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
EPHESUS was a survey concerning the under-treatment of hypercholesterolemia in patients with hyperlipidemias. The registry lists an enrollment of 6000 participants and identifies as its primary endpoint the number of patients in the secondary prevention group receiving appropriate lipid-lowering therapy at 2 months.
| Feature | EPHESUS |
|---|---|
| Acronym | EPHESUS |
| Official title | A Survey on the Under-treatment of hypercholeSterolemia (EPHESUS Trial) |
| Condition | Hyperlipidemias |
| Enrollment | 6000 |
| Status | UNKNOWN |
| Start date | 2015-11 |
| Primary completion date | 2016-09 |
| Lead sponsor | Muğla Sıtkı Koçman University |
| Sponsor type | OTHER |
| ClinicalTrials.gov | NCT02608645 |
2. Clinical Question
The registry-defined clinical question centers on the treatment of hypercholesterolemia in patients with hyperlipidemias, specifically examining whether patients in a secondary prevention group are receiving appropriate lipid-lowering therapy.
Population
The registry identifies the condition under study as hyperlipidemias and specifies a secondary prevention group within the primary endpoint.
Intervention
No randomized intervention is identified in the registry information summarized here. The study is described as a survey on under-treatment of hypercholesterolemia.
Comparator
No comparator group is specified in the registry information summarized here.
Primary question
How many patients in the secondary prevention group are receiving appropriate lipid-lowering therapy at 2 months?
3. Trial Design
The registry identifies EPHESUS as a survey concerning the under-treatment of hypercholesterolemia. The planned enrollment is 6000 participants, with a start date of 2015-11 and a primary completion date of 2016-09.
The registry information does not specify randomization, masking, treatment allocation, study arms, or a comparator. Those design features should therefore not be assumed when interpreting the primary endpoint.
4. Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Primary endpoint | The number of patients with secondary prevention group receiving appropriate lipid lowering therapy | 2 months |
The endpoint is expressed as a number of patients, rather than as a hazard ratio, mean change, median time, or other continuous or time-to-event measure. Statistically, that distinction determines the appropriate framework for summarizing and interpreting the endpoint.
5. Statistical Methodology
What the primary endpoint measures
The primary endpoint counts patients in the secondary prevention group who are receiving appropriate lipid-lowering therapy at the specified 2-month assessment. The most direct descriptive quantity is therefore the number meeting the endpoint definition.
Because the registry defines the endpoint as a number, interpretation begins with the observed count. If the denominator of the relevant secondary prevention group is available, the corresponding proportion can also provide clinically useful context.
Count versus proportion
A count is dependent on the size of the population being evaluated. For example, the same number of appropriately treated patients can represent very different levels of treatment coverage if the underlying secondary prevention populations differ in size. A proportion addresses this by relating the count to the number of eligible patients:
where x is the number of patients receiving appropriate therapy and n is the number of patients in the relevant secondary prevention group.
The registry information identifies the primary endpoint as a number but does not provide the observed count or the size of the secondary prevention denominator.
Confidence intervals for a proportion
If an analysis reports the proportion of the secondary prevention group receiving appropriate therapy, a binomial confidence interval would ordinarily be appropriate. The interval quantifies uncertainty around the estimated proportion; it does not describe variability among individual patients.
Here, X represents the number of patients meeting the endpoint among n eligible patients, while p represents the underlying probability of meeting the endpoint under the chosen statistical framework.
Comparison of treatment coverage
If the study had prespecified comparisons between distinct groups, the statistical analysis would depend on the way those groups were defined. For independent groups, a comparison of proportions could be used. For paired observations on the same patients, a paired binary method would be more appropriate. The registry information summarized here does not identify such comparison groups.
Regression analysis
If patient-level covariates were available, logistic regression could be used to examine factors associated with receipt of appropriate lipid-lowering therapy. Such an analysis would estimate associations rather than establish that a particular patient characteristic causes treatment under-use.
Multiplicity
Multiplicity becomes important when investigators conduct several formal statistical tests or evaluate multiple primary or co-primary endpoints. EPHESUS has one registered primary endpoint in the registry information summarized here, so there is no basis here for describing a more complex primary-endpoint multiplicity strategy.
6. Statistical Methods Explained
Why is the primary endpoint a count?
The registry defines the primary endpoint as the number of patients receiving appropriate lipid-lowering therapy in the secondary prevention group. A count directly answers the operational question of how many patients meet the specified treatment criterion.
Why might a proportion be more informative than a count?
A count alone does not account for the size of the population eligible for assessment. Dividing the number meeting the endpoint by the number of eligible patients produces a treatment-coverage proportion, which is easier to compare across populations of different sizes.
What would a confidence interval add?
A confidence interval would show the statistical uncertainty around an estimated proportion or other parameter. A narrow interval indicates greater statistical precision, while a wider interval indicates greater uncertainty under the specified model and sampling framework.
Would a hazard ratio be appropriate here?
Not for the registered primary endpoint as stated. A hazard ratio is designed for a time-to-event comparison, whereas this endpoint is a count of patients receiving appropriate therapy at a specified 2-month time frame.
Could logistic regression be used?
Yes, if individual-level information on the binary endpoint and relevant explanatory variables were available. Logistic regression could model the odds of receiving appropriate therapy. The resulting association would depend on the covariates, model specification, and study design.
Why does the denominator matter?
The denominator defines the population to which the treatment-coverage estimate applies. For a secondary prevention endpoint, the number of patients in that group is essential for converting the registered count into a proportion and for assessing statistical precision.
7. Planned Analysis
The registry's primary endpoint is the number of patients in the secondary prevention group receiving appropriate lipid-lowering therapy at 2 months. For an endpoint of this type, the natural primary descriptive analysis is a count of patients meeting the endpoint definition, accompanied where appropriate by the proportion of eligible patients meeting the criterion.
If inferential analysis were prespecified, a binomial framework could be used for the proportion, with an appropriate confidence interval. If comparisons between independent groups were part of the statistical plan, methods for comparing proportions could be used. If repeated measurements on the same patients were involved, paired binary methods would be appropriate. If patient characteristics were examined as predictors of treatment receipt, logistic regression could provide adjusted associations.
8. Interpretation of the Primary Endpoint
The primary endpoint is designed to quantify treatment coverage within the secondary prevention group: specifically, how many patients are receiving appropriate lipid-lowering therapy at 2 months.
A count alone would not establish why a patient was or was not receiving appropriate therapy. It would also not by itself establish a causal effect of a particular treatment strategy because the registry information does not identify a randomized intervention or comparator.
The registry specifies 2 months as the time frame for the primary endpoint. Treatment coverage measured at a defined time point is a snapshot of the treatment status at that point rather than a time-to-event estimate.
The registry lists an enrollment of 6000, but the primary endpoint specifically concerns patients in the secondary prevention group. The total enrollment therefore should not automatically be substituted for the number of patients eligible for the primary endpoint.
9. What a Complete Statistical Report Would Need to Show
A complete report of this primary endpoint would ideally distinguish the overall enrolled population from the subset forming the secondary prevention group. It would then report the number receiving appropriate lipid-lowering therapy and, if applicable, the corresponding proportion with a confidence interval.
| Statistical component | Purpose |
|---|---|
| Eligible secondary prevention population | Defines the denominator for the primary endpoint. |
| Number receiving appropriate therapy | Directly reports the registered primary endpoint. |
| Proportion receiving appropriate therapy | Places the count in the context of the eligible population size. |
| Confidence interval | Quantifies statistical uncertainty around an estimated proportion. |
| Prespecified subgroup analyses | Could describe whether treatment coverage differed across clinically defined groups if such analyses were planned. |
| Missing-data accounting | Shows how patients without an evaluable 2-month assessment were handled. |
The registry information summarized here does not provide these observed results or a posted statistical analysis plan. Consequently, the registered endpoint is best understood as the target measurement rather than as evidence of a particular observed treatment-coverage level.
10. Missing Data and Analysis Population
For a 2-month treatment-status endpoint, missingness can affect both the numerator and denominator. Patients without an evaluable assessment may require explicit handling rules so that the reported treatment-coverage estimate is interpretable.
Common approaches include defining an analysis population before examining outcomes, reporting the number of patients with evaluable assessments, and conducting sensitivity analyses when missingness could materially affect the conclusion. The appropriate method depends on the prespecified study plan and the mechanism generating missing observations.
Complete-case analysis
Uses patients with an evaluable primary endpoint. Its interpretation depends on whether excluded patients differ systematically from those with observed assessments.
Missingness sensitivity
Alternative assumptions about patients without an evaluable assessment can show how robust the treatment-coverage estimate is to missing observations.
11. Statistical Design Issues
Randomization
The registry information does not specify randomization. Therefore, the primary endpoint should not be interpreted as a randomized treatment comparison.
Non-inferiority
No non-inferiority hypothesis or non-inferiority margin is identified in the registry information summarized here. Non-inferiority methodology is therefore not part of the documented primary-endpoint framework.
Factorial design
No factorial allocation is identified. A factorial analysis should not be assumed from the registry information.
Bayesian methods
No Bayesian analysis is identified in the registry information summarized here. The primary endpoint can be described using frequentist binomial methods if inferential analysis is required, but that represents a general statistical framework rather than a documented EPHESUS analysis.
Interim analysis
No interim statistical analysis is identified in the registry information summarized here. The study dates alone do not establish an interim-monitoring strategy.
12. Study Timeline
Study start
The registry lists 2015-11 as the EPHESUS start date.
Primary completion
The registry lists 2016-09 as the primary completion date.
Primary endpoint time frame
The primary endpoint is assessed at 2 months: the number of patients in the secondary prevention group receiving appropriate lipid-lowering therapy.
13. Trial Status and Registry Context
The registry lists the study status as UNKNOWN. It also identifies Muğla Sıtkı Koçman University as the lead sponsor and lists the sponsor type as OTHER.
| Registry field | EPHESUS |
|---|---|
| Status | UNKNOWN |
| Start date | 2015-11 |
| Primary completion date | 2016-09 |
| Lead sponsor | Muğla Sıtkı Koçman University |
| Sponsor type | OTHER |
| Enrollment | 6000 |
14. Limitations
- No posted statistical analyses: the ClinicalTrials.gov record does not contain formal statistical analyses for the registered primary endpoint.
- No posted endpoint result: the registry does not provide an observed number of patients receiving appropriate lipid-lowering therapy at 2 months.
- Denominator not reported: the registry information summarized here does not provide the number of patients comprising the secondary prevention group.
- No comparator specified: the available registry information does not identify a comparator group for the primary endpoint.
- No randomized intervention identified: the study is described as a survey on under-treatment rather than a randomized treatment comparison.
- Missing-data framework not specified: the available information does not describe how patients without an evaluable 2-month assessment were handled.
- Limited statistical detail: the available registry information does not identify a prespecified inferential model, confidence-interval method, subgroup strategy, or multiplicity procedure for the primary endpoint.
15. Why This Trial Matters Statistically
EPHESUS illustrates an important distinction in clinical-trial statistics: not every clinical research question is a comparison of two randomized treatments or a time-to-event problem. The registered primary endpoint is a treatment-coverage measure expressed as a patient count at a fixed time point.
| Concept | How it applies to EPHESUS |
|---|---|
| Descriptive statistics | The primary endpoint is a count of patients receiving appropriate lipid-lowering therapy. |
| Proportions | The count can be placed in context by dividing by the number of eligible patients in the secondary prevention group. |
| Binomial inference | A natural framework for estimating uncertainty around a treatment-coverage proportion. |
| Denominator definition | Essential for interpreting a treatment-coverage percentage. |
| Missing data | Patients without an evaluable 2-month assessment can affect the estimated treatment-coverage rate. |
| Logistic regression | Could be used to study associations between patient characteristics and receipt of appropriate therapy if patient-level covariates were available. |
| Causal inference | A treatment-coverage survey does not automatically establish that a specific treatment strategy caused an observed outcome. |
| Fixed-time assessment | The 2-month time frame differs fundamentally from a time-to-event endpoint such as survival or progression. |
16. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The registered primary endpoint is a fixed-time binary treatment-status concept summarized as a patient count. Its natural statistical extension is a proportion among the eligible secondary prevention population, with uncertainty quantified using an appropriate binomial framework when inferential analysis is required.
Clinical interpretation
The clinical meaning concerns the extent to which patients in the secondary prevention group are receiving appropriate lipid-lowering therapy at the specified assessment time. Understanding that treatment coverage requires knowing both the number treated and the population eligible for the endpoint.
17. Statistical Perspective on Treatment Under-Use
A treatment-under-use survey can be informative even without a randomized intervention. The statistical task is first to define the eligible population precisely, then to determine how many patients meet the treatment criterion, and finally to quantify uncertainty and investigate relevant sources of variation.
The distinction between how many patients are treated and what fraction of eligible patients are treated is particularly important. The former is the registry's stated endpoint; the latter provides a standardized measure of treatment coverage when the appropriate denominator is known.
This ratio converts the registered count into a population-level treatment-coverage measure. Its interpretation depends critically on how eligibility and missing assessments are defined.
18. What Cannot Be Concluded From the Registered Endpoint Alone
The registered endpoint does not by itself establish whether a particular therapy improves clinical outcomes. It measures treatment receipt, not a clinical outcome such as cardiovascular events, mortality, or another time-to-event endpoint.
Likewise, a treatment-coverage percentage would not by itself identify the causes of under-treatment. Differences could reflect clinical eligibility, treatment decisions, patient factors, access, adherence, contraindications, or other factors. Establishing those relationships would require additional variables and an analysis designed for that purpose.
Finally, the total enrollment of 6000 should not automatically be interpreted as the denominator for the primary endpoint. The endpoint specifically concerns the secondary prevention group, whose size is not reported in the available registry information.
19. Sources
- ClinicalTrials.gov: EPHESUS, NCT02608645.
Continue through the Clinical Biostats clinical trial library
Explore additional clinical trial analyses and statistical methods across the Clinical Biostats library.
20. Record Summary
EPHESUS is registered as a survey concerning the under-treatment of hypercholesterolemia in patients with hyperlipidemias, with an enrollment of 6000. Its registered primary endpoint is the number of patients in the secondary prevention group receiving appropriate lipid-lowering therapy at 2 months. Statistically, this is a fixed-time treatment-coverage endpoint rather than a randomized time-to-event comparison. The most informative analysis would distinguish the eligible secondary prevention population from total enrollment, report the number meeting the endpoint, and, where appropriate, express that count as a proportion with a suitable measure of statistical uncertainty.