This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Online Social Networking for HIV Prevention (NCT00725959), a randomized na study. The primary outcome is Proportion of Condom Usage.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT00725959 |
| Phase | NA |
| Condition | HIV |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 1578 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Colorado School of Public Health |
| Results posted | Yes |
2. Clinical Question
Population
The trial concerns people recruited through Facebook social networks for an HIV-prevention study. The registry lists HIV as the condition and records enrollment of 1578. Participant characteristics, eligibility details and the population represented by the posted endpoint estimates are not reported in the registry record.
Intervention
The intervention was online social networking for HIV prevention. Participants in the Facebook arm were to encounter dynamic, interactive and regularly updated prevention messages. The brief description says intervention content would be developed with input from the target audience. The record does not describe the precise messages or how consistently participants encountered them.
Comparator
The control arm received static information about HIV prevention that was already available online. The trial description says social networks were randomly assigned to intervention or control status, with the network as the level of assignment. The registry does not report how exposure to either type of message was measured.
Primary question
The primary outcome was the proportion of condom use, defined as the number of times a condom was used divided by the number of times a person had sex in the past 60 days. The statistical question is whether the posted arm-level estimates support a clear comparison. The registry reports means and confidence intervals, but no inferential test or arm comparison.
3. Trial Design
This was a randomized, parallel-group behavioral trial. The registry describes random assignment at the level of Facebook social networks: each network was assigned to control or intervention status. That cluster-level assignment matters when interpreting participant-level measurements, because members of one network may share experiences or influence one another. The record does not explain whether the posted analyses accounted for such within-network similarity.
The intervention involved dynamic, interactive HIV-prevention messages on Facebook, while the control involved static online HIV-prevention information. The record identifies the study as unmasked. In a behavioral messaging study, participants would generally know what kind of information they encountered; the registry does not describe any masking procedures for outcome assessment or analysis.
The registry records enrollment of 1578 and lists the study as completed. The study start is recorded as 2008-09 and primary completion as 2010-12. The phase is listed as NA, and the sponsor is Colorado School of Public Health. These details describe the trial record but do not establish how many participants contributed data to each analysis.
The brief description refers to baseline and 3 and 6 month assessments of condom use and abstinence. The posted primary outcome, however, specifies a timeframe of two months and defines the measure using sexual activity in the past 60 days. The registry does not identify which assessment occasion corresponds to the arm-level values. Nor does it report the number of networks assigned, network sizes, deviations from the assignment plan, or participant flow by arm.
4. Participant Flow and Accounting
The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.
Overall Study
| Stage | Intervention Arm | Control Arm |
|---|---|---|
| STARTED | 942 | 636 |
| COMPLETED | 653 | 439 |
| NOT COMPLETED | 289 | 197 |
5. Baseline Characteristics
The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.
| Characteristic | Unit | Arm 1 | Arm 2 | Total |
|---|---|---|---|---|
| Participants analysed | 942 | 636 | 1578 | |
| Age, Continuous | years | 19.7 (SD 2.0) | 20.2 (SD 2.4) | 19.8 (SD 2.1) |
| Sex: Female, Male: Female | Participants | 521 | 391 | 912 |
| Sex: Female, Male: Male | Participants | 421 | 245 | 666 |
6. Primary Endpoint
The registry lists the primary outcome as Proportion of Condom Usage, measured over two months.
| Element | Registry entry |
|---|---|
| Result type | Mean |
| Measure of spread | 95% Confidence Interval |
| Unit | proportion of condom use |
| Endpoint classification | Continuous Endpoint |
The primary endpoint is titled Proportion of Condom Usage. It is a continuous measure expressed as a proportion: for each person, condom-use occasions are divided by occasions of sex during the past 60 days. A value closer to one represents a higher share of reported sexual occasions involving condom use; a value closer to zero represents a lower share. This interpretation concerns the measure, not a clinical conclusion about either arm.
Using a proportion places condom use in relation to sexual activity rather than treating the number of condom uses alone as the outcome. That can make the measure more informative when participants report different numbers of sexual occasions. At the same time, its meaning depends on the underlying reports and on how participants with no sexual activity or incomplete responses were handled. Those rules are not reported in the registry record.
The record gives the endpoint timeframe as two months and describes the denominator as sexual occasions in the past 60 days. The brief study description separately mentions baseline and 3 and 6 month assessments. The registry does not specify which assessment generated the posted means. The measure therefore has a clear general definition, but its exact assessment timing and handling of unavailable or undefined proportions remain unclear.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The posted results give a mean proportion of 0.68 for the control arm, with a 95% confidence interval from 0.56 to 0.74. The other arm has a mean of 0.56, with a 95% confidence interval from 0.48 to 0.61. The registry identifies OG000 as the control arm; its posted outcome count is 653. The other arm's posted outcome count is 439. These counts describe the posted endpoint entries and should not be read as the total enrollment or as proof of why the groups have different counts.
The arm estimates describe the average reported condom-use proportion in the data represented by each result. They are not percentages of participants who always used condoms: the outcome is a proportion of sexual occasions, averaged as a continuous measure. The confidence intervals communicate uncertainty around the reported arm means. They do not describe the range of individual participants' values, and they do not by themselves establish whether the arms differ.
No arm-level contrast, p-value, adjusted estimate or inferential analysis is posted. The record lists zero analyses. Consequently, the estimates should be read descriptively rather than as evidence of a statistically tested difference. It would be inappropriate to infer statistical significance from the two means or their confidence intervals alone. No unadjusted recalculation is needed to describe the arm results, and the record does not identify a sponsor analysis that can be compared with a recalculation.
There is a limitation in linking these values to the study timeline. The endpoint says two months and refers to the past 60 days, while the study summary describes baseline and 3 and 6 month assessments. The record does not say which measurement occasion produced the posted values. Without that detail, the temporal interpretation of the means is uncertain. The summary's mention of abstinence also does not make abstinence a posted primary result; the displayed primary outcome concerns condom-use proportion.
These data alone do not establish whether the intervention changed behavior, whether any difference was caused by message exposure, or whether an observed difference would be meaningful to participants. The record does not provide a comparison estimate, an analysis plan, or sufficient detail about missing outcome data and network clustering to resolve those questions. The defensible reading is limited to the reported arm means, confidence intervals and outcome counts, together with the absence of a posted statistical comparison.
9. Statistical Methods Explained
The primary outcome is a continuous measure because it is expressed as a proportion for each participant and the registry identifies the posted parameter as a mean. A mean summarizes the center of the recorded values across participants contributing to an arm's result. It does not mean that every participant had that value, nor does it show how widely individual responses varied. The registry does not post a separate measure of participant-level spread.
A confidence interval describes uncertainty in an estimate under the assumptions of the method used to construct it. In general, a narrower interval indicates greater precision and a wider interval indicates less precision. A 95% confidence interval is not a range expected to contain 95% of individual outcomes, and it does not mean there is a 95% probability that the fixed true mean lies inside the particular calculated interval. Its interpretation depends on the sampling and analysis assumptions.
The control-arm mean is 0.68 with a 95% confidence interval of 0.56 to 0.74. The other arm mean is 0.56 with a 95% confidence interval of 0.48 to 0.61. These intervals describe uncertainty around separate arm means. They are not a confidence interval for the difference between arms. A direct contrast would require a defined comparison and its uncertainty, neither of which is reported in the registry record.
Random assignment at the network level creates a clustered design. As a general statistical point, participants within the same network may have more similar exposures or outcomes than participants in different networks. Analyses that treat all participant observations as independent can understate uncertainty when within-network correlation is present. Appropriate methods often account for the assignment clusters, but the registry record does not describe the method used for these posted summaries or whether a cluster-adjusted analysis was performed.
The endpoint is itself a ratio: condom-use occasions divided by sexual occasions for each person. Such measures can have different precision depending on the denominator and may behave differently among people with very few reported occasions. A general analysis would need to specify how zero denominators, missing reports and unusual values were handled. The registry does not report those details. Nor does it clarify whether the posted mean was calculated from all individual proportions or through another aggregation approach.
The stated confidence intervals are the only uncertainty summaries attached to the posted arm means. The registry's dispersion field identifies 95% confidence intervals, while no separate standard deviation or other participant-level distribution summary is posted. The interval limits should therefore not be interpreted as minimum and maximum observed outcomes. They summarize uncertainty in the means, not the full variability of responses.
No inferential method, model specification, covariate adjustment, standard-error calculation or hypothesis test is described in the registry record. The record also lists zero analyses. Therefore, it is not possible to determine whether the posted confidence intervals incorporate network clustering, repeated assessments, baseline values or any other design feature. The arm summaries may be read as reported, but their statistical construction cannot be independently assessed from the registry information.
For a randomized trial, the treatment comparison usually concerns an arm contrast, with an estimate and uncertainty calculated according to the randomization and analysis plan. A general interpretation would ask whether the contrast is compatible with no difference and how large a range of differences remains plausible. Here, no contrast or interval for a contrast is posted. Comparing individual arm intervals informally is not a substitute for a direct analysis, particularly in a clustered study.
Planning calculations based on an observed effect, when presented, are illustrations for future study design rather than evidence about the results of this trial. Such calculations depend on assumptions about the effect to detect, significance level, allocation and statistical model. The registry record does not report the trial's sample-size justification or statistical planning assumptions. No planning illustration can establish observed power or replace a reported analysis. The primary conclusion from the statistical record is thus about reporting: arm-level descriptive estimates are available, while the inferential comparison and methods needed to interpret it are not.
10. Multiplicity, Missing Data and Analysis Populations
The registry identifies one primary outcome, proportion of condom use, and lists Condom and Abstinence Norms, Attitudes, Self-efficacy and Intentions as a secondary outcome. It also records two outcome measures and one posted secondary outcome. The brief study description mentions condom use and abstinence assessments, but posted results for abstinence are not included in the record summarized by these registry entries.
When a study evaluates multiple outcomes, time points or comparisons, the chance of a false-positive finding can increase if each is tested separately. General approaches include designating a primary outcome in advance, limiting confirmatory tests, or adjusting interpretation for multiple testing. The registry record does not report a multiplicity strategy, a hierarchy among secondary measures, or which analyses were intended to be confirmatory. Because no inferential analyses are posted, there is no reported set of p-values to assess for multiplicity.
Missing-data methods and analysis populations are also not reported in the registry record. The outcome entries include different arm counts, but the reason for that difference is not stated; it cannot be assumed to represent loss to follow-up or any particular exclusion. The record does not describe participant flow, incomplete questionnaires, missing sexual-activity reports, or how undefined proportions were handled.
The brief description refers to repeated assessments, but the record does not say whether all randomized participants were analyzed, whether analysis followed assignment regardless of message exposure, or whether results came from a subset with available measurements. It also does not report how repeated measures or network-level assignment were incorporated. These omissions limit assessment of selection bias, comparability across arms, and the robustness of the posted means.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Arm 1 | 0 | 942 | 0.0% |
| Arm 2 | 0 | 636 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Arm 1 | 0 of 942 |
| Arm 2 | 0 of 636 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Arm 1 | 0 of 942 |
| Arm 2 | 0 of 636 |
The registry reports serious adverse event counts separately for the two arms. For the control arm, 942 participants are listed as at risk, with zero serious adverse events, zero other adverse events in the reported fields, and zero deaths. For the other arm, 636 participants are listed as at risk, with zero serious adverse events, zero other adverse events in the reported fields, and zero deaths. No top serious-event terms are listed.
These are event counts as recorded in the registry, not a guarantee that no adverse experiences occurred. The counts should be read as no events entered in the specified registry fields, rather than proof of absence of all harms. The record does not explain event collection procedures, the observation period for safety reporting, definitions used for the categories, or completeness of reporting.
The numbers listed as at risk differ from the outcome counts attached to the condom-use means. The registry does not explain the difference, and the two sets of counts should not be treated as interchangeable denominators. These safety entries are descriptive and do not provide a statistical comparison of risk between arms.
12. What the Result Does, and Does Not, Mean
The posted outcome means describe the average condom-use proportion recorded in each arm, with a 95% confidence interval for each mean. They provide a summary of the available endpoint data, not a direct estimate of the effect of assignment to dynamic messages compared with static information. The record does not report an arm contrast or a statistical test.
The values do not show how individual participants responded, how much their responses varied, or whether changes occurred between baseline and later assessments. The measurement occasion associated with the posted results is not reported in the registry record. Although the brief description mentions condom use and abstinence assessments, the posted primary result is a condom-use proportion; no result for abstinence is included in the registry record.
Because the trial assigned networks, the analysis should account for the possibility that outcomes among members of the same network are related. The registry does not state whether that feature was incorporated into the reported intervals. Without the analysis method, missing-data rules, and a direct comparison, the precision and interpretation of a treatment-effect estimate cannot be evaluated.
The results therefore do not establish that dynamic Facebook messages changed condom use, nor do they establish that the approaches were equivalent. Absence of a posted test is not evidence of no difference. Similarly, the reported means should not be framed as evidence that either intervention is effective or ineffective for HIV prevention. They are descriptive values from a behavioral outcome, not a clinical recommendation or a measure of HIV incidence.
A fuller interpretation would require a clearly identified assessment time, a comparison estimate with uncertainty, and methods that address the network assignment and outcome data. It would also benefit from participant-flow information and a transparent account of the planned and completed analyses. Until those details are available, the safest interpretation remains confined to the arm means and confidence intervals that the registry reports.
13. Why This Trial Matters Statistically
This trial is statistically notable because randomization occurred at the social-network level while the primary outcome is a participant-level behavioral measure. That combination requires attention to the relationship among participants within each assigned network. In general, when people share a network, their exposure and outcomes may not be independent. If an analysis ignores that structure, its uncertainty estimates can be misleading. The registry record does not say whether the posted intervals account for clustering.
The endpoint also illustrates how a proportion can be informative while relying on details that need to be reported. It relates condom use to sexual occasions rather than simply counting condom use. But interpretation depends on the denominator and on handling people whose reports are incomplete or whose denominator is zero. Those rules are not available in the registry record.
A further statistical lesson is the distinction between arm summaries and a treatment comparison. Means and confidence intervals describe each arm, but a randomized trial's central question concerns the contrast between assigned groups. Without a contrast, its uncertainty and an analysis method, the reported summaries cannot answer whether the observed arm difference is compatible with chance variation or reflects a real difference.
Finally, the study description and outcome record do not fully align on timing: the brief summary describes baseline and 3 and 6 month assessments, while the primary endpoint names a two-month period based on the past 60 days. Clear linkage between outcome, measurement occasion, analysis population and model would make the results easier to interpret and reproduce. The record's omissions make these reporting issues central to statistical understanding.
14. Limitations
- The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
- Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
Several trial-specific limitations affect interpretation. First, the registry does not identify which assessment occasion produced the posted primary outcome means. The brief description mentions baseline and 3 and 6 month assessments, while the endpoint specifies two months and refers to the past 60 days. Without a link between those descriptions and the reported estimates, the timing of the outcome is unclear.
Second, the study randomized networks, but the record does not report the number of networks, their sizes, or whether the analysis accounted for clustering. Participant counts alone cannot show how many independent assignment units contributed to the estimates. The registry also does not describe participant recruitment, eligibility, demographic characteristics, message exposure, or adherence, limiting assessment of generalizability and intervention contrast.
Third, the posted record gives arm means and confidence intervals but no direct comparison, p-value, model, or analysis plan. Missing-data rules, analysis populations, and handling of the repeated assessments are not reported. The different arm counts attached to the endpoint are unexplained, so their cause cannot be inferred.
Finally, the outcome is based on self-reported condom use and sexual activity, and the registry does not describe measurement validation or procedures to reduce reporting error. The record does not report results for abstinence or the listed secondary outcome. These limitations do not invalidate the posted descriptive estimates, but they restrict conclusions about causal effects, precision, timing and broader relevance.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Online Social Networking for HIV Prevention — NCT00725959.
- Registry record last updated 2017-11-07; status last verified 2017-10.
- Results first posted 2017-11-07.
- Data retrieved from ClinicalTrials.gov on 2026-11-09.
- Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
The registry record describes NCT00725959, Online Social Networking for HIV Prevention, a completed randomized, parallel-group behavioral trial sponsored by Colorado School of Public Health. It records enrollment of 1578, a study start of 2008-09 and primary completion in 2010-12. Assignment was described at the Facebook-network level, with dynamic prevention messages compared with static online HIV-prevention information; masking was listed as none.
The primary outcome was proportion of condom use, defined using condom-use occasions divided by sexual occasions in the past 60 days. The posted control mean is 0.68, with a 95% confidence interval of 0.56 to 0.74; the other arm mean is 0.56, with a 95% confidence interval of 0.48 to 0.61. The registry lists zero analyses and does not report a statistical comparison, analysis method, missing-data strategy, or the assessment occasion associated with these means. Serious adverse event fields report zero events in both arms.