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
VIEW 2 was a randomized phase 3 trial evaluating aflibercept injection against ranibizumab in participants with macular degeneration. The registered primary endpoint was the percentage of participants who maintained vision at week 52, analyzed using a non-inferiority framework.
| Feature | VIEW 2 |
|---|---|
| Phase | Phase 3 |
| Condition | Macular Degeneration |
| Brief title | Vascular Endothelial Growth Factor (VEGF) Trap-Eye: Investigation of Efficacy and Safety in Wet Age-Related Macular Degeneration (AMD) |
| Design | Randomized, parallel-group |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 1,240 |
| Arms | 4 |
| Start | April 2008 |
| Primary completion | September 2010 |
| Lead sponsor | Bayer |
| ClinicalTrials.gov | NCT00637377 |
2. Clinical Question
The central statistical question was whether aflibercept injection could maintain vision at week 52 at a rate that was not unacceptably worse than ranibizumab 0.5mg Q4. The registry specified a non-inferiority margin of 10 and a 90% power target in the study protocol's sample-size estimation.
Population
Participants enrolled in the phase 3 VIEW 2 trial for wet age-related macular degeneration.
Reference treatment
Ranibizumab 0.5mg Q4 served as the comparator for the three primary aflibercept comparisons.
Investigational treatment
Aflibercept injection was evaluated at 2mg Q4, 0.5mg Q4, and 2mg Q8 dosing schedules.
Primary question
For each aflibercept regimen, was the probability of maintaining vision at week 52 sufficiently close to that observed with ranibizumab to satisfy the prespecified non-inferiority criterion?
3. Trial Design
Ranibizumab 0.5mg Q4
- Ranibizumab 0.5mg
- Q4 dosing schedule
- Reference treatment for the three primary comparisons
Aflibercept 2mg Q4
- Aflibercept injection
- 2mg Q4 dosing schedule
- Primary non-inferiority comparison versus ranibizumab 0.5mg Q4
Aflibercept 0.5mg Q4
- Aflibercept injection
- 0.5mg Q4 dosing schedule
- Primary non-inferiority comparison versus ranibizumab 0.5mg Q4
Aflibercept 2mg Q8
- Aflibercept injection
- 2mg Q8 dosing schedule
- Primary non-inferiority comparison versus ranibizumab 0.5mg Q4
4. Primary Endpoint
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Percentage of Participants Who Maintained Vision at Week 52 - LOCF | Maintenance of vision was defined as a loss of < 15 letters in the ETDRS letter score in the study eye. The defined study baseline range of ETDRS Best Corrected Visual Acuity letter score was 73 to 25, corresponding to acuity of 20/40 to 20/320. A higher score represents better functioning. | At week 52 |
The registered calculation expressed the percentage as the number of participants who maintained vision multiplied by 100, divided by the number of participants analyzed. The endpoint is therefore binary at the participant level: each analyzed participant either met the maintenance criterion or did not.
5. Statistical Methodology
Primary analysis: risk difference
The primary comparisons used the risk difference, calculated as the percentage maintaining vision in the ranibizumab 0.5mg Q4 group minus the percentage maintaining vision in the corresponding aflibercept group.
A negative value therefore favors the corresponding aflibercept regimen because the observed percentage maintaining vision is higher in the aflibercept group.
Wald / z-test
The registry analysis used a Wald / z-test for the primary binary endpoint. The null hypothesis was expressed as a non-inferiority hypothesis: for each aflibercept regimen, the probability of maintaining vision had to be sufficiently close to the ranibizumab probability, with the allowable difference defined by the prespecified margin.
Non-inferiority margin
The non-inferiority margin was set to 10. The registry states that the study protocol's sample-size estimation targeted 90% power.
For the 2mg Q4 comparison, for example, the null hypothesis was that the probability of maintaining vision under aflibercept 2mg Q4 was no more than the ranibizumab probability minus the non-inferiority margin. The same structure was used for aflibercept 0.5mg Q4 and aflibercept 2mg Q8.
Here πᵢ is the probability of maintaining vision under the aflibercept regimen, πc is the probability under ranibizumab 0.5mg Q4, and δ is the non-inferiority margin.
Confidence intervals
The registry describes a two-sided 95% confidence interval using a normal approximation for the difference in percentages. Because the reported difference is defined as ranibizumab minus aflibercept, the interpretation of the confidence interval must be made using that direction rather than treating the estimate as an unsigned distance.
Analysis population and missing data
The primary endpoint was analyzed in the Per-Protocol Set (PPS) using Last Observation Carried Forward (LOCF). This combination is important statistically because non-inferiority trials commonly distinguish the per-protocol population from broader efficacy populations when assessing whether protocol deviations could affect the treatment comparison.
Multiplicity
The primary comparisons were handled using a conditional sequence of statistical hypotheses with alpha = 0.05. The registry's limitation states that the conditional sequence had to stop after testing the first superiority hypothesis. Consequently, all further statistical tests are described in the registry as exploratory tests.
6. Primary Results
The registry contains three formal primary- endpoint analyses. All three compare an aflibercept regimen with ranibizumab 0.5mg Q4 using the risk difference defined as ranibizumab minus aflibercept.
Aflibercept 2mg Q4 versus Ranibizumab 0.5mg Q4
Difference in percentage maintaining vision
95% CI: −4.86 to 2.46
Risk difference = Ranibizumab 0.5mg Q4 minus Aflibercept 2mg Q4
Non-inferiority margin: 10
The estimated risk difference of −1.20 means that the percentage maintaining vision was estimated to be 1.20 percentage points higher with aflibercept 2mg Q4 than with ranibizumab 0.5mg Q4, because the registry defines the difference in the reverse direction.
The 95% confidence interval extends from −4.86 to 2.46. Thus, the estimate is compatible with a modest advantage for aflibercept as well as with a modest advantage for ranibizumab under the reported statistical model. The confidence interval is also evaluated against the prespecified non-inferiority margin rather than against zero alone.
This is a non-inferiority question, so the key issue is whether the confidence interval excludes differences that would be worse than the allowed margin. The p-value is not an effect-size measure and does not describe the magnitude of the difference or the clinical importance of a difference.
The analysis uses a per-protocol population and LOCF. That matters because conclusions from a non-inferiority analysis can be sensitive to deviations from assigned treatment and assumptions about missing observations.
Aflibercept 0.5mg Q4 versus Ranibizumab 0.5mg Q4
Difference in percentage maintaining vision
95% CI: −5.40 to 1.71
Risk difference = Ranibizumab 0.5mg Q4 minus Aflibercept 0.5mg Q4
Non-inferiority margin: 10
The estimated risk difference of −1.84 corresponds to an estimated 1.84 percentage-point higher percentage maintaining vision with aflibercept 0.5mg Q4 than with ranibizumab 0.5mg Q4 under the registry's direction convention.
The 95% confidence interval ranges from −5.40 to 1.71. It therefore includes zero, meaning the interval is compatible with either direction of a small difference. For non-inferiority, however, the relevant reference is the prespecified margin of 10 rather than zero by itself.
The estimate should not be interpreted as proof that the two treatments have identical effects. Non-inferiority means that the observed evidence is evaluated against a prespecified acceptable loss of efficacy; it is not equivalent to demonstrating exact equality.
As with the other primary comparisons, the analysis population was the per-protocol set and missing observations were handled with LOCF. The reported confidence interval reflects uncertainty under that analysis framework.
Aflibercept 2mg Q8 versus Ranibizumab 0.5mg Q4
Difference in percentage maintaining vision
95% CI: −4.81 to 2.55
Risk difference = Ranibizumab 0.5mg Q4 minus Aflibercept 2mg Q8
Non-inferiority margin: 10
The estimated risk difference of −1.13 means that the estimated percentage maintaining vision was 1.13 percentage points higher with aflibercept 2mg Q8 than with ranibizumab 0.5mg Q4, using the registry's stated subtraction order.
The 95% confidence interval of −4.81 to 2.55 spans zero, so the data are compatible with a small difference in either direction. The interval is substantially narrower than the non-inferiority margin, which is the relevant scale for the prespecified non-inferiority question.
The confidence interval describes uncertainty around the estimated group difference; it does not describe the range of responses that individual patients could experience. Similarly, a p-value, where reported, would test a particular null hypothesis rather than quantify the clinical magnitude of the treatment effect.
The interpretation is conditional on the per-protocol analysis, LOCF handling of missing observations, and the prespecified hypothesis sequence.
7. Secondary Endpoint Results: BCVA
The registry also reports secondary analyses of mean change from baseline in Best Corrected Visual Acuity (BCVA), measured by the ETDRS letter score at week 52. These analyses used ANCOVA in the Full-Analysis Set with LOCF.
| Comparison | Estimate | 95% CI | P-value |
|---|---|---|---|
| Aflibercept 2mg Q4 vs Ranibizumab 0.5mg Q4 | −1.9484 | −4.1009 to 0.2040 | 0.076 |
| Aflibercept 0.5mg Q4 vs Ranibizumab 0.5mg Q4 | −0.0620 | −2.2398 to 2.1158 | 0.9555 |
| Aflibercept 2mg Q8 vs Ranibizumab 0.5mg Q4 | −0.9014 | −3.0615 to 1.2587 | 0.4131 |
For these analyses, the difference was calculated as Aflibercept minus Ranibizumab, so a positive value favors the corresponding aflibercept regimen. The ANCOVA model included treatment group as a fixed factor across all four treatment groups and the baseline ETDRS letter score as a covariate.
The registry describes the treatment comparison as a contrast from an ANCOVA model containing treatment group and baseline ETDRS letter score. Adjusting for baseline visual acuity can improve precision by accounting for baseline differences in the outcome being analyzed.
8. Secondary Endpoint Results: Participants Gaining at Least 15 Letters
A second binary endpoint evaluated the percentage of participants who gained at least 15 letters of vision in the ETDRS letter score in the study eye at week 52. These analyses used the Full-Analysis Set with LOCF and a chi-squared method.
| Comparison | Risk difference | 95% CI | P-value |
|---|---|---|---|
| Aflibercept 2mg Q4 vs Ranibizumab 0.5mg Q4 | −4.57 | −12.02 to 2.88 | 0.229 |
| Aflibercept 0.5mg Q4 vs Ranibizumab 0.5mg Q4 | 0.78 | −6.91 to 8.46 | 0.843 |
| Aflibercept 2mg Q8 vs Ranibizumab 0.5mg Q4 | −2.65 | −10.18 to 4.88 | 0.490 |
For these secondary analyses, the registry defines the difference as Aflibercept minus Ranibizumab, so a positive value favors the corresponding aflibercept regimen. The confidence intervals include zero for all three comparisons, and the estimates should be interpreted as secondary superiority analyses rather than as replacements for the primary non-inferiority endpoint.
9. Secondary Endpoint Results: NEI VFQ-25
The National Eye Institute 25-item Visual Function Questionnaire (NEI VFQ-25) total score was analyzed as a continuous endpoint using ANCOVA, with baseline NEI VFQ-25 total score as the covariate and LOCF for missing observations.
| Comparison | Estimate | 95% CI | P-value |
|---|---|---|---|
| Aflibercept 2mg Q4 vs Ranibizumab 0.5mg Q4 | −2.7885 | −4.9012 to −0.6757 | 0.0097 |
| Aflibercept 0.5mg Q4 vs Ranibizumab 0.5mg Q4 | −0.9320 | −3.0658 to 1.2019 | 0.3917 |
| Aflibercept 2mg Q8 vs Ranibizumab 0.5mg Q4 | −1.9470 | −4.0659 to 0.1718 | 0.0717 |
The difference was calculated as Aflibercept minus Ranibizumab, so negative estimates favor the ranibizumab group under the registry's stated direction. The confidence interval for the 2mg Q4 comparison excludes zero, whereas the other two intervals include zero. These are secondary superiority analyses and therefore must also be interpreted in the context of the trial's conditional hypothesis-testing sequence.
10. Secondary Endpoint Results: Choroidal Neovascularization Area
Mean change from baseline in Choroidal Neovascularization (CNV) Area at week 52 was analyzed using ANCOVA with baseline CNV area as a covariate. The Full-Analysis Set with assessment for the outcome measure was used, with LOCF.
| Comparison | Estimate, mm2 | 95% CI | P-value |
|---|---|---|---|
| Aflibercept 2mg Q4 vs Ranibizumab 0.5mg Q4 | −1.180 | −1.979 to −0.382 | 0.0038 |
| Aflibercept 0.5mg Q4 vs Ranibizumab 0.5mg Q4 | 0.170 | −0.632 to 0.972 | 0.6784 |
| Aflibercept 2mg Q8 vs Ranibizumab 0.5mg Q4 | −0.733 | −1.534 to 0.068 | 0.0727 |
The registry defines the difference as Aflibercept minus Ranibizumab. Consequently, the negative estimate for the 2mg Q4 comparison favors aflibercept for this endpoint, whereas the positive estimate for the 0.5mg Q4 comparison favors ranibizumab under the stated direction convention. Again, the statistical role of each test depends on the prespecified multiplicity sequence.
11. Secondary Results: Consolidated Statistical View
| Endpoint | 2mg Q4 | 0.5mg Q4 | 2mg Q8 |
|---|---|---|---|
| BCVA mean change: estimate | −1.9484 | −0.0620 | −0.9014 |
| BCVA 95% CI | −4.1009 to 0.2040 | −2.2398 to 2.1158 | −3.0615 to 1.2587 |
| BCVA P-value | 0.076 | 0.9555 | 0.4131 |
| ≥15-letter gain: risk difference | −4.57 | 0.78 | −2.65 |
| ≥15-letter gain 95% CI | −12.02 to 2.88 | −6.91 to 8.46 | −10.18 to 4.88 |
| ≥15-letter gain P-value | 0.229 | 0.843 | 0.490 |
| NEI VFQ-25: estimate | −2.7885 | −0.9320 | −1.9470 |
| NEI VFQ-25 95% CI | −4.9012 to −0.6757 | −3.0658 to 1.2019 | −4.0659 to 0.1718 |
| NEI VFQ-25 P-value | 0.0097 | 0.3917 | 0.0717 |
| CNV area: estimate | −1.180 | 0.170 | −0.733 |
| CNV area 95% CI | −1.979 to −0.382 | −0.632 to 0.972 | −1.534 to 0.068 |
| CNV area P-value | 0.0038 | 0.6784 | 0.0727 |
This table is most useful when read vertically by endpoint and horizontally by regimen. The three aflibercept schedules were not evaluated with one universal effect measure: binary outcomes were expressed as risk differences, while continuous outcomes were expressed as differences in least-squares means from ANCOVA.
12. Safety: Serious Adverse Events
The registry reports serious adverse events by treatment arm as affected participants divided by participants at risk.
| Treatment arm | Participants with serious adverse events | Participants at risk |
|---|---|---|
| Ranibizumab 0.5mg Q4 | 65 | 291 |
| Aflibercept 2mg Q4 | 81 | 309 |
| Aflibercept 0.5mg Q4 | 72 | 297 |
| Aflibercept 2mg Q8 | 81 | 307 |
These figures describe the number affected and the corresponding number at risk in each treatment arm. They should not be interpreted as a formal comparative safety hypothesis test because no effect estimate, confidence interval, or P-value for the between-arm serious-adverse-event comparison is reported here.
13. Statistical Methods Explained
Why was a risk difference used for the primary endpoint?
The primary endpoint is binary: each participant either maintained vision at week 52 according to the prespecified ETDRS criterion or did not. A risk difference expresses the absolute difference between the percentages meeting that criterion. In VIEW 2, the subtraction order was ranibizumab minus aflibercept, so negative values correspond to a higher observed percentage in the aflibercept group.
Why is the non-inferiority margin more important than whether the confidence interval crosses zero?
For a superiority analysis, zero is often the central reference for a difference. Non-inferiority asks a different question: whether the investigational treatment could be worse than the comparator by more than a prespecified clinically acceptable amount. VIEW 2 used a margin of 10. Therefore, the relevant boundary is the lower side of the confidence interval relative to that margin, not simply whether zero is contained in the interval.
What does a negative primary risk difference mean?
Because VIEW 2 defined the primary difference as ranibizumab minus aflibercept, a negative estimate means the observed percentage maintaining vision was higher in the aflibercept group. For example, an estimate of −1.20 corresponds to an estimated 1.20 percentage-point higher percentage in the aflibercept group.
Why was ANCOVA used for BCVA and other continuous endpoints?
ANCOVA provides a treatment comparison while adjusting for the baseline value of the same outcome. In the BCVA analysis, baseline ETDRS letter score was included as a covariate. For NEI VFQ-25, baseline total score was the covariate, and for CNV area, baseline CNV area was the covariate. The treatment comparison was reported as a difference in least-squares means.
What does LOCF contribute to these analyses?
LOCF, or Last Observation Carried Forward, uses a participant's last available observation when the week 52 assessment is missing. This creates a complete analysis value under the specified imputation rule, but it relies on an assumption that the carried-forward observation is an appropriate representation of the missing endpoint. The interpretation of the reported estimate therefore includes the limitations of that imputation strategy.
Why is the per-protocol population important in a non-inferiority trial?
Non-inferiority trials can be particularly sensitive to treatment deviations because departures from assigned treatment can make groups appear more similar. VIEW 2's primary analysis used the Per-Protocol Set, making adherence to the protocol part of the primary analysis framework. The distinction between the per-protocol population and the Full-Analysis Set is therefore central to understanding the analysis.
What does the conditional hypothesis sequence accomplish?
The conditional sequence provides a prespecified order for hypothesis testing. This helps control the overall false-positive rate when several hypotheses are tested. The registry states that the sequence stopped after testing the first superiority hypothesis, after which further statistical tests were exploratory. That distinction is important when interpreting the secondary P-values.
14. Confidence Intervals and Statistical Precision
The primary confidence intervals are two-sided 95% intervals for the risk difference. A confidence interval provides information about statistical precision: narrower intervals indicate more concentrated estimates, whereas wider intervals indicate greater uncertainty.
| Primary comparison | Estimate | 95% CI width |
|---|---|---|
| Aflibercept 2mg Q4 | −1.20 | −4.86 to 2.46 |
| Aflibercept 0.5mg Q4 | −1.84 | −5.40 to 1.71 |
| Aflibercept 2mg Q8 | −1.13 | −4.81 to 2.55 |
The three intervals have similar widths, which is consistent with the three primary analyses having broadly similar statistical precision. None of these confidence intervals should be interpreted as a range of individual patient outcomes. They quantify uncertainty around the estimated population-level risk difference under the stated analysis framework.
A risk difference is an absolute measure. An estimate of −1.20 means a 1.20 percentage-point difference under the subtraction convention used by the registry. This differs from a relative risk, odds ratio, or hazard ratio, which describe treatment effects on different scales.
15. Why P-values Do Not Measure Effect Size
The secondary analyses include P-values ranging from 0.0097 to 0.9555. A P-value describes the compatibility of the observed data with a specified null hypothesis under the statistical model; it does not measure the size or clinical importance of the treatment difference.
For example, the NEI VFQ-25 analysis for aflibercept 2mg Q4 reports an estimate of −2.7885 with a 95% confidence interval of −4.9012 to −0.6757 and a P-value of 0.0097. The estimate and confidence interval describe the magnitude and precision of the observed contrast. The P-value addresses the corresponding hypothesis test. These are related but different pieces of information.
16. Non-Inferiority Logic in Detail
Non-inferiority designs require a prespecified margin because "not statistically significant" is not itself evidence of non-inferiority. The margin defines how much worse the investigational treatment could be while still being considered acceptably close to the comparator for the trial's stated purpose.
For VIEW 2, the registry specified a margin of 10 and constructed a two-sided 95% confidence interval for the ranibizumab-minus-aflibercept difference.
This framework also explains why the sign convention matters. A negative value favors aflibercept under the primary endpoint's subtraction order, but the non-inferiority question concerns whether the lower end of the interval is compatible with a loss greater than the allowed margin.
Non-inferiority should also be interpreted with attention to the analysis population. VIEW 2 used the Per-Protocol Set for its primary endpoint, with LOCF. The statistical conclusion is therefore tied to that prespecified analysis strategy rather than being a generic statement that the two therapies have equivalent effects.
17. Multiplicity and the Conditional Testing Sequence
The trial included multiple formal statistical comparisons. The registry specifies a conditional sequence of statistical hypotheses with alpha = 0.05. This is a design feature intended to manage the probability of false-positive conclusions across the prespecified sequence.
| Feature | Registry-supported interpretation |
|---|---|
| Primary endpoint | Percentage maintaining vision at week 52 |
| Primary comparisons | Three aflibercept regimens versus ranibizumab 0.5mg Q4 |
| Primary test framework | Wald / z-test |
| Confidence interval | Two-sided 95% |
| Non-inferiority margin | 10 |
| Alpha | 0.05 |
| Multiplicity procedure | Conditional sequence of statistical hypotheses |
| Registry caveat | The sequence had to stop after testing the first superiority hypothesis; all further statistical tests are exploratory |
This caveat is especially relevant for the secondary analyses. Their numerical P-values remain useful descriptive statistics, but the registry's stated testing sequence limits how those P-values should be interpreted as confirmatory evidence.
18. Analysis Populations and Missing Data
Per-Protocol Set
The primary endpoint analyses used the Per-Protocol Set with LOCF. This focuses the primary non-inferiority comparison on participants included in the protocol-defined analysis population.
Full-Analysis Set
Secondary analyses of BCVA, ≥15-letter gain, NEI VFQ-25, and CNV area used the Full-Analysis Set or the Full-Analysis Set with assessment for the specific outcome.
LOCF
Last Observation Carried Forward was used for missing endpoint observations in the reported analyses. The method imposes assumptions about the relationship between the last observed value and the missing week 52 value.
Interpretive consequence
The same endpoint can produce different estimates under different analysis populations or missing-data strategies. The reported population and imputation method are therefore part of the result, not incidental metadata.
19. What the Primary Risk Difference Does — and Does Not — Mean
A primary risk difference of −1.20 does not mean that the treatment reduces an individual's risk by 1.20%. It means that, under the registry's subtraction convention, the estimated percentage maintaining vision differed by 1.20 percentage points between the two randomized groups.
A 95% CI of −4.86 to 2.46 describes uncertainty around the estimated population-level risk difference under the specified statistical model. It does not describe the distribution of individual patient responses.
The central non-inferiority question is whether the evidence rules out a difference worse than the prespecified margin of 10. This is a different question from asking whether the two treatment groups have exactly equal outcomes.
20. Why This Trial Matters Statistically
VIEW 2 is a useful clinical-trial statistics case because it combines a binary non-inferiority endpoint with multiple treatment regimens, per-protocol analysis, LOCF handling of missing data, ANCOVA for continuous outcomes, categorical-data testing, and a conditional multiplicity strategy.
| Concept | How it appears in VIEW 2 |
|---|---|
| Randomization | Randomized, parallel-group phase 3 design |
| Blinding | Quadruple masking |
| Non-inferiority | Primary endpoint evaluated against a margin of 10 |
| Risk difference | Primary binary effect measure |
| Wald / z-test | Primary categorical-data testing method |
| Per-protocol analysis | Primary endpoint analysis population |
| LOCF | Imputation technique for primary and secondary analyses |
| ANCOVA | Continuous secondary endpoints including BCVA, NEI VFQ-25, and CNV area |
| Covariate adjustment | Baseline value included as an ANCOVA covariate |
| Chi-squared test | Secondary analysis of ≥15-letter vision gain |
| Multiplicity | Conditional sequence with alpha = 0.05 |
| Exploratory testing | Further tests after the conditional sequence stopped are described as exploratory |
21. Limitations
- Non-inferiority interpretation: non-inferiority depends on the prespecified margin and the analysis framework; absence of a statistically significant difference is not by itself evidence of non-inferiority.
- Per-protocol primary analysis: the primary endpoint was analyzed in the Per-Protocol Set, so the result is specifically tied to that population.
- LOCF: carrying the last observation forward requires assumptions about missing week 52 observations and may not represent the participant's unobserved value.
- Multiplicity: the registry states that the conditional hypothesis-testing sequence stopped after the first superiority hypothesis, making subsequent statistical tests exploratory.
- Secondary endpoints: the secondary outcomes use different statistical models and effect measures, so their estimates should not be interpreted as interchangeable measures of the same treatment effect.
- Confidence intervals: intervals describe uncertainty around estimates; they do not quantify the probability that an individual participant will benefit.
- Serious adverse events: the registry provides affected and at-risk counts by arm, but the ClinicalTrials.gov record does not provide comparative effect estimates or P-values for those serious-adverse-event counts.
22. Related Tutorials
Learn more about the methods used in this trial:
23. Related Calculators
24. Sources
- ClinicalTrials.gov: NCT00637377 — VIEW 2.
- Publication: PMID 23084240.
- Publication: PMID 24084500.
- Publication: PMID 36932356.
- Publication: PMID 31826853.
- Publication: PMID 25439429.
Continue through the Clinical Biostats knowledge graph
Connect the trial's endpoints and statistical methods to deeper biostatistics tutorials and practical statistical calculators.
25. Record Summary
VIEW 2 provides a detailed example of how a randomized phase 3 trial can evaluate several dosing regimens against a common comparator using a prespecified non-inferiority framework. The primary endpoint was binary and analyzed using a risk difference, a Wald / z-test, a per-protocol population, LOCF, a two-sided 95% confidence interval, and a non-inferiority margin of 10. Secondary analyses then used ANCOVA and chi-squared methods for continuous and binary outcomes, respectively.
The most informative statistical reading combines the effect estimate, confidence interval, non-inferiority margin, analysis population, missing-data method, and multiplicity procedure. The reported secondary P-values also require attention to the registry's statement that the conditional testing sequence stopped after the first superiority hypothesis, leaving subsequent tests exploratory.