The DEB Pilot Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The DEB Pilot Study (NCT03190993), a randomized na study. The primary outcome is Change From Baseline in Body Weight.

2. Clinical Question

3. Trial Design

The DEB Pilot Study (NCT03190993) is recorded as a randomized, single-masked, parallel-group trial. Its status is completed. The registry lists a start date of 2017-07-21 and primary completion on 2018-06-21. The phase is listed as NA. Pennington Biomedical Research Center is identified as the sponsor. The record lists enrollment as 20.

Participants were assigned to one of two treatment arms: a sugar-sweetened solid treatment or a sugar-sweetened beverage treatment. The solid treatment was made from sugar-sweetened beverage syrup concentrate and gelatin and described as equivalent to one 20 oz. soft drink. The beverage treatment was a 20 oz. carbonated sugar-sweetened beverage with ~2.1g whey protein powder listed as an added ingredient. The registry does not identify which party was masked, or describe the allocation sequence or concealment procedures.

The primary outcome was change from baseline in body weight at 1 month, measured in kg. The registry lists one outcome measure and one analysis. It does not post secondary outcome results. The outcome population field is blank, so the specific analysis population is not reported in the registry record.

The posted primary outcome has 10 participants in each arm. The arm-level mean change was -0.325 kg for the solid-treatment group and 0.23 kg for the beverage-treatment group. The arm-level dispersion is reported as a standard error of 0.307 kg in each group. These are group summaries; they do not provide individual participant data or establish how missing measurements were handled.

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 Change From Baseline in Body Weight, measured over 1 month.

The primary endpoint is change from baseline in body weight at 1 month, expressed in kg. It describes the direction and average magnitude of weight change during the study period for each treatment arm. The arm means are -0.325 kg for the solid treatment and 0.23 kg for the beverage treatment. The registry reports standard errors of 0.307 kg for both means.

Change from baseline is used to express how the measured outcome shifts over the specified period, rather than reporting only a body-weight measurement at the end. Comparing the mean changes between randomized groups addresses whether the average change differed between the two assigned food forms. The recorded comparison is the beverage group relative to the solid group, with a recalculated mean difference of 0.555 kg. A positive value in that direction means the estimated mean change was higher in the beverage group than in the solid group; it does not by itself establish a reliable difference.

The registry does not give the primary outcome population description, details on measurement procedures, or the confidence level for the recalculated interval. Those omissions matter when judging how the measure was collected and which participants contributed to the analysis. The endpoint is a continuous measure, summarized by arm means and standard errors, rather than a count of participants reaching a specified threshold.

7. Results Posted in the Registry

• Sugar Sweetened Beverage Treatment: mean 0.23 versus -0.33 on control; difference 0.56 (95% CI -0.30 to 1.41), standardized effect d = 0.57; z-test p = 0.201.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: body-weight change is reported as a mean in kg. A continuous endpoint can take a range of values, so the analysis compares average changes between groups. The registry identifies the reported method as linear regression and describes its estimate as a net mean difference. The record does not provide the regression specification, covariates, or other details about that model.

What does the mean difference represent?

The comparison is oriented as beverage treatment relative to solid treatment. The recalculated unadjusted mean difference is 0.555 kg, based on the arm means of 0.23 kg and -0.325 kg. A positive difference in this direction indicates that the estimated average change was higher in the beverage arm. The comparison is a contrast between group averages; it is not a count of participants who gained or lost weight, and it does not describe each participant's response.

The registry also reports a sponsor analysis using linear regression, with a net mean difference of 0.56 and p=0.22. The recalculated comparison is an unadjusted re-analysis from posted summary statistics. It is useful as a check and a way to describe the contrast, but it is not a substitute for details about the sponsor's analysis. The record does not report the model terms or whether any adjustments were made.

What does the standard error describe?

The standard error measures the statistical uncertainty in an estimated mean, rather than the spread of individual participant measurements. The arm-level standard errors are both 0.307 kg. A smaller standard error would generally indicate a more precise estimate of a group mean, while a larger one would indicate greater uncertainty. The registry does not post the individual observations or a standard deviation for each group. The recalculation uses the posted summary statistics, and no approximate conversion from standard error to standard deviation was needed in the reported calculation.

General explanation: standard errors depend on both variability in measurements and the amount of information available. They should not be read as showing that every participant's change was close to the arm mean. The summary values alone cannot reveal the distribution of individual changes, unusual observations, or differences in variability that might matter to a fuller analysis.

How should the interval be read?

A confidence interval expresses a range of effect sizes reasonably compatible with the data and the statistical assumptions used to form it. The recalculated interval for the difference extends from below zero to 1.406 kg. Since it includes zero, the observed data do not tightly distinguish a difference in one direction from no difference or a difference in the other direction. The registry does not report the confidence level for this recalculated interval, and it does not report an interval for the sponsor's regression estimate.

General explanation: an interval is not a guarantee that the true effect lies within its endpoints, nor does it assign probabilities to individual values after the data have been observed. Its width conveys precision: a broader interval leaves more uncertainty about the effect's size. The interval must be interpreted alongside the design, the outcome definition, and the analysis used.

What does a p value convey?

The recalculated unadjusted p value is 0.201, while the registry's sponsor analysis reports p=0.22. A p value describes how unusual the data, or data more extreme, would be under a model that assumes no difference and the analysis assumptions. It is not the probability that the no-difference assumption is true, and it does not measure the probability that a result will replicate. The registry does not report the primary analysis significance threshold, so the p values should be read descriptively rather than as a complete decision rule.

General explanation: a p value is sensitive to the amount of information as well as the size of the estimated contrast. A comparatively large p value cannot establish equivalence, and a small p value would not by itself establish practical importance. For this pilot, the estimate and its interval are important because they show the scale and imprecision of the contrast, not just a test statistic.

What is the role of randomization?

Random allocation is intended to make assigned groups comparable on average, including with respect to measured and unmeasured factors. It does not guarantee that groups in a small study will be balanced on every characteristic. The registry identifies the design as randomized and parallel, but does not report sequence generation, allocation concealment, stratification, or baseline characteristics. The record also labels the trial single-masked without identifying who was masked. These omissions limit evaluation of how the design was implemented and how risk of bias was addressed.

The sponsor's reported analysis and the summary-statistic recalculation are not described with identical levels of detail. Since the record does not flag a discrepancy, the recalculated values can be presented as unadjusted checks alongside the sponsor's reported result. They should not be taken as evidence that the sponsor used the same model or assumptions. The record's original statistical plan, any planned interim analyses, and the handling of missing data are not reported in the registry record.

10. Confidence Intervals and Precision

The recalculated mean difference is 0.555 kg for beverage relative to solid treatment. Its interval runs from below zero to 1.406 kg. This interval spans the no-difference value and includes estimates on either side of it. Accordingly, the result leaves considerable uncertainty about whether the average change was higher or lower between groups, and about the size of any difference. The confidence level for this recalculated interval is not reported in the registry record.

Precision is limited in part by the small posted arm sizes: 10 participants in each group. The standard error is 0.307 kg for each arm. The observed arm means and their standard errors do not describe the range of individual outcomes, and the registry does not post participant-level data. The interval's width is more informative about uncertainty in the group contrast than the point estimate alone.

The sponsor's own analysis reports a net mean difference of 0.56 and p=0.22 but no confidence interval. The recalculated unadjusted analysis reports p=0.201 and an interval. These results are similar in broad direction, but their p values are not identical; the record does not provide enough model detail to attribute that difference to a particular analytical choice. The record does not flag a discrepancy between the sponsor analysis and the recalculation.

The interval should not be interpreted as demonstrating equivalence or as ruling out meaningful differences. It is a measure of uncertainty under the calculation's assumptions. Because the study is a pilot and the estimate is imprecise, the observed contrast is a limited basis for drawing conclusions about the expected effect in a larger study.

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 planning figures are participants per arm, not total events. The illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its reference and considers detecting the stated fraction of that effect. These are planning assumptions for the illustration; the trial's own original design assumptions are not reported in the registry record.

For planning around 100% of the observed effect, the listed requirement is 49 participants per arm for 80% power and 65 participants per arm for 90% power. For planning around 75% of the observed effect, the corresponding figures are 86 participants per arm for 80% power and 115 participants per arm for 90% power. For planning around 50% of the observed effect, the figures are 193 participants per arm for 80% power and 258 participants per arm for 90% power.

The 80% power column describes the planned chance of detecting the chosen target effect under the stated statistical assumptions if that effect is the true one. The 90% power column uses a higher target chance of detection under those same assumptions. Smaller target effects require more participants per arm because they are harder to distinguish from random variation. The curve should be read as the same planning relationship across effect assumptions, not as a record of outcomes collected by the trial.

This is not observed power. The figures are forward-looking planning illustrations based on the observed effect, not evidence that the completed study had a particular probability of detecting an effect and not a description of the original design justification.

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 = 0.555; pooled standard deviation σ = 0.971.
2. n = 2 × (1.96 + 0.842)² × 0.971² / 0.555² = 48.0, rounded up to 49 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 <- 0.555                      # observed difference in means
sigma <- 0.970819241671693                     # 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(3190993)
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 = 49 (80% power), 65 (90% power)
 75% of observed effect: n per arm = 86 (80% power), 115 (90% power)
 50% of observed effect: n per arm = 193 (80% power), 258 (90% power)
power.t.test: 49 
Simulated power at n = 49 per arm: 0.804 (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. Analyses Reported by the Sponsor

The registry reports the sponsor's primary analysis as linear regression, with a net mean difference of 0.56 and p=0.22. It does not give a confidence interval for this analysis or describe the regression model in detail. The sponsor result is the study's reported analysis and should be distinguished from the unadjusted recalculation made from the posted arm summaries.

The recalculated comparison is 0.555 kg, with an interval extending from below zero to 1.406 kg and an unadjusted p value of 0.201. The recalculation uses the arm-level means and standard errors. It is not a reconstruction of an unspecified regression model. The registry does not flag a discrepancy, so both results can be reported with their methods identified; the recalculation is a supplementary check rather than a replacement for the sponsor's analysis.

The point estimates, 0.56 in the sponsor result and 0.555 in the recalculation, are close as reported, while the p values differ. The record does not provide sufficient details to say why. Differences in model specification, adjustment, or calculation could affect results in general, but the reason in this trial is not reported in the registry record. No confidence interval for the sponsor analysis is available to compare directly with the recalculated interval.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and one analysis, and no secondary outcomes are posted. This limits what can be assessed about multiple outcome testing, but it does not establish that no additional analyses were conducted. The record does not report a multiplicity adjustment, a prespecified hierarchy of outcomes, or an analysis plan specifying how multiple comparisons would be handled. Whether multiplicity was a concern in the original analysis is therefore not reported in the registry record.

The primary outcome population field is blank. The record does not say whether the analysis included all randomized participants, only participants with an observed endpoint, or another analysis set. It also does not report the number of missing weight measurements, reasons for missingness, or any imputation method. The posted arm sizes for the outcome are 10 in each group, but those summaries alone do not establish how missing data were handled or whether all randomized participants contributed.

The record identifies the design as randomized and single-masked, but does not specify who was masked or whether masking was maintained. It does not report a protocol-level statistical analysis plan, interim analysis procedures, or the handling of deviations from assigned treatment. Those details matter because exclusions, unplanned analyses, and outcome availability can affect interpretation even when group summaries and a test result are posted.

These omissions do not invalidate the reported numbers, but they narrow the claims that can be made from them. The clearest description is the posted mean change in each arm and the reported between-group analysis, accompanied by uncertainty about the precise analysis population, missing data, and any unreported analytical decisions.

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.

The registry reports serious adverse events for each treatment group. For the solid-treatment group, 10 participants were listed as at risk, 0 as affected by a serious adverse event, and 0 deaths. For the beverage-treatment group, 10 were listed as at risk, 0 were affected by a serious adverse event, and 0 deaths. No serious adverse event terms are listed.

The record also reports 0 participants affected by other adverse events in each group, with 10 listed as at risk in each group. These are registry counts for the recorded categories and do not describe the nature, timing, or severity of symptoms that were not classified as adverse events. Zero recorded events should be read as a statement about what the registry reports during this study, not as proof that no possible adverse events could occur.

The small study and its limited follow-up at 1 month restrict the scope of safety conclusions. The registry does not provide additional detail on event collection procedures or the period used for adverse-event surveillance. The figures therefore describe the posted safety record without establishing broader safety conclusions.

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

The results describe average body-weight change over 1 month in two randomized groups receiving sugar-sweetened foods in different forms. The solid-treatment group had a mean change of -0.325 kg, and the beverage-treatment group had a mean change of 0.23 kg. The unadjusted recalculated contrast, beverage relative to solid, is 0.555 kg. The sponsor's regression analysis reports a net mean difference of 0.56.

The interval for the recalculated contrast extends from below zero to 1.406 kg, and the recalculated p value is 0.201. The sponsor's reported p value is 0.22. The interval's inclusion of zero and its span indicate that the study estimates the contrast imprecisely. These results do not settle whether the average change differed, nor do they pin down the magnitude of a possible difference.

The results do not show that either treatment works or fails to work. The trial compares two sugar-sweetened treatment forms; it does not include a comparison group without those treatments in the posted primary analysis. The data therefore do not establish what weight change would have been without either treatment. The contrast also does not establish effects for different foods, treatment durations, or participant populations.

The pilot size, limited registry detail, and absence of posted secondary outcomes further constrain interpretation. The population for the primary outcome is not described, and handling of missing measurements is not reported. The study is most informative as an estimate of the observed group contrast and the uncertainty around it, rather than as a definitive answer about the broader effects of food form on body weight.

17. Why This Trial Matters Statistically

This trial is statistically interesting because its question compares food form while describing the treatments as identical in carbohydrate, or sugar. The contrast is between a solidified food made from beverage syrup concentrate and gelatin and a carbonated sugar-sweetened beverage. The primary outcome is a continuous measure, change in body weight at 1 month, so the central statistical quantity is a difference in group means.

The pilot also illustrates why an estimated effect should be read alongside its uncertainty. The recalculated contrast is 0.555 kg, but the interval extends from below zero to 1.406 kg. The point estimate alone can suggest a clear direction; the interval makes clear that the data are consistent with a wider range of possibilities. The sponsor's reported p value and the recalculated p value are useful descriptions, but neither replaces the effect estimate and its precision.

The posted planning illustrations demonstrate how target effect size affects sample requirements. When the planning target is a smaller fraction of the observed effect, the listed number of participants per arm rises. This illustrates a general design principle: detecting smaller differences generally requires more information. It does not show the original rationale for the trial's enrollment or establish the effect that a future study should expect.

The study also highlights reporting questions that shape interpretation: which participants were included, how missing outcomes were handled, who was masked, and how the regression model was specified. Those details are not reported in the registry record. For clinical researchers, distinguishing the posted sponsor analysis from an unadjusted re-calculation and separating planning illustrations from observed results are essential to reading a pilot study accurately.

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 posted primary outcome analysis is based on 10 participants in each arm. This limits precision, as reflected by the interval for the recalculated contrast, which extends from below zero to 1.406 kg. Pilot results can be especially sensitive to chance variation, so the observed arm means should not be treated as stable estimates of what a larger study would find.

The registry record does not specify the outcome population, provide participant-level data, or describe missing measurements and their handling. It does not give baseline characteristics, weight-measurement procedures, adherence information, or details about how the intervention was delivered. Consequently, it is not possible from the registry record to assess group comparability beyond the randomized design label or to examine patterns of response within the arms.

Although the trial is labeled single-masked, the masked party is not identified. The allocation sequence and concealment procedures are not described. The sponsor analysis is labeled linear regression, but covariates and other model details are not reported; no interval accompanies that result. These gaps limit direct evaluation of the analysis and its assumptions.

The endpoint covers change at 1 month, and no secondary outcome results are posted. The record does not establish whether effects persist beyond that timeframe or apply to other populations and food forms. The sample-size illustrations are based on the observed effect rather than the original design plan, which is not reported. These limitations support a cautious reading focused on the specific, short-term comparison and the uncertainty of its estimate.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: The DEB Pilot Study — NCT03190993.
• Registry record last updated 2020-10-19; status last verified 2020-10.
• Results first posted 2020-05-13.
• 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 DEB Pilot Study (NCT03190993) is recorded as a completed randomized, single-masked, parallel-group study sponsored by Pennington Biomedical Research Center. It compared a sugar-sweetened solid treatment with a sugar-sweetened beverage treatment. The primary endpoint was change from baseline in body weight at 1 month, measured in kg.

The posted mean changes were -0.325 kg for the solid group and 0.23 kg for the beverage group, with standard errors of 0.307 kg in both arms. The sponsor's linear-regression analysis reports a net mean difference of 0.56 and p=0.22. The recalculated unadjusted difference is 0.555 kg, with an interval extending from below zero to 1.406 kg and p=0.201.

The interval indicates substantial uncertainty about the contrast. The registry does not report the primary outcome population, missing-data handling, the masked party, or detailed regression specifications. No secondary outcomes are posted, and the adverse-event record lists zero serious events and zero other adverse events in each arm.