A Safety and Efficacy Comparison of BLI4700 Bowel Preparation Versus an FDA-approved Comparator in Adult Subjects Prior to Colonoscopy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Safety and Efficacy Comparison of BLI4700 Bowel Preparation Versus an FDA-approved Comparator in Adult Subjects Prior to Colonoscopy (NCT03261960), a randomized phase 3 study. The primary outcome is Number and Percentage of Subjects With Successful Bowel Cleansing.

2. Clinical Question

The primary endpoint counts participants whose preparation was rated successful. The description says that colonoscopists used a four-point scale from poor to excellent; it does not state the exact score threshold that counted as success. The results report 175 successful preparations among 190 participants in the BLI4700 arm and 174 among 198 in the comparator arm.

The posted summary comparison expresses the unadjusted difference in success proportions as a risk difference of 0.042, with an interval from -0.017 to 0.102. This is a recalculation from posted summary statistics, not the sponsor's reported test. The record also reports a separate Cochran-Mantel-Haenszel non-inferiority analysis with p < 0.001. Since the two analyses are not the same calculation and a discrepancy is flagged, the sponsor's own analysis takes precedence. The registry record does not provide enough detail to reconstruct its non-inferiority criterion.

3. Trial Design

This was a Phase 3, randomized, single-masked, parallel-group study. It compared BLI4700 bowel preparation with an FDA-approved bowel preparation before colonoscopy in adults. The study objective describes each preparation as a 2-day, split-dose regimen. The registry identifies Braintree Laboratories as the sponsor and gives the study status as completed.

The recorded enrollment is 389. The primary outcome's efficacy-population results list 190 participants in the experimental arm and 198 in the control arm. These arm-specific outcome denominators should be understood in their stated outcome context; the registry record does not explain their relationship to the overall enrollment. The safety record uses different at-risk counts for the control arm. Reasons for these differences are not reported in the registry record.

Random allocation supports a comparative design, while masking can reduce some forms of assessment or participant-related bias. The record labels masking as single, but does not identify who was masked. It also does not report the allocation method, allocation concealment procedures, site count, or whether colonoscopists were masked to preparation assignment. These details therefore cannot be used to assess how allocation and masking were implemented.

The recorded study start was 2017-08-08, and primary completion was 2018-03-26. The primary endpoint was assessed on the day of colonoscopy. Further visit timing, follow-up duration, recruitment details, and the original design assumptions are not reported in the registry record. Although the study is described as Phase 3, that designation alone does not establish what effect size, non-inferiority margin, or statistical power guided the original design.

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

Analysis population for the primary outcome, as defined in the registry: Efficacy population

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.

Population: Safety population - all patients that took preparation.

6. Primary Endpoint

The registry lists the primary outcome as Number and Percentage of Subjects With Successful Bowel Cleansing, measured over Day of colonoscopy.

The primary endpoint is the number and percentage of participants with successful bowel cleansing, assessed on the day of colonoscopy. The outcome description says that the colonoscopist rated bowel preparation on a four-point scale, with scores ranging from poor to excellent. The registry does not state the score threshold used to classify a preparation as successful, so that operational detail cannot be assessed from the record.

Preparation quality is relevant to the procedure because the trial's stated objective was to compare bowel preparations before colonoscopy. A binary success outcome provides a direct count of participants meeting the registry's success criterion and allows the two randomized groups to be compared on the proportion meeting that criterion. The endpoint is therefore interpretable at the participant level, although reducing a four-point rating to success or not successful can conceal differences among ratings within either category.

The recorded outcome belongs to the efficacy population. The BLI4700 arm reports 175 participants with successful cleansing among 190, and the comparator arm reports 174 among 198. The registry does not provide the full distribution across the four rating categories, the success threshold, or details about handling participants whose ratings were unavailable. Those omissions limit independent assessment of how ratings became the binary endpoint and whether the posted counts represent all randomized participants.

The endpoint captures cleansing success as judged by the colonoscopist; it is not, by itself, a complete measure of safety, tolerability, or other aspects of bowel preparation. The registry record lists five outcome measures but posts no secondary results. Accordingly, interpretation of the primary endpoint should not be extended to unreported outcomes.

7. Results Posted in the Registry

• Experimental Arm: 92.1% (175/190) versus 87.9% (174/198) on control; risk difference 4.2 percentage points (95% CI -1.7 to 10.2), risk ratio 1.05 (0.98–1.12); unadjusted two-proportion z-test p = 0.166.

8. Statistical Interpretation

9. Statistical Methods Explained

The outcome is binary in the posted comparison: each participant is counted as having successful bowel cleansing or not. For a binary outcome, the observed proportion is the count meeting the outcome divided by the stated group denominator. Comparing such proportions can be expressed in different ways, including a risk difference, risk ratio, or odds ratio. Each measure describes the contrast on a different scale, so the measure and its interval should be identified rather than treated as interchangeable.

What does the risk difference describe?

The recalculated risk difference is 0.042. It is the unadjusted difference between the success proportions, expressed on the proportion scale, with the BLI4700 arm first in the comparison. A positive value indicates a higher observed success proportion in that arm; a value of zero would indicate equal observed proportions. The interval reported for this estimate runs from -0.017 to 0.102. Because it includes zero, the posted interval is compatible with no difference as well as with differences in either direction. It does not identify a single effect size as certain.

Why are other effect measures also useful?

The recalculated risk ratio is 1.048, with an interval from 0.981 to 1.12. A risk ratio compares the proportions as a ratio rather than subtracting them. The recalculated odds ratio is 1.609, with an interval from 0.817 to 3.171. An odds ratio compares odds, which are not the same as probabilities. When an outcome is common, an odds ratio can look numerically more distant from one than a risk ratio, so it should not be read as a risk ratio. These measures are unadjusted recalculations from the posted summary data.

What does a p-value contribute?

A p-value describes how compatible the observed data are with a specified null hypothesis under the assumptions of the statistical test. It is not the probability that the null hypothesis is true, and it does not measure the clinical importance of an effect. The unadjusted recalculation reports p = 0.166. The sponsor's reported analysis uses a Cochran-Mantel-Haenszel method and is labeled non-inferiority, with p < 0.001. These p-values arise from different analyses and cannot be substituted for one another.

For a non-inferiority analysis, the question is whether the evidence excludes a difference worse than a prespecified margin in the relevant direction. The margin is central to interpretation, as is the estimate and uncertainty under the analysis plan. The margin, adjusted estimate, confidence interval, and detailed hypotheses are not reported in the registry record. Consequently, the sponsor's p-value cannot by itself show how large a difference was ruled out or how the margin was applied.

What does the Cochran-Mantel-Haenszel method do?

In general, the Cochran-Mantel-Haenszel method can combine evidence across strata when a comparison is made while accounting for a stratification variable. Its precise use depends on the study's strata, outcome coding, and specified test. Those implementation details are not reported in the registry record, so the specific adjustment cannot be reconstructed. The method label identifies an analysis approach but does not disclose the stratum definitions or the effect estimate.

Over-dispersion is a general concern when observations vary more than a model or test assumes. If it is present and not appropriately addressed, uncertainty can be understated. The registry does not report a model-based variance assessment or an over-dispersion analysis for this binary endpoint; it would be inappropriate to assume either that over-dispersion occurred or that a particular correction was used.

Why does the discrepancy matter?

The recalculation uses the posted arm counts and denominators to produce unadjusted comparisons. The registry also reports a sponsor analysis with a different method and non-inferiority framing, and the record flags a discrepancy. This means the recalculation is descriptive and should not be presented as an alternative sponsor result. The sponsor's own analysis takes precedence. The registry does not report enough information to determine the precise source of the discrepancy, such as differences in analysis population, adjustment, coding, or testing procedure.

In general, a randomized comparison is strongest when its analysis follows prespecified definitions and accounts for the design. The original statistical analysis plan, analysis populations beyond the primary outcome label, missing-data procedures, and any interim analyses are not reported in the registry record. Their absence prevents a fuller assessment of whether the posted analysis was planned in advance and how all available observations were handled.

10. Confidence Intervals and Precision

The interval around the recalculated risk difference extends from -0.017 to 0.102, with a point estimate of 0.042. The interval includes zero, so the recalculation does not pin down a direction with high precision. Its lower endpoint represents a negative difference on the stated scale, while its upper endpoint represents a positive difference. The range therefore accommodates different possible magnitudes and directions of the unadjusted contrast.

A confidence interval is a general statistical way to express uncertainty around an estimate. Its width reflects, among other factors, the amount of information in the data and the variability under the analysis assumptions. It should not be interpreted as a range containing every plausible value with certainty, nor as a direct measure of clinical importance. The registry record does not specify the confidence level for the recalculated interval, so it should not be assigned a level that is not reported.

The recalculated risk ratio is 1.048, with an interval from 0.981 to 1.12; the recalculated odds ratio is 1.609, with an interval from 0.817 to 3.171. Both intervals include their respective no-difference values, one for ratios. The risk-difference interval is often easier to relate to an absolute change in the outcome proportion, while ratio measures express relative contrasts. All three estimates are unadjusted recalculations, not replacements for the sponsor's analysis.

Precision also depends on which participants and outcome observations were included. The registry identifies the primary outcome population as the efficacy population but does not report its full definition, missing-outcome handling, or the derivation of the denominators. The reported sponsor analysis has no posted estimate or interval, limiting assessment of its precision. The non-inferiority p-value does not supply that missing magnitude information.

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. They are illustrations based on fractions of the observed effect, with separate columns for a target of 80% power and a target of 90% power. In plain terms, power is the chance that a specified statistical procedure will detect the effect assumed for planning, if that effect is present and the model assumptions hold. A higher target power generally requires a larger planned sample under otherwise unchanged assumptions.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect used for each illustration is the stated fraction of the observed effect. For 100% of the observed effect, the estimates are 791 participants per arm for 80% power and 1,058 for 90% power. For 75% of the observed effect, they are 1,472 and 1,970 per arm. For 50% of the observed effect, they are 3,458 and 4,629 per arm.

These figures are not the original trial's design calculations. The original design assumptions, including its planned effect and target power, are not reported in the registry record. Because the illustrations scale the effect being targeted, smaller effects need more participants to reach the same power under the stated assumptions. This is not observed power; it is a planning illustration based on an observed effect and specified assumptions, not a measure of the probability that the completed trial succeeded.

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 = 0.879; treatment proportion p₁ = 0.921; average p̄ = 0.900.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4244 = 0.8318.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4234 = 0.3563.
4. n = (0.8318 + 0.3563)² / (0.0423)² = 790.3, rounded up to 791 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 <- 0.878787878787879                          # control proportion
p1 <- 0.921052631578947                          # 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(3261960)
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 = 791 (80% power), 1,058 (90% power)
 75% of observed effect: n per arm = 1,472 (80% power), 1,970 (90% power)
 50% of observed effect: n per arm = 3,458 (80% power), 4,629 (90% power)
power.prop.test: 790.3 
Simulated power at n = 791 per arm: 0.808 (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 own primary analysis as a Cochran-Mantel-Haenszel non-inferiority test, with p < 0.001. It does not provide the sponsor's effect estimate, confidence interval, non-inferiority margin, or enough detail to reconstruct the analysis. The record separately flags a discrepancy between that result and the posted summary data.

The recalculated comparison uses the reported success counts and denominators to estimate an unadjusted risk difference of 0.042, with an interval from -0.017 to 0.102, and an unadjusted p-value of 0.166. These calculations answer a different, unadjusted comparison question. They do not reproduce the sponsor's Cochran-Mantel-Haenszel non-inferiority analysis and should not be used to overturn it.

When a discrepancy is flagged, the sponsor's own analysis takes precedence. At the same time, the registry record's missing margin and estimate limit independent evaluation of what the sponsor's non-inferiority test establishes. The reported p-value should be read as the result of that named analysis, not as information about effect size or as a substitute for the unadjusted recalculation.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists five outcome measures and reports one analysis. The primary outcome is successful bowel cleansing on the day of colonoscopy. No secondary results are posted in the registry record. The record therefore does not permit a broader comparison across the other listed outcomes or an assessment of whether findings across multiple outcomes were consistent.

Multiplicity arises in general when several outcomes, time points, subgroups, or statistical comparisons are tested. Conducting many tests can increase the chance of finding at least one apparently notable result by chance. Statistical plans may address this in different ways, but the registry record does not report a multiplicity adjustment or a testing hierarchy. It also does not report whether additional analyses were conducted but not posted. No claim about the trial's multiplicity procedure can therefore be made.

The primary result is labeled as belonging to the efficacy population. The definition of that population, the rules for inclusion, and its relationship to all randomized participants are not reported. The primary outcome denominators are 190 in the BLI4700 arm and 198 in the comparator arm. The registry does not report how missing colonoscopist ratings were handled or whether participants with missing ratings were excluded, classified in a particular way, or handled by another method.

Missing-data methods can affect a binary outcome comparison if availability of ratings differs between groups or relates to preparation success. This is a general consideration, not evidence that missing ratings occurred in this trial. The registry record does not provide the number of missing primary outcomes, the reasons for any missingness, or sensitivity analyses addressing it.

The statistical analysis plan, interim-analysis details, and subgroup-analysis plans are not reported in the registry record. The single posted analysis count does not establish that no other analyses were performed; it describes what is recorded. These omissions, together with the flagged discrepancy between the unadjusted recalculation and sponsor analysis, limit independent evaluation of the complete analysis process.

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 3% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cholelithiasis (0.5%).

The safety record reports serious adverse events for 190 participants at risk in the BLI4700 arm, with 0 affected, and for 199 participants at risk in the comparator arm, with 1 affected. Deaths are reported as 0 in both arms. Other adverse events are reported for 5 participants in the BLI4700 arm and 2 in the comparator arm, using at-risk counts of 190 and 199 respectively.

These are counts, not an adjusted comparison of safety risk. The registry does not provide event-level details for all events, follow-up duration, or a statistical comparison for the safety outcomes. The serious-event count is small in the posted record, so the counts alone provide limited information about the range or frequency of possible harms. They also do not establish whether an event was caused by a preparation.

The listed serious adverse-event term is Cholelithiasis, reported as 0.5%. The record does not state its relationship to a specific arm in the serious-event term listing. The safety denominators differ from the primary efficacy denominators, including a control-arm at-risk count of 199 rather than the primary outcome denominator of 198. The reason is not reported in the registry record. Safety and efficacy figures should therefore be read using their respective stated populations and denominators.

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

The registry record describes a randomized comparison of BLI4700 with an FDA-approved bowel preparation before colonoscopy in adults. The primary outcome was successful cleansing as judged by a colonoscopist on the day of the procedure. The posted efficacy counts are 175 of 190 in the BLI4700 arm and 174 of 198 in the comparator arm. The unadjusted recalculation gives a risk difference of 0.042 and an interval from -0.017 to 0.102.

Those figures describe the posted binary endpoint and its uncertainty under an unadjusted comparison. The interval includes zero, and the recalculated p-value is 0.166. These statistics do not establish that the preparations are equivalent. Nor do they establish that one is better or worse. A non-significant unadjusted comparison is not, by itself, evidence of equivalence or non-inferiority.

The sponsor's reported analysis is separately labeled non-inferiority and has p < 0.001 using a Cochran-Mantel-Haenszel method. Since the registry flags a discrepancy, the sponsor's own analysis takes precedence over the recalculation. But the registry does not report the non-inferiority margin, adjusted effect estimate, or its interval. The information available therefore does not support a full independent interpretation of the sponsor's non-inferiority claim.

The result concerns successful cleansing as defined by the trial, but the exact scoring threshold is not reported. The record does not provide the complete distribution of colonoscopist ratings or secondary outcome results. It also does not explain missing-outcome handling or provide enough detail about the efficacy population to evaluate all analysis decisions.

The safety counts are a separate component of the record. They should not be used to infer causality or to make broad claims about relative safety. More generally, a registry result is not a substitute for details about the full protocol, statistical analysis plan, or complete study report. The record supports describing the measured outcome and the posted statistical results, while leaving important questions about definitions, precision, and analysis procedures unresolved.

17. Why This Trial Matters Statistically

This trial is statistically interesting because it combines a binary efficacy endpoint with a sponsor analysis framed as non-inferiority. These are related to the same clinical comparison, but the analyses answer questions differently. An unadjusted risk difference describes the observed absolute contrast in the posted proportions. A non-inferiority test asks whether evidence meets a prespecified criterion tied to a margin. Interpretation of that test depends on the margin and analysis details, which are not reported in the registry record.

The registry also flags a discrepancy between the sponsor's analysis and a recalculation from the posted counts. That makes it important to distinguish a transparent summary-data recalculation from the sponsor's own analysis. The recalculation offers a direct description of the arm counts and uncertainty on common effect scales, while the sponsor's analysis takes precedence for the stated non-inferiority result. Neither the recalculation nor the reported p-value should be stretched beyond the question it addresses.

The endpoint itself illustrates a common measurement issue: a four-point clinician rating is summarized as a binary success outcome. Binary outcomes can be clear to communicate, but the threshold determining success matters. Since the registry record does not state that threshold or show the complete rating distribution, readers cannot assess how much information the categorization retains.

Finally, the sample-size illustrations show how planning requirements depend on the effect chosen for detection and the target power. They are not the original design rationale and are not observed power. Keeping that distinction clear prevents a planning calculation based on the observed effect from being mistaken for evidence about the trial's design quality or the certainty of its results.

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 details needed for independent assessment are not reported in the registry record. These include the exact comparator product and dosing instructions, the success threshold on the four-point cleansing scale, and the definition of the efficacy population. The record also does not identify who was masked, describe allocation concealment, or provide the original protocol and statistical analysis plan.

The discrepancy between the posted summary-data recalculation and the sponsor's reported non-inferiority analysis is a particular limitation. The sponsor's own analysis takes precedence, but the registry does not post its effect estimate, interval, non-inferiority margin, or sufficient implementation details for replication. The p-value alone cannot convey the estimated magnitude or the margin's practical meaning.

The primary outcome record gives the group denominators but does not explain how these relate to enrollment or how missing ratings were handled. The efficacy denominators and safety at-risk counts are not identical in the comparator arm, and the reason is not reported. Five outcome measures are listed, but no secondary results are posted. The record does not report multiplicity procedures, interim analyses, or subgroup-analysis plans.

The safety information is limited to event counts and one serious-event term. Follow-up duration, detailed event descriptions, and statistical comparisons are not reported. These limitations do not show that a particular problem occurred; they describe what cannot be established from the registry record. They constrain interpretation of both the efficacy comparison and the overall balance of outcomes.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Safety and Efficacy Comparison of BLI4700 Bowel Preparation Versus an FDA-approved Comparator in Adult Subjects Prior to Colonoscopy — NCT03261960.
• Registry record last updated 2021-09-08; status last verified 2021-08.
• Results first posted 2021-09-08.
• 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 registry record describes NCT03261960, a completed Phase 3 randomized, single-masked, parallel-group study sponsored by Braintree Laboratories. It compared BLI4700 bowel preparation with an FDA-approved comparator before colonoscopy in adults. The study objective describes 2-day, split-dose preparations. The recorded start was 2017-08-08, and primary completion was 2018-03-26.

The primary endpoint was successful bowel cleansing, assessed by colonoscopist rating on the day of colonoscopy. The efficacy results report 175 of 190 participants in the BLI4700 arm and 174 of 198 in the comparator arm. The recalculated unadjusted risk difference is 0.042, with an interval from -0.017 to 0.102. The sponsor's reported Cochran-Mantel-Haenszel non-inferiority analysis has p < 0.001; the registry flags a discrepancy, and the sponsor's own analysis takes precedence.

The record lists five outcome measures but posts no secondary results. Safety figures include 0 serious adverse events in the BLI4700 arm and 1 in the comparator arm, with 0 deaths in both arms. Important analysis details, including the non-inferiority margin and missing-data procedures, are not reported in the registry record.