This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
SHEA is a completed phase 4 randomized clinical trial evaluating the effectiveness and safety of a cream containing spent grain wax extract, argan oil, and shea butter after trichloroacetic acid 15% peel in aging disorder. The registry describes a controlled, quadruple-masked, split-face clinical trial with a factorial design.
| Feature | SHEA |
|---|---|
| Phase | Phase 4 |
| Condition | Aging Disorder |
| Design | Randomized, controlled, quadruple-masked, split-face clinical trial |
| Allocation | Randomized |
| Design model | Factorial |
| Primary purpose | Treatment |
| Enrollment | 27 |
| Intervention | Ezerra cream (drug) |
| Lead sponsor | Universitas Indonesia |
| Status | Completed |
| ClinicalTrials.gov | NCT06255431 |
2. Clinical Question
The central clinical question is whether treatment with Ezerra cream following trichloroacetic acid 15% peel affects objective measures of skin barrier function, hydration-related skin capacitance, and erythema over the registered assessment period in participants with aging disorder.
Population
Participants in a phase 4 clinical trial for aging disorder following trichloroacetic acid 15% peel.
Intervention
Ezerra cream, registered as the study intervention.
Comparator
The trial is described as randomized and controlled with a split-face design. The registry information available here does not provide a separate named comparator intervention.
Primary question
How do transepidermal water loss, skin capacitance, and erythema index change across the prespecified assessment times following the peel and study treatment?
3. Trial Design
Intervention
Ezerra cream
- Registered intervention: Ezerra cream (drug)
- Trial context: post-trichloroacetic acid 15% peel
- Trial population: aging disorder
Within-trial comparison
- Randomized controlled study
- Split-face clinical trial
- Quadruple masking
4. Endpoints
The registry identifies three primary endpoints. All three are assessed at the same three registered time points: baseline, defined as 30 minutes after peeling, followed by day 3 and day 7.
| Primary endpoint | Registered time frame | Statistical role |
|---|---|---|
| Transepidermal water loss | Baseline (30 minutes after peeling), day 3, and day 7 | Primary endpoint |
| Skin Capacitance | Baseline (30 minutes after peeling), day 3, and day 7 | Primary endpoint |
| Erythema index | Baseline (30 minutes after peeling), day 3, and day 7 | Primary endpoint |
Why the baseline definition matters
The baseline is not a pre-peel measurement. The registry explicitly defines baseline as 30 minutes after peeling. Consequently, changes from baseline should be interpreted as changes occurring after that post-peel reference point rather than as the total change caused by the peel itself.
This distinction is important for both statistical interpretation and clinical interpretation. A treatment comparison can address differences between study conditions after the intervention while leaving the separate biological effect of the initial peel outside the treatment contrast unless a pre-peel measurement was incorporated into the analysis.
5. Planned Analysis
Transepidermal water loss
Transepidermal water loss is measured repeatedly at baseline, day 3, and day 7. A suitable analysis would account simultaneously for treatment condition, time, and the repeated observations within each participant. In a split-face design, the two sides of the same participant should also be treated as paired observations rather than as independent observations.
Skin Capacitance
Skin capacitance follows the same registered assessment schedule. A repeated-measures model can estimate the average difference between treatment conditions while accounting for correlation among measurements from the same participant. The treatment-by-time interaction is particularly informative when the scientific question concerns whether the trajectory differs between conditions.
Erythema index
The erythema index is also measured at baseline, day 3, and day 7. Its analysis should preserve the longitudinal and paired structure of the split-face design. Depending on the distribution and measurement characteristics, an appropriate model could use a continuous repeated-measures framework or another method suitable for the observed data.
The treatment term represents the average treatment contrast, time represents change across assessment occasions, and the treatment-by-time interaction evaluates whether the treatment contrast changes over time. The participant-level correlation term acknowledges that repeated observations from the same participant are related.
What the registry establishes
- The study is randomized.
- The design model is identified as factorial.
- The trial is quadruple-masked.
- The study is a split-face clinical trial.
- There are two arms.
- Three primary endpoints are registered.
- Each primary endpoint is measured at baseline 30 minutes after peeling, day 3, and day 7.
The registry information does not specify the exact statistical model, covariance structure, transformation strategy, missing-data method, multiplicity procedure, or prespecified contrast for the three primary endpoints. Those choices can materially affect inference and would ordinarily be defined in a statistical analysis plan.
6. Statistical Methodology
Repeated measurements and within-participant correlation
The defining statistical feature of this study is that outcome measurements are obtained repeatedly over time and the trial is explicitly described as split-face. Both features create correlation that must be respected by the analysis.
Measurements from the same participant tend to be statistically related. A model that assumes independence for all observations can underestimate or otherwise misrepresent uncertainty when this correlation is ignored.
Change from post-peel baseline
One natural estimand is the difference in change from the registered baseline between treatment conditions. For a given participant and side, the basic quantity can be represented as:
The registered baseline is 30 minutes after peeling. Therefore, this change describes movement away from that post-peel reference measurement.
Change-score analysis is not the only possible approach. A repeated-measures model can instead adjust directly for baseline and estimate treatment contrasts at subsequent time points. The choice should be specified prospectively because the resulting estimand and assumptions are not identical.
Treatment-by-time interaction
Because the endpoints are assessed at three occasions, a treatment-by-time interaction can be useful. A statistically meaningful interaction would indicate that the treatment contrast is not constant across the assessment schedule. This is different from asking whether there is simply an average treatment difference across all time points.
Main treatment effect
Addresses the average treatment contrast across the modeled assessment period, depending on the prespecified model and estimand.
Time effect
Addresses whether the outcome changes across the registered assessment occasions regardless of treatment condition.
Treatment × time
Addresses whether the treatment contrast changes over the registered follow-up period.
Participant effect
Accounts for the fact that repeated measurements from the same participant are correlated rather than independent.
Paired structure of a split-face trial
The split-face structure adds another layer of dependence. If the two sides of a participant receive different conditions, observations from those sides are naturally paired. The statistical model should therefore retain participant as a source of clustering or otherwise explicitly model the within-participant pairing.
Factorial design
ClinicalTrials.gov identifies the design model as factorial. In a conventional factorial trial, treatment factors and their levels are explicitly defined so that main effects and interactions can be estimated. The registry information available here does not identify those factors or levels. Consequently, the factorial label establishes the registered design classification but does not by itself provide enough information to specify a factor-by-factor statistical model.
7. Statistical Methods Explained
Why does the split-face design change the analysis?
A split-face trial creates paired observations within the same participant. The two sides are exposed to the same participant-level environment, so observations from the two sides are correlated. The analysis should use that pairing rather than treating each side as an independent observation from a separate person.
Why are repeated measurements important?
The same primary endpoint is assessed at baseline, day 3, and day 7. These observations provide information about both the level of the outcome and its trajectory. Because measurements come from the same participant, the observations are correlated. Repeated-measures methods explicitly account for this structure.
What does a treatment-by-time interaction tell us?
It asks whether the treatment contrast changes across the assessment schedule. For example, the difference between conditions could be small at baseline and larger at a later assessment, or it could follow another pattern. A single overall treatment effect can obscure such changes.
Why is the baseline defined as 30 minutes after peeling?
The registry defines baseline as 30 minutes after peeling. This means that baseline represents the post-peel reference state. Consequently, a subsequent change estimate describes what happens after that reference point rather than measuring the complete effect of the peel itself.
Why is participant-level correlation important?
Repeated observations from one participant share biological and measurement characteristics. The same is true for paired sides in a split-face design. Accounting for this correlation is important because standard errors, confidence intervals, and hypothesis tests depend on the assumed covariance structure.
What does the factorial designation tell us?
It tells us that ClinicalTrials.gov classifies the study using a factorial design model. It does not, by itself, identify the factors, levels, or interaction terms. Those details are necessary before a conventional factorial treatment-effect analysis can be specified.
8. Interpreting the Three Primary Endpoints
| Endpoint | What it contributes | Key statistical issue |
|---|---|---|
| Transepidermal water loss | Repeated objective assessment of a skin-related physiological measure | Repeated observations and split-face pairing |
| Skin Capacitance | Repeated objective assessment of skin capacitance | Repeated observations and split-face pairing |
| Erythema index | Repeated objective assessment of erythema | Repeated observations and split-face pairing |
The three endpoints are related to the same post-peel clinical context but represent different measurements. They should not automatically be collapsed into a single composite endpoint. Each endpoint has its own measurement scale, variability, and scientific interpretation.
The presence of multiple primary endpoints also raises a multiplicity question. If each endpoint is tested independently at the same nominal significance level, the probability of at least one false-positive result can exceed the nominal level. A confirmatory analysis would ordinarily specify how the family of primary hypotheses is handled.
9. Masking and Bias Control
The trial is described as quadruple-masked. Masking can reduce the risk that expectations about treatment assignment influence treatment administration, participant behavior, outcome assessment, or other trial procedures.
Why masking matters
Objective measurements can still be affected by procedural decisions, measurement conditions, or expectations. Masking helps reduce these potential sources of bias.
Why objective endpoints help
Transepidermal water loss, skin capacitance, and erythema index are instrument-based measurements, providing a more standardized outcome framework than purely subjective assessments.
Masking is not randomization
Masking primarily addresses bias in trial conduct and assessment. Randomization addresses confounding in treatment comparisons. They solve different statistical problems.
Split-face and masking
The split-face structure can strengthen within-participant comparison, while masking can reduce differential behavior or assessment associated with knowledge of treatment assignment.
10. Limitations and Interpretation Issues
- No posted statistical analyses: the ClinicalTrials.gov record identifies the study as completed but does not provide formal statistical analyses for the primary endpoints.
- Small enrollment: the registered enrollment is 27.0, so estimates from the study would be expected to have substantial sampling uncertainty.
- Split-face dependence: analyses must account for the paired nature of observations from the same participant.
- Repeated-measures dependence: measurements at baseline, day 3, and day 7 are not independent observations.
- Multiple primary endpoints: three primary endpoints create a multiplicity issue unless a prespecified error-control strategy is used.
- Factorial specification: the registry identifies a factorial design model but does not provide the individual factors and levels needed to reconstruct a full factorial analysis.
- Baseline timing: baseline is 30 minutes after peeling, so changes from baseline should not be interpreted as the total effect of the peel.
- Missing-data considerations: the available registry information does not specify a missing-data or imputation strategy. In a longitudinal study, differential loss of measurements can affect treatment comparisons and precision.
- Model specification: the available record does not specify the covariance structure, transformation strategy, or exact inferential model for the repeated measurements.
- Generalizability: the trial concerns a specific post-peel setting and aging disorder population, so statistical findings would need to be interpreted in the context of the enrolled study population.
11. Why This Trial Matters Statistically
SHEA is a useful statistical teaching example because several important clinical-trial concepts occur simultaneously in a compact randomized study: randomization, quadruple masking, factorial design classification, split-face pairing, repeated measurements, multiple primary endpoints, and post-intervention baseline definition.
| Concept | How it appears in SHEA |
|---|---|
| Randomization | The registry identifies randomized allocation. |
| Masking | The trial is described as quadruple-masked. |
| Split-face design | The study is explicitly described as a split-face clinical trial. |
| Factorial design | ClinicalTrials.gov classifies the design model as factorial. |
| Repeated measures | Each primary endpoint is measured at baseline, day 3, and day 7. |
| Within-participant correlation | Repeated measurements and split-face observations create dependence within participant. |
| Multiple primary endpoints | Transepidermal water loss, skin capacitance, and erythema index are all registered as primary endpoints. |
| Baseline definition | Baseline is specifically defined as 30 minutes after peeling. |
| Small-sample inference | Enrollment is 27, making uncertainty and model stability important considerations. |
The statistical lesson
The central lesson is that the analysis must follow the experimental structure. A conventional two-group comparison that ignores the split-face pairing, repeated measurements, or participant-level correlation would fail to use the information encoded by the design.
At the same time, sophisticated modeling should not be confused with stronger evidence by itself. A mixed-effects model, generalized estimating equation, or another repeated-measures method is only as appropriate as its assumptions, estimand, covariance specification, and handling of missing observations. With a small study, those choices can materially affect the resulting inference.
12. Related Tutorials
Learn more about the methods used in this trial:
13. Related Statistical Calculators
14. Sources
- ClinicalTrials.gov: NCT06255431.
Continue through the Clinical Biostats knowledge graph
Explore the broader statistical methods that underlie randomized clinical-trial design, masking, and factorial studies.
15. Record Summary
SHEA is a completed phase 4 randomized, controlled, quadruple-masked, split-face clinical trial in aging disorder. The registry identifies Ezerra cream as the intervention, a factorial design model, two arms, and 27.0 enrolled participants. Its three primary endpoints—transepidermal water loss, skin capacitance, and erythema index—are assessed at baseline 30 minutes after peeling, day 3, and day 7.
From a statistical perspective, the defining challenge is the combination of within-participant pairing and repeated measurements. An appropriate analysis should preserve those dependencies rather than treating every measurement as independent. The three primary endpoints also create a multiplicity consideration, while the post-peel definition of baseline determines precisely what a change-from-baseline analysis would represent. The registry identifies the study as completed, but formal statistical analyses and results are not posted in the available record.