A Study to Assess the Safety and Efficacy of ECF843 vs Vehicle in Subjects With Dry Eye Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Assess the Safety and Efficacy of ECF843 vs Vehicle in Subjects With Dry Eye Disease (NCT04391894), a randomized phase 2 study. The primary outcome is Part 1: Change From Baseline in Symptom Assessment in Dry Eye (SANDE) Score.

2. Clinical Question

The trial concerned dry eye disease and included a total enrollment of 718. The posted primary outcome is change in the Symptom Assessment in Dry Eye (SANDE) score during Part 1. SANDE records symptom frequency and severity, and combines the two scores into an overall score on a 0-to-100 scale. The registry describes 100 as the maximal amount of dry eye symptoms and 0 as the minimal amount.

The clinical question is therefore about the average change in reported symptoms over a short randomized-treatment period, rather than a claim about longer-term outcomes. The record also lists corneal fluorescein staining outcomes and treatment-emergent adverse events among the secondary outcomes. The posted information does not establish how any change relates to an individual participant’s experience.

3. Trial Design

This was a completed, randomized, parallel-group Phase 2 study of ECF843 versus vehicle in subjects with dry eye disease. The registry lists 718 enrolled participants. The study began on 2020-10-06, and primary completion was recorded as 2021-05-13. The registry masking field says triple masking, while the study summary describes both parts as double-masked. The record therefore contains both descriptions.

The study summary says the study was planned in two parts. Part 1 was intended to assess efficacy and safety, followed by Part 2 with additional exploratory assessments. The primary endpoint and the posted comparisons discussed in this analysis concern Part 1. Randomization was described as stratified for subjects with Sjogren's Syndrome. Further details of the allocation process, the masking procedures, and how the two parts were linked are not reported in the registry record.

Part 1 included three ECF843 arms: 0.45 mg/mL TID, 0.15 mg/mL TID, and 0.15 mg/mL BID. The corresponding vehicle comparisons were vehicle TID and vehicle BID. The primary endpoint analysis includes arm sizes from 109 to 114, with the exact number differing by arm. The primary outcome timeframe was up to 28 days, from baseline to the end of randomized treatment.

For the primary outcome, baseline was the last available value collected prior to or on the first administration of randomized study treatment. The analysis population required both a baseline and a post-baseline value. The registry record does not report the original sample-size justification, detailed statistical analysis plan, interim analyses, or full eligibility criteria.

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.

Before assignment: Approximately 800 subjects were planned to be screened in Part 1 to have approximately 680 subjects randomized into the 56-day treatment period. A total of 970 subjects were screened, of which 558 subjects received randomized treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set 1 including subjects with a value at both Baseline and post-baseline visit. Baseline is defined as the last available value collected prior to or on the first administration of randomized study treatment.

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 Part 1: Change From Baseline in Symptom Assessment in Dry Eye (SANDE) Score, measured over Up to 28 days (Baseline (BL) to end of randomized treatment).

The primary endpoint was change from baseline in the Symptom Assessment in Dry Eye (SANDE) score during Part 1, measured up to 28 days through the end of randomized treatment. SANDE uses a 100 mm visual analog scale. Participants rated the frequency and severity of ocular discomfort over the past 24 hours on separate scoring lines. The frequency line ranges from “Rarely” to “All the Time,” and the severity line from “Very Mildly” to “Very Severely uncomfortable.”

The overall SANDE score is calculated as the square root of the product of the frequency and severity scores. Its range is 0 to 100, with 100 representing the maximal amount of dry eye symptoms and 0 the minimal amount. The diary was completed electronically at screening and then every evening before bedtime during the study. The endpoint captures participant-reported symptoms, rather than an examination finding.

Change from baseline is useful for describing whether symptom scores moved during the treatment period, while comparison with vehicle helps place that movement in the context of the randomized comparator. The registry specifies that negative change indicates improvement. This direction convention is important: a more negative within-arm change is not by itself the same as a favorable difference between randomized arms. The comparison of interest is the difference in change between an ECF843 arm and its vehicle arm, interpreted with its uncertainty.

7. Results Posted in the Registry

• ECF843 0.45 mg/mL TID (Part 1): mean -6.00 versus -4.90 on control; difference -1.10 (95% CI -6.24 to 4.04), standardized effect d = -0.06; z-test p = 0.675.
• ECF843 0.15 mg/mL TID (Part 1): mean -8.10 versus -4.90 on control; difference -3.20 (95% CI -8.33 to 1.93), standardized effect d = -0.16; z-test p = 0.221.
• ECF843 0.15 mg/mL BID (Part 1): mean -1.80 versus -4.90 on control; difference 3.10 (95% CI -2.11 to 8.31), standardized effect d = 0.16; z-test p = 0.244.
• ECF843 Vehicle BID (Part 1): mean -6.10 versus -4.90 on control; difference -1.20 (95% CI -6.31 to 3.91), standardized effect d = -0.06; z-test p = 0.646.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Part 1: Change From Baseline in Central Corneal Fluorescein Staining

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • ECF843 0.45 mg/mL TID (Part 1): mean -0.20 versus -0.20 on control; difference 0.00 (95% CI -0.22 to 0.22), standardized effect d = 0.00; z-test p = 1.000.
  • ECF843 0.15 mg/mL TID (Part 1): mean -0.30 versus -0.20 on control; difference -0.10 (95% CI -0.31 to 0.11), standardized effect d = -0.12; z-test p = 0.356.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between ECF843 0.45 mg/mL TID (Part 1) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between ECF843 0.15 mg/mL TID (Part 1) and the control. The interval gives the range of effects that remain compatible with the data.

Part 1: Change From Baseline in Inferior Corneal Fluorescein Staining

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • ECF843 0.45 mg/mL TID (Part 1): mean -0.20 versus -0.20 on control; difference 0.00 (95% CI -0.22 to 0.22), standardized effect d = 0.00; z-test p = 1.000.
  • ECF843 0.15 mg/mL TID (Part 1): mean -0.30 versus -0.20 on control; difference -0.10 (95% CI -0.33 to 0.13), standardized effect d = -0.11; z-test p = 0.393.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between ECF843 0.45 mg/mL TID (Part 1) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between ECF843 0.15 mg/mL TID (Part 1) and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary continuous outcomes concern changes from baseline in central and inferior corneal fluorescein staining. At the arm-summary level, changes are generally small, and the comparisons described in the registry do not establish a consistent separation from vehicle across the posted staining measures. The record also lists ocular and non-ocular treatment-emergent adverse events among secondary outcomes, but the outcome summaries do not provide a full secondary-outcome analysis plan.

These findings are related to, but distinct from, the primary SANDE symptom outcome. SANDE is participant-reported symptom frequency and severity, whereas corneal staining is a different measure. Similarity or difference across these types of outcomes cannot be assumed to indicate the same underlying effect. The posted secondary summaries do not provide a basis for replacing the primary endpoint or drawing a unified conclusion across all measures.

Secondary comparisons commonly carry less inferential weight when many outcomes, time points, or treatment contrasts are examined. In general, testing multiple hypotheses increases the chance of at least one apparently unusual result occurring by chance if no correction or prespecified hierarchy is used. The registry record does not report a multiplicity adjustment or detailed hierarchy for secondary outcomes. Therefore, the secondary results should be treated as descriptive or exploratory unless the analysis plan establishes otherwise; that plan is not reported in the registry record.

The results also have varying participant counts across outcomes and arms. Those counts should be kept in mind when comparing summaries. The registry record does not fully explain the outcome-specific missingness or whether the different analysis populations were prespecified. The secondary evidence is consequently best read as contextual information alongside the primary result, not as independent confirmation.

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

• Part 1: Change From Baseline in Central Corneal Fluorescein Staining
• Part 1: Change From Baseline in Inferior Corneal Fluorescein Staining
• Part 1: Percentage of Participants With Ocular and Non-ocular Treatment Emergent Adverse Events (AEs)

9. Statistical Interpretation

10. Statistical Methods Explained

Statistical methods need to be read in light of the information the registry record actually reports. It identifies a mixed-model analysis for the three posted primary comparisons and gives least-squares mean differences, 95% confidence intervals, and p-values. More detailed model specifications and the complete analysis plan are not reported in the registry record.

What does a least-squares mean represent?

In general, a least-squares mean is a model-based estimate of an outcome mean. It is not necessarily the same as a simple arithmetic average, because the model may account for factors or repeated observations specified in its analysis. The registry calls the posted comparison quantity a least-squares mean and identifies the method as mixed models analysis. It does not describe the model terms, covariance structure, handling of repeated measurements, or adjustment factors, so those details cannot be inferred from the label alone.

In general, a mixed model can account for correlations among repeated measurements or other structured data, depending on how it is specified. That can be relevant when participants contribute observations at multiple visits. Here, the primary endpoint description says SANDE was collected at screening and then every evening during the study, but the registry record does not state exactly which observations entered the model or how their correlation was handled. The posted model-based result should therefore be distinguished from a simple comparison reconstructed from arm summaries.

How should the recalculations be read?

The recalculated estimates are simple, unadjusted comparisons based on the posted arm-level summary statistics. They are intended as a transparent check on the direction and approximate scale of contrasts, not as a reproduction of the sponsor’s full analysis. The recalculation uses arm means, sample sizes, and standard errors; the pooled standard deviation is approximate when derived from those summaries. It does not reproduce unreported model features, missing-data procedures, or any adjustment used in the sponsor analysis.

That distinction matters in this record because the recalculated BID comparison differs from the sponsor’s posted mixed-model comparison. Where that discrepancy exists, the sponsor’s own analysis takes precedence. The recalculation should not be used to replace the sponsor’s estimate or its reported inference. For the TID comparisons, the sponsor and simple recalculation also have different p-values and intervals, reinforcing that the summaries cannot reconstruct the full model.

What does a confidence interval communicate?

In general, a confidence interval describes the statistical uncertainty around an estimated quantity under the assumptions of the analysis. A narrower interval indicates greater precision than a wider interval on the same scale. An interval that includes zero is compatible with no between-group difference as well as with differences in either direction at the stated confidence level. It does not mean that all values inside the interval are equally plausible, nor does it establish the probability that a particular effect is true.

The sponsor’s reported primary intervals for the two TID comparisons include zero. The reported BID interval does not include zero, but its direction and interpretation remain tied to the defined comparison and the model used. A p-value addresses compatibility with a specified null hypothesis under analysis assumptions; it is not the probability that the null hypothesis is true and does not measure clinical importance.

Why can over-dispersion matter?

In general, over-dispersion means observed variability is greater than a statistical model expects. If a model underestimates variability, its standard errors can be too small and its intervals too narrow. This is a general statistical concern, not a reported property of this trial: the registry record does not report an over-dispersion assessment or enough model detail to determine whether it was relevant. For the continuous SANDE outcome, the more immediate limitation is that the detailed model specification and its assumptions are not reported.

The primary endpoint is continuous, and the registry reports least-squares mean changes with standard errors at the arm level. Standard errors express uncertainty in estimated means; they are not the same as standard deviations, which describe participant-level spread. The arm-level primary results list standard errors, while the secondary staining summaries list standard deviations. Those measures should not be interchanged.

What is not documented?

The registry record does not report the original sample-size justification, the planned treatment-effect target, a full multiplicity procedure, interim analyses, detailed missing-data methods, or a complete statistical analysis plan. The recorded Full Analysis Set 1 definition requires a baseline and post-baseline value, but does not explain how missing values were otherwise handled. It also does not provide enough detail to assess model diagnostics, deviations from the analysis plan, or whether additional covariates were used. These omissions limit independent reconstruction and make the sponsor’s posted model-based results the primary source for its analyses.

11. Confidence Intervals and Precision

Precision is conveyed most directly by the confidence intervals around the sponsor’s primary comparisons. For ECF843 0.45 mg/mL TID versus vehicle TID, the reported difference is -1.1 and the 95% confidence interval runs from -5.0 to 2.8. For ECF843 0.15 mg/mL TID versus vehicle TID, the difference is -3.2 and the interval runs from -7.01 to 0.7. Both intervals include zero and include values on either side of it, so the direction and magnitude of the between-arm difference are not tightly determined by those estimates.

For ECF843 0.15 mg/mL BID versus vehicle BID, the sponsor reports a difference of 4.3, with a 95% confidence interval from 0.3 to 8.3. This interval is entirely positive, but spans a range of possible differences. The registry’s result should be read using the sponsor’s analysis, particularly because the simple recalculation differs for this comparison.

In general, a 95% confidence interval is a range produced by a method that, under its assumptions, has 95% coverage over repeated studies. It is not a guarantee that the true effect lies in this particular interval, and it does not express the probability that the treatment effect is within the interval. Wider intervals indicate less precision; intervals crossing zero leave the sign of the contrast unresolved by that interval.

Arm-level standard errors are also posted for the primary changes. They describe uncertainty around the estimated arm means, not the spread of individual SANDE scores. The record gives no clinically meaningful threshold for interpreting differences on SANDE, so statistical precision alone cannot establish whether an estimated difference is important to participants.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm for 80% power and participants per arm for 90% power. They are not total events: the total-events convention applies to time-to-event calculations, and the primary outcome is continuous. The planning illustrations use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They use the observed effect as a reference and consider stated fractions of that observed effect.

The 80% power column gives the planned number of participants per arm under those assumptions for a design intended to have an 80% chance of detecting the specified effect if that effect is present and the assumptions hold. The 90% power column uses the same general interpretation with a higher target chance of detection. These are design-planning quantities, not a guarantee about the outcome of a particular trial. The illustrations do not describe the original design: the registry’s own sample-size assumptions are not reported.

Rows that plan for 75% or 50% of the observed effect use smaller target effects than the row based on 100% of the observed effect. Smaller effects are harder to distinguish from random variation, so the planning illustrations require more participants per arm. When the observed contrast is small, the corresponding planning numbers can be very large. These calculations are planning illustrations based on the observed effect, not observed power.

The figures should therefore be read as conditional examples, not as evidence that a trial with a particular sample size would necessarily have found a result. They assume the stated effect, allocation, significance level, and approximation, and they do not incorporate unreported design features or the sponsor’s model specifications.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -1.100; pooled standard deviation σ = 19.587.
2. n = 2 × (1.96 + 0.842)² × 19.587² / -1.100² = 4,977.4, rounded up to 4,978 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -1.1                      # observed difference in means
sigma <- 19.5873996805177                     # 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(4391894)
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 = 4,978 (80% power), 6,664 (90% power)
 75% of observed effect: n per arm = 8,849 (80% power), 11,847 (90% power)
 50% of observed effect: n per arm = 19,910 (80% power), 26,654 (90% power)
power.t.test: 4978.4 
Simulated power at n = 4,978 per arm: 0.801 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary comparisons as mixed-model analyses of least-squares mean differences. Those analyses are the appropriate reference for the sponsor’s reported results. The simple recalculations use posted arm summaries and do not reproduce the full model. Consequently, they can differ because they use a different, unadjusted calculation rather than the sponsor’s model-based approach.

The reported TID comparisons are -1.1 for the 0.45 mg/mL dose and -3.2 for the 0.15 mg/mL dose, compared with their vehicle TID arm. Their sponsor-reported p-values are 0.585 and 0.107, respectively. For the BID comparison, the sponsor reports 4.3, with a 95% confidence interval from 0.3 to 8.3 and p = 0.033. The corresponding simple recalculation is 3.1, with a 95% confidence interval from -2.112 to 8.312 and p = 0.244.

Because the registry record flags a discrepancy, the sponsor’s own analysis takes precedence. The recalculated figures can illustrate how posted summary statistics compare under a simple unadjusted approach, but they should not be used to overturn the reported model result. Detailed explanations for the differences are not reported in the registry record.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when analyses test more than one hypothesis, such as several doses, outcomes, or treatment contrasts. In general, the more comparisons made, the greater the chance that at least one result will appear statistically notable by chance alone. Procedures such as adjusted significance thresholds or a prespecified testing hierarchy can address this issue, but the registry record does not report whether such a procedure was planned or used for the posted primary and secondary analyses.

The primary outcome includes multiple ECF843-versus-vehicle contrasts: two TID dose comparisons and a BID comparison. The record labels the sponsor analyses as superiority comparisons but does not state whether the p-values were adjusted across those comparisons. The p-value of 0.033 for the BID comparison should therefore be read as the posted value, without assuming it has or has not been corrected for multiplicity. The complete statistical analysis plan and any interim-analysis procedure are not reported in the registry record.

The analysis population for the primary outcome is described as Full Analysis Set 1, including subjects with both a baseline and post-baseline value. Baseline is the last available value collected prior to or on the first administration of randomized study treatment. This definition makes clear that participants without the required outcome values are not included in that analysis population. It does not specify how other missing observations were handled, whether any imputation was used, or how missingness was assessed.

Missing data can affect estimated treatment differences if the participants with observed values differ from those without them. In general, the direction and size of any such effect depend on why data are missing and on the analysis assumptions. The record does not report those assumptions, sensitivity analyses, or reasons for missing outcome measurements. It is therefore not possible to determine from the registry record how robust the primary result is to alternative missing-data approaches.

The registry also does not report the sponsor’s original sample-size justification, interim analyses, or a complete analysis plan. These omissions limit the ability to evaluate whether the reported analysis followed prespecified procedures and how the primary and secondary tests were coordinated.

16. 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: pelvic mass (0.9%), ischaemic stroke (0.9%).

The registry’s Part 1 serious adverse event figures show zero serious events in each of the three ECF843 arms, one in vehicle TID, and one in vehicle BID. No deaths are listed in any of the five arms. The registry also lists Pelvic mass and Ischaemic stroke among the most frequent serious-event terms, each at 0.9%. Those terms are presented as registry figures; the record excerpt does not give enough detail to connect either term to a particular treatment arm or to establish causality.

These counts should be read against the numbers at risk recorded for each arm, which range from 109 to 114. Serious adverse events are uncommon events, and small counts provide limited information about possible differences between groups. Absence of a listed serious event in an arm does not establish absence of risk. Conversely, a serious event occurring during a study does not by itself show that study treatment caused it.

The registry also records other, non-serious adverse events by arm, but those counts are separate from the serious-event figures. A full safety interpretation would require information about event definitions, exposure, timing, relatedness, and follow-up; those details are not reported in the registry record. These figures describe the posted record and are not medical advice.

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

The results describe changes in reported dry eye symptoms and other recorded outcomes during Part 1, with the primary endpoint measured up to 28 days. They provide estimates of between-arm differences and their statistical uncertainty for the randomized comparisons. The two sponsor-reported TID confidence intervals include zero, while the sponsor-reported BID interval is positive and excludes zero. The BID comparison also differs from the simple recalculation, so the sponsor’s model-based result is the primary reported estimate.

Statistical evidence is not the same as evidence of clinical importance. The registry record does not report a threshold for a clinically meaningful SANDE difference, and a p-value does not quantify the importance of a difference. The endpoint definition establishes the direction of symptom change within an arm, but the between-arm estimate must be interpreted separately. A change in one group alone cannot establish that the change is attributable to the randomized treatment.

The results do not establish how participants would fare beyond the short recorded timeframe, how results apply to people outside the study population, or whether the same findings would occur in another study. The registry describes Part 2 as including additional exploratory assessments, but the posted primary comparisons are from Part 1. The record does not provide the full details needed to assess all aspects of the study’s analysis plan, missing data, or multiplicity.

Nor do the posted summaries establish that an individual participant would experience the arm-average change. Group estimates describe averages and do not show the distribution of individual responses in full. Safety counts likewise describe recorded events and cannot establish causality or rule out uncommon risks. The appropriate interpretation is limited to the posted estimates, intervals, analysis population, and time period.

18. Why This Trial Matters Statistically

This trial illustrates why a randomized comparison should be interpreted through the contrast between groups rather than by looking only at whether scores improved within each arm. All five primary-outcome arm summaries have negative changes, but the clinical question concerns how each ECF843 arm compares with its corresponding vehicle arm. The sponsor’s estimates, confidence intervals, and analysis method provide that comparison.

It also demonstrates the distinction between a model-based analysis and a simple reconstruction from summary statistics. The registry reports mixed-model least-squares mean comparisons, while the recalculations use unadjusted posted summaries. Their differences are especially clear for the BID contrast, and the sponsor’s own analysis takes precedence because the record flags a discrepancy. Recalculation can be informative, but it cannot reproduce model details that are not reported.

Precision matters alongside statistical significance. The intervals communicate which directions and magnitudes remain compatible with the estimates under the sponsor analysis. The registry does not give a clinical-importance threshold for SANDE, so statistical significance alone cannot answer whether a difference is meaningful to participants.

Finally, the trial highlights limits in registry-level interpretation: the original sample-size rationale, complete analysis plan, multiplicity procedures, and detailed missing-data methods are not reported. Recognizing what is and is not documented is part of sound statistical reading, especially when multiple doses and outcomes are involved.

19. 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 primary endpoint timeframe is short: up to 28 days from baseline to the end of randomized treatment. The posted results therefore do not describe longer-term symptom patterns. The registry identifies a Part 2 with additional exploratory assessments, but the primary comparisons described are from Part 1, and the relationship between the parts is not detailed.

There is a discrepancy between the sponsor’s reported model-based results and the simple recalculations from arm summaries. The record does not explain the source of that discrepancy. The sponsor’s own analysis takes precedence, but the difference reinforces that the recalculations should be treated only as unadjusted checks rather than as a substitute for the reported analysis.

Some design and analysis details are missing from the registry record. The original sample-size justification, full statistical analysis plan, interim analyses, model specification, multiplicity approach, and detailed missing-data procedures are not reported. The record gives a primary analysis population requiring baseline and post-baseline measurements, but does not fully explain outcome-specific missingness or sensitivity analyses.

The primary outcome is participant-reported, and the record does not provide a clinical-importance threshold for the scale. Arm-level averages do not describe the full distribution of individual responses. The reported safety counts are also limited by the small number of serious events and the lack of further details on exposure, timing, and relatedness. These limitations restrict how far the posted estimates can be generalized or interpreted beyond the recorded comparisons and timeframe.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Assess the Safety and Efficacy of ECF843 vs Vehicle in Subjects With Dry Eye Disease — NCT04391894.
• Registry record last updated 2025-01-28; status last verified 2025-01.
• Results first posted 2022-05-26.
• 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.

23. Record Summary

This completed Phase 2 randomized parallel-group study evaluated ECF843 and vehicle in subjects with dry eye disease. The registry records 718 enrolled participants and a primary completion date of 2021-05-13. Part 1 included ECF843 arms at 0.45 mg/mL TID, 0.15 mg/mL TID, and 0.15 mg/mL BID, with vehicle comparisons.

The primary outcome was change from baseline in SANDE score up to 28 days. The sponsor reported mixed-model comparisons for the three ECF843-versus-vehicle contrasts. The two TID confidence intervals include zero; the reported BID comparison is positive with a 95% confidence interval from 0.3 to 8.3. Because the registry record flags a discrepancy with the unadjusted recalculation, the sponsor’s analysis takes precedence.

The record also lists corneal staining and adverse events among the secondary outcomes. The original sample-size rationale, complete analysis plan, multiplicity procedure, and detailed missing-data methods are not reported in the registry record.