This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical results on this page are restricted to the trial data reported in that registry record.
1. Trial at a Glance
VIKING-3 was a phase 3, single-group trial evaluating dolutegravir in HIV-infected subjects with treatment failure on an integrase inhibitor-containing regimen. The registry reports 183 enrolled participants, one intervention arm, and three posted statistical analyses for primary endpoints.
| Feature | VIKING-3 |
|---|---|
| Trial name | VIKING-3 |
| NCT ID | NCT01328041 |
| Phase | Phase 3 |
| Condition | Infection, Human Immunodeficiency Virus |
| Population | HIV-infected subjects with treatment failure on an integrase inhibitor-containing regimen |
| Design model | Single-group |
| Allocation | NA |
| Masking | None |
| Primary purpose | Treatment |
| Intervention | Dolutegravir |
| Enrollment | 183 |
| Lead sponsor | ViiV Healthcare |
| Sponsor type | Industry |
| Status | Completed |
2. Clinical Question
The clinical question was whether dolutegravir could produce a measurable change in plasma HIV-1 RNA shortly after treatment initiation and whether participants could achieve HIV-1 RNA less than 50 copies/mL at Week 24 and Week 48.
Population
HIV-infected subjects with treatment failure on an integrase inhibitor-containing regimen.
Intervention
Dolutegravir. The serious-adverse-event summary identifies the study arm as dolutegravir 50 mg BID.
Comparator
There was no separate comparator arm. The registry classifies the design model as single-group and allocation as NA.
Primary questions
What was the mean change in plasma HIV-1 RNA at Day 8, and what percentage of participants had HIV-1 RNA less than 50 copies/mL at Weeks 24 and 48?
3. Trial Design
No randomized comparator arm was reported.
The registry identifies the study as unmasked.
The registry does not specify an allocation ratio because this was a single-group design.
The trial was registered with treatment as its primary purpose.
The statistical structure is fundamentally different from a two-arm randomized superiority trial. Because VIKING-3 has one study arm, the Week 24 and Week 48 estimates describe the observed response proportion within the treated population rather than a treatment-versus-control contrast.
4. Endpoints
The registry lists seven primary endpoints, including the three endpoints for which formal statistical analyses are posted. The endpoint definitions and registered time frames are reproduced below in abbreviated form while preserving their statistical meaning.
| Registered primary endpoint | Time frame | Endpoint type |
|---|---|---|
| Mean Change From Baseline in Plasma HIV-1 RNA at Day 8 | Baseline and Day 8 | Other / unclear; analyzed as continuous |
| Number of Participants With HIV-1 RNA Less Than 50 Copies/mL at Week 24 | Week 24 | Binary |
| Number of Participants With HIV-1 RNA Less Than 50 Copies/mL at Week 48 | Week 48 | Binary |
| Number of Participants With Any Adverse Event (AE) or Any Serious Adverse Event (SAE) | From the day of the first dose of study drug until end of treatment visit for each participant, up to Week 180 (median of 758 days)) | Binary |
| Number of Participants With Adverse Events of the Indicated Severity, Per the Division of Acquired Immune Deficiency Syndrome (DAIDS) Grading Scale | From the day of the first dose of study drug until end of treatment visit for each participant, up to Week 180 (median of 758 days)) | Binary |
| Number of Participants With the Maximum Post-Baseline-emergent Clinical Chemistry Toxicities of the Indicated Grade | From the day of the first dose of study drug until end of treatment visit for each participant, up to Week 180 (median of 758 days)) | Binary |
| Number of Participants With the Maximum Post-Baseline-emergent Hematology Toxicities of the Indicated Grade | From the day of the first dose of study drug until end of treatment visit for each participant, up to Week 180 (median of 758 days)) | Binary |
Day 8 endpoint definition
Mean change from baseline in plasma HIV-1 RNA at Day 8 was calculated as the Day 8 value minus the baseline value. The registry states that the last observation was carried forward if a participant missed the Day 8 visit, and the baseline observation was carried forward if a participant discontinued treatment before Day 8.
Week 24 and Week 48 endpoint definition
The Week 24 and Week 48 endpoints were based on the Food and Drug Administration's Snapshot algorithm. The registry states that participants without HIV-1 RNA data at the visit of interest because of missing data or discontinuation of investigational product before the visit window were treated as nonresponders. Participants who switched their treatment were also handled according to the Snapshot framework.
5. Analysis Populations and Missing Data
The Day 8 statistical analysis used the Intent-to-Treat-Exposed (ITT-E) Population, defined as all participants who received at least one dose of study drug. The analysis description adds that only participants who had Day 8 observations were considered for the analysis.
| Analysis | Population | Important handling rule |
|---|---|---|
| Day 8 mean change | ITT-E | Only participants with Day 8 observations were considered; the endpoint definition specifies last-observation-carried-forward rules for missed Day 8 visits and baseline carry-forward after discontinuation before Day 8. |
| Week 24 HIV-1 RNA <50 copies/mL | ITT-E | FDA Snapshot algorithm; participants without required HIV-1 RNA data or who discontinued before the visit window were treated as nonresponders. |
| Week 48 HIV-1 RNA <50 copies/mL | ITT-E | FDA Snapshot algorithm; participants without required HIV-1 RNA data or who discontinued before the visit window were treated as nonresponders. |
This distinction illustrates an important statistical point: an analysis population and a missing-data rule answer different questions. The ITT-E definition specifies who enters the analysis, while the endpoint-specific rules determine how missing or unavailable measurements affect the outcome classification.
6. Statistical Methodology
Day 8 t-test
The registry reports a 2-sided t-test for the mean change from baseline in plasma HIV-1 RNA at Day 8. The reported outcome unit is log10 copies/milliliter (mL).
95% CI: -1.520 to -1.343 · P < 0.001
Because VIKING-3 has a single study arm, the t-test is not a treatment-versus-control comparison. The registry reports the method and the resulting mean change, but does not state the reference value or null hypothesis in the ClinicalTrials.gov record. Therefore, the safest interpretation is that the analysis evaluates the reported mean change within the single treated group rather than estimating a between-arm treatment difference.
Confidence intervals
All three posted primary statistical analyses include 95% two-sided confidence intervals. For the Day 8 continuous endpoint, the interval quantifies uncertainty around the estimated mean change. For the Week 24 and Week 48 binary endpoints, the interval quantifies uncertainty around the reported percentage of participants meeting the HIV-1 RNA threshold.
A 95% confidence interval is a statement about the statistical procedure and repeated-sampling framework; it is not a probability statement that the fixed true parameter has a 95% chance of being inside this particular interval.
Intention-to-treat analysis
The Day 8 analysis explicitly identifies the ITT-E population and the Week 24 and Week 48 analyses identify the ITT-E population as well. In a single-group trial, this preserves the prespecified study population rather than creating a treatment-versus-control comparison.
Binary responder analysis
The Week 24 and Week 48 endpoints are binary responder outcomes: each participant is classified according to whether HIV-1 RNA was less than 50 copies/mL at the specified visit under the Snapshot algorithm. The registry reports the resulting percentage and confidence interval but does not report a formal statistical-analysis method for these two endpoints.
7. Results: Mean Change From Baseline in Plasma HIV-1 RNA at Day 8
The first posted primary analysis evaluates the mean change from baseline in plasma HIV-1 RNA at Day 8. The outcome was expressed in log10 copies/milliliter (mL) and was analyzed in the ITT-E population.
Mean change from baseline
95% CI: -1.520 to -1.343 · P < 0.001
Outcome unit: log10 copies/milliliter (mL)
| Feature | Reported result |
|---|---|
| Endpoint | Mean Change From Baseline in Plasma HIV-1 RNA at Day 8 |
| Time frame | Baseline and Day 8 |
| Analysis population | Intent-to-Treat-Exposed (ITT-E) |
| Group | Dolutegravir 50 mg BID |
| Method | t-test, 2 sided |
| Estimate | -1.432 |
| 95% CI | -1.520 to -1.343 |
| P-value | <0.001 |
The reported estimate of -1.432 log10 copies/mL means that the average change from baseline in the analyzed dolutegravir group was negative 1.432 on the log10 HIV-1 RNA scale. The negative sign therefore indicates a decrease from baseline on that scale.
The estimate does not represent a percentage reduction, a hazard ratio, or a treatment effect relative to a control group. Because the trial has a single study arm, it cannot by itself establish how the observed change compares with an untreated or alternative-treatment population.
The 95% CI from -1.520 to -1.343 describes statistical uncertainty around the estimated mean change. Its relatively narrow span provides a more informative description of precision than the point estimate alone.
The P < 0.001 result is evidence against the null hypothesis used by the reported two-sided t-test, but the p-value does not measure the magnitude of the observed change. Effect size and statistical evidence are separate concepts.
The interpretation also depends on the registered Day 8 missing-data rules and the restriction to participants with Day 8 observations in the stated analysis population. Those rules should be considered when generalizing the estimate to all 183 enrolled participants.
8. Results: HIV-1 RNA Less Than 50 Copies/mL at Week 24
The second posted primary analysis reports the percentage of participants with HIV-1 RNA less than 50 copies/mL at Week 24 using the FDA Snapshot algorithm.
Week 24 response
95% CI: 62% to 76%
Analysis population: ITT-E
| Feature | Reported result |
|---|---|
| Endpoint | Number of Participants With HIV-1 RNA Less Than 50 Copies/mL at Week 24 |
| Time frame | Week 24 |
| Analysis population | ITT-E Population |
| Effect measure | Percentage of participants |
| Estimate | 69 |
| 95% CI | 62 to 76 |
| Formal method | Not reported |
The reported estimate means that 69% of participants met the registered HIV-1 RNA threshold of less than 50 copies/mL at Week 24 under the FDA Snapshot algorithm.
This is a within-group response proportion. It is not a relative risk, risk difference, odds ratio, or comparison against a control treatment because no comparator arm was registered.
The 95% CI of 62% to 76% describes statistical uncertainty around the estimated percentage. It does not mean that individual participants have response probabilities confined to this interval, nor does it describe the range of responses among individual patients.
No formal analysis method is reported for this endpoint in the registry statistical-analysis record. Therefore, the confidence interval should be interpreted as the registry-reported inferential interval without assigning an unreported statistical procedure to it.
The Snapshot algorithm is especially important: participants lacking the required HIV-1 RNA data because of missing data or discontinuation before the relevant visit window were classified as nonresponders. Thus, the 69% figure is not equivalent to the percentage of evaluable laboratory measurements below 50 copies/mL.
9. Results: HIV-1 RNA Less Than 50 Copies/mL at Week 48
The third posted primary analysis reports the percentage of participants with HIV-1 RNA less than 50 copies/mL at Week 48 using the same registered FDA Snapshot framework.
Week 48 response
95% CI: 56% to 70%
Analysis population: ITT-E
| Feature | Reported result |
|---|---|
| Endpoint | Number of Participants With HIV-1 RNA Less Than 50 Copies/mL at Week 48 |
| Time frame | Week 48 |
| Analysis population | ITT-E Population |
| Effect measure | Percentage of participants |
| Estimate | 63 |
| 95% CI | 56 to 70 |
| Formal method | Not reported |
The reported estimate means that 63% of participants met the HIV-1 RNA threshold of less than 50 copies/mL at Week 48 under the registered Snapshot algorithm.
Again, this is a descriptive response proportion within the single dolutegravir study arm. It does not establish a treatment difference because VIKING-3 does not contain a separate randomized comparator group.
The 95% CI of 56% to 70% provides an interval measure of precision around the reported percentage. A confidence interval is not a range of outcomes that individual participants should be expected to experience.
No formal inferential method is reported for this endpoint in the ClinicalTrials.gov record. Consequently, this page does not infer an unreported hypothesis test or confidence-interval construction method.
The Week 48 result also should not be interpreted independently of the endpoint's Snapshot definition. Missing or otherwise unavailable Week 48 HIV-1 RNA data under the specified conditions are handled as nonresponse rather than simply removed from the denominator.
10. Comparing the Three Primary Results
The three posted primary analyses address different statistical quantities and therefore should not be collapsed into a single measure of treatment effect.
| Endpoint | Estimate | 95% CI | P-value | Method |
|---|---|---|---|---|
| Mean change from baseline in plasma HIV-1 RNA at Day 8 | -1.432 log10 copies/mL | -1.520 to -1.343 | <0.001 | 2-sided t-test |
| HIV-1 RNA less than 50 copies/mL at Week 24 | 69% | 62% to 76% | Not reported | Not reported |
| HIV-1 RNA less than 50 copies/mL at Week 48 | 63% | 56% to 70% | Not reported | Not reported |
The Day 8 endpoint is a continuous change score, whereas the Week 24 and Week 48 endpoints are binary responder outcomes. The Day 8 analysis therefore summarizes a mean change, while the later endpoints summarize the proportion classified as responders according to the registered Snapshot algorithm.
11. Safety Results
Safety was included among the registered primary endpoints. The trial data provide a serious-adverse-event summary for the dolutegravir 50 mg BID arm: 46/183.
Serious adverse events
Dolutegravir 50 mg BID
| Safety measure | Dolutegravir 50 mg BID |
|---|---|
| Participants with serious adverse events | 46/183 |
| Study design context | Single-group |
| Comparator arm | None reported |
The registry also lists primary safety endpoints covering any AE or SAE, AE severity according to the DAIDS grading scale, clinical chemistry toxicities, and hematology toxicities. The ClinicalTrials.gov record does not provide formal comparative analyses for these safety endpoints.
12. Statistical Methods Explained
Why was a t-test used for the Day 8 endpoint?
The Day 8 outcome is a continuous mean change in plasma HIV-1 RNA, expressed in log10 copies/milliliter. The registry specifically reports a two-sided t-test. In a single-group setting, a t-test can be used to evaluate a sample mean against a specified reference value.
What does a mean change of -1.432 mean?
It means the estimated average change from baseline was negative 1.432 log10 copies/mL. The sign communicates direction on the logarithmic measurement scale. It is not itself a percentage, a hazard ratio, or a relative risk.
What does the 95% CI of -1.520 to -1.343 tell us?
It describes the uncertainty around the estimated mean change using the reported statistical framework. The interval is narrower than the set of all possible individual participant changes because it concerns the estimated population mean rather than individual variability.
Why does P < 0.001 not tell us how large the effect is?
A p-value measures evidence against a null hypothesis under the specified statistical model. It depends on both the magnitude of the observed result and its statistical uncertainty. Consequently, a small p-value does not provide a standardized measure of clinical magnitude. The estimate and its confidence interval are needed to understand the size and precision of the result.
What does the 69% Week 24 estimate represent?
It represents the percentage of participants with HIV-1 RNA less than 50 copies/mL at Week 24 under the FDA Snapshot algorithm. It is a response proportion, not a relative treatment effect, because there is only one study arm.
Why is the Snapshot algorithm important?
The endpoint definition determines who is classified as a responder. Under the registered rules, participants without the required HIV-1 RNA data because of missing data or discontinuation before the visit window are treated as nonresponders, as are participants handled as nonresponders because of treatment switching under the algorithm. Therefore, the analysis incorporates more than observed laboratory values among completers.
Why can the Week 24 and Week 48 percentages not simply be compared as a hypothesis test?
The registry reports estimates and confidence intervals for each time point, but the ClinicalTrials.gov record does not report a formal method for comparing the two percentages. The difference between two descriptive estimates should not be converted into a claim of statistically significant change without an appropriate paired or longitudinal analysis and its corresponding assumptions.
13. Understanding the Single-Group Design
One of the most important statistical features of VIKING-3 is its single-group design. In a randomized controlled trial, the central estimand is usually a contrast between treatment groups. Here, the registry instead reports outcomes within the dolutegravir group.
What the design can estimate
Mean change in HIV-1 RNA and the percentage of participants meeting the prespecified HIV-1 RNA threshold at specified time points.
What the design does not estimate
A randomized treatment-versus-control difference, relative risk versus a comparator, or hazard ratio between treatment groups.
Why this matters
Without a concurrent comparator, changes over time cannot be separated as cleanly from secular trends, background treatment effects, or other changes occurring during follow-up.
How to read the reported superiority label
The registry labels the posted analyses as superiority analyses, but the design model remains single-group. The label should therefore not be interpreted as evidence of superiority over an unreported comparator.
14. Missing Data and Carry-Forward Rules
VIKING-3 provides an unusually useful teaching example for distinguishing different approaches to missing data because the registered endpoint definitions explicitly describe different rules.
| Endpoint | Registered missing-data approach |
|---|---|
| Day 8 mean change | Last observation was carried forward if a participant missed the Day 8 visit; baseline was carried forward if treatment was discontinued before Day 8. |
| Week 24 responder status | FDA Snapshot algorithm; participants without HIV-1 RNA data at the visit of interest due to specified missingness/discontinuation conditions were treated as nonresponders. |
| Week 48 responder status | FDA Snapshot algorithm; participants without HIV-1 RNA data at the visit of interest due to specified missingness/discontinuation conditions were treated as nonresponders. |
These approaches have different statistical implications. Carrying forward an earlier observation substitutes an observed value for an unobserved later measurement, whereas Snapshot classification converts the absence of the required outcome into a responder/nonresponder classification. Neither should be described simply as "ignoring missing data."
15. Confidence Intervals Across the Primary Endpoints
The three primary analyses illustrate how confidence intervals apply to different types of estimates.
| Endpoint | Estimate | 95% CI | What the interval describes |
|---|---|---|---|
| Day 8 mean change | -1.432 | -1.520 to -1.343 | Uncertainty around the estimated mean change in log10 HIV-1 RNA. |
| Week 24 response | 69% | 62% to 76% | Uncertainty around the estimated response percentage. |
| Week 48 response | 63% | 56% to 70% | Uncertainty around the estimated response percentage. |
The intervals should not be compared solely by their amount of overlap or width. They arise from different endpoints and different statistical quantities. A formal comparison between time points or between parameters would require its own analysis.
16. P-values and Statistical Evidence
Only the Day 8 mean-change analysis has a p-value in the ClinicalTrials.gov record: P < 0.001. The Week 24 and Week 48 analyses report estimates and 95% confidence intervals but no p-values.
What the p-value addresses
For the Day 8 analysis, the reported two-sided t-test supplies a measure of evidence relative to its unspecified null/reference value.
What it does not address
It does not quantify clinical importance, establish a treatment-versus-control difference, or describe the probability that the null hypothesis is true.
The absence of a reported p-value for Week 24 and Week 48 is not evidence that those endpoints were statistically significant or nonsignificant. It simply means the ClinicalTrials.gov record does not provide a formal p-value for those estimates.
17. Multiplicity, Interim Analysis, and Other Design Features
The ClinicalTrials.gov record identifies seven registered primary endpoints and three posted primary- endpoint statistical analyses. They do not provide a multiplicity-adjustment procedure, interim-analysis boundary, alpha-spending plan, non-inferiority margin, crossover scheme, factorial structure, stratification factors, or Bayesian analysis.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Multiplicity | Seven registered primary endpoints are listed; no multiplicity-adjustment method is reported in the ClinicalTrials.gov record. |
| Interim analysis | No interim-analysis procedure is reported in the ClinicalTrials.gov record. |
| Non-inferiority margin | Not reported; the registry classifies the posted analyses as superiority. |
| Crossover | No crossover design is reported. |
| Factorial design | No factorial structure is reported. |
| Stratification | No stratification factors are reported in the ClinicalTrials.gov record. |
| Bayesian methods | No Bayesian method is reported. |
These omissions are important because statistical interpretation should follow the documented analysis rather than fill gaps with assumptions. In particular, the presence of multiple registered primary endpoints does not by itself establish whether or how the type I error was controlled.
18. Statistical Analysis Summary
| Analysis | Population | Method | Estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| Day 8 mean change | ITT-E; participants with Day 8 observations | 2-sided t-test | -1.432 | -1.520 to -1.343 | <0.001 |
| Week 24 HIV-1 RNA <50 copies/mL | ITT-E | Not reported | 69% | 62% to 76% | Not reported |
| Week 48 HIV-1 RNA <50 copies/mL | ITT-E | Not reported | 63% | 56% to 70% | Not reported |
19. Why the Endpoint Scale Matters
The Day 8 and later virologic endpoints use different scales. Day 8 measures a continuous change on the log10 scale, while Weeks 24 and 48 classify participants using a threshold of less than 50 copies/mL.
Continuous endpoint
The Day 8 analysis preserves quantitative information about the magnitude of change in HIV-1 RNA and summarizes it with a mean.
Binary endpoint
The Week 24 and Week 48 analyses convert viral-load information into responder versus nonresponder categories using a clinically specified threshold.
These endpoints answer complementary questions. A continuous change endpoint asks how much the measured outcome changed on average, while a binary endpoint asks how many participants crossed a prespecified threshold by a specified time.
20. Longitudinal Interpretation Without Overreach
The reported Day 8, Week 24, and Week 48 estimates can be read as three snapshots of the trial's prespecified outcomes, but they should not be treated as if they were one repeated-measures model.
Mean virologic change
The reported mean change was -1.432 log10 copies/mL, with a 95% CI of -1.520 to -1.343 and P < 0.001 using a 2-sided t-test.
Threshold-based response
69% of participants were reported as having HIV-1 RNA less than 50 copies/mL, with a 95% CI of 62% to 76%.
Threshold-based response
63% of participants were reported as having HIV-1 RNA less than 50 copies/mL, with a 95% CI of 56% to 70%.
The change in the reported percentages from Week 24 to Week 48 is descriptive. The ClinicalTrials.gov record does not provide a formal longitudinal model, paired comparison, or hypothesis test for the difference between those two time points.
21. Limitations
- Single-group design: without a concurrent comparator arm, the reported outcomes do not estimate a randomized treatment-versus-control effect.
- No formal method reported for Week 24 and Week 48: the registry provides percentages and confidence intervals but does not identify the statistical procedure used to produce those analyses in the ClinicalTrials.gov record.
- Endpoint-specific missing-data rules: Day 8 and later virologic endpoints use different handling rules, which affects interpretation of the resulting estimates.
- Day 8 analysis population: although the analysis is identified as ITT-E, only participants with Day 8 observations were considered for the reported analysis, according to the registry-reported analysis description.
- Carry-forward assumptions: the Day 8 endpoint uses last-observation-carried-forward and baseline-carried-forward rules in specified circumstances. Such rules rely on assumptions about unobserved measurements.
- Multiplicity: seven primary endpoints are registered, while the ClinicalTrials.gov record does not identify a multiplicity-adjustment strategy.
- Limited safety analysis information: the ClinicalTrials.gov record provides 46/183 serious adverse events in the dolutegravir 50 mg BID arm but do not provide a comparator-based safety analysis.
- Restricted source data: this analysis intentionally does not add numerical results, baseline characteristics, subgroup analyses, or other findings that are absent from the ClinicalTrials.gov record.
22. Why This Trial Matters Statistically
VIKING-3 is a useful statistical teaching case because it demonstrates how much the interpretation of a result depends on study architecture and endpoint definition. The registry contains a continuous endpoint analyzed with a t-test, binary virologic endpoints analyzed under a Snapshot framework, explicit missing-data rules, and a single-group design.
| Concept | How it appears in VIKING-3 |
|---|---|
| Single-group analysis | One study arm with no separate randomized comparator. |
| Continuous endpoint | Mean change from baseline in plasma HIV-1 RNA at Day 8. |
| Logarithmic measurement scale | Day 8 outcome reported in log10 copies/milliliter. |
| t-test | 2-sided t-test reported for the Day 8 mean change. |
| Confidence interval | 95% two-sided intervals reported for all three posted primary analyses. |
| P-value | P < 0.001 reported for the Day 8 analysis. |
| Binary endpoint | HIV-1 RNA less than 50 copies/mL at Weeks 24 and 48. |
| Snapshot algorithm | Missing/discontinued participants are incorporated into responder classification according to the registered rules. |
| ITT-E | Primary analyses identify the Intent-to-Treat-Exposed population. |
| Missing-data handling | Carry-forward rules are specified for the Day 8 endpoint. |
| Safety endpoint | Serious adverse events reported as 46/183 in the dolutegravir 50 mg BID arm. |
23. Statistical Methods vs Clinical Conclusions
What the statistics show
The registry reports a mean Day 8 HIV-1 RNA change of -1.432 log10 copies/mL, a Week 24 response estimate of 69%, and a Week 48 response estimate of 63%, each with a 95% confidence interval.
What they do not show
They do not provide a randomized comparison with another treatment, because VIKING-3 is a single-group study.
This distinction is central to interpreting single-arm clinical trials. A large observed response proportion can be clinically informative, but causal attribution to the intervention is stronger when a concurrent comparator establishes what would have happened under an alternative treatment or control condition.
24. Related Tutorials
Learn more about the methods used in this trial:
25. Related Statistical Calculators
26. Sources
- ClinicalTrials.gov: VIKING-3, NCT01328041.
- PubMed: Publication indexed under PMID 24446523.
- PubMed: Publication indexed under PMID 42703616.
Continue through the Clinical Biostats statistical library
Explore tutorials and calculators for confidence intervals, intention-to-treat analysis, p-values, and t-tests—the core statistical concepts illustrated by VIKING-3.
27. Record Summary
VIKING-3 illustrates how to interpret clinical-trial results when the design is single-group rather than randomized and comparative. The registry reports three primary statistical analyses: a Day 8 mean change in plasma HIV-1 RNA of -1.432 log10 copies/mL with a 95% CI of -1.520 to -1.343 and P < 0.001; a Week 24 HIV-1 RNA response estimate of 69% with a 95% CI of 62% to 76%; and a Week 48 response estimate of 63% with a 95% CI of 56% to 70%. The Day 8 analysis used a reported 2-sided t-test, while the formal methods for the Week 24 and Week 48 estimates were not reported in the ClinicalTrials.gov record.
The most important statistical lesson is that an estimate must always be interpreted in the context of its design, endpoint definition, analysis population, missing-data rules, and comparator structure. VIKING-3's single-group design means its reported percentages and mean changes describe outcomes within the treated cohort rather than randomized treatment effects.