Direct Swallowing Training and Oral Sensorimotor Stimulation in Preterm Infants: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Direct Swallowing Training and Oral Sensorimotor Stimulation in Preterm Infants (NCT02508571), a randomized na study. The primary outcome is Days From Start to Independent Oral Feeding.

2. Clinical Question

The registry describes a randomized controlled trial of oral feeding performance in preterm infants. Its primary outcome is a continuous duration measure, expressed in days. The posted results give a group mean and standard deviation for each arm. The comparisons presented for the intervention groups are mean differences relative to control, with negative values indicating a lower mean number of days than in control.

The question is specifically about time to the stated definition of independent oral feeding, not about every aspect of feeding development or later health. Other listed outcomes include days to first full oral feeding, days to complete full oral feeding, measures of milk transfer and volume loss, an oral motor assessment, and length of hospital stay. Findings for those outcomes are not included in the posted primary results, and should not be inferred from the primary endpoint.

3. Trial Design

The registry identifies NCT02508571 as a completed, randomized, triple-masked, parallel-group trial. It lists Seoul National University Hospital as the sponsor and an enrollment of 189. The listed start was 2015-07, and primary completion was 2020-09. The phase is recorded as NA. These registry entries describe the trial’s design and status; they do not establish how masking was implemented or which trial roles were masked.

Three groups are described. The control group received two 15-minute sham sessions per day, five days a week. The DST group received one DST session and one sham session per day, five days a week. The DST+OSMS group received one DST session and one OSMS session per day, five days a week. The record classifies all three interventions as “OTHER.” The detailed procedures, timing in relation to feeding, duration of the intervention course, and criteria for changing or stopping sessions are not reported in the registry record.

For the primary outcome, the registry lists 63 participants in control, 63 in DST, and 60 in DST+OSMS. These are the group counts attached to the posted outcome summaries. The record separately lists safety denominators, which should be read in their own context rather than assumed to be identical to the outcome counts. The registry does not report the allocation sequence method, concealment procedures, or a detailed description of the trial setting and participant eligibility.

The primary endpoint is reported as a mean with standard deviation. The registry includes several secondary outcomes and records 14 posted secondary outcomes. It does not include analysis records. Details about the trial’s original statistical analysis plan, analysis populations, missing-data handling, and any adjustment for multiple comparisons are 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.

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 Days From Start to Independent Oral Feeding, measured over From date of starting oral feeding until the date of independent, full oral feeding, an expected average of 3 weeks.

The primary endpoint is the number of days from the start of oral feeding until independent, full oral feeding. The registry defines independent oral feeding as two days in a row with no adverse events that do not self-resolve, and uses the first successful day as the endpoint date. The stated timeframe begins on the date oral feeding starts and ends on the date of independent, full oral feeding; the registry gives an expected average of 3 weeks for this timeframe.

This endpoint turns progress toward independent feeding into a continuous duration measure. Reporting the mean number of days describes the average time in each group, while the standard deviation describes how dispersed individual times are around that mean. A lower group mean represents fewer recorded days to the endpoint on average, but it does not by itself describe the experience of every infant or the spread of times within the group.

The endpoint depends on both the definition of success and the dates used to mark its start and completion. The registry supplies a definition of independent feeding, but does not report further details about how endpoint assessment was conducted, whether assessors were unaware of group assignment, or how infants who did not reach the endpoint were handled in the analysis. The other feeding and hospital outcomes may capture different aspects of care, but they do not replace the primary endpoint.

7. Results Posted in the Registry

• DST Group: mean 17.20 versus 21.10 on control; difference -3.90 (95% CI -8.40 to 0.60), standardized effect d = -0.30; z-test p = 0.089.
• DST+OSMS Group: mean 14.80 versus 21.10 on control; difference -6.30 (95% CI -10.76 to -1.84), standardized effect d = -0.49; z-test p = 0.006.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous, and the registry reports it using the mean and standard deviation. The mean is the arithmetic average of the recorded days in a group. The standard deviation describes how spread out those days are around the group mean. A larger standard deviation indicates more variability in the observations; it does not indicate uncertainty in the mean by itself. Precision of a mean difference also depends on the number of observations and on the variability within groups.

What does a mean difference represent?

A mean difference subtracts one group’s mean from another group’s mean according to the comparison’s stated direction. In the reported comparisons, control is the reference. Thus, the estimate of -3.9 days for DST means the DST group mean was lower than the control mean by the stated amount in the unadjusted recalculation. The estimate of -6.3 days for DST+OSMS has the same direction relative to control. These estimates describe differences in averages, not a change experienced by every participant.

The comparisons are marked as unadjusted recalculations based on the posted group summaries. They use the reported group means, standard deviations, and outcome group counts to estimate a mean difference and its uncertainty. The registry record does not report a model-based or covariate-adjusted comparison, so such an analysis should not be assumed. The registry also does not report its original statistical analysis plan or a sponsor analysis for this endpoint.

How should the intervals be understood?

A confidence interval is a range of values generated by a statistical procedure to express uncertainty around an estimate. In general, a wider interval indicates less precision, while a narrower interval indicates greater precision, under the method’s assumptions. For these comparisons, the reported ranges are -8.4 to 0.6 for DST versus control and -10.76 to -1.836 for DST+OSMS versus control. The confidence level and the exact interval method are not reported in the registry record, so the ranges should be interpreted as the posted intervals without assigning them an unreported confidence level.

For a mean difference, zero represents no difference in group means. The DST interval includes zero; the DST+OSMS interval does not. This distinction describes the intervals and does not establish clinical importance. An interval can include values that may matter in different directions, and statistical uncertainty is separate from the question of whether a difference would be important in practice. The record does not specify a threshold for a clinically meaningful difference.

What do the p-values say?

The posted p-values are unadjusted: 0.089 for DST versus control and 0.006 for DST+OSMS versus control. A p-value measures how unusual data at least as discrepant as the observed data would be under a specified null model and its assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of a difference. Since the registry does not report the trial’s multiplicity plan or confirmatory analysis, these values should be read as unadjusted comparisons rather than as a complete account of the trial’s inferential plan.

General statistical explanation: when several outcomes or group comparisons are examined, the chance of obtaining at least one small p-value can increase if each is assessed separately. Multiplicity procedures can address that issue, but the appropriate approach depends on the prespecified analysis plan. Whether such a procedure was used in this trial is not reported in the registry record. The posted p-values therefore should not be described as multiplicity-adjusted.

Why does variation matter?

General statistical explanation: greater variability among individual observations makes a group mean harder to estimate precisely when other features of the design are held constant. Standard deviations of 15.3 in control, 9.9 in DST, and 9.4 in DST+OSMS describe variation in the recorded days within the respective groups. They are not confidence limits for the group means. The interval around a mean difference incorporates sampling uncertainty and is the more direct summary of precision for the comparison.

The registry reports a parallel randomized design, but it does not report details such as baseline adjustment, handling of censoring if an infant did not reach the endpoint, or how missing observations were treated. These choices can affect an analysis of time until an event. The record’s summary does not establish whether a time-to-event method or a simple comparison of means was used by the trial investigators. The posted recalculations are unadjusted comparisons from summary statistics, not a reconstruction of unreported individual-level analyses.

General statistical explanation: over-dispersion refers to variation greater than a statistical model expects. It can make standard errors and inferential conclusions unreliable if the model’s variance assumptions are not appropriate. The registry does not identify a model that would allow over-dispersion to be assessed for this endpoint. The available standard deviations show within-group spread, but do not on their own diagnose a model problem.

10. Confidence Intervals and Precision

The intervals give a sense of how precisely the mean differences are estimated by the unadjusted recalculations. For DST versus control, the estimate is -3.9 days and the reported interval extends from -8.4 to 0.6. The range crosses zero, so the summary statistics do not rule out a zero mean difference under the interval procedure. They also allow negative and positive values, which represent different directions of the group-mean contrast.

For DST+OSMS versus control, the estimate is -6.3 days and the interval extends from -10.76 to -1.836. This entire reported range is below zero. The estimate is therefore more precise in direction than the DST comparison under this unadjusted calculation, while still allowing a range of possible mean differences. The interval is not a prediction of the time for an individual infant.

The registry does not report the confidence level or interval method. The ranges should therefore be quoted as posted and not labeled with a confidence level that is absent from the record. They also do not address potential systematic uncertainty, such as bias from missing data, endpoint assessment, or analysis choices. The absence of reported details on those matters limits how fully the statistical precision can be judged.

The standard deviations—15.3 in control, 9.9 in DST, and 9.4 in DST+OSMS—show that individual recorded durations varied within each group. A mean difference can be estimated even when observations vary widely, but the estimate should not be read as a uniform change across participants. The posted group summaries do not show the full distribution of times or how many observations were missing or censored, so those features cannot be assessed from the registry record.

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 participants across both arms and not total events. For the DST comparison, planning for 100% of the observed effect gives 172 participants per arm for 80% power and 230 participants per arm for 90% power. Planning for 75% of the observed effect gives 305 and 408 participants per arm, respectively; planning for 50% gives 686 and 918. For DST+OSMS, the corresponding figures are 65 and 87 for 100% of the observed effect, 115 and 154 for 75%, and 258 and 346 for 50%.

These are planning illustrations based on the observed effect, not the trial’s original design assumptions. The stated illustration assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. The trial’s own sample-size justification and design assumptions are not reported in the registry record.

In general, power is the chance that a planned statistical test will detect an effect of a specified size under the assumptions used for planning. The 80% and 90% columns indicate target power levels for detecting the selected effect size; they are not probabilities that the observed result is correct. A 90% target uses more participants per arm than an 80% target for each listed comparison and effect fraction. Smaller target effects require more participants per arm because they are harder to distinguish from sampling variation. These values are planning illustrations, not observed power.

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 = -3.900; pooled standard deviation σ = 12.886.
2. n = 2 × (1.96 + 0.842)² × 12.886² / -3.900² = 171.4, rounded up to 172 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 <- -3.9                      # observed difference in means
sigma <- 12.8860389569487                     # 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(2508571)
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 = 172 (80% power), 230 (90% power)
 75% of observed effect: n per arm = 305 (80% power), 408 (90% power)
 50% of observed effect: n per arm = 686 (80% power), 918 (90% power)
power.t.test: 172.3 
Simulated power at n = 172 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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists the primary outcome as days from start to independent oral feeding and also identifies numerous secondary outcomes, including measures of feeding progress, milk transfer, oral motor assessment, and hospital stay. It records 15 outcome measures and 14 posted secondary outcomes. The available primary comparison summaries include two intervention-versus-control contrasts. When multiple outcomes and comparisons are considered, the interpretation of individual p-values depends on whether analyses were prespecified and whether multiplicity was addressed.

The registry record does not report a multiplicity strategy, a hierarchy of outcomes, or which comparisons were designated confirmatory. The displayed p-values are unadjusted. It is therefore not possible from the registry record to determine whether the trial’s own analysis accounted for testing across outcomes or comparisons. The p-values should not be treated as adjusted for multiplicity.

The registry lists no analysis records. It does not report the analysis population for the primary endpoint, whether all randomized infants were included, or how protocol deviations were handled. The primary outcome group counts are 63 for control, 63 for DST, and 60 for DST+OSMS. Safety denominators are reported separately. These details should not be used to infer unreported inclusion rules or analysis populations.

Missing-data handling is also not reported in the registry record. For a duration outcome, it matters whether every participant reached the defined endpoint and how observations without an endpoint date were treated. The record does not state whether any observations were missing, censored, excluded, or imputed, and does not describe sensitivity analyses. These omissions limit assessment of potential bias and reproducibility. They do not show that missing data occurred; they mean the registry record does not report how the issue was handled.

General statistical explanation: prespecifying the primary analysis and documenting missing-data decisions help readers distinguish planned analyses from exploratory ones. Without those details, the posted means, intervals, and unadjusted p-values can be described, but the full inferential framework of the trial cannot be reconstructed.

14. 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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: death (1.6%).

The registry reports serious adverse events affecting 0 of 64 participants at risk in control, 0 of 64 at risk in DST, and 1 of 61 at risk in DST+OSMS. These figures are counts of participants affected and the corresponding at-risk denominators as entered in the safety record. They are separate from the group counts reported for the primary outcome, and should be interpreted using the safety denominators shown.

The record also lists deaths as 0 in control, 0 in DST, and 1 in DST+OSMS. Among the top serious adverse event terms, Death is listed at 1.6%. The registry does not provide additional narrative detail in the entries about timing, attribution, or relationship to an intervention. The count alone does not establish causation.

Other adverse events are listed for 9 of 64 at risk in control, 5 of 64 in DST, and 5 of 61 in DST+OSMS. These are not the serious adverse event counts. The registry record does not provide enough detail to assess event type, severity for the non-serious entries, or differences in surveillance and reporting between groups. These safety counts describe what the registry reports; they do not determine whether any intervention is safe or unsafe.

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

The primary outcome summaries show lower average numbers of days to independent oral feeding in both intervention groups than in control: 17.2 days for DST and 14.8 days for DST+OSMS, compared with 21.1 days for control. The unadjusted recalculated mean differences are -3.9 days for DST versus control and -6.3 days for DST+OSMS versus control. The interval for DST includes zero, whereas the interval for DST+OSMS lies below zero. These are descriptions of the posted group summaries and their unadjusted comparisons.

The results do not establish that the interventions caused the observed group differences. Randomization is part of the reported design, but the registry record does not provide the original analysis plan, analysis populations, missing-data handling, or a sponsor analysis for the primary comparison. The reported p-values are unadjusted, and multiplicity handling is not reported. Those limitations matter when assessing the strength of evidence across the trial’s outcomes and comparisons.

The mean differences do not describe each infant’s outcome. The within-group standard deviations show considerable spread in recorded days, and the registry record does not give the individual outcome distributions. Nor do the primary results establish effects on the listed secondary outcomes, such as length of hospital stay or oral motor assessment. Results for those outcomes should be evaluated separately.

The registry defines the endpoint and gives an expected average timeframe, but does not report a threshold for what would count as a clinically meaningful difference. Statistical evidence about a mean difference is not the same as evidence that the difference matters to infants, families, or care decisions. The findings are specific to the trial population and recorded outcome; details about participant characteristics and setting that would support broader generalization are not reported in the registry record.

Accordingly, the most defensible reading is descriptive: the posted averages favor fewer days in the intervention groups, with different uncertainty in the two comparisons. The registry record does not provide enough information to make broader claims about efficacy, safety, or effects beyond this endpoint.

16. Why This Trial Matters Statistically

This trial illustrates why statistical interpretation should separate the size and direction of an estimate from its uncertainty. The two unadjusted mean differences are both negative, but their intervals differ: the DST interval includes zero, while the DST+OSMS interval does not. Reporting the estimates alongside their intervals conveys more information than relying on p-values alone.

It also illustrates why group averages need to be read alongside measures of variation. The outcome is a duration in days, and the standard deviations indicate that individual times varied within all three groups. A mean difference is a comparison of average durations; it cannot be translated into a uniform change for every infant or a prediction for an individual.

The trial has three randomized parallel groups, which creates two intervention-versus-control comparisons in the posted primary summaries. The registry does not report whether multiplicity was handled or whether the comparisons were prespecified as confirmatory. This makes the distinction between an unadjusted summary-statistic comparison and a complete trial analysis especially important. Statistical significance in one comparison should not be considered in isolation from the broader outcome and comparison structure.

Finally, the record demonstrates the value of transparent reporting. Details about analysis populations, missing observations, endpoint assessment, the original sample-size justification, and the statistical analysis plan are not reported in the registry record. Their absence limits the ability to evaluate robustness, but does not itself establish that the methods were inappropriate. Clear reporting of these elements would help readers assess how closely the posted estimates reflect the randomized design and how much weight to give the comparisons.

17. Limitations

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

The registry provides limited information about the infants’ baseline characteristics, eligibility criteria, and clinical setting. The results are therefore difficult to place in context or to judge for applicability beyond the enrolled preterm infants. The record identifies the trial population broadly but does not report characteristics such as gestational age or baseline feeding status in the primary outcome entry.

Intervention schedules are described, including the daily sessions and weekly frequency, but the registry does not provide detailed procedures, the length of the intervention course, adherence information, or rules for modifying treatment. The record identifies triple masking but does not state which roles were masked or how masking was maintained. Assessment procedures for the primary endpoint are also not described beyond the endpoint definition.

The registry reports the primary outcome as a mean and standard deviation, but does not provide the original analysis plan, analysis populations, missing-data procedures, or the method used in the trial’s own analysis. It includes no analysis records. The posted comparisons are unadjusted recalculations from summary statistics, so they do not account for unreported covariates or individual-level data. The confidence level and calculation method for the reported intervals are not stated.

The registry lists 189 as enrollment, while the primary outcome summaries report group counts of 63, 63, and 60; those entries describe different record fields and should not be reconciled by inference. The safety record uses at-risk denominators of 64, 64, and 61. The record does not explain differences among these reported counts. This further supports keeping outcome and safety summaries distinct rather than assuming they use the same analysis set.

Finally, the posted primary results do not establish findings for the secondary outcomes. The registry record does not report a multiplicity strategy or a clinically meaningful threshold for the primary difference. These gaps limit conclusions about statistical robustness and practical importance.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Direct Swallowing Training and Oral Sensorimotor Stimulation in Preterm Infants — NCT02508571.
• Registry record last updated 2025-04-03; status last verified 2025-03.
• Results first posted 2025-04-03.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

NCT02508571 is a completed randomized, triple-masked, parallel-group trial of direct swallowing training and oral sensorimotor stimulation in preterm infants. The registry lists Seoul National University Hospital as sponsor, enrollment of 189, a start date of 2015-07, and primary completion in 2020-09.

The primary outcome is days from starting oral feeding to independent oral feeding. The posted means were 21.1 days in control, 17.2 in DST, and 14.8 in DST+OSMS. Unadjusted recalculated comparisons versus control were -3.9 days for DST and -6.3 days for DST+OSMS, with reported intervals of -8.4 to 0.6 and -10.76 to -1.836, respectively.

The registry record does not report the original analysis plan, multiplicity approach, analysis populations, missing-data handling, or detailed endpoint assessment procedures. The comparisons should be read as unadjusted summaries rather than a complete account of the trial’s analysis.