Prazosin for Treatment of Patients With Alcohol Dependence (AD) and Post Traumatic Stress Disorder (PTSD).: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Prazosin for Treatment of Patients With Alcohol Dependence (AD) and Post Traumatic Stress Disorder (PTSD). (NCT00744055), a randomized phase 3 study. The primary outcome is Number of Drinking Days.

2. Clinical Question

3. Trial Design

The registry record describes a randomized, triple-masked, parallel-group phase 3 trial. Its listed enrollment is 96. The study compared a prazosin arm with a placebo arm, and the posted primary outcome includes 50 participants in the prazosin group and 46 in the placebo group. The registry lists Yale University as the sponsor and gives the conditions as alcohol dependence and post-traumatic stress disorder.

The prazosin arm is described as prazosin 16mg/day. Placebo was described as being packaged in identical-looking capsule blister packs. Triple masking indicates that the registry classifies the trial as having masking at three levels, but the record does not identify those roles or explain how masking was maintained or assessed. It also does not report allocation concealment details, treatment adherence, or whether the analysis retained participants according to their randomized assignment.

The record lists the study as completed. Its start is recorded as 2009-01, and its primary completion is recorded as 2014-10. These dates describe the registry record and do not, by themselves, explain how long any participant was followed beyond the primary outcome timeframe.

The primary endpoint was the number of drinking days during 12 weeks. The posted arm results are means with standard deviations. The comparison is expressed as a mean difference between prazosin and placebo, with placebo as the reference group. The record does not report the trial's original sample-size justification, detailed protocol, statistical analysis plan, or analysis population definition.

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: Randomization was conducted by the pharmacy using a 1:1 randomization in blocks of 4 and stratified by site, gender. Completers were defined as subjects for whom we had complete data at the end of the treatment period (week 12) whether they remained on medication or not.

Overall Study

Reasons for not completing this period:

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: Veterans

6. Primary Endpoint

The registry lists the primary outcome as Number of Drinking Days, measured over 12 weeks.

The primary endpoint is titled Number of Drinking Days, measured in days, with a 12-week timeframe. It is a continuous outcome reported as a mean and standard deviation for each arm. The record says that the Timeline Follow Back method was used. It describes this as a calendar method for assessing drug and alcohol use.

Counting drinking days focuses on the frequency of drinking over the specified assessment period. It does not, on its own, describe the amount consumed on drinking days, the severity of alcohol-related problems, or changes in PTSD symptoms. The mean gives the average number of recorded drinking days within an arm, while the standard deviation describes how dispersed individual recorded values are around that arm's mean. A relatively large standard deviation signals substantial variation among participants, but does not explain its causes.

The posted means are 11.04 days for prazosin and 9.21 days for placebo. The reported mean difference is 1.83 days, defined in the comparison as prazosin relative to placebo. Because a higher count represents more recorded drinking days, the direction of this contrast is important when describing the result. The registry record does not report details such as how missing calendar entries were handled or how the endpoint was analyzed beyond the posted summary statistics.

7. Results Posted in the Registry

• Prazosin: mean 11.04 versus 9.21 on control; difference 1.83 (95% CI -5.27 to 8.93), standardized effect d = 0.10; z-test p = 0.614.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant's recorded number of drinking days is summarized with a mean and standard deviation. The registry record labels the parameter as a mean and gives the unit as days. The posted summaries are 11.04 with a standard deviation of 18.86 for prazosin, and 9.21 with a standard deviation of 16.64 for placebo. These describe the observed outcome values in the listed groups; they are not percentages or rates.

What does the mean difference represent?

The comparison is defined as prazosin relative to placebo. Subtracting the placebo mean from the prazosin mean gives the reported mean difference of 1.83 days. This is an unadjusted comparison based on the posted arm summaries. A positive value means that the prazosin arm's mean is higher on this endpoint; because the endpoint counts drinking days, that direction is not a reduction in the mean count. The estimate does not by itself show why the groups differed or establish a causal effect.

What do the standard deviations describe?

A standard deviation describes spread among individual observations around a group mean. In general, a larger standard deviation indicates more variation in the measured values. The reported standard deviations, 18.86 and 16.64 days, show that individual values varied substantially within both groups relative to their means. They do not describe uncertainty in the estimated mean difference; that is the role of a standard error and an interval estimate.

How do standard errors relate to precision?

The comparison lists standard errors of 2.453 for the placebo mean and 2.667 for the prazosin mean. In general, a standard error describes the sampling uncertainty of an estimate: smaller standard errors usually correspond to greater precision, all else being equal. Standard errors are not measures of variation among individual participants. The registry record does not describe the procedure used to calculate these standard errors or whether any adjustments were made.

What does the p-value say?

The posted unadjusted p-value is 0.614. In general, a p-value describes how compatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, and it does not quantify the size or clinical importance of a difference. The registry record does not report the test used to obtain this p-value, the complete model specification, or a multiplicity adjustment. It should therefore be read as the posted unadjusted statistic rather than as a complete account of the analysis.

How should the interval be read?

The reported limits for the mean difference are -5.273 and 8.933 days. In general, an interval estimate represents a range of values compatible with the data and the method used to construct it. Here, the interval includes zero and extends to both negative and positive differences. That means the estimate does not pin down the direction of the contrast precisely. The confidence level and construction method are not reported in the registry record, so these should not be inferred from the limits alone.

What is the role of the pooled standard deviation?

The comparison reports a pooled standard deviation of 17.83. A pooled standard deviation combines information about outcome spread across groups for use in some standardized comparisons and planning calculations. It is not another estimate of the mean difference, and it does not replace the arm-specific standard deviations. The registry record does not specify the complete method or assumptions behind this pooled value.

Overall, the posted analysis is a comparison of group means, accompanied by measures of spread and uncertainty. The record does not report covariate adjustment, methods for repeated observations, distributional checks, imputation procedures, or sensitivity analyses. Those methods should not be assumed from the trial's randomized and masked design alone.

10. Confidence Intervals and Precision

The reported mean difference is 1.83 days, with interval limits from -5.273 to 8.933 days. The interval is broad enough to include both a negative and a positive difference. Thus, although the estimate itself is positive, the interval does not identify the direction of the underlying contrast with precision. Its lower limit is compatible with fewer drinking days in the prazosin arm, while its upper limit is compatible with more.

A confidence interval is generally used to express uncertainty around an estimate, but the confidence level and method used to construct this interval are not reported in the registry record. The limits should therefore be described as the reported interval rather than assigned an unreported confidence level or method. The unadjusted p-value is 0.614; it is another summary of the comparison and does not replace the interval's information about the range of estimates compatible with the analysis.

The arm standard deviations, 18.86 days for prazosin and 16.64 days for placebo, indicate substantial dispersion in the outcome. In general, variability among observations and the amount of information available affect how precisely a group contrast can be estimated. The posted interval reflects uncertainty in the mean difference, not a range expected to contain individual participants' drinking-day counts.

The interval's breadth limits how specifically the result can be interpreted. It does not establish equivalence, and it does not rule out effects in either direction. The registry record does not provide additional analyses that might clarify robustness or precision, such as adjusted estimates, sensitivity analyses, or results under different assumptions about missing outcomes.

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 event counts. They are illustrations computed from the observed effect rather than the trial's original design assumptions; those original assumptions are not reported in the registry record. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect.

The 80% power column gives the planned number of participants per arm for a calculation with an 80% chance of detecting the chosen effect under those assumptions, if that effect and the assumptions are correct. The 90% power column gives the corresponding per-arm planning figure for a 90% chance. These are design calculations, not estimates of what happened during this trial. This is not observed power.

The planning rows consider 100%, 75%, and 50% of the observed effect. A smaller effect is harder to distinguish from random variation, so the calculation requires more participants per arm to achieve the same planned power. The displayed figures become larger as the assumed fraction of the observed effect becomes smaller. They should be read as conditional illustrations, not as recommendations or as evidence that the trial was designed to detect those effects.

The registry record does not report the original sample-size justification, nor does it establish that the trial's own design used these calculations. The planning assumptions attached to this illustration should not be attributed to the study's protocol.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 1.830; pooled standard deviation σ = 17.832.
2. n = 2 × (1.96 + 0.842)² × 17.832² / 1.830² = 1,490.5, rounded up to 1,491 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
delta <- 1.83                      # observed difference in means
sigma <- 17.8317528165528                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(744055)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 1,491 (80% power), 1,996 (90% power)
 75% of observed effect: n per arm = 2,650 (80% power), 3,548 (90% power)
 50% of observed effect: n per arm = 5,962 (80% power), 7,982 (90% power)
power.t.test: 1491.4 
Simulated power at n = 1,491 per arm: 0.801 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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

The registry record lists 2 outcome measures, but it reports 0 secondary outcomes and 0 analyses. The primary drinking-days result is therefore the only posted outcome comparison available in these results. Although the study objective mentions both alcohol consumption and PTSD symptoms, the record does not post a PTSD symptom analysis. Its absence should not be interpreted as evidence of no change in PTSD symptoms.

Multiplicity arises when researchers make several statistical comparisons or assess several outcomes: the chance of at least one apparently unusual result can increase as the number of tests increases. In general, analysis plans may address that issue by defining a primary comparison or applying an adjustment. The registry record does not report a multiplicity strategy, a complete list of planned analyses, or enough detail to determine whether other analyses were conducted. The posted p-value is explicitly unadjusted.

The record also does not report how missing outcome data were handled, whether all randomized participants were included in the primary analysis, or how the analysis population was defined. It does not say whether participants were analyzed according to randomized assignment or whether exclusions occurred after randomization. These details matter because missing observations and analysis-set decisions can affect the estimate and its uncertainty.

No imputation method, sensitivity analysis, or assessment of missingness is reported. The registry record also does not state whether observations were adjusted for baseline values or other characteristics. These methods must remain unknown rather than being inferred from the randomized design. Without them, the posted summaries give a useful account of the reported means and contrast but not a complete statistical analysis plan or a full account of robustness.

14. Safety Results

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

Other (non-serious) adverse events

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

Most frequent non-serious adverse event terms

The adverse-event record lists 0 serious adverse events among 50 participants at risk in the prazosin group and 0 among 46 at risk in the placebo group. These are counts of participants affected by serious adverse events as recorded for the groups. The registry record does not list any serious adverse-event terms.

For other adverse events, the record lists 11 affected participants among 50 at risk in the prazosin group and 8 among 46 at risk in the placebo group. These are counts, not percentages. The record does not provide further detail about event types, severity beyond the serious-versus-other classification, timing, or whether events were judged related to treatment.

Deaths are recorded as null for both groups, which means death counts are not reported in the registry record; null should not be read as zero. Adverse-event counts alone do not establish whether an intervention is safe or unsafe. The record's limited detail, including the absence of event terms and fuller reporting context, constrains comparisons between groups.

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

The posted primary result shows a mean of 11.04 drinking days in the prazosin group and 9.21 in the placebo group over 12 weeks. The reported contrast is 1.83 days higher in the prazosin arm, with an interval from -5.273 to 8.933 days and an unadjusted p-value of 0.614. Taken together, these statistics describe an estimate that points toward a higher mean in the prazosin group but remains imprecise about the direction and size of the underlying difference.

This result does not establish that prazosin reduced drinking days, nor does it establish that prazosin increased them. It also does not establish equivalence between prazosin and placebo. The interval includes both directions, and the posted analysis does not rule out potentially meaningful differences. The p-value is not a measure of the probability that one treatment is better, nor a measure of clinical importance.

The primary result concerns drinking days, not every outcome named in the study objective. The objective mentions PTSD symptoms, but a PTSD symptom result is not reported in the registry record. No secondary outcome results are posted. Conclusions about PTSD symptoms or other unreported outcomes cannot be drawn from the drinking-days comparison.

The summaries also do not describe how missing observations, adherence, or analysis populations were handled. The record does not provide the confidence level or construction method for the interval, or the trial's detailed statistical plan. These omissions limit assessment of the analysis and its robustness.

Accordingly, the most defensible reading is limited to the posted descriptive means and unadjusted comparison. They characterize the registry results for the specified endpoint; they do not support medical advice or a general conclusion about treatment for people with alcohol dependence and PTSD.

16. Why This Trial Matters Statistically

This trial illustrates why an estimated group difference should be read together with its uncertainty. The mean difference of 1.83 days is easy to interpret as a point estimate, but the reported interval extends from -5.273 to 8.933 days. That range makes clear that the point estimate alone does not provide a precise account of the contrast. Focusing only on the direction of the mean difference would overstate what the posted data can resolve.

The endpoint also shows why measures of spread matter. The arm-specific standard deviations, 18.86 and 16.64 days, indicate substantial variation in drinking-day counts. When an outcome varies considerably across participants, a group mean can conceal heterogeneous individual patterns. The registry record does not provide distributions or participant-level results, so the summaries cannot show how the observed values were shaped.

Another statistical lesson is the difference between a study result and a planning calculation. The sample-size figures are illustrations based on fractions of the observed effect under stated assumptions; they are not the trial's original design rationale or a measure of the trial's observed power. Keeping those purposes separate avoids treating a planning scenario as evidence about the results.

Finally, a registry record can be incomplete for interpretation even when it posts a primary comparison. Here, secondary results and analysis entries are absent, and important details about missing data, analysis populations, multiplicity, and interval construction are not reported. Those gaps do not erase the posted findings, but they define how far statistical conclusions can responsibly extend.

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 primary result is a mean comparison, but the registry record does not provide participant-level data or distributional summaries beyond the arm means and standard deviations. The standard deviations are large relative to the means, so the group summaries alone cannot reveal whether values were symmetric, concentrated, or influenced by unusually high counts. No distributional checks or alternative summaries are reported.

The record does not report the method used to calculate the interval, its confidence level, or the statistical test underlying the unadjusted p-value. It also does not provide an adjusted analysis, sensitivity analyses, or model details. The posted estimate should therefore be interpreted as the reported comparison rather than as a fully documented analysis.

Details about missing outcome data, adherence, and the analysis population are not reported in the registry record. It is also not clear from the posted results whether all participants with randomized assignments contributed outcome data or whether any post-randomization exclusions occurred. These uncertainties matter for judging how the observed summaries relate to the randomized comparison.

The objective includes PTSD symptoms, but no PTSD symptom findings or secondary outcome results are posted. The adverse-event record gives affected-participant counts but no serious-event terms, other-event descriptions, or reported death counts. These limits leave important aspects of efficacy and safety uncharacterized in the posted record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Prazosin for Treatment of Patients With Alcohol Dependence (AD) and Post Traumatic Stress Disorder (PTSD). — NCT00744055.
• Registry record last updated 2020-03-19; status last verified 2020-03.
• Results first posted 2016-01-18.
• 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

The registry record describes a completed, randomized, triple-masked, parallel-group phase 3 trial sponsored by Yale University. It lists enrollment of 96 and compares prazosin 16mg/day with placebo in patients with alcohol dependence and post-traumatic stress disorder. The primary endpoint is mean number of drinking days over 12 weeks, assessed using the Timeline Follow Back method.

The posted means are 11.04 days for prazosin and 9.21 days for placebo. The reported mean difference is 1.83 days, with interval limits from -5.273 to 8.933 days and an unadjusted p-value of 0.614. The interval spans differences in either direction, so the estimate is not precise about the direction or size of the contrast.

The record does not report secondary outcome results or analysis entries. It also does not report the trial's original sample-size rationale, detailed statistical plan, missing-data methods, analysis populations, or the interval's confidence level and construction method.