Perfluorohexylcotane (NOV03) for the Treatment of Signs and Symptoms of Dry Eye Disease Associated With Meibomian Gland Dysfunction (Gobi Study): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Perfluorohexylcotane (NOV03) for the Treatment of Signs and Symptoms of Dry Eye Disease Associated With Meibomian Gland Dysfunction (Gobi Study) (NCT04139798), a randomized phase 3 study. The primary outcome is Change From Baseline (CFB) in Total Corneal Fluorescein Staining (tCFS) at Day 57.

2. Clinical Question

3. Trial Design

The Gobi Study is recorded as a Phase 3, randomized, double-masked, parallel-group, saline-controlled trial. Its stated purpose was to evaluate the efficacy, safety, and tolerability of NOV03 ophthalmic solution compared with saline for signs and symptoms of dry eye disease associated with meibomian gland dysfunction. Bausch & Lomb Incorporated is listed as the sponsor, and the registry status is completed.

The registry lists enrollment of 599. The study start date is 2020-07-20, and the primary completion date is 2021-03-12. The treatment arms were NOV03 4 times daily (QID) and placebo 4 times daily (QID); the arm descriptions specify 100% perfluorohexyloctance solution and saline solution (0.6% sodium chloride solution), respectively. The posted primary outcome is change from baseline in total corneal fluorescein staining at Day 57. The endpoint counts in the posted outcome results are 289 for NOV03 and 279 for placebo. These endpoint counts are not the registry's overall enrollment figure.

Randomization and double masking describe important design features: allocation is intended to create comparable groups, while masking is intended to limit the influence of treatment knowledge on participants or trial conduct. The registry record does not report the randomization procedure, the people masked, allocation concealment details, treatment adherence, or the original statistical analysis plan. The trial's original sample-size assumptions and justification are not reported in the registry record. The planning assumptions used for the posted sample-size illustration are separate from those original design details.

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 Change From Baseline (CFB) in Total Corneal Fluorescein Staining (tCFS) at Day 57, measured over Assessed from Baseline to Day 57.

The primary endpoint is change from baseline in total corneal fluorescein staining (tCFS) at Day 57, reported as a mean in units on a scale. The registry describes tCFS as an assessment of corneal staining using the NEI scale. The cornea is divided into central, inferior, superior, nasal, and temporal subregions. Each subregion is scored from 0-3, where 0 corresponds to no staining and 3 to maximum staining; the total is the sum of the five subregion scores and ranges from 0-15.

Change from baseline expresses how the score at follow-up differs from the starting score. In general, this can make a comparison more directly focused on change over the assessment period than a follow-up score alone, although interpretation still depends on the baseline values and the analysis method. On the described scale, lower scores mean less staining, so a negative change indicates a decline in the score. The posted mean changes are -2.0 for NOV03 and -1.0 for placebo.

Corneal staining is a clinical sign measured on a defined scale. It is one component of the trial's stated focus on dry eye signs and symptoms, not a complete account of every symptom, safety outcome, or patient experience. The registry record does not report a threshold for a clinically important change or additional details about how this endpoint was assessed beyond the scale description and timeframe.

7. Results Posted in the Registry

• NOV03 4 Times Daily (QID): mean -2.00 versus -1.00 on control; difference -1.00 (95% CI -1.44 to -0.56), standardized effect d = -0.38; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean change describe?

The primary outcome is continuous: it is recorded as a numeric score on the tCFS scale. For each group, the posted mean change is an average of the individual changes from baseline to Day 57. The NOV03 mean change is -2.0, and the placebo mean change is -1.0. A mean is useful for describing the central tendency of a group, but it does not imply that each participant had that change. Some observations can be above the mean and some below it. The registry reports standard deviations of 2.6 for NOV03 and 2.7 for placebo, which describe the spread of the observed changes within each group.

How is the group contrast read?

The posted mean difference is -1.0 units on a scale. It represents the reported contrast between mean change in the NOV03 arm and mean change in the placebo arm, with negative values indicating a more negative mean change for NOV03 under the stated ordering. Since negative change corresponds to a decline in staining score on the scale description, the direction of the contrast is consistent with a larger average score reduction in the NOV03 group. This is a statement about the group averages, not a claim that every individual had a greater reduction.

The comparison also reports a pooled standard deviation of 2.65 and a standardized effect estimate of -0.377. In general, a standardized effect expresses a difference relative to a measure of variability, which can help describe magnitude on a scale that is less dependent on the original measurement units. It does not replace the original-unit mean difference, which is usually easier to relate to the tCFS score. These statistics do not establish clinical importance; the registry record does not state a threshold for interpreting the size of a meaningful change.

What does the confidence interval contribute?

The confidence interval for the mean difference runs from -1.436 to -0.564. In general, a confidence interval gives a range of values compatible with the observed data and the statistical procedure, under its assumptions. The endpoints quantify uncertainty in the estimated group contrast; they are not bounds on the range of individual participant changes. The fact that the reported interval does not include zero is consistent with the unadjusted p-value of <0.001 for a null comparison of no mean difference. Neither result addresses whether the effect is clinically important or whether the assumptions of the analysis are appropriate.

What does the p-value say?

In general, a p-value assesses how unusual the data would be under a specified null hypothesis and statistical model. The posted value is unadjusted and reported as <0.001. It does not give the probability that the null is true, the probability that the observed estimate will be repeated, or the probability that a treatment is beneficial. It is also not a measure of effect size: the mean difference and its confidence interval describe the estimate and its precision more directly.

The registry lists no posted analyses, despite listing two outcome measures. The record does not report whether the p-value comes from a prespecified primary analysis, what covariates or model were used, or whether there were adjustments for multiple comparisons. A simple unadjusted recalculation from the posted summary statistics is not a substitute for the sponsor's analysis plan or any analysis not reported in the registry. The registry record does not identify those details.

Why can over-dispersion matter?

In general, over-dispersion occurs when observed variability is greater than a statistical model expects. It matters because a model that understates variability can produce standard errors that are too small and intervals that are too narrow. For this continuous endpoint, the posted summaries include standard deviations, but the registry record does not describe a model or report an over-dispersion assessment. There is therefore no basis in the registry record to say whether over-dispersion was evaluated or relevant to the sponsor's analysis.

For a continuous outcome, analysis choices can include comparing change scores or adjusting follow-up scores for baseline values. These approaches can yield different estimates and precision depending on the design and data. This is general methodological context, not a claim about the method used in the Gobi Study: the specific analysis model, covariate adjustment, variance method, and handling of missing observations are not reported in the registry record.

10. Confidence Intervals and Precision

The estimated mean difference in change from baseline is -1.0 units on a scale, with a confidence interval from -1.436 to -0.564. The interval communicates uncertainty around the estimated between-group difference: values closer to either endpoint remain compatible with the data and procedure used to calculate the interval. It is not an interval containing a specified proportion of individual participant changes, and it does not show how much participants varied from one another.

The confidence interval lies below zero, while the unadjusted p-value is <0.001. Together, these describe a statistically clear direction for the posted mean contrast under the calculation's assumptions. Statistical precision and clinical importance are different questions. The registry record does not report a threshold for a clinically important tCFS difference, so the interval cannot be classified as showing a clinically meaningful effect using a trial-specific criterion.

The group standard deviations, 2.6 for NOV03 and 2.7 for placebo, indicate that individual changes vary around their respective group means. These standard deviations are not the uncertainty intervals for the mean difference. The comparison also reports standard errors of 0.153 for NOV03 and 0.162 for placebo; these characterize uncertainty in the respective group means in the posted summary, not the spread of individual observations. No confidence intervals for the two group means are reported.

Precision depends on the number of observations, the variability of the outcome, and the analysis method. The registry gives endpoint counts of 289 and 279, but does not report missing-data handling, analysis population details, or the model used for the sponsor's analysis. Thus, the reported interval is the best available summary of precision for this contrast in the registry record, while the analysis details needed to independently assess its construction are not reported.

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 sample-size figures are planning illustrations based on fractions of the observed effect, not the trial's original design calculations. Each figure is the number of participants per arm, rather than a total across arms or a total number of events. The assumptions listed for the illustration are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The registry record does not report the original design assumptions or sample-size justification.

The column for 80% power gives the planned participants per arm for an illustration designed to have an 80% chance of detecting the chosen effect, if the assumptions hold and the effect is real. The 90% power column uses a higher target chance of detecting that effect. These are planning concepts, not guarantees about a future study. The stated effect is a fraction of the observed effect: 100%, 75%, or 50% of the observed effect. For example, planning for 75% of the observed effect gives 196 participants per arm for 80% power and 263 participants per arm for 90% power.

Smaller effects are harder to distinguish from random variation, so they require more participants under the same planning framework. The illustration therefore lists 441 participants per arm for 80% power and 591 participants per arm for 90% power when planning for 50% of the observed effect. The figures are not observed power and should not be read as a retrospective measure of the probability that this completed trial detected an effect. They are conditional planning estimates based on the stated assumptions and the observed effect used as a reference.

12. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Observed difference Δ = -1.000; pooled standard deviation σ = 2.650.
2. n = 2 × (1.96 + 0.842)² × 2.650² / -1.000² = 110.2, rounded up to 111 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                      # observed difference in means
sigma <- 2.64958827390246                     # 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(4139798)
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 = 111 (80% power), 148 (90% power)
 75% of observed effect: n per arm = 196 (80% power), 263 (90% power)
 50% of observed effect: n per arm = 441 (80% power), 591 (90% power)
power.t.test: 111.2 
Simulated power at n = 111 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.

13. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates multiple outcomes, time points, treatment contrasts, or analysis choices. In general, the more statistical tests considered, the greater the chance that at least one result appears unusual by chance if every test is treated as a separate confirmatory test. A prespecified testing strategy or adjustment can address this issue, but the appropriate approach depends on the trial's aims and analysis plan.

The registry lists two outcome measures and zero posted analyses. It identifies the Day 57 change in tCFS result as the primary outcome, but does not report a multiplicity procedure, a hierarchy of outcomes, or whether the unadjusted p-value was part of a prespecified testing sequence. The posted p-value is explicitly unadjusted. The record therefore does not establish whether or how multiplicity was addressed in the trial's broader analysis.

Important analysis details are also not reported in the registry record. These include the analysis population, the treatment of missing outcome values, whether participants were analyzed according to randomized assignment, the number of participants with missing Day 57 measurements, and the statistical model used for the primary comparison. The registry does not provide baseline endpoint summaries or an account of exclusions from the endpoint analysis. Without these details, it is not possible to assess attrition patterns or to determine how sensitive the estimate might be to alternative missing-data assumptions.

These gaps do not demonstrate that a method was not used; they mean that the method is not reported in the registry record. The randomized, double-masked design is an important part of the trial description, but design labels alone do not answer questions about analysis populations, missing data, or multiplicity. The reported mean difference and confidence interval can be interpreted as posted, while a full appraisal of the analysis strategy requires information not present in the registry record.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events affecting 0 participants in the NOV03 group, among 303 at risk, and 0 participants in the placebo group, among 294 at risk. It reports 0 deaths in each group. These figures describe the serious adverse event counts recorded for the listed groups; they do not establish that serious events were impossible or that the treatments have identical safety profiles.

The registry separately reports other adverse events affecting 16 participants in the NOV03 group and 11 in the placebo group, with 303 and 294 at risk, respectively. These are not the serious adverse event figures and should not be combined with them. The record does not list serious adverse event terms, since no top serious terms are posted, and does not provide details such as event timing, severity breakdowns, or attribution to study treatment.

Safety counts depend on the recorded definitions, observation period, and completeness of follow-up. The registry record does not report enough detail to assess those elements or to compare event rates adjusted for exposure. The counts should be read as descriptive registry results, not as proof of absence of risk or as a treatment recommendation.

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

The posted result describes a difference between group means in change from baseline in tCFS at Day 57. The NOV03 group mean change was -2.0 units on a scale, compared with -1.0 units on a scale for placebo. The mean difference was -1.0 units on a scale, with a confidence interval from -1.436 to -0.564 and an unadjusted p-value of <0.001. On the scale described in the registry, lower scores correspond to less staining, and negative changes indicate a reduction from baseline. Thus, the posted endpoint result indicates a larger average reduction in the NOV03 group than in the placebo group.

This is a result for one measured sign at one stated assessment time. It does not summarize every symptom or outcome in dry eye disease, and the registry record does not post secondary outcome analyses. The result also does not establish how an individual participant's score changed: group means conceal variation, reflected in the reported standard deviations of 2.6 and 2.7.

Statistical evidence and clinical importance are not interchangeable. The confidence interval and p-value describe the estimated contrast and its statistical uncertainty under the analysis assumptions. The registry record does not report a clinically important difference threshold for tCFS, so the posted statistics alone cannot determine whether the size of the difference matters to patients. The analysis population, model, missing-data approach, and multiplicity strategy are also not reported in the registry record.

The safety figures are separate from the efficacy endpoint. No serious adverse events or deaths are recorded in either group in the posted safety results, while other adverse events are reported for both groups. These counts do not establish the absence of risk or provide a complete safety evaluation. Overall, the record supports a description of the reported Day 57 mean-change contrast and its precision, but broader conclusions about symptoms, safety, clinical value, or durability require information not reported in the registry record.

16. Why This Trial Matters Statistically

This trial illustrates how a continuous clinical endpoint can be presented in several complementary ways. The group means describe average change on the original tCFS scale; the mean difference expresses the comparison between randomized groups; the confidence interval communicates precision; and the p-value addresses the compatibility of the data with a null comparison under the statistical procedure. These quantities answer related but distinct questions and should not be treated as interchangeable evidence.

The scale description also matters for interpreting direction. A negative change is a decline from baseline, and lower tCFS scores correspond to less staining. This makes the negative mean difference interpretable as a greater average decline in score in the NOV03 group than in the saline group. Yet the registry does not state a threshold for clinical importance. Statistical evidence for a difference does not by itself say whether its magnitude is meaningful in patient experience or practice.

The record provides a useful example of the limits of registry-level statistical reporting. It includes group summaries and an interval for the primary comparison, while listing no posted analyses. Details about the analysis population, missing data, model, multiplicity, and original sample-size rationale are not reported in the registry record. Those omissions matter because they shape how readers judge whether the estimate is robust and whether the p-value should be interpreted within a wider testing strategy.

Finally, the planning illustration shows how sample-size requirements depend on the effect chosen for planning: smaller effects require more participants under fixed assumptions. Those calculations are not observed power and are distinct from the original design rationale. Keeping design planning, observed estimates, statistical precision, and clinical importance separate leads to a more careful reading of the trial's results.

17. Limitations

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

The registry record leaves several trial-specific details unavailable. It does not report the original sample-size justification, planned statistical analysis, analysis population, randomization procedure, or the people included in double masking. The posted analysis count is zero, and the record does not explain whether other analyses were conducted or where they might be reported. These gaps limit assessment of whether the posted comparison follows a prespecified plan.

The primary outcome result does not include baseline tCFS means, participant-level data, or a breakdown of missing observations. The endpoint counts are reported, but the registry does not state how missing Day 57 measurements were handled, whether exclusions occurred, or how adherence and exposure were considered. It is therefore not possible from the registry record alone to evaluate attrition or reproduce the sponsor's analysis.

The record gives a confidence interval and an unadjusted p-value but does not describe the calculation method or any multiplicity strategy. It also does not report a clinical-importance threshold for tCFS. The effect estimate can be interpreted statistically, but its clinical meaning and the robustness of its inference cannot be fully assessed from the posted details.

Safety reporting is also limited: serious adverse event counts and deaths are listed, along with counts for other adverse events, but event terms, timing, severity details, exposure duration, and attribution are not reported in the registry record. These limitations describe what is missing from the record; they do not establish that the trial did not collect or analyze the information.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Perfluorohexylcotane (NOV03) for the Treatment of Signs and Symptoms of Dry Eye Disease Associated With Meibomian Gland Dysfunction (Gobi Study) — NCT04139798.
• Registry record last updated 2024-10-23; status last verified 2024-08.
• Results first posted 2024-10-23.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The Gobi Study (NCT04139798) is a completed Phase 3, randomized, double-masked, parallel-group, saline-controlled trial sponsored by Bausch & Lomb Incorporated. The registry lists enrollment of 599, a start date of 2020-07-20, and a primary completion date of 2021-03-12.

The primary result is change from baseline in total corneal fluorescein staining at Day 57. Mean change was -2.0 for NOV03 and -1.0 for placebo; the mean difference was -1.0, with a confidence interval from -1.436 to -0.564 and an unadjusted p-value of <0.001. The registry also reports no serious adverse events or deaths in either group, alongside other adverse event counts of 16 and 11.

Analysis-plan details, missing-data methods, multiplicity procedures, and the original sample-size rationale are not reported in the registry record.