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
FACTS 001 was a randomized phase 3 prevention trial evaluating tenofovir gel versus universal placebo gel in women. The registry identifies two primary endpoints at 30 months: effectiveness, defined through incidence of HIV-1 infection, and safety, defined through grade 2, 3, and 4 clinical and laboratory adverse events.
| Feature | FACTS 001 |
|---|---|
| Trial acronym | FACTS 001 |
| ClinicalTrials.gov identifier | NCT01386294 |
| Phase | Phase 3 |
| Status | Completed |
| Condition | HIV Prevention |
| Enrollment | 2059 |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Prevention |
| Number of arms | 2 |
| Lead sponsor | CONRAD |
| Sponsor type | Other |
| Start date | 2011-10 |
| Primary completion date | 2014-08 |
2. Clinical Question
The central clinical question was whether tenofovir gel could affect the incidence of HIV-1 infection in women over a 30-month period, while also evaluating the safety of the intervention through prespecified clinical and laboratory adverse-event grades.
Population
Women participating in a phase 3 trial for HIV prevention. The registry identifies the condition as HIV Prevention.
Intervention
Tenofovir gel.
Comparator
Universal placebo gel.
Primary questions
What is the incidence of HIV-1 infection over 30 months, and what is the safety profile based on grade 2, 3, and 4 clinical and laboratory adverse events?
3. Trial Design
Tenofovir gel
- Intervention: tenofovir gel
- Randomized allocation
- Parallel-group design
- Quadruple-masked trial
Universal placebo gel
- Comparator: universal placebo gel
- Randomized allocation
- Parallel-group design
- Quadruple-masked trial
The combination of randomized allocation, a parallel design, and quadruple masking is important statistically because the primary comparison is defined prospectively between randomized groups while masking can reduce the potential influence of knowledge of assignment on trial conduct and assessment.
Randomization and masking
The registry describes the allocation as randomized and the masking as quadruple. The registry does not report the randomization ratio, the specific stratification factors, or the identities of the four masked parties in the information presented here.
4. Endpoints
| Primary endpoint | Time frame | Registry definition |
|---|---|---|
| Effectiveness | 30 months | Incidence of HIV-1 infection: HIV incidence will be determine by detection of HIV antibodies using two HIV rapid tests (of which one will be FDA approved) according to algorithm in protocol. One of the rapid tests will detect both HIV-1 and HIV-2; the other will be specific for HIV-1. All endpoints will be reviewed by an expert committee (the Endpoint Adjudication Committee). In carrying out this review, the Committee will use guidelines prepared by the protocol committee for this purpose and recorded in the Manual of Procedures. |
| Safety | 30 months | Grade 2, 3, and 4 clinical and laboratory adverse events as defined by the DAIDS toxicity table. |
Effectiveness endpoint
The effectiveness endpoint is an incidence endpoint: whether HIV-1 infection occurs during the 30-month assessment period. The registry specifies an HIV-antibody detection algorithm using two HIV rapid tests, with one test detecting both HIV-1 and HIV-2 and the other specific for HIV-1.
The registry also specifies review by an Endpoint Adjudication Committee. This is an important design feature because endpoint adjudication creates a structured mechanism for reviewing endpoint determinations using guidelines prepared by the protocol committee and recorded in the Manual of Procedures.
Safety endpoint
The safety endpoint is defined as grade 2, 3, and 4 clinical and laboratory adverse events according to the DAIDS toxicity table. This definition focuses the primary safety assessment on adverse events reaching the specified severity grades rather than on every possible event regardless of grade.
5. Statistical Methodology
The registry identifies what will be measured but does not post a formal statistical analysis plan or statistical results. For the effectiveness endpoint, the natural statistical framework is an analysis of incident HIV-1 infections over follow-up. Depending on the prespecified analysis plan, this type of endpoint can be summarized using cumulative incidence, incidence rates, or time-to-event methods.
For a randomized prevention trial, the key comparison is the difference in HIV-1 infection experience between the tenofovir-gel and placebo-gel groups. The most useful effect measure depends on the analysis specification. A risk ratio compares cumulative risks, an incidence-rate ratio compares event rates over person-time, and a hazard ratio compares instantaneous event rates under a time-to-event model.
An RR below 1 would describe a lower observed cumulative risk in the tenofovir-gel group relative to the placebo-gel group. It would not by itself establish why the difference occurred or describe the experience of every individual participant.
Time-to-event analysis
If participants have different amounts of observed follow-up, time-to-event methods can use the available follow-up while accounting for censoring. A Kaplan-Meier analysis is one standard approach for estimating the probability of remaining free of an event over time, while a Cox model can be used when a hazard-ratio framework is prespecified.
For HIV prevention, a survival-style representation can be interpreted as the probability of remaining free of the defined HIV-1 infection event through time t.
Safety analysis
The primary safety endpoint is based on grade 2, 3, and 4 clinical and laboratory adverse events over 30 months. A typical safety analysis would summarize the number and proportion of participants experiencing at least one event in the specified grades and could additionally describe event counts or exposure-adjusted rates when those quantities are prespecified and available.
The distinction between participants with at least one adverse event and the total number of adverse-event occurrences is statistically important. A participant experiencing several events should not automatically be treated as several independent participants. The appropriate unit of analysis depends on the prespecified safety analysis.
Endpoint adjudication
Because the registry specifies an Endpoint Adjudication Committee for the endpoints, the endpoint classification process is an important part of the statistical chain. The committee's role is to review endpoints using protocol-committee guidelines recorded in the Manual of Procedures.
6. Planned Analysis
The registry identifies two primary endpoints to be assessed over 30 months: incidence of HIV-1 infection for effectiveness and grade 2, 3, and 4 clinical and laboratory adverse events for safety. The registry does not post formal statistical analyses or numerical results for these endpoints.
Effectiveness analysis
HIV-1 infection incidence would typically be compared between randomized groups using an appropriate risk, incidence-rate, or time-to-event framework specified before unblinding. The endpoint definition uses the prespecified rapid-test algorithm and expert endpoint review.
Safety analysis
Grade 2, 3, and 4 clinical and laboratory adverse events would typically be summarized by randomized treatment group, with participant-level event incidence and relevant event characteristics defined according to the safety analysis plan.
7. Statistical Methods Explained
Why is HIV-1 incidence an appropriate endpoint for a prevention trial?
Incidence measures the occurrence of new infections during a defined period. For a prevention study, this directly addresses whether participants develop the infection being prevented during follow-up. The FACTS 001 registry defines the primary effectiveness timeframe as 30 months.
Why might time-to-event analysis be useful?
Participants in longitudinal trials may not all contribute exactly the same amount of observable follow-up. Time-to-event methods use both whether an infection occurred and the time information available before infection or censoring. This can be more informative than treating every participant as if they had identical follow-up.
What does a risk ratio describe?
A risk ratio compares cumulative event probabilities between two groups. An RR of 1 would correspond to equal cumulative risk, while an RR below 1 would indicate lower cumulative risk in the numerator group. The numerical value is a relative measure; it does not provide the absolute risk difference.
What is the difference between incidence and prevalence?
Incidence concerns new events occurring during a period of observation. Prevalence describes the proportion of a population with a condition at a particular point or period. The FACTS 001 primary effectiveness endpoint is incidence of HIV-1 infection, making new infection during follow-up the relevant event.
Why does the safety endpoint use adverse-event grades?
Adverse-event grading provides a standardized way to distinguish different levels of clinical or laboratory severity. The registry specifically defines the primary safety endpoint as grade 2, 3, and 4 clinical and laboratory adverse events according to the DAIDS toxicity table.
Why does endpoint adjudication matter?
Endpoint adjudication provides a structured review process for determining whether an observed event satisfies the protocol's endpoint definition. In FACTS 001, the registry states that all endpoints will be reviewed by an expert committee using guidelines prepared by the protocol committee.
What would a confidence interval add to an effectiveness estimate?
A confidence interval communicates statistical uncertainty around an estimated effect. It is different from the effect estimate itself and should be considered alongside the absolute event experience, the analysis population, follow-up, and prespecified testing framework.
8. Understanding the Effectiveness Endpoint Statistically
The primary effectiveness endpoint is unusual in one important respect: it is not a continuous laboratory measurement or a simple cross-sectional proportion. It concerns the occurrence of new HIV-1 infections during longitudinal follow-up.
| Statistical quantity | What it describes | Interpretive question |
|---|---|---|
| Cumulative risk | Probability of experiencing HIV-1 infection over a defined period | How many participants experienced the event by a specified time? |
| Risk ratio | Relative cumulative risk between randomized groups | How does cumulative risk compare between groups? |
| Risk difference | Absolute difference in cumulative risk | How much does the observed event probability differ between groups? |
| Incidence rate | Events relative to accumulated person-time | How frequently did new infections occur relative to observed time? |
| Hazard ratio | Relative instantaneous event rate under a time-to-event model | How do event rates compare over follow-up under the fitted model? |
These measures are related but not interchangeable. A hazard ratio is not a risk ratio, and neither one directly states the absolute number of infections prevented. A complete statistical interpretation would therefore identify the effect measure used by the prespecified analysis rather than substituting one measure for another.
The most important statistical distinction is between relative and absolute measures. A relative measure can describe how event experience differs between groups, while an absolute measure describes the magnitude of the difference in infection probability or event rate. Both perspectives are useful when evaluating a prevention endpoint.
The uncertainty interval around an effect estimate would describe statistical precision under the chosen model. It would not describe the range of effects experienced by individual women. Likewise, a p-value, if reported, would address evidence against a specified null hypothesis; it would not measure the size or clinical importance of an effect.
9. Understanding the Safety Endpoint Statistically
The safety endpoint includes grade 2, 3, and 4 clinical and laboratory adverse events. This creates several possible statistical summaries, and the distinction between them matters.
| Safety summary | Question answered |
|---|---|
| Participants with at least one qualifying event | What proportion of participants experienced one or more grade 2, 3, or 4 events? |
| Total qualifying events | How many qualifying event occurrences were recorded? |
| Event rate per person-time | How frequently did events occur relative to accumulated observation? |
| Event severity distribution | How were qualifying events distributed across grades? |
For a participant-level safety comparison, the proportion experiencing at least one qualifying event is often especially intuitive. However, recurrent events can contain additional information, and the appropriate handling of multiple events depends on the prespecified safety analysis.
Safety should not be reduced to a single number without identifying the denominator and unit of analysis. "Participants with an event" and "number of events" answer different questions. Exposure time can also matter when follow-up differs between treatment groups.
10. Endpoint Verification and Adjudication
The effectiveness endpoint uses two HIV rapid tests according to an algorithm in the protocol. One rapid test detects both HIV-1 and HIV-2, while the other is specific for HIV-1. The registry states that one of the rapid tests will be FDA approved.
From a statistical perspective, adjudication can reduce ambiguity in endpoint classification by applying a common decision framework. This is particularly important when a trial's primary outcome is defined by diagnostic evidence rather than by a directly observed continuous measurement.
11. What the Registry Reports About Trial Timing
Trial start
The registry lists 2011-10 as the study start date.
Primary endpoint timeframe
Both the effectiveness and safety primary endpoints use a 30-month time frame.
Primary completion
The registry lists 2014-08 as the primary completion date.
Registry status
The ClinicalTrials.gov record identifies the study status as completed.
12. Design Features With Statistical Consequences
| Design feature | Statistical relevance |
|---|---|
| Randomized allocation | Creates the primary basis for comparing outcomes between the two intervention groups while reducing systematic allocation differences in expectation. |
| Parallel design | Participants remain in their assigned study group rather than switching between intervention conditions as part of the design. |
| Quadruple masking | Masking can reduce the influence of treatment knowledge on trial conduct, assessment, and reporting. |
| 30-month endpoint timeframe | Defines the principal period over which the primary effectiveness and safety endpoints are assessed. |
| Endpoint adjudication | Provides an expert review mechanism for endpoint classification under protocol-defined guidelines. |
| Two-arm comparison | Creates a direct randomized comparison between tenofovir gel and universal placebo gel. |
Features not specified in the registry record described here
The registry information does not report a randomization ratio, stratification factors, a non-inferiority margin, a crossover procedure, a factorial design, an interim-analysis plan, a missing-data or imputation strategy, a Bayesian analysis, or a multiplicity-adjustment strategy. These details should not be inferred from the randomized phase 3 designation alone.
13. Interpreting a Potential Effect Estimate
Because no formal statistical estimates are posted in the ClinicalTrials.gov record, the interpretation below illustrates how the principal quantities would be read if they were reported, without assigning any numerical treatment effect to FACTS 001.
An effect estimate summarizes the observed difference between the randomized groups according to a specified statistical measure. For an HIV-incidence endpoint, this might be expressed as a risk ratio, risk difference, incidence-rate ratio, or hazard ratio depending on the prespecified analysis.
A confidence interval describes statistical uncertainty around an estimated effect under the chosen sampling and modeling framework. A narrow interval indicates greater statistical precision than a wide interval, all else equal. It does not describe the distribution of individual treatment effects.
A p-value measures compatibility of the observed data with a specified null hypothesis under the statistical model. It does not measure effect size, clinical importance, or the probability that the null hypothesis is true.
For a prevention endpoint, the absolute difference in cumulative infection risk can be especially informative because it describes the difference in event probability directly. Relative measures and absolute measures answer different questions.
14. Missing Data and Censoring
Longitudinal HIV-prevention studies can encounter participants who contribute incomplete follow-up. A participant who is not observed through the full 30-month period may be censored or otherwise handled according to the prespecified analysis rules.
The statistical consequences depend on why follow-up is incomplete and how the analysis treats the resulting observations. In time-to-event analysis, censoring is informative about follow-up duration even though it does not constitute the infection event itself. If censoring is related to prognosis or treatment experience, assumptions behind a conventional analysis can become important.
The ClinicalTrials.gov record described here does not report a missing-data or imputation strategy. Consequently, the appropriate handling of incomplete follow-up should be taken from the prespecified statistical analysis documentation rather than assumed from the trial's general design description.
15. Safety Analysis: Statistical Considerations
The safety endpoint combines clinical and laboratory adverse events and includes grades 2, 3, and 4. This means the analysis must define both what constitutes a qualifying event and how participants with multiple events are counted.
Participant-level incidence
Counts each participant once for the question of whether at least one qualifying adverse event occurred.
Event-level frequency
Can describe the total number of adverse-event occurrences, but repeated events from one participant require appropriate handling.
Severity
The registry restricts the primary safety definition to grades 2, 3, and 4 under the DAIDS toxicity table.
Follow-up
The primary safety timeframe is 30 months, so the duration of observation is relevant to interpretation.
For safety, statistical significance is not the only consideration. The frequency, severity, clinical consequences, timing, and pattern of adverse events can all be relevant. A participant-level proportion and an event-level count can therefore provide complementary information rather than interchangeable alternatives.
16. Why Randomization Matters
Randomization establishes the treatment assignment mechanism before outcomes are observed. In expectation, it balances both measured and unmeasured baseline characteristics across groups, making the randomized comparison the principal basis for estimating the causal effect of assignment to tenofovir gel versus universal placebo gel.
Randomization does not guarantee identical groups in every observed characteristic, nor does it remove all uncertainty. It establishes the design framework; the statistical analysis then quantifies the observed difference and its uncertainty.
The actual statistical estimator may be expressed as a risk difference, ratio, rate ratio, hazard ratio, or another prespecified measure. The definition of the estimator determines exactly what this comparison means.
17. Why Quadruple Masking Matters
The registry identifies the trial as quadruple-masked. Masking is especially relevant when trial outcomes or trial processes could be influenced by knowledge of treatment assignment.
For an HIV-incidence endpoint based on specified diagnostic testing, the laboratory and adjudication procedures provide an additional structured framework for determining the endpoint. For safety, masking can also reduce the potential influence of treatment knowledge on symptom reporting, clinical assessment, and event attribution.
Masking does not replace randomization or statistical analysis. Instead, it is a design feature intended to reduce avoidable sources of bias during trial conduct and assessment.
18. Limitations
- No posted statistical analyses: the ClinicalTrials.gov record does not post formal statistical analyses or numerical results for the primary endpoints.
- Endpoint-analysis details: the registry identifies the endpoints and their time frame but does not provide the statistical estimator, hypothesis-testing framework, confidence-interval method, or analysis population used for the primary analyses.
- Incomplete design detail: the registry information described here does not specify the randomization ratio or stratification factors.
- Missing-data strategy: the registry does not report how incomplete follow-up or missing observations were handled statistically.
- Multiplicity: the registry information described here does not report a multiplicity-control strategy for the two primary endpoints.
- Interim analysis: no interim-analysis framework is specified in the registry information described here.
- Safety interpretation: the safety endpoint aggregates grade 2, 3, and 4 clinical and laboratory adverse events, so the detailed pattern of individual events would be needed for a fuller safety assessment.
- Endpoint classification: HIV-1 incidence depends on the specified testing algorithm and endpoint-adjudication process, making those procedures integral to the definition of the outcome.
19. Why This Trial Matters Statistically
FACTS 001 provides a useful teaching example because the primary effectiveness endpoint is a longitudinal infection endpoint rather than a conventional continuous or cross-sectional outcome. The trial also illustrates how endpoint definition, diagnostic confirmation, adjudication, randomization, masking, follow-up, and safety grading fit together in a prevention study.
| Concept | How it appears in FACTS 001 |
|---|---|
| Randomization | The trial uses randomized allocation between tenofovir gel and universal placebo gel. |
| Parallel design | The registry identifies a parallel design model with two arms. |
| Masking | The trial is quadruple-masked. |
| Incidence | The primary effectiveness endpoint is incidence of HIV-1 infection over 30 months. |
| Diagnostic endpoint definition | HIV incidence is determined through two HIV rapid tests according to an algorithm in the protocol. |
| Endpoint adjudication | An expert Endpoint Adjudication Committee reviews all endpoints. |
| Safety grading | The primary safety endpoint includes grade 2, 3, and 4 clinical and laboratory adverse events. |
| Longitudinal follow-up | The primary endpoints have a 30-month timeframe. |
| Endpoint-specific analysis | Effectiveness and safety require different statistical summaries because they represent different types of outcomes. |
20. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The randomized comparison defines the principal treatment contrast. The effectiveness endpoint concerns incident HIV-1 infection over 30 months, while the safety endpoint concerns grade 2, 3, and 4 clinical and laboratory adverse events. The appropriate effect measure and inferential procedure must be tied to the prespecified statistical analysis.
Clinical interpretation
The clinical meaning of the study depends on both prevention outcomes and safety. Relative measures alone are insufficient to describe the practical magnitude of an HIV-prevention effect, and safety requires consideration of the frequency and severity of qualifying adverse events.
21. Reading the Primary Endpoint as a Time-Based Outcome
The 30-month timeframe is not merely a reporting date. It defines the period over which the primary effectiveness and safety endpoints are framed.
For HIV-1 infection, a participant who remains free of infection throughout observation contributes information about remaining infection-free. A participant who develops HIV-1 infection contributes an event at the relevant time. A participant whose follow-up ends before 30 months without an event creates a censoring or incomplete-follow-up issue that must be addressed by the statistical analysis plan.
The statistical analysis then compares the distributions of T between the randomized treatment groups using the prespecified method.
This framework explains why a simple end-of-study proportion can sometimes discard information. If one participant experiences infection early and another experiences infection late, a purely binary analysis at 30 months treats them identically with respect to the event indicator, while a time-to-event analysis retains their different event times.
22. What Is Known and What Remains Unspecified
| Element | Registry record |
|---|---|
| Trial phase | Phase 3 |
| Enrollment | 2059 |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary effectiveness endpoint | Incidence of HIV-1 infection over 30 months |
| Primary safety endpoint | Grade 2, 3, and 4 clinical and laboratory adverse events over 30 months |
| Endpoint adjudication | Expert Endpoint Adjudication Committee |
| Statistical estimates | Not posted in the ClinicalTrials.gov record described here |
| Formal statistical analysis | Not posted in the ClinicalTrials.gov record described here |
This distinction is important for statistical reading. A registry record can provide a detailed definition of what a trial intended to measure without providing the numerical analysis needed to estimate the treatment effect. The design information remains useful even when the statistical results are not available in the registry record.
23. Sources
- ClinicalTrials.gov: NCT01386294 — FACTS 001.
- PubMed: PubMed record.
Continue through the Clinical Biostats clinical trial library
Explore additional clinical trial analyses and statistical methods across randomized studies, endpoints, and study designs.
24. Record Summary
FACTS 001 is a randomized phase 3, parallel-group, quadruple-masked prevention trial with 2059 participants and two intervention groups: tenofovir gel and universal placebo gel. Its two primary endpoints are defined over 30 months: incidence of HIV-1 infection for effectiveness and grade 2, 3, and 4 clinical and laboratory adverse events for safety.
Statistically, the trial illustrates the importance of matching the analysis to the endpoint. HIV-1 incidence is a longitudinal event outcome for which cumulative risk, incidence rates, or time-to-event methods can provide different but complementary descriptions. Safety is defined through graded clinical and laboratory adverse events and requires clear distinction between participant-level incidence and event-level frequency.
The registry also places substantial emphasis on endpoint ascertainment: HIV-1 incidence is determined using two rapid tests according to a protocol-defined algorithm, and all endpoints are reviewed by an expert Endpoint Adjudication Committee. These features form part of the statistical foundation because the validity of an analysis depends on consistent and prespecified outcome classification.