Pilot Study of Pharmaceutical and Behavioral Interventions to Treat Anxiety Disorders: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pilot Study of Pharmaceutical and Behavioral Interventions to Treat Anxiety Disorders (NCT01631682), a randomized phase 4 study. The primary outcome is Change From Baseline Skin Conductance Response.

2. Clinical Question

3. Trial Design

The registry identifies this as a randomized, parallel-group study with no masking and a listed phase of PHASE4. The sponsor was Massachusetts General Hospital. The study is marked completed, with enrollment of 186. Its title describes a pilot study of pharmaceutical and behavioral interventions to treat anxiety disorders, while the official title focuses on the psychophysiology of delayed extinction and reconsolidation in humans.

The brief summary describes a four-day fear-conditioning and delayed-extinction protocol. A conditioned stimulus was paired with an aversive unconditioned stimulus, described as potentially an electric shock, until the conditioned stimulus elicited a conditioned response. Skin conductance response was measured across different phases. The stated research aim was to test interventions intended to affect reconsolidation, and to develop a paradigm that would allow detection of partial changes in conditioned response rather than a floor effect. These are aims and hypotheses, not evidence that an intervention produced a particular effect.

The listed intervention arms were propranolol, reactivation with time delay, mifepristone, and intranasal oxytocin. The arm descriptions specify the dose and timing for each drug arm and the timing of reactivation and extinction in the behavioral arm. The registry describes reactivation with time delay as the procedure for participants not receiving propranolol on visit 2. Further details about allocation across all arms, visit scheduling beyond the described procedures, and any additional treatment or control conditions are not reported in the registry record.

The primary outcome was a continuous mean skin conductance difference score in microSiemens at 48hrs. The posted outcome summaries include 24 participants in the propranolol arm, 21 in reactivation with time delay, 13 in mifepristone, and 5 in intranasal oxytocin. The analysis population and reasons that these outcome counts differ from the overall enrollment are 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: 186 individuals initially signed a consent form. Of these 186 individuals who gave signed consent, 14 had unmeasureable skin conductance levels and were not able to participate in the study, leaving 172 individuals who participated.

Overall Study

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 Change From Baseline Skin Conductance Response, measured over 48hrs.

The primary endpoint is the change-from-baseline skin conductance response, expressed as a difference score in microSiemens. The registry defines the score as response to the treated conditioned stimulus (CS+R) minus response to a non-treated conditioned stimulus (CS+N), using day-three data. Day three is 48hrs after fear conditioning and is identified as the primary measure for assessing whether treatment affected the conditioned response.

Skin conductance response is a physiological change in skin conductance after a stimulus. In this experiment, it serves as a measurable response to conditioned cues. Comparing responses to a treated and non-treated conditioned stimulus creates a within-participant contrast, while comparing the resulting mean scores across arms addresses whether the groups differed on this laboratory measure.

The outcome is continuous, and the registry reports arm means and standard deviations. A positive or negative difference score reflects the direction of the contrast as defined by the registry, but interpretation depends on the experimental conditions and the score’s construction. The record does not report a clinically meaningful threshold for this measure. It also does not establish that a change in this laboratory response corresponds to a change in anxiety symptoms, daily functioning, or another patient-centered outcome. The primary endpoint therefore answers a focused psychophysiological question within the fear-conditioning task.

7. Results Posted in the Registry

• Reactivation With Time Delay: mean 0.17 versus 0.06 on control; difference 0.11 (95% CI -0.12 to 0.34), standardized effect d = 0.28; z-test p = 0.353.
• Mifepristone: mean -1.27 versus 0.06 on control; difference -1.33 (95% CI -2.25 to -0.41), standardized effect d = -1.30; z-test p = 0.005.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean difference describe?

A mean is the average value of a continuous outcome within a group. Here, the arm means summarize the day-three skin conductance difference score. A mean difference compares two group averages, with the direction depending on which arm is subtracted from the other. The recalculations use propranolol as the reference: the reactivation-with-time-delay mean is compared with the propranolol mean, and the mifepristone mean is compared with the propranolol mean. The resulting estimates are unadjusted comparisons based on the posted summary statistics, not estimates adjusted for participant characteristics or other variables.

The outcome is a difference score already defined from two conditioned-stimulus responses. It should not be confused with a difference between the intervention arms themselves. The arm-level mean difference score describes the average within-participant contrast; the recalculated mean difference describes the contrast between arm averages.

Why does the standard deviation matter?

A standard deviation describes how dispersed participant measurements are around the group mean. A larger standard deviation indicates more variability in individual scores, which generally makes a group mean less precise when the number of observations is held fixed. In the posted summaries, standard deviations are 0.34 for propranolol, 0.44 for reactivation with time delay, 1.68 for mifepristone, and 0.27 for intranasal oxytocin. The mifepristone standard deviation is notably larger than the other listed values, so its participants’ scores are more dispersed in this sample. This is descriptive and does not explain why the variability differs.

For an unadjusted comparison of two independent groups, the recalculation combines information about the group means, standard deviations, and sample sizes. The pooled standard deviation is a scale estimate used in the standardized effect calculation. The standardized effect expresses a mean difference relative to that pooled variability. It can help describe magnitude across scales, but it does not supply a clinical interpretation or remove limitations from small groups and variable data.

What does a p-value tell the reader?

In general, a p-value describes how incompatible the observed data are with a specified null hypothesis, under the assumptions of the statistical test. It is not the probability that the null hypothesis is true, and it does not measure the probability that a result will replicate. A p-value also does not describe effect size or clinical importance. The recalculated values are unadjusted, so they do not account for testing multiple interventions or outcomes. The sponsor’s posted p-value entries are threshold statements rather than exact values.

The registry lists ANOVA for its reported analyses, but details of the model and contrasts are not reported in the registry record. The recalculated pairwise comparisons are not the same as a full description of the sponsor’s analysis. Because the record identifies a discrepancy, the sponsor’s own analysis takes precedence when describing the trial’s reported inferential result. The recalculations can still be read as a separate, simple check based on posted arm summaries.

Why can over-dispersion matter?

In general, over-dispersion means that observed variability is greater than a model assumes. If a model understates variability, its standard errors can be too small and its intervals too narrow. The record does not establish that over-dispersion was assessed or present for this continuous endpoint. The relatively large mifepristone standard deviation is a reason to attend to dispersion in the descriptive data, not a basis for diagnosing a modeling problem.

What cannot be determined from the analysis details?

The registry does not report the sponsor’s full statistical analysis plan, the ANOVA specification, covariate adjustment, handling of missing data, analysis population, or any multiplicity procedure. It also does not report whether interim analyses were planned or conducted, or the trial’s original sample-size justification. These omissions prevent a full assessment of whether the posted analyses followed a prespecified plan. The recalculations are limited to unadjusted estimates from the available summary statistics and should not be mistaken for the sponsor’s complete analysis.

10. Confidence Intervals and Precision

A confidence interval expresses a range of values compatible with an estimate under the method used to construct it. In general, a narrower interval indicates greater precision, while a wider interval indicates more uncertainty. An interval that includes zero for a mean difference is compatible with no difference as well as with differences in either direction; an interval entirely on one side of zero indicates that the interval does not include a zero difference. These statistical descriptions do not determine whether a difference is important in clinical or experimental terms.

The recalculated interval for reactivation with time delay versus propranolol includes zero. The recalculated interval for mifepristone versus propranolol lies below zero. The registry does not report a confidence level for these recalculated intervals, so no confidence percentage is assigned. These are unadjusted recalculations, not sponsor-reported intervals. The registry’s reported analyses do not include interval estimates.

Precision depends on both variability and the number of observations. The posted outcome counts are 24 for propranolol, 21 for reactivation with time delay, 13 for mifepristone, and 5 for intranasal oxytocin. The standard deviations differ across arms, with the largest listed value in the mifepristone arm. The small number of outcome observations in some arms and the differing spread of scores limit how precisely their means and contrasts can be estimated.

The posted intervals do not account for uncertainty about missing data, the analysis population, or multiplicity. No interval or recalculated contrast is posted for intranasal oxytocin. The sponsor-recalculation discrepancy also matters: the recalculated intervals are not a replacement for the sponsor’s analysis, which takes precedence according to the record. Interpretation should therefore distinguish descriptive arm estimates, recalculated unadjusted comparisons, and sponsor-reported threshold results.

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

The 80% power column gives the illustrated number of participants per arm for an 80% chance of detecting the specified effect under those assumptions, if that effect and the assumptions were correct. The 90% power column gives the corresponding number per arm for a 90% chance. These are prospective planning quantities, not evidence about the probability that the completed trial detected an effect and not observed power.

For reactivation with time delay compared with propranolol, the illustration lists 198 participants per arm for 80% power and 264 for 90% power when planning for 100% of the observed effect. For mifepristone compared with propranolol, the corresponding figures are 10 and 13. The illustrations also show the consequences of planning for 75% and 50% of the observed effect. In both comparisons, the smaller effect assumptions require more participants per arm than the full observed-effect assumption. This is a general consequence of needing more information to distinguish a smaller signal from sampling variability.

These figures should be interpreted cautiously because the observed effects come from small outcome groups and the registry flags a discrepancy with the sponsor’s analysis. A planning illustration based on an observed estimate can be especially sensitive to that estimate. It does not establish the sample size that would be appropriate for a future study without a justified target effect and a prespecified design.

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 = 0.110; pooled standard deviation σ = 0.390.
2. n = 2 × (1.96 + 0.842)² × 0.390² / 0.110² = 197.0, rounded up to 198 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 <- 0.11                      # observed difference in means
sigma <- 0.38971665318208                     # 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(1631682)
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 = 198 (80% power), 264 (90% power)
 75% of observed effect: n per arm = 351 (80% power), 469 (90% power)
 50% of observed effect: n per arm = 789 (80% power), 1,056 (90% power)
power.t.test: 198 
Simulated power at n = 198 per arm: 0.798 (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. Analyses Reported by the Sponsor

The registry’s own analysis entries list ANOVA for each arm. They report p > .05 for propranolol, reactivation with time delay, and intranasal oxytocin, and p < .05 for mifepristone. The record does not give the corresponding ANOVA estimates, exact p-values, confidence intervals, or enough model detail to reconstruct those analyses.

The recalculated results are simple, unadjusted comparisons of posted means and standard deviations for reactivation with time delay versus propranolol and mifepristone versus propranolol. Their estimates and p-values do not constitute a reproduction of the sponsor’s full analysis. The record explicitly flags a discrepancy between the sponsor’s results and the recalculations. In that situation, the sponsor’s own analysis takes precedence for reporting the trial’s analysis result.

The recalculations may help readers understand how the posted summary statistics behave under a straightforward two-group comparison, but they cannot explain why the analyses differ. The registry does not report enough information to determine whether the discrepancy reflects different model specifications, populations, assumptions, or another cause. The reported statistical threshold should therefore be kept distinct from the recalculated pairwise estimates.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists four intervention arms and one primary outcome measure, and it includes four sponsor-reported analysis entries. The analyses are described as ANOVA, but details of the comparisons and model are not reported in the registry record. No multiplicity adjustment or procedure for controlling error across multiple comparisons is reported. The recalculated comparisons are unadjusted, so their p-values do not account for the fact that more than one arm or contrast is being examined.

The registry identifies one outcome measure and no posted secondary outcomes. This does not establish that no other analyses were planned or performed; the analysis plan and any additional outcomes are not reported in the registry record. Interim analyses and stopping rules are also not reported. The original sample-size justification is not reported, so the relationship between enrollment planning and the reported analyses cannot be assessed.

The outcome counts in the arm summaries are smaller than overall enrollment, but the registry does not explain the difference. It does not report how missing endpoint values were handled, whether participants were excluded from analyses, or whether any imputation or sensitivity analysis was used. The analysis population—such as whether results were based on all randomized participants or another subset—is not reported in the registry record.

These gaps affect interpretation of both the estimated effects and their uncertainty. Missing data can change the composition of the analyzed groups, while choices about analysis populations can affect the estimand. Multiplicity procedures can affect how inferential evidence is interpreted across multiple comparisons. Without these details, readers cannot fully assess whether the sponsor’s analyses followed a prespecified plan or how robust the results are to alternative analytical choices.

15. 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 posted safety figures show zero serious adverse events among the participants at risk in each listed arm: 24 in propranolol, 21 in reactivation with time delay, 13 in mifepristone, and 5 in intranasal oxytocin. The record lists no serious adverse-event terms. These counts describe the registry’s reported safety results; they do not show that an event is impossible or establish comparative safety.

For other adverse events, the registry lists zero affected participants in propranolol, reactivation with time delay, and intranasal oxytocin. It lists 2 affected participants in mifepristone. The number at risk in that arm was 13. The record does not provide further details about event type, severity, timing, or attribution in the posted safety summary.

Deaths are not reported in the registry record. This should not be interpreted as a reported count of zero deaths. The arm sizes are small, particularly for intranasal oxytocin, so the absence of listed serious events cannot establish that the interventions have no safety risks. The safety counts should be read as descriptive observations from this study, not as a basis for clinical advice or a definitive safety comparison.

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

The results describe skin conductance responses measured during a fear-conditioning experiment. The arm means show the average difference between responses to treated and non-treated conditioned stimuli, as defined in the registry. The sponsor-reported analyses give threshold results for each listed arm, while the recalculations provide two unadjusted comparisons with propranolol. The record flags a discrepancy, so the sponsor’s analysis takes precedence.

These data do not establish that an intervention works or does not work. A p-value above a threshold does not prove that there is no effect, and a p-value below a threshold does not by itself establish a meaningful or reliable effect. Effect estimates, uncertainty, sample size, analysis choices, and the laboratory context all matter. The registry’s limited analysis details and absent missing-data information further restrict interpretation.

The primary outcome is not a direct measure of anxiety symptoms, posttraumatic stress symptoms, functioning, or quality of life. A difference in a physiological response during a conditioning task cannot by itself establish a change in those clinical outcomes. The registry also does not report a clinically meaningful threshold for the skin conductance score.

The trial was described as a pilot study intended to develop a paradigm for testing interventions and to examine partial conditioned-response reduction. That purpose is distinct from demonstrating clinical effectiveness. Its results can inform questions about measuring conditioned responses under the described experimental procedures, but they do not settle whether any intervention is beneficial in clinical practice or how the findings generalize beyond the studied task and participants.

The small and unequal outcome counts, varying standard deviations, absence of detailed participant characteristics, and lack of a full analysis plan all counsel against broad conclusions. The posted summaries are best read as limited evidence about this specific laboratory endpoint in the participants represented by the reported outcome counts.

17. Why This Trial Matters Statistically

This trial raises statistical questions that are common in experimental biomedicine. Its primary measure is a constructed contrast between responses to treated and non-treated conditioned stimuli. That distinction makes it important to separate the participant-level difference score from the between-arm mean difference. Clear definitions of the outcome and comparison are essential for interpreting direction and magnitude.

The study also illustrates the tension between a pilot objective and inferential claims. The brief summary emphasizes developing a paradigm that can avoid floor effects and detect partial response changes. Such methodological aims concern the behavior and sensitivity of the measurement procedure, while the posted arm comparisons concern estimates from the observed participants. A pilot dataset can be informative for designing later work, but estimates from small groups may be imprecise and sensitive to individual observations.

The posted data make the role of variability visible. The mifepristone arm has a larger standard deviation than the other listed arms, and the intranasal oxytocin outcome count is 5. These features affect precision and make it especially important not to read an observed mean as a stable estimate. Planning illustrations based on observed effects also depend strongly on how credible those effect estimates are.

Finally, the registry’s discrepancy flag highlights why transparent reporting of model specifications, analysis populations, missing-data handling, and multiplicity procedures matters. Without those details, a simple recalculation can provide context but cannot substitute for the sponsor’s analysis or explain differences between analyses. The case reinforces the value of reporting estimates and uncertainty alongside test thresholds.

18. 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 outcome results cover 24 participants in propranolol, 21 in reactivation with time delay, 13 in mifepristone, and 5 in intranasal oxytocin, compared with overall enrollment of 186. The registry does not explain which participants contributed to the primary endpoint or why the outcome counts differ from enrollment. This makes it difficult to assess the representativeness of the analyzed participants and the possible effect of missing outcomes.

The arms have unequal outcome counts and differing standard deviations. In particular, the mifepristone arm has a standard deviation of 1.68, and the intranasal oxytocin arm has only 5 participants in the posted results. Estimates from groups of these sizes can be unstable. The registry provides no detailed participant characteristics or information about baseline comparability, so the potential influence of group differences cannot be evaluated from the posted record.

The registry lists randomized allocation and no masking, but it does not report the allocation process, concealment details, or additional safeguards against bias. It does not provide a full statistical analysis plan, model specifications, missing-data methods, analysis populations, multiplicity procedure, interim-analysis information, or original sample-size justification. The sponsor’s reported analyses also lack effect estimates and confidence intervals.

The recalculated comparisons are unadjusted and available for only two intervention arms against propranolol. They cannot explain the discrepancy flagged in the record, and the sponsor’s own analysis takes precedence. No recalculated contrast is listed for intranasal oxytocin. The endpoint is a laboratory physiological response, with no reported clinically meaningful threshold or direct clinical outcome. Together, these limits constrain both the precision of the comparisons and their relevance to broader clinical questions.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Pilot Study of Pharmaceutical and Behavioral Interventions to Treat Anxiety Disorders — NCT01631682.
• Registry record last updated 2017-02-23; status last verified 2017-01.
• Results first posted 2017-02-23.
• 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.

22. Record Summary

NCT01631682 is a completed, randomized, unmasked, parallel-group study with listed phase PHASE4 and enrollment of 186. It examined propranolol, reactivation with time delay, mifepristone, and intranasal oxytocin in a fear-conditioning and delayed-extinction protocol. The primary outcome was a day-three skin conductance difference score in microSiemens, measured 48hrs after fear conditioning.

The posted outcome counts range from 5 to 24 participants across the four arms. The sponsor-reported analyses list p < .05 for mifepristone and p > .05 for the other three arms. Two unadjusted recalculations compare arms with propranolol, and the registry flags a discrepancy; the sponsor’s analysis takes precedence. Details on missing data, analysis populations, multiplicity, and the original sample-size justification are not reported in the registry record.