ACTIV-6: COVID-19 Study of Repurposed Medications: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for ACTIV-6: COVID-19 Study of Repurposed Medications (NCT04885530), a randomized phase 3 study. The primary outcome is Total Number of Participants Randomized Within Each Appendix.

2. Clinical Question

3. Trial Design

ACTIV-6 is listed as a phase 3, randomized, double-masked, parallel study of repurposed medications for outpatient COVID-19. The registry status is completed. The study started on 2021-06-08, and primary completion is recorded as 2024-04-29. Enrollment is listed as 10956. These describe the overall registry entry; the primary measure also refers to participant counts within individual appendices.

The study was designed to evaluate medications in people with mild to moderate COVID-19 who were not hospitalized. Participants were to receive a study drug or placebo and self-report new or worsening symptoms or medical events while taking their assigned product. The record says the study was intended to be entirely remote unless the study team considered an in-person visit to be in a participant’s best interest.

The intervention list and arm descriptions show several medication regimens, including two ivermectin arms and two fluvoxamine arms, as well as fluticasone, montelukast and metformin. Placebo descriptions are also listed. The record explains that additional medication arms could be activated and that arm-specific outcome data would be reported under associated registry entries. Consequently, the overall enrollment figure should not be read as the size of any single medication comparison.

The numerical contrasts posted for the binary measure use Arm A - Ivermectin 400 as their comparator. That is distinct from the placebo groups described in the arm information. No separate randomization ratio, allocation sequence method, or masking procedure details are reported in the registry record. The trial’s original sample-size justification and its intended statistical analysis plan are also not reported in the registry record.

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.

Before assignment: Numbers reported represent those who were eligible for study drug and randomized. Participants who were randomized to a placebo group may have participated in more than one study arm due to the nature of the study design.

Only the first 6 of 7 groups are shown.

Substudy A

Substudy B

Substudy C

Substudy D

Analysis population for the primary outcome, as defined in the registry: Individuals who met inclusion criteria and consented for study drug for each study arm. Given that participants were not randomized at this point, one number is reported for each study arm as a whole.

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.

Population: Numbers reported represent those who were eligible for study drug and randomized to each appendix.

Only the first 6 of 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Total Number of Participants Randomized Within Each Appendix, measured over Up to approximately 2 years 10 months.

The stated primary measure is the total number of participants randomized within each appendix, expressed as a participant count. Its timeframe is listed as up to approximately 2 years 10 months. The registry describes a population of individuals who met inclusion criteria and consented for study drug for each arm, and says one number is reported for each study arm as a whole because participants were not randomized at that point.

A count of participants can describe the scale of enrollment or participation in an appendix. It is not, by itself, a measure of symptom severity, time to recovery, hospitalization, or another clinical outcome. The record’s study purpose refers to reducing symptoms, and its brief description says participants self-reported symptoms and medical events. However, the stated primary measure does not quantify those symptoms or events.

The numerical arm results include counts and denominators that are used for binary contrasts. The record does not clarify how those binary counts relate to the stated primary measure, nor does it identify a symptom-based primary endpoint among the posted outcome measures. Appendix-specific outcomes are said to be reported under associated registry entries. The distinction matters: a participant count describes how many people were counted, whereas a clinical endpoint would describe an outcome experienced by participants. These should not be treated as interchangeable.

7. Results Posted in the Registry

• Arm B - Fluvoxamine: 34.7% (1331/3837) versus 52.1% (1800/3457) on control; risk difference -17.4 percentage points (95% CI -19.6 to -15.1), risk ratio 0.67 (0.63–0.70); unadjusted two-proportion z-test p = <0.001.
• Arm C - Fluticasone: 42.4% (1407/3315) versus 52.1% (1800/3457) on control; risk difference -9.6 percentage points (95% CI -12.0 to -7.3), risk ratio 0.82 (0.77–0.86); unadjusted two-proportion z-test p = <0.001.
• Arm D - Ivermectin 600: 66.0% (1459/2212) versus 52.1% (1800/3457) on control; risk difference 13.9 percentage points (95% CI 11.3 to 16.5), risk ratio 1.27 (1.21–1.32); unadjusted two-proportion z-test p = <0.001.
• Arm E - Fluvoxamine 100: 72.9% (1208/1657) versus 52.1% (1800/3457) on control; risk difference 20.8 percentage points (95% CI 18.1 to 23.5), risk ratio 1.40 (1.34–1.46); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a risk difference describe?

For a binary outcome, a risk is the proportion of participants in a group counted as having the outcome. A risk difference compares two such proportions by subtracting the comparator risk from the intervention risk. A value above zero means the coded outcome was more frequent in the intervention group; a value below zero means it was less frequent. This is a general interpretation of the statistic, not a conclusion about whether the outcome is desirable. The outcome needs a clear clinical definition before direction can be called favorable or unfavorable.

The posted contrasts use Arm A - Ivermectin 400 as comparator and give proportions for each arm. They are reported as unadjusted risk differences. For example, the risk difference for fluvoxamine is -0.174, while the difference for metformin is 0.347. The record does not identify the binary outcome underlying these contrasts in a way that connects it to a symptom endpoint. Therefore, the statistical direction can be described, but its clinical meaning remains uncertain.

How should a confidence interval be read?

A confidence interval expresses uncertainty around an estimated effect under the statistical model used. A narrower interval indicates greater statistical precision than a wider interval, all else equal. If an interval for a risk difference lies entirely on one side of zero, the data are inconsistent with a zero difference under the assumptions behind that interval. This does not establish that the outcome is clinically important or that the model assumptions are appropriate.

Each posted interval is for an unadjusted contrast. For fluvoxamine, the interval runs from -0.196 to -0.151; for ivermectin 600, it runs from 0.113 to 0.165. Those intervals exclude zero, as do the other reported intervals. This statistical feature is not a substitute for endpoint interpretation, multiplicity adjustment, or evaluation of missing data and analysis populations.

What do the p-values say?

A p-value describes how compatible the observed data are with a specified null hypothesis under the analysis assumptions. Here, each reported unadjusted p-value is p < 0.001. It is not the probability that the null hypothesis is true, nor the probability that a result occurred “by chance.” It also does not measure the size or clinical importance of an effect. Because the figures are unadjusted and several contrasts are listed, they do not account for multiplicity.

How do the counts relate to the endpoint?

The stated primary measure is a count of participants randomized within each appendix. The arm entries also include a count and a denominator, and the binary comparisons use those numbers to report proportions and risk differences. The registry does not explain whether the binary count is a symptom measure, an event outcome, or another classification. It also says appendix-specific outcome measures will be reported under associated registry entries. The statistical calculations can summarize the posted counts, but they cannot supply an outcome definition missing from the record.

What can randomization and masking establish?

Randomization is a general method for balancing measured and unmeasured characteristics across assigned groups in expectation. Double masking is intended to reduce the influence of knowledge of assignment on participant behavior, care, or outcome assessment. These design labels are reported for ACTIV-6, but the registry record does not provide details of sequence generation, concealment, or masking implementation. The presence of randomization and masking labels does not resolve the mismatch between the stated count endpoint and the binary contrasts, or explain why the posted comparison uses an ivermectin arm rather than placebo.

What does over-dispersion mean?

Over-dispersion is a general issue in count models: observed variability can exceed what a model assumes. It matters when analyzing counts because standard errors can be too small if extra variability is ignored. The primary measure is described as a participant count, but the record reports no count-model analysis or dispersion estimate. The listed contrasts are risk differences for binary data, so over-dispersion is not presented as part of their reported method.

The registry does not report the trial’s original statistical plan, adjustments for covariates, handling of missing outcomes, or prespecified multiplicity strategy. These details cannot be inferred from the reported p-values and intervals. The comparisons should therefore be understood as unadjusted summaries of the posted binary counts, with their interpretation limited by the endpoint information recorded.

10. Confidence Intervals and Precision

The confidence intervals quantify statistical uncertainty around the posted unadjusted risk differences. For fluvoxamine, the estimate is -0.174 and the 95% interval is -0.196 to -0.151. For fluticasone, the estimate is -0.096 and the interval is -0.120 to -0.073. For ivermectin 600, the estimate is 0.139 and the interval is 0.113 to 0.165. For fluvoxamine 100, it is 0.208, with an interval of 0.181 to 0.235.

For montelukast, the risk difference is 0.355, with an interval of 0.332 to 0.378. For metformin, it is 0.347, with an interval of 0.327 to 0.367. Each reported interval excludes zero. In general, an interval that excludes zero is consistent with a nonzero difference under the calculation’s assumptions. It does not establish the clinical relevance or desirability of the binary outcome.

The intervals are relatively concentrated around their estimates, but precision is only one part of interpretation. The record’s primary endpoint is a participant count, and it does not explain the binary outcome used for these contrasts. The estimates are unadjusted, and no multiplicity procedure is reported. Confidence intervals also do not address potential bias from missing observations, changes in eligibility or assignment, or differences between the recorded and intended analysis populations. Those issues are not detailed in the registry record, so the intervals should be read as statistical uncertainty around the posted calculations rather than as a complete account of uncertainty in the study’s clinical conclusions.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. They are illustrations based on the observed risk differences, not the original design justification for ACTIV-6. The listed assumptions are a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect to detect set to the stated fraction of the observed effect. The trial’s own sample-size assumptions are not reported in the registry record.

For each contrast, the 80% power column gives the estimated number of participants per arm for a future two-arm study to have an 80% chance of detecting the specified effect under those planning assumptions, if that effect and the assumptions are correct. The 90% power column gives the corresponding estimate for a higher chance of detection. These are design calculations, not observed outcomes and not a measure of the probability that the posted result is true.

Rows labelled as planning for 100% of the observed effect use the full observed difference as the target. Rows for 75% and 50% use smaller target effects. Detecting a smaller effect generally requires more participants, which is reflected in the larger planning estimates for smaller fractions. For example, the fluvoxamine illustration lists 127 participants per arm for 80% power at 100% of the observed effect, 228 at 75%, and 518 at 50%. The corresponding 90% figures are 169, 305 and 693 participants per arm.

These estimates are hypothetical planning illustrations and are not observed power. They depend on using the observed effect as the starting point, and may be sensitive to uncertainty in that estimate. They do not incorporate any trial-specific design features not included in the stated assumptions.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.521; treatment proportion p₁ = 0.347; average p̄ = 0.434.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7009 = 1.3737.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6900 = 0.5807.
4. n = (1.3737 + 0.5807)² / (-0.1738)² = 126.5, rounded up to 127 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.520682672837721                          # control proportion
p1 <- 0.346885587698723                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(4885530)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 127 (80% power), 169 (90% power)
 75% of observed effect: n per arm = 228 (80% power), 305 (90% power)
 50% of observed effect: n per arm = 518 (80% power), 693 (90% power)
power.prop.test: 126.5 
Simulated power at n = 127 per arm: 0.801 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

Several medication contrasts are listed against the same comparator. When multiple hypotheses are tested, the chance of at least one apparently unusual result can increase. Statistical methods for addressing multiplicity may adjust significance thresholds or use another prespecified strategy. The registry record does not report a multiplicity strategy, and the posted p-values are unadjusted. They should not be read as multiplicity-adjusted evidence across all listed contrasts.

The record reports one primary outcome measure, a count of participants randomized within each appendix, and no secondary outcome measures. It also states that appendix-specific outcome data will be reported under associated registry entries. The outcome description does not define the binary event underlying the posted risk differences. This limits the interpretation of both the individual comparisons and any apparent pattern across medications.

Missing-data handling is not reported in the registry record. The record does not specify how incomplete symptom reports, participants without an outcome classification, or withdrawals were handled in the comparisons. The listed arm denominators and counts do not by themselves establish whether all randomized participants were included or whether the analysis followed an intention-to-treat principle. Analysis-population definitions are not reported in the registry record.

The record identifies the study as randomized and double-masked, but does not provide the statistical analysis plan or detailed methods for allocation, outcome assessment, or covariate adjustment. It therefore is not possible to establish from this entry whether the posted contrasts were prespecified, whether other outcomes were considered, or whether any interim analyses occurred. Those details are not reported in the registry record. The p-values and intervals should be treated as unadjusted summaries of the posted figures rather than as a complete confirmatory analysis.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry lists zero serious adverse events, zero other adverse events and zero deaths for each displayed medication group. The numbers at risk for serious and other adverse events are the same within each group: 1800 for ivermectin 400, 1331 for fluvoxamine, 1407 for fluticasone, 1459 for ivermectin 600, 1208 for fluvoxamine 100, 1453 for montelukast and 3214 for metformin. The listed death count is zero in every group.

These entries describe the adverse-event counts recorded for the displayed groups. They do not establish that no events occurred among all enrolled participants, outside the reported groups, or beyond the recorded reporting period. The brief study description says participants would self-report medical events, but additional details about adverse-event definitions, ascertainment, follow-up completeness and analysis populations are not reported in the registry record. The record lists no top serious-event terms. Read the zero counts as the posted safety figures, not as a general guarantee of safety or as proof that adverse events were absent.

15. What the Result Does, and Does Not, Mean

The registry record establishes that ACTIV-6 was a completed phase 3 randomized, double-masked outpatient study with several repurposed medication arms. It records a total enrollment of 10956 and describes remote participant reporting of symptoms and medical events. The primary measure named in the record is a participant count within each appendix, rather than a measure of symptom change or another clinical outcome.

The posted binary comparisons show unadjusted risk differences relative to Arm A - Ivermectin 400, with reported p-values of p < 0.001 and confidence intervals excluding zero for each listed contrast. Those statistics describe the posted binary counts. The endpoint definition needed to translate those differences into a clinical interpretation is not reported in the registry record. The comparisons are also not presented as medication-versus-placebo contrasts, even though placebo arms are described.

Accordingly, these figures alone do not establish that any medication improves symptoms, worsens symptoms, or changes another clinically meaningful outcome. Statistical evidence of a difference in a coded binary measure is not the same as evidence of a beneficial or harmful clinical effect. The unadjusted p-values also do not address multiplicity across the listed comparisons. The registry record does not report missing-data procedures or analysis-population definitions.

The record’s stated intention to report appendix-specific outcomes under associated registry entries is important context. A complete interpretation of a medication evaluation would require a clearly defined clinical endpoint and its associated analysis. The safety entries list zero reported serious events, other events and deaths in the displayed groups, but do not establish that no events occurred in every participant or reporting context. The conclusions should remain limited to what this registry entry records: study design labels, participant counts, binary contrast statistics, and posted adverse-event counts.

16. Why This Trial Matters Statistically

ACTIV-6 is statistically notable as a multi-arm outpatient study with several repurposed medications and a completed phase 3 registry record. Its entry illustrates why study design labels, enrollment counts and small p-values are not enough on their own to interpret an intervention’s clinical effect. The endpoint definition and the relationship between the stated primary measure and the reported contrasts are equally important.

The primary measure is described as a count of participants randomized within each appendix. The listed comparisons, by contrast, are binary risk differences against the ivermectin 400 arm. Because the record does not identify the clinical meaning of the binary event, readers cannot confidently interpret whether a positive or negative contrast represents a desirable outcome. The placebo arms described in the intervention information also differ from the comparator used for the numerical contrasts.

The example underscores the role of multiplicity. Several medication contrasts are presented, but the p-values are unadjusted and no multiplicity plan is reported. Confidence intervals communicate precision for each contrast, but they do not resolve the endpoint-definition issue or account for multiplicity and unreported missing-data methods.

Finally, the registry’s plan to report appendix-specific outcomes under associated records signals that the umbrella entry is not a complete account of every medication evaluation. For clinical researchers, careful linkage between the trial question, endpoint, comparison group and analysis population is central to distinguishing descriptive registry counts from interpretable evidence about an outcome.

17. 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.

The main limitation specific to this record is the gap between the study’s symptom-focused purpose and its stated primary measure. The primary measure counts participants within each appendix, while the numerical comparisons are binary risk differences. The record does not define the binary outcome or explain how it relates to the participant count, symptoms, or medical events. This prevents a confident clinical interpretation of the contrast directions.

The posted contrasts use Arm A - Ivermectin 400 as comparator, although placebo groups are described for the medication arms. The reason for this comparator choice is not reported in the registry record. Appendix-specific outcome data are said to appear under associated registry entries, so the overall record may not capture each medication’s full outcome reporting.

The comparisons are unadjusted and several are listed, without a reported multiplicity strategy. The record does not report missing-data methods, analysis populations, the original sample-size justification, detailed allocation procedures, or the statistical analysis plan. It also does not report whether interim analyses occurred. These omissions mean that the posted estimates cannot be evaluated against a complete prespecified analysis framework using this record alone.

The safety entries list zero serious adverse events, other adverse events and deaths in the displayed groups, but provide limited detail about event definitions, ascertainment and completeness. The record describes self-reporting of events but does not specify the full safety assessment process. These limitations call for restraint in interpreting both efficacy-related contrasts and safety counts from this entry.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: ACTIV-6: COVID-19 Study of Repurposed Medications — NCT04885530.
• Registry record last updated 2025-11-17; status last verified 2025-11.
• Results first posted 2025-11-17.
• 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.

21. Record Summary

ACTIV-6 is listed as a completed phase 3 randomized, double-masked, parallel outpatient study of repurposed medications for mild to moderate COVID-19. Overall enrollment is 10956. The registry describes several medication arms, placebo groups and remote self-reporting of symptoms and medical events.

The stated primary measure is the number of participants randomized within each appendix. The posted binary comparisons are unadjusted risk differences against Arm A - Ivermectin 400, but the binary outcome’s clinical definition is not reported in the registry record. Appendix-specific outcomes are directed to associated records. The displayed safety entries list zero serious adverse events, zero other adverse events and zero deaths in each group. Interpretation of clinical effects is limited by the endpoint description, comparator choice and unreported analysis details.