Testing Doxazosin to Treat Stress Mechanisms in Alcoholism: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Testing Doxazosin to Treat Stress Mechanisms in Alcoholism (NCT02989493), a randomized phase 2 study. The primary outcome is Startle Potentiation During Stress Reactivity Task.

2. Clinical Question

The study focuses on stress-related responses rather than reporting a single clinical outcome in the posted primary results. Its task included no-shock, predictable-shock, and unpredictable-shock conditions. Predictable shocks occurred at a consistent, known time; unpredictable shocks were fully unpredictable. The registry explains that higher startle potentiation means a higher stress-reactivity response for the condition being assessed.

The central statistical question would be whether the measured responses differed between the randomized treatment groups, and how precisely any difference was estimated. However, the registry record does not provide the group means, standard deviations, effect estimate, confidence interval, or test result for this outcome. It also does not reduce the several task categories or time points to one comparison. The research question can therefore be described, but the posted record does not establish the magnitude or precision of a treatment-group difference.

3. Trial Design

This was a randomized, parallel-group phase 2 trial sponsored by the University of Wisconsin, Madison. The registry lists 61 participants as enrolled and gives the condition as alcoholism. The trial is marked completed. Its recorded start date is 2017-04-12, and its primary completion date is 2020-03-13.

Participants were assigned to doxazosin or placebo. The doxazosin arm description specifies 8 weeks of treatment with an 8mg target dose; the comparator was matched placebo for 8 weeks. The registry lists masking as triple. Its brief summary describes the study as double-blind and placebo-controlled. The record does not explain which roles were masked or clarify the difference between those masking descriptions.

The primary outcome was startle potentiation during a stress-reactivity task, assessed at 4 weeks. That assessment was distinct from the full 8-week treatment period. The registry says all 61 participants were able to provide TLFB data through the end of 8 weeks, but the task required attendance at a 4-week study visit. It reports that 24 participants did not complete that visit and therefore had missing data for the primary outcome. The record's accompanying parenthetical counts and trailing text do not provide a clear accounting of all those missing observations.

The registry lists two outcome measures and zero analyses. It does not report the randomization procedure, allocation concealment, the detailed masking procedures, or the planned statistical analysis. A sample-size justification and any interim-analysis plan are not reported in the registry record. The enrollment count describes the trial overall; it should not be treated as the number with observed primary-outcome data.

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: 61 participants were consented, passed medical screening, and were assigned to a drug group.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Although all 61 participants were able to provide TLFB data through the end of 8 weeks, the NPU task required that they attend a study visit at 4 weeks. 24 participants (10 in doxazosin, 13 in placebo) did not complete that study visit and so have data missing for this outcome measure. 1 (doxazosin) attended the visit but chose not to complete the NPU task. 2 participants (both placebo) completed NPU at this visit, but their NPU data was unusable for this analysis due to technical issues.

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.

6. Primary Endpoint

The registry lists the primary outcome as Startle Potentiation During Stress Reactivity Task, measured over 4 weeks.

The primary endpoint is titled “Startle Potentiation During Stress Reactivity Task.” It is a continuous measure reported in microvolts, with the mean and standard deviation specified as the summary statistics. The recorded timeframe is 4 weeks.

The measure came from the no-shock, predictable-shock, and unpredictable-shock conditions of the NPU task. The registry describes the NPU task as a commonly used, well-validated laboratory stressor task for studying anxiety and fear. In its predictable condition, shocks were fully predictable and occurred at a consistent, known time. In its unpredictable condition, shocks were fully unpredictable. The no-shock condition is also listed, but further details about its task procedures are not reported in the registry record.

According to the record, higher startle potentiation represents a higher stress-reactivity response for the condition being considered. This makes the outcome a laboratory response measure, not a direct count of drinking behavior or a report of long-term clinical outcomes. The record does not report the actual means or standard deviations by treatment group, nor does it provide a treatment contrast or uncertainty interval. It also notes that the endpoint involved several task categories or time points and was not reduced to one comparison. As a result, the endpoint's purpose and scale are described, but the posted primary results do not show how the groups performed.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The record identifies a continuous primary outcome summarized using a mean and standard deviation. These are descriptive statistics: the mean gives a measure of the group’s central value, while the standard deviation describes how dispersed individual measurements are around that mean. They do not, on their own, quantify the uncertainty in a comparison between treatment groups. For that, an analysis would ordinarily report an estimated contrast and an uncertainty measure, such as a confidence interval. No such primary-outcome analysis is reported in the registry record.

What does a confidence interval express?

As a general statistical explanation, a confidence interval gives a range of values compatible with the data and the method used to calculate it. A narrower interval indicates greater precision than a wider interval, while the interval's position indicates which effect values remain compatible with the data. A confidence interval is not a guarantee that the true effect lies within a particular computed range. Because the registry does not post an effect estimate or interval for this endpoint, neither the direction nor the precision of a group difference can be assessed.

Why does the task structure matter?

The NPU task includes no-shock, predictable-shock, and unpredictable-shock conditions. A statistical comparison could be defined in several ways, such as comparing groups within a particular condition or assessing a pattern across conditions. These are not interchangeable questions. The registry says that the primary outcome has several categories or time points and was not reduced to one comparison. It does not report which contrast, if any, was used for an analysis. Without a defined contrast and the corresponding estimates, a reader should not treat the endpoint title as evidence of one specific treatment effect.

General statistical practice distinguishes a prespecified primary contrast from multiple possible comparisons. When many outcomes, conditions, or time points are examined, the chance of finding an apparently unusual result among them can increase if each is tested separately without an appropriate plan. Methods for addressing multiplicity depend on the number and relationship of the comparisons and the intended inferential question. The registry record does not describe a multiplicity procedure or an analysis plan for the task conditions. It therefore does not support a claim about how multiplicity was handled.

What can the sample-size illustration establish?

The sample-size assumptions describe a planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. The illustration is described as detecting a stated fraction of the observed effect. This is not the trial's own design justification, and its values must not be interpreted as observed power. A sample-size calculation is a conditional planning exercise: it depends on the effect size and statistical assumptions chosen. It does not show whether the trial actually observed an effect, nor does it replace an estimate and confidence interval from the trial.

The record does not report the trial's sample-size justification. It also does not provide the primary outcome estimate needed to assess any planning illustration based on an observed effect. The illustration therefore cannot fill the gap in the posted results or establish how precisely the trial estimated the task outcome.

How should missing observations be understood?

In general, missing outcome data reduce the information available for an analysis and may affect how representative the analyzed participants are of everyone enrolled. The consequences depend on why data are missing and on the assumptions and methods used to handle them. A complete-case analysis, for example, uses participants with observed outcomes; other methods can use additional information but require their own assumptions. The registry reports missed 4-week visits but does not say what missing-data method was used, whether any imputation was performed, or what analysis population was intended.

Randomization is a design feature intended to support a fair comparison between assigned groups. It does not itself resolve missing outcome data, define the treatment contrast, or guarantee that a particular statistical analysis is unbiased. Details such as allocation procedures, covariate adjustment, model choice, and handling of repeated or condition-specific measurements are not reported in the registry record. Since there are zero listed analyses and no numerical primary results, the methods that would be needed to interpret an estimated treatment contrast cannot be checked against posted analysis output.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome involves several NPU task conditions or time points, and the registry explicitly indicates that it was not reduced to one comparison. The task description names no-shock, predictable-shock, and unpredictable-shock conditions. Without a single reported contrast, there is no one primary numerical comparison to interpret. The registry record does not report whether the conditions were analyzed separately, jointly, or under another plan, and it does not describe any adjustment for multiple comparisons.

Missingness is also relevant. The record says 24 participants did not attend the 4-week visit required for the task outcome. It lists 10 participants in the doxazosin group and 13 in the placebo group, then includes the incomplete fragment “1 (doxazosin).” Those details do not clearly reconcile the stated total, so the group-specific missing-data accounting is uncertain from the record. The report that all 61 participants could provide TLFB data through 8 weeks concerns a different measure and does not mean the task outcome was observed for all participants.

The registry does not report whether the analysis used only participants with observed task data, imputed missing values, or applied another method. The intended analysis population—such as all randomized participants or a subset defined by outcome availability—is not reported in the registry record. The reasons for missed visits beyond noncompletion, and any sensitivity analyses exploring the effect of missing data, are also not reported. These omissions prevent assessment of whether missing observations or multiple task comparisons could affect the interpretation of a primary-outcome analysis.

11. 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 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events for 0 of 29 participants at risk in the doxazosin group and 0 of 32 participants at risk in the placebo group. It also reports 0 deaths in each group. These are counts recorded in the registry; they do not establish that no adverse experiences of any kind occurred.

For other adverse events, the record lists 24 of 29 participants in the doxazosin group and 28 of 32 participants in the placebo group. These figures are counts, not a measure of severity or a comparison adjusted for differences in exposure or follow-up. The record does not list the terms for serious adverse events, and it does not provide details about the types, timing, severity, or duration of the other events in the information available for this analysis.

Safety counts should be read separately from the primary stress-task outcome. They do not provide evidence about the task response, and they do not explain the missing 4-week task visits. The reported denominators are the numbers at risk for the adverse-event summaries; they should not be substituted for the number of participants with observed primary-outcome data. These descriptive counts alone do not establish a causal safety difference between groups.

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

The registry record describes a randomized comparison of an 8-week doxazosin regimen with matched placebo and specifies a 4-week laboratory stress-reactivity endpoint. It explains the measure's units, summary form, task conditions, and direction: higher startle potentiation indicates a higher stress-reactivity response in the condition assessed. Those details make the intended outcome understandable.

What the posted record does not show is the primary result. It contains no group means, standard deviations, effect estimate, confidence interval, or p-value for startle potentiation. It also does not identify one comparison across the task's several categories or time points. Accordingly, the posted data do not quantify a treatment-group difference or its precision. They cannot support a conclusion about whether responses differed or about the size or direction of any difference.

The trial's focus on stress mechanisms does not make a task measure equivalent to a clinical outcome. The trial summary refers to stress reactivity and clinical outcomes, but the registry information contains no numerical primary outcome analysis and no clinical-outcome comparison. It would be inappropriate to infer a clinical benefit or lack of benefit from the design, from the endpoint description, or from adverse-event counts.

Nor does the enrollment count alone indicate how informative the endpoint analysis was. The record reports missed 4-week visits and does not fully reconcile its accompanying missing-participant details. It does not describe how missing data were handled, which analysis population was used, or how task conditions and time points were compared. These gaps limit what can be concluded from the registry record, independent of the trial's randomized design.

The defensible reading is therefore narrow: the registry documents a completed randomized trial and a defined stress-reactivity measure, but the posted primary results do not allow a statistical assessment of the treatment comparison. The absence of reported estimates is not evidence that the groups were the same; it means the comparison's result is not reported in the registry record.

13. Why This Trial Matters Statistically

This trial raises statistical questions that are common in studies using laboratory measures of behavior or stress response. A continuous task measure can be described with means and standard deviations, but meaningful group interpretation requires a clearly specified contrast and uncertainty around its estimate. Here, the endpoint includes multiple task conditions or time points, yet the registry does not reduce them to one comparison. That makes the distinction between the measurement and the inferential question especially important.

The study also illustrates why a randomized design does not remove every challenge in analysis. Random assignment establishes the intended basis for comparing groups, but the primary task required attendance at a 4-week visit. The record reports that some participants did not complete that visit and does not explain the missing-data method or analysis population. Readers therefore need both the design information and the outcome-analysis details to understand how an estimate was obtained.

Another statistical distinction concerns planning and results. A sample-size illustration based on an observed effect and specified assumptions is not the same as a trial's original sample-size justification, nor does it establish observed power. It cannot replace the actual effect estimate and its precision. The registry identifies assumptions for such an illustration but does not report the trial's own design justification or the primary outcome analysis.

These points matter because the posted record can describe a study's question without answering it numerically. Keeping those two things separate avoids treating an endpoint definition, enrollment count, or study design as if it were evidence about the treatment contrast. In this case, the primary statistical result is not reported in the registry record.

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.

The main limitation for interpreting this trial statistically is that the registry does not report numerical primary-outcome results. There are no group-level means or standard deviations, no estimated contrast, no confidence interval, and no test result. The record lists zero analyses. As a result, the magnitude and precision of any treatment-group difference in startle potentiation cannot be assessed.

The outcome itself includes several NPU task conditions or time points, but the registry does not specify one comparison for interpretation. It does not report whether the conditions were evaluated separately or jointly, or whether any multiplicity procedure was used. The analysis plan, statistical model, and choice of analysis population are also not reported in the registry record.

Availability of the primary outcome was limited by the need to attend a study visit at 4 weeks. The registry says 24 participants missed that visit, but the accompanying group counts and trailing “1 (doxazosin)” fragment do not clearly account for the stated total. The record does not explain reasons for missed visits, the missing-data method, or whether sensitivity analyses were conducted. The reported ability of all 61 participants to provide TLFB data through 8 weeks does not resolve missingness for the task endpoint.

The record lists triple masking, while its brief summary describes a double-blind trial; it does not explain the difference or identify masked roles. It also does not report randomization procedures, allocation concealment, or a sample-size justification. These reporting gaps do not establish that the procedures were absent, but they limit what can be verified from the registry record.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Testing Doxazosin to Treat Stress Mechanisms in Alcoholism — NCT02989493.
• Registry record last updated 2021-03-18; status last verified 2021-03.
• Results first posted 2021-03-18.
• 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.

18. Record Summary

NCT02989493 is a completed phase 2 randomized, parallel-group trial sponsored by the University of Wisconsin, Madison. It enrolled 61 participants and compared 8 weeks of doxazosin, with an 8mg target dose, against 8 weeks of matched placebo. The registry lists triple masking, while the brief summary calls the trial double-blind.

The primary endpoint was startle potentiation during a stress-reactivity task at 4 weeks, measured in microvolts and described using means and standard deviations. The registry says the task required a study visit and reports missing task data among participants who did not attend. It lists two outcome measures and zero analyses. Numerical primary outcome results, a single comparison across task conditions or time points, and details of the analysis are not reported in the registry record.

The adverse-event summary records zero serious adverse events and zero deaths in both groups, alongside counts for other adverse events. Those descriptive safety counts do not substitute for the unreported primary outcome analysis.