posted contrast is education plus mindfulness compared with education alone, so a negative estimate indicates that the posted mean was lower in the mindfulness group. The estimate describes the contrast in the analyzed endpoint summaries; it does not, by itself, explain why the groups differed or whether any difference is important in practice.

The group means are 70.4 for education plus mindfulness and 72.2 for education alone. The posted mean difference is -1.8. The comparison is presented as an unadjusted mean difference. The registry record does not report whether the trial's primary analysis used a different model or adjustment, so the posted contrast should not be assumed to represent a covariate-adjusted analysis.

What does the standard deviation describe?

A standard deviation describes the spread of observations around a group mean. It is not the uncertainty in the estimated mean difference. The posted standard deviations are 21.1 for education plus mindfulness and 23.5 for education alone. These values indicate variation among the summarized participant-level endpoint values; they do not tell the reader how that variation was distributed or whether the data contained unusual observations.

As a general statistical concept, the standard error describes how much an estimate would vary across repeated samples under a specified model. It is used in constructing confidence intervals and tests. The standard error of a group mean is not interchangeable with the standard deviation of individual observations. The registry does not describe the exact analysis method used to obtain its interval or p-value.

How should the confidence interval be read?

A confidence interval gives a range of values compatible with the data and the statistical procedure used to construct it. In repeated use of a method under its assumptions, intervals at a stated confidence level would cover the true parameter at that rate. It is not a range that assigns a probability to the true value after the observed data have been collected. The primary interval runs from -9.72 to 6.12. Because it spans zero, it is compatible with no mean difference as well as with differences in either direction. Its width also shows that the data do not pin down a single precise value.

As a general explanation, a confidence interval can be affected by sample size, variability, and the analysis method. It does not account for design or measurement problems that the analysis fails to address. In this trial, the registry does not report the handling of repeated measurements, missing endpoint observations, or any adjustment for baseline factors. Those omissions make it difficult to assess the assumptions behind the posted interval.

What does the p-value add?

The unadjusted p-value for the primary contrast is 0.656. A p-value describes how unusual data at least as incompatible with a specified null hypothesis would be, if that hypothesis and the test assumptions were correct. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The posted value does not provide strong evidence against a no-difference null under the unadjusted comparison. It should be read alongside the estimate and confidence interval, not as a standalone verdict.

The sample-size illustration uses a two-sided significance level of 0.05. That is a planning assumption for the illustration, not evidence that the original trial used the same decision rule. The trial's original significance level and analysis plan are not reported in the registry record.

Why can over-dispersion matter?

As a general statistical matter, over-dispersion occurs when the data vary more than a model assumes. If a model understates variability, its standard errors can be too small and its confidence intervals too narrow. Repeated observations from the same participant are often correlated rather than independent; treating them as independent can also make uncertainty appear too small. This trial collected repeated daily ketone measures at multiple timepoints, but the registry does not report how that dependence was handled or whether over-dispersion was assessed. No conclusion about the presence of over-dispersion can be drawn from the posted summaries alone.

The comparison is an unadjusted contrast of group summaries. The registry does not report an intention-to-treat analysis, a per-protocol definition, a missing-data method, or a model for repeated measures. Those details matter because the analytic approach determines which participants and observations contribute and how the uncertainty is calculated. The posted numerical results can be described, but the record does not provide enough methodological detail to reproduce or fully evaluate the analysis.

11. Confidence Intervals and Precision

The primary estimate is -1.8 percentage points, with a 95% confidence interval from -9.72 to 6.12. This interval crosses zero and allows for a range of possible between-group differences in both directions. It does not tightly isolate the size or direction of the contrast. The unadjusted p-value is 0.656, which is consistent with the interval's lack of strong evidence against a no-difference hypothesis under the posted comparison.

Precision is about how narrowly the data estimate a quantity, not about whether the estimate is important. A narrow interval would constrain the range of differences compatible with the data more closely; a wider interval leaves more uncertainty. The registry does not report a threshold for a difference considered meaningful for this endpoint, so the interval cannot be classified against such a threshold. It also does not report the original trial's sample-size justification, making it impossible to assess from the record whether the trial was designed to estimate this endpoint with a particular precision.

The posted primary endpoint includes participants with ketone measurements, with arm summaries of 62 and 60. The registry does not report how many daily measurements were missing or whether participant contributions varied across timepoints. Consequently, the precision conveyed by the posted interval may not capture uncertainty arising from missingness or the structure of repeated measurements if those issues were not incorporated into the analysis. The interval should be read as the uncertainty reported for the posted comparison, not as a guarantee that every source of uncertainty has been accounted for.

The secondary carbohydrate-intake contrast is also imprecise: its interval runs from -18.4 to 6.001 grams per day. As with the primary result, the interval includes no difference. The registry does not provide a complete inferential account for the full set of secondary outcomes, so the precision of the posted contrasts should not be generalized to outcomes without detailed results.

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.

Read each planning row as an illustration of the participants per arm needed to detect a stated fraction of the observed effect, under the listed planning assumptions. The continuous-outcome figures are participants per arm, not total events. The event-count convention for time-to-event outcomes does not apply to these continuous comparisons.

The illustration assumes a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each row uses the stated fraction of the observed effect as the effect to detect. The columns for 80% and 90% power describe the number of participants per arm associated with those target probabilities of detecting an effect of that size under the assumptions. The 90% target requires more participants per arm than the 80% target for the same effect size. Likewise, planning for a smaller effect requires more participants, because a smaller signal is harder to distinguish from sampling variation.

The observed primary estimate is small relative to the variation in the endpoint, so the planning figures based on that observed estimate are very large. For example, the row using 100% of the observed primary effect gives 2,421 participants per arm for 80% power and 3,241 participants per arm for 90% power. The rows using 75% and 50% of the observed effect require still more participants per arm. These are planning illustrations, not the trial's original design calculations; the original sample-size assumptions are not reported in the registry record. This is not observed power.

These figures depend on the observed effect used as the starting point and on the stated statistical assumptions. They should not be interpreted as a recommendation for a future trial without considering its intended endpoint, design, analysis, and target effect.

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 = -1.800; pooled standard deviation σ = 22.352.
2. n = 2 × (1.96 + 0.842)² × 22.352² / -1.800² = 2,420.7, rounded up to 2,421 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 <- -1.8                      # observed difference in means
sigma <- 22.3522258399471                     # 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(3837405)
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 = 2,421 (80% power), 3,241 (90% power)
 75% of observed effect: n per arm = 4,304 (80% power), 5,762 (90% power)
 50% of observed effect: n per arm = 9,683 (80% power), 12,963 (90% power)
power.t.test: 2421.6 
Simulated power at n = 2,421 per arm: 0.805 (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 does not report whether the primary endpoint was tested under a prespecified hierarchy, whether secondary outcomes were adjusted for multiple comparisons, or whether any interim analyses were planned or conducted. It lists six secondary outcome measures and reports an unadjusted p-value for the detailed secondary carbohydrate-intake comparison. Without a stated multiplicity strategy, individual secondary p-values should be interpreted cautiously: testing several outcomes can increase the chance of apparently notable results arising by chance.

The record also does not describe missing-data procedures or define the analysis population beyond stating that the primary endpoint includes trial participants with ketone measures. It does not say whether all randomized participants were included in the primary analysis, how participants without ketone measures were handled, or how missing daily observations and missed timepoints affected each participant's summary. Those omissions are important for a repeated-measure endpoint because participants may contribute different amounts of data.

The registry does not report whether the primary analysis followed an intention-to-treat approach, used a per-protocol population, or applied another rule for inclusion. It also does not provide a protocol-based definition of adherence or describe how adherence-related groups in the later safety record were formed. The arm summaries therefore should not be assumed to include every enrolled participant or to represent a particular analysis population beyond the endpoint description.

Finally, the record does not report the statistical model, covariate adjustment, assumptions checks, handling of repeated observations, or procedures for multiplicity. These missing details do not invalidate the posted summaries, but they limit independent assessment of how robustly the estimates and uncertainty reflect the trial data. The results are best read as the registry's posted comparisons with those methodological qualifications in mind.

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 non-serious adverse event terms

The registry reports zero serious adverse events among the participants at risk in each listed safety group. In Period 1, the diet-education group lists 63 at risk and zero serious events; the education-plus-mindfulness group lists 62 at risk and zero serious events. The Period 2 groups also list zero serious events. These counts describe the registry's group entries; they do not by themselves establish that the risk of serious events is the same between interventions.

The safety record also lists deaths in two Period 2 assessment-only groups: one in the education-plus-mindfulness assessment-only group and one in the diet-education assessment-only group. No deaths are listed for the other groups. The record does not report causes or timing of these deaths, or whether they were considered related to an intervention. Period 2 entries include maintenance and assessment-only categories, so they are not a simple continuation of the two initial randomized-arm summaries.

Other adverse events are listed for several groups, but the record does not give event terms in the safety information. Group counts should therefore be read as registry counts, not as a comparison of event types or severity. The record does not provide enough detail to draw a causal safety conclusion.

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

The posted primary result is an estimate of the difference between two behavioral intervention groups on a ketone-based measure of diet adherence. The estimated mean difference is -1.8 percentage points, and its confidence interval ranges from -9.72 to 6.12. The result is compatible with differences in either direction and does not precisely determine the size of any contrast. The unadjusted p-value of 0.656 does not provide strong evidence against no mean difference under that comparison.

This is not proof that the interventions have identical effects. A result that does not provide strong evidence of a difference is not the same as a demonstration of equivalence or non-inferiority. Those conclusions require a prespecified margin and appropriate analysis, neither of which is reported in the registry record. Nor does the result show that either approach did or did not change an individual's diet, ketone measurements, or health.

The primary measure is an operational adherence endpoint based on the percentage of ketone readings meeting a threshold. It does not capture every element of dietary behavior, and the registry does not explain how missing readings or repeated measurements were handled. The secondary carbohydrate-intake comparison is also uncertain, and detailed results are not available for all listed secondary measures.

The trial compares diet education alone with the same education plus mindfulness components; it does not compare either strategy with no intervention. The registry descriptions therefore support interpretation of the added mindfulness components relative to diet education, not a conclusion about the effects of diet education itself. The record's limited analytic detail, including missing information about the original statistical plan and analysis populations, further restricts the strength of conclusions that can be drawn from the posted results.

17. Why This Trial Matters Statistically

This trial is statistically interesting because its primary endpoint is based on repeated behavioral and biochemical measurements over an extended follow-up period. Summarizing daily readings across multiple study timepoints raises questions about within-person correlation, missing measurements, and how participant-level summaries are formed. Those issues affect both what the estimate represents and how precisely it can be estimated. The registry gives the measurement schedule but not the analysis details needed to evaluate these choices.

The posted primary estimate is small in relation to the reported spread, and its confidence interval remains broad enough to include differences in either direction. This illustrates why a point estimate and a p-value should not be interpreted alone. The interval conveys that the data leave substantial uncertainty about the magnitude of the contrast, while the unadjusted test does not show strong evidence against a no-difference hypothesis.

The planning illustration also highlights how sample-size requirements depend on the effect one aims to detect. When the assumed target effect is smaller, more participants per arm are needed under the stated assumptions. Because the illustration uses the observed effect rather than the trial's own design assumptions, it helps explain precision and planning concepts but cannot establish whether the original enrollment target was adequate.

Finally, a trial with a primary outcome and multiple secondary domains needs a transparent account of its prespecified analysis plan. The registry does not report the multiplicity strategy, missing-data approach, or detailed analyses for all secondary outcomes. That makes the distinction between confirmatory and exploratory findings especially important when interpreting a behavioral intervention study with several kinds of outcomes.

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.

The registry's enrollment figure of 125 differs from the brief summary's plan to randomize 120 people. The record does not explain the difference. The primary endpoint summaries include 62 participants in the education group and 60 in the education-plus-mindfulness group, but the registry does not provide a participant flow accounting for enrollment, randomization, follow-up, and endpoint availability.

The primary outcome depends on repeated ketone measurements starting at week 5, with daily measurement over 7 days at five timepoints. The record does not explain how incomplete measurement periods were handled, whether the participant-level percentages were weighted equally, or how dependence among repeated readings was incorporated. Without those details, the precise estimand and uncertainty calculation are difficult to assess.

The intervention descriptions are incomplete in the registry record. It identifies diet education and the additional mindfulness components, including use of the Eat Right Now platform, but further details are not reported in the registry record. The record also does not identify who was masked or give details of allocation concealment, the analysis plan, or the original sample-size justification.

Only limited detailed secondary results are available despite multiple listed secondary outcomes. The registry does not report multiplicity handling or provide a full account of missing data and analysis populations. Safety entries include several Period 2 maintenance and assessment-only groups, which makes direct comparison with the initial randomized groups difficult without further explanation. These limitations concern the completeness and interpretability of the registry record; they do not, by themselves, determine the direction of the trial results.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: (R33 Phase) DELISH Study: Diabetes Education to Lower Insulin, Sugars, and Hunger — NCT03837405.
• Registry record last updated 2026-06-17; status last verified 2026-05.
• Results first posted 2026-06-17.
• 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

The DELISH study is a completed randomized, single-masked, parallel-group trial in people with type 2 diabetes. It compared diet education focused on a carbohydrate-restricted diet with the same education plus mindfulness components. The registry lists enrollment of 125, while the brief summary describes a plan to randomize 120 participants.

The primary endpoint was the mean percentage of ketone measurements at or above 0.5 mmol/L, measured from week 5 through month 12. The posted mean difference for education plus mindfulness compared with education alone was -1.8 percentage points, with a 95% confidence interval from -9.72 to 6.12 and an unadjusted p-value of 0.656. This estimate is uncertain and does not establish equivalence or a meaningful difference.

The record lists six secondary outcome measures but does not provide detailed results for all of them. The statistical analysis plan, original sample-size justification, multiplicity approach, and missing-data methods are not reported in the registry record.