Impact of Emergency Department Probiotic (LGG) Treatment of Pediatric Gastroenteritis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Impact of Emergency Department Probiotic (LGG) Treatment of Pediatric Gastroenteritis (NCT01773967), a randomized phase 2, phase 3 study. The primary outcome is Number of Participants With Modified Vesikari Scale Score >=9.

2. Clinical Question

The primary endpoint is a binary event: whether a participant reached the score threshold. A score of at least 9 indicates moderate-to-severe gastroenteritis; higher scores indicate worse severity. The endpoint includes features of diarrhea, vomiting, and fever, along with use of health care resources. It therefore represents a severity outcome rather than a direct measure of one symptom alone.

The registry record describes the study as examining LGG efficacy and side-effect profile. The specific primary comparison concerns the severity threshold, while the posted secondary outcomes include diarrhea duration and the number of participants with LGG bacteremia. The registry record does not report detailed eligibility criteria or further information about the population beyond pediatric patients with acute gastroenteritis.

3. Trial Design

This was a randomized, parallel-group trial recorded as Phase 2 and Phase 3. The registry lists quadruple masking. That label describes the recorded masking category; the registry record does not identify which roles were masked or provide information about whether masking was maintained throughout the trial. Washington University School of Medicine is listed as the sponsor. The study is marked completed.

The registry reports enrollment of 971 participants. One group received LGG and the other received micro-crystalline cellulose as placebo. Both treatments were described as oral and twice daily for 5 days. The LGG arm description specifies LGG at 10^10 cfu; the placebo arm description names micro-crystalline cellulose. Other details about treatment delivery, adherence, or changes to treatment are not reported in the registry record.

The recorded start was 2014-07, and primary completion was 2018-08. The primary outcome was the number of participants with a Modified Vesikari Scale score of at least 9 during 14 days. The posted results describe 55 such participants in the LGG group, with 468 participants in that arm, and 60 in the placebo group, with 475 participants in that arm. The registry also lists diarrhea duration and LGG bacteremia as secondary outcomes. Detailed procedures for randomization, allocation concealment, follow-up, and outcome ascertainment 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

Reasons for not completing this period:

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 Number of Participants With Modified Vesikari Scale Score >=9, measured over 14 days.

The primary endpoint counts participants whose Modified Vesikari Scale score reached at least 9 within 14 days. The registry describes this as a validated gastroenteritis severity score. It combines information on the duration and frequency of diarrhea, vomiting, and fever with use of health care resources. The threshold marks moderate-to-severe gastroenteritis, and higher scores represent worse severity.

Using a threshold turns the outcome into a binary measure: each participant either meets the specified severity criterion or does not. This allows the groups to be compared by the proportion of participants experiencing the event. Such an endpoint can be easier to interpret as a count of participants than a comparison of the full range of possible scores, though it does not show how far scores lie above or below the threshold.

The registry record reports the event counts and arm sizes for this outcome. It does not report the distribution of individual scores, the exact score for each participant, or additional details about how the scale was assessed. The primary endpoint should therefore be understood as a comparison of participants crossing the recorded threshold, not as a comparison of average severity scores or of each component symptom separately.

7. Results Posted in the Registry

• Lactobacillus Rhamnosus GG: 11.8% (55/468) versus 12.6% (60/475) on control; risk difference -0.9 percentage points (95% CI -5.1 to 3.3), risk ratio 0.93 (0.66–1.31); unadjusted two-proportion z-test p = 0.680.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary because it records whether each participant did or did not have a Modified Vesikari Scale score of at least 9 within 14 days. For this kind of outcome, a group event proportion describes the fraction of participants who experienced the event. The registry posts counts and arm sizes, and the recalculated comparisons use these summary data without adjustment.

What does the risk difference describe?

The risk difference compares the event proportions between the LGG and placebo groups by subtracting one group’s proportion from the other. Its estimate is -0.009, with the comparison defined as LGG relative to placebo. A value of zero represents no difference in event proportions. A negative estimate means the observed proportion was lower in the LGG group, while a positive estimate would mean it was higher. This point estimate is not a statement of certainty: its confidence interval, from -0.051 to 0.033, includes zero.

The risk difference is on the proportion scale and expresses an absolute contrast. It can be useful for understanding how far apart the observed group proportions are on that scale. Its interpretation depends on the endpoint definition and the populations represented by the analyzed groups. The registry record does not report additional adjustment for participant characteristics in the recalculated result.

How do relative measures differ?

The risk ratio compares the event proportions as a ratio, while the odds ratio compares the odds of the event. Both are recalculated from the posted counts and arm sizes. The risk ratio is 0.930, with a confidence interval from 0.660 to 1.311; the odds ratio is 0.921, with a confidence interval from 0.623 to 1.361. For each relative measure, a value of one represents no difference. Both intervals include one. Risk differences and relative measures use different scales, so their point estimates should not be compared as though they were interchangeable.

These relative estimates are also unadjusted. The registry's own reported analysis is identified as Mantel-Haenszel, but the record does not report the analysis parameter, confidence interval, or details of any stratification. The Mantel-Haenszel method can combine comparisons across strata. General statistical context: stratification can account for defined groupings in an analysis, but the registry record does not specify which groupings, if any, were used for this result. The sponsor's reported analysis remains the trial's own result.

What does the p-value tell the reader?

The sponsor reports p = 0.83 for the superiority analysis. The unadjusted recalculation gives p = 0.68. A p-value measures how unusual data at least as inconsistent with a specified null hypothesis would be if that null model and the analysis assumptions were true. It is not the probability that the null hypothesis is true, nor does it measure the clinical importance of a difference. A larger p-value does not prove that the groups are equivalent.

General statistical context: p-values are affected by the amount of information in the data as well as the size and variability of an observed effect. They should be read alongside the estimate and its confidence interval. Here, the unadjusted interval for the risk difference includes zero and spans both negative and positive values, so it communicates uncertainty about direction as well as magnitude.

Why can analyses differ?

The Mantel-Haenszel result and the recalculated unadjusted result need not be identical because they are not described as the same analysis. The recalculation uses only posted group counts and sizes, whereas the reported method is Mantel-Haenszel. Differences in adjustment or stratification can change estimates or uncertainty. The registry record does not report enough detail to identify the specific reason for the difference. The discrepancy flag is false, and the recalculation is not presented as a correction to the sponsor's result.

All recalculated comparisons are simple, unadjusted calculations based on the posted summary statistics. They do not recover participant-level information and cannot account for details absent from those summaries. No additional assumptions about the protocol, statistical plan, or analysis populations can be inferred from the listed results.

10. Confidence Intervals and Precision

The recalculated risk difference is -0.009, with a confidence interval from -0.051 to 0.033. A confidence interval describes a range of values reasonably compatible with the data under the method used; it is not a guarantee that the true value lies within those particular endpoints. The interval includes zero, the value representing no absolute difference, and it extends to both sides of zero. Thus, the estimate does not tightly determine the direction of the difference.

The interval also shows that the point estimate is not the only plausible value under this analysis. The endpoints are separated enough to leave uncertainty about the size of the absolute contrast. Reading only the point estimate would conceal that uncertainty. The registry's reported analysis does not include a confidence interval in the posted result, so the precision discussion relies on the unadjusted recalculation rather than attributing that interval to the sponsor's analysis.

The recalculated risk ratio is 0.930, with a confidence interval from 0.660 to 1.311, and the recalculated odds ratio is 0.921, with a confidence interval from 0.623 to 1.361. For these relative measures, one represents no difference; both intervals include one. General statistical context: confidence intervals for relative measures are interpreted on their own ratio scale and should not be read as absolute changes in event proportions.

The precision available from these summary results is limited to the reported endpoint and the stated unadjusted methods. Details about the sponsor analysis's uncertainty, including its confidence interval, are not reported in the registry record. The interval estimates do not establish equivalence, and they do not resolve the significance or meaning of secondary outcomes.

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 event counts. They are illustrations based on the observed effect, not the trial's own design justification. The listed 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 set to a stated fraction of the observed effect. The registry record does not report the trial's original sample-size assumptions or justification.

For planning at 100% of the observed effect, the figures are 21,728 participants per arm for 80% power and 29,087 participants per arm for 90% power. At 75% of the observed effect, they are 38,927 participants per arm for 80% power and 52,112 participants per arm for 90% power. At 50% of the observed effect, they are 88,259 participants per arm for 80% power and 118,153 participants per arm for 90% power.

In plain terms, the 80% power column gives the estimated number of participants per arm for a planning scenario designed to have an 80% chance of detecting the specified effect under the listed assumptions. The 90% power column gives the corresponding planning number for a 90% chance. These are not probabilities that the observed result is correct. Smaller effects require more participants to detect under otherwise comparable assumptions, as reflected by the larger planning figures for the smaller fractions of the observed effect. Any curve associated with these illustrations should be read as showing how the planning requirement changes with the effect being targeted; it is not a record of the trial's observed outcomes. This is 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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.126; treatment proportion p₁ = 0.118; average p̄ = 0.122.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4627 = 0.9069.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4627 = 0.3894.
4. n = (0.9069 + 0.3894)² / (-0.0088)² = 21,727.2, rounded up to 21,728 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
p0 <- 0.126315789473684                          # control proportion
p1 <- 0.117521367521368                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1773967)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 21,728 (80% power), 29,087 (90% power)
 75% of observed effect: n per arm = 38,927 (80% power), 52,112 (90% power)
 50% of observed effect: n per arm = 88,259 (80% power), 118,153 (90% power)
power.prop.test: 21727.2 
Simulated power at n = 21,728 per arm: 0.799 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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 superiority analysis as Mantel-Haenszel, with p = 0.83. The registry does not post an estimate or confidence interval for that analysis. Separately, an unadjusted recalculation from the posted event counts and arm sizes gives a risk difference of -0.009 and p = 0.68, with a confidence interval from -0.051 to 0.033.

The recalculation is not the same as the sponsor's analysis. It uses the aggregate counts and does not reproduce any possible stratification or other analysis details. The registry record does not report enough information to explain the difference between the p-values. The discrepancy flag is false, and the sponsor's own analysis takes precedence as the trial's reported result. The recalculated result is useful as a transparent, simple comparison of the posted counts, but it should not be described as replacing or correcting the reported analysis.

Both results should be interpreted in context: the sponsor's p-value is not a measure of treatment importance, and the unadjusted estimate comes with uncertainty spanning values on either side of no difference. The registry record does not report a sponsor confidence interval for comparison with the recalculated interval.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists three outcome measures and posts two secondary outcomes: the number of participants with LGG bacteremia and diarrhea duration. It reports two analyses. The registry entries do not identify every analysis in relation to every outcome, or indicate whether all planned outcomes and analyses are posted. The record therefore does not allow a complete reconstruction of the outcome-by-analysis structure.

Multiplicity arises when several outcomes, comparisons, or analyses are considered. General statistical context: testing multiple questions can increase the chance of at least one apparently notable result by chance, depending on the testing strategy. The registry record does not report whether the trial used a multiplicity adjustment, a prespecified testing sequence, or another procedure to address this issue. No adjustment should be assumed from the listed outcomes or analyses.

The registry record does not report the handling of missing primary-outcome data, the number of participants with unavailable outcomes, or the assumptions used for any missing-data procedure. It also does not specify the analysis population, such as whether the primary analysis followed all randomized participants or a narrower group. The outcome arm sizes are posted, but those counts alone do not establish the analysis population or explain how participants without outcome data were handled.

Other details not reported in the registry record include the original statistical analysis plan, interim analyses, methods for protocol deviations, and sensitivity analyses. These absences limit assessment of whether the posted analysis followed a prespecified plan and how robust the result is to alternative assumptions. They do not establish that any particular procedure was or was not performed; they mean only that the relevant details are not reported in the registry record.

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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: worsening diarrhea (0.8%), lower respiratory infection (0.4%), dehydration (0.4%), ileus (0.2%), seizure (0.2%).

The serious adverse event figures list 8 affected participants among 483 at risk in the LGG group and 13 among 488 at risk in the placebo group. The record lists zero deaths in each group. These are reported counts, and they should be read with their stated at-risk group sizes. They do not, on their own, establish whether an event was caused by the assigned treatment.

The most frequently listed serious event terms include worsening diarrhea at 0.8%, lower respiratory infection at 0.4%, dehydration at 0.4%, ileus at 0.2%, seizure at 0.2%, and lethargy at 0.2%. These percentages are quoted as recorded. The registry record does not provide event-specific participant counts by treatment group for each term in this list, or further detail about timing, severity, or causality.

The other adverse-event figures list 55 affected participants among 483 at risk in the LGG group and 49 among 488 at risk in the placebo group. These figures are separate from the serious-event counts. The posted safety summary is descriptive; it does not establish that the groups differ in safety, or that any listed event resulted from treatment. Further details about adverse-event collection and assessment are not reported in the registry record.

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

The primary outcome compares the number of participants who met the Modified Vesikari Scale threshold for moderate-to-severe gastroenteritis during 14 days. The posted counts are 55 in the LGG group and 60 in the placebo group. The sponsor's reported Mantel-Haenszel analysis has p = 0.83. The unadjusted recalculation produces a risk difference of -0.009, with a confidence interval from -0.051 to 0.033, and p = 0.68.

These results do not show a statistically clear difference in the primary outcome under the reported analysis. They also do not demonstrate that the treatments are equivalent. A failure to find a clear difference is not proof that no difference exists: the recalculated interval includes zero and allows values on either side of it. The estimate and interval should be considered together rather than treating the p-value as a yes-or-no verdict.

The primary endpoint captures crossing a score threshold, not every aspect of gastroenteritis. It does not show the full distribution of scores or isolate the contribution of diarrhea, vomiting, fever, or health care use. Diarrhea duration and LGG bacteremia are listed as secondary outcomes, but their detailed results are not reported among the registry entries. The primary result cannot answer those separate questions.

The reported safety counts and serious-event terms provide descriptive information, not proof of causation or a complete account of risk. The registry record does not include the additional methods and participant-level details needed to evaluate all sources of uncertainty. Accordingly, the result should be read as evidence about this particular recorded endpoint and analysis, with the stated precision and reporting limits, rather than as a broad conclusion about every possible outcome or patient group.

17. Why This Trial Matters Statistically

This trial illustrates why effect estimates and uncertainty intervals matter alongside p-values. The sponsor's reported p-value is 0.83, while the unadjusted recalculation gives p = 0.68. Neither number describes the size of the group difference. The recalculated risk difference of -0.009 and its interval from -0.051 to 0.033 make the estimated contrast and its uncertainty more visible. Because the interval includes zero and spans both directions, it cautions against interpreting the point estimate as a precise account of the underlying difference.

The endpoint also demonstrates the trade-off involved in using a thresholded severity score. Counting participants who reach at least 9 on the Modified Vesikari Scale creates a clear binary outcome, but it does not retain the full information in each score or show how individual symptoms contribute. Statistical interpretation depends on keeping the endpoint definition in view rather than treating the count as a complete description of illness severity.

The registry's reporting provides useful information about randomization, masking category, treatment arms, primary outcome, and the sponsor's analysis, while leaving important analytic details unreported. That distinction is central to sound statistical reading: results can be interpreted as posted without filling gaps about the protocol, missing data, or multiplicity. The planning illustrations further show that sample-size requirements depend on the effect targeted and the assumptions used, but they are not evidence about the study's original design rationale.

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.

Several trial-specific reporting limits affect interpretation. The registry record does not report detailed eligibility criteria, age limits, or other characteristics of the pediatric population. It also does not provide participant-level outcome distributions, so the primary result can be interpreted only as a count above the specified score threshold, not as a full comparison of score severity or component symptoms.

The sponsor's reported method is identified as Mantel-Haenszel, but the analysis estimate, confidence interval, and details of any stratification are not reported. The unadjusted recalculation from the posted counts is not a substitute for those missing details. The registry record also does not report how missing outcome data were handled, which participants were included in each analysis, whether analyses followed a prespecified plan, or whether multiplicity was addressed.

The secondary outcomes are named, but detailed results for diarrhea duration and LGG bacteremia are not included in the registry record for this interpretation. The safety summary gives group counts and selected serious-event terms, but does not report further event-specific detail, timing, or causality assessment. Consequently, neither the primary endpoint nor the safety summary provides a complete description of all outcomes relevant to the study.

The record reports quadruple masking but does not specify which parties were masked or whether masking was maintained. It lists a completed study and dates for start and primary completion, but does not provide further information about treatment adherence, protocol deviations, or follow-up beyond the stated primary timeframe. These are limits of what is reported, not evidence that particular procedures were absent.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Impact of Emergency Department Probiotic (LGG) Treatment of Pediatric Gastroenteritis — NCT01773967.
• Registry record last updated 2019-09-06; status last verified 2019-08.
• Results first posted 2019-09-06.
• 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

NCT01773967 is a completed randomized, parallel-group Phase 2 and Phase 3 trial sponsored by Washington University School of Medicine. The registry records quadruple masking, enrollment of 971, a start of 2014-07, and primary completion in 2018-08. The study compared oral LGG with a micro-crystalline cellulose placebo, with both treatments described as twice daily for 5 days.

The primary outcome was the number of participants with a Modified Vesikari Scale score of at least 9 within 14 days. The posted counts were 55 among 468 participants in the LGG group and 60 among 475 in the placebo group. The sponsor's reported Mantel-Haenszel superiority analysis gives p = 0.83. An unadjusted recalculation gives a risk difference of -0.009, with a confidence interval from -0.051 to 0.033, and p = 0.68. The registry also lists diarrhea duration and LGG bacteremia as secondary outcomes.