A Study in Pediatric Participants With Generalized Anxiety Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study in Pediatric Participants With Generalized Anxiety Disorder (NCT01226511), a randomized phase 3 study. The primary outcome is Change From Baseline to 10-Week Endpoint in the Pediatric Anxiety Rating Scale (PARS) Severity Score Evaluated for Symptoms Identified on the Generalized Anxiety Subsection of the PARS Symptom Checklist.

2. Clinical Question

This is a treatment comparison, not a comparison of duloxetine with another active medicine. The central statistical question concerns the difference between groups in average change, rather than whether either group changed from baseline in isolation. The registry record identifies the study as randomized, double-blind, and Phase 3. It records 281 participants enrolled, while the primary analysis population is defined by the availability of baseline and post-baseline PARS measurements and the stated site exclusion. The population actually used for each outcome therefore matters when interpreting any reported estimate.

The primary question is specific to the blinded acute-treatment period and the designated PARS score. It does not, by itself, answer whether changes persisted during the extension or apply to people outside the trial population. Other outcomes address response, other symptom and functioning measures, and safety; they provide related but distinct information. Eligibility details beyond the broad pediatric GAD population are not reported in the registry record.

3. Trial Design

The registry record describes a randomized, parallel-group Phase 3 study with quadruple masking. The title identifies a double-blind comparison, while the design record uses the broader masking label; the identities of the parties masked are not reported in the registry record. Participants were allocated to duloxetine or placebo for the 10-week acute-treatment period. Duloxetine used flexible dosing from 30-120 mg once daily by mouth; placebo was administered once daily.

After the blinded treatment period, the registry describes an 18-week extension. Participants assigned to duloxetine could participate in the extension, and placebo participants received duloxetine during it. The primary outcome, however, is measured from baseline to the 10-week endpoint, so the main comparison concerns the acute-treatment phase rather than the extension. The registry record lists the study as completed and records 281 participants enrolled. Start and primary completion are listed as 2011-06 and 2013-02.

The primary analysis population is described as randomized participants with a baseline and at least 1 post-baseline PARS severity score during acute treatment, with 9 participants at 1 site excluded because of major quality issues. The reported arm counts for the primary outcome are 135 for duloxetine and 133 for placebo. These counts describe that analysis, not necessarily every population used for every safety or secondary outcome.

The registry record does not report the original sample-size justification, the planned allocation ratio, the identities of masked parties, or detailed randomization procedures. It also does not report an interim-analysis plan. These details should not be inferred from the observed group sizes or from the methods used to analyze the primary endpoint.

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: This study had 4 periods: Screening period (1-week), acute treatment period (10-week, double-blind period with flexible duloxetine dosing), extension treatment (18-week period, of which 16 weeks were open-label treatment with flexible duloxetine dosing), and a taper period (2 weeks recommended at discontinuation from study any point after Week 2).

Acute Treatment Period

Reasons for not completing this period:

Extension Treatment Period

Reasons for not completing this period:

Taper Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Randomized participants with a baseline and at least 1 post-baseline PARS severity score for GAD during the acute treatment period, excluding 9 participants from 1 site with major quality 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.

Population: All randomized participants, excluding 9 participants from 1 site with major quality issues.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline to 10-Week Endpoint in the Pediatric Anxiety Rating Scale (PARS) Severity Score Evaluated for Symptoms Identified on the Generalized Anxiety Subsection of the PARS Symptom Checklist, measured over Baseline, 10 weeks.

The primary endpoint is change from baseline to the 10-week endpoint in the PARS severity score for symptoms identified in the generalized anxiety subsection of the PARS symptom checklist. The endpoint summarizes symptom severity, impairment, and interference. Its score is derived by summing 5 of 7 severity, impairment, and interference items: items 2, 3, 5, 6, and 7. Each item ranges from 0 (none) to 5 (extreme), and the resulting score ranges from 0 (none) to 25 (extreme severity). A score of 15 indicates moderate illness severity.

Change is used to describe how the measured symptom burden shifts during treatment. Because the outcome is measured at baseline and at the 10-week endpoint, the comparison focuses on average change over the acute-treatment period. A negative change represents a reduction in the score, and the group difference indicates how the average change in the duloxetine group compares with the average change in the placebo group. The endpoint does not directly measure every aspect of GAD or functioning, so the other registered outcomes address complementary questions.

The registry reports least-squares means from a mixed-effects model for repeated measures. These are model-adjusted average changes, not simple raw averages. The adjustment variables are described in the statistical interpretation and methods discussion. The registry record does not report a minimum clinically important difference for this score; the observed numerical contrast should therefore be interpreted as a difference on the scale, not as proof of a particular level of clinical importance.

7. Results Posted in the Registry

• Duloxetine (Acute Treatment): mean -9.70 versus -7.05 on control; difference -2.65 (95% CI -4.04 to -1.26), standardized effect d = -0.46; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Response Rate at Endpoint for Generalized Anxiety Disorder (GAD) Using Pediatric Anxiety Rating Scale (PARS) Severity Score for GAD

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Duloxetine (Acute Treatment): 51.1% (69/135) versus 36.8% (49/133) on control; risk difference 14.3 percentage points (95% CI 2.5 to 26.0), risk ratio 1.39 (1.05–1.83); unadjusted two-proportion z-test p = 0.019.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Duloxetine (Acute Treatment) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline to 10-Week Endpoint on the Pediatric Anxiety Rating Scale (PARS) Severity Total Score Evaluated for All Symptoms Identified on the PARS Symptom Checklist Symptoms

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Duloxetine (Acute Treatment): mean -9.15 versus -6.36 on control; difference -2.79 (95% CI -4.11 to -1.47), standardized effect d = -0.50; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Duloxetine (Acute Treatment) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline to 10-Week Endpoint on the Clinical Global Impression of Severity (CGI-S) Scale

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Duloxetine (Acute Treatment): mean -1.93 versus -1.38 on control; difference -0.55 (95% CI -0.86 to -0.24), standardized effect d = -0.42; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Duloxetine (Acute Treatment) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted numeric secondary results cover a response outcome and additional symptom-severity outcomes. In broad direction, these results align with the primary finding: the duloxetine group has a higher reported response percentage and larger reported reductions on the additional continuous scales. These measures address related aspects of the clinical question, but they are distinct endpoints with different definitions and scales; they should not be treated as repetitions of the primary endpoint.

The registry record lists 13 secondary outcomes, including response, remission, functioning, global severity, treatment-emergent suicidal ideation and behavior, and a later PARS assessment during the extension. Numerical results for outcomes beyond those with posted numeric results are not reported in the registry record. The list of planned or registered measures therefore should not be mistaken for a complete set of numerical findings.

Unadjusted secondary comparisons carry less evidential weight than the designated primary result when many outcomes are examined. In general, testing multiple outcomes increases the chance that at least one comparison will appear statistically notable even if there is no true difference for that outcome. A multiplicity procedure can address this issue, but one is not reported in the registry record. The available secondary results are therefore best viewed as supportive context, not as a set of independent confirmations.

The registry record does not report a hierarchy for testing secondary outcomes, adjustment of their p-values, or a single confirmatory interpretation across them. Response, symptom scores, functioning, safety measures, and extension outcomes also answer different questions and may have different analysis populations. Without a reported multiplicity plan and complete numerical results, the overall secondary pattern cannot establish the certainty of every individual outcome. It is most appropriately read alongside, rather than in place of, the primary PARS comparison.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Response Rate at Endpoint for Generalized Anxiety Disorder (GAD) Using Pediatric Anxiety Rating Scale (PARS) Severity Score for GAD
• Change From Baseline to 10-Week Endpoint on the Pediatric Anxiety Rating Scale (PARS) Severity Total Score Evaluated for All Symptoms Identified on the PARS Symptom Checklist Symptoms
• Change From Baseline to 10-Week Endpoint on the Clinical Global Impression of Severity (CGI-S) Scale
• Remission Rate at Endpoint for Generalized Anxiety Disorder (GAD) Using Clinical Global Impressions of Severity (CGI-S) Scale
• Change From Baseline to 10-Week Endpoint in the Children's Global Assessment Scale (CGAS)
• Percentage of Participants During the 10-Week Period With Treatment-Emergent (New or Worsening) Suicidal Ideation as Assessed by the Columbia-Suicide Severity Rating Scale (C-SSRS)
• Percentage of Participants During the 10-Week Period With Treatment-Emergent (New or Worsening) Suicidal Behavior as Assessed by the Columbia-Suicide Severity Rating Scale (C-SSRS)
• Change From 10-Week to 28-Week Endpoint in the Pediatric Anxiety Rating Scale (PARS) Severity Score Evaluated for Symptoms Identified on the Generalized Anxiety Subsection of the PARS Symptom Checklist

9. Statistical Interpretation

10. Statistical Methods Explained

The primary analysis used a mixed-effects model for repeated measures (MMRM). In general, this type of model analyzes measurements collected at multiple visits while accounting for their relationship within a participant. It can use participants' available outcome measurements without requiring every participant to have a complete set of visits. The method depends on assumptions about the outcome data and the pattern of missingness; the registry record does not report a missing-data sensitivity analysis or a specific imputation procedure.

What does the adjusted mean represent?

The registry reports least-squares (LS) mean changes. In general, an LS mean is a model-based estimate adjusted for variables included in the statistical model. It is not necessarily the same as the arithmetic average of observed participant scores. Here, the reported model included baseline, pooled investigator, age category, visit, treatment, treatment-by-visit, age-category-by-visit, and baseline-by-visit terms. This lets the estimated treatment contrast account for the listed factors and for how some of them relate to visit.

The treatment-by-visit term allows the modeled treatment contrast to vary across visits. The baseline-by-visit term allows the relationship between baseline score and outcome to vary across visits. In general, including such terms can reflect the repeated measurement structure and the planned adjustment, but the registry record does not provide the full model specification, covariance structure, estimation details, or diagnostic checks. Those details are not reported in the registry record.

How should the group contrast be read?

The primary result is a mean difference in change: the modeled change for duloxetine compared with the modeled change for placebo. The reported difference is -2.65. The negative sign follows the score-change coding; because reductions in PARS severity are negative changes, the estimate indicates a larger modeled reduction in the duloxetine group. It is not a difference in the number of participants responding, and it is not a ratio.

The registry also gives standard errors for the two arm-level LS mean changes. In general, a standard error describes the estimated sampling uncertainty of an estimate; it is not the spread of individual participant outcomes. The reported confidence interval for the group contrast is more directly useful for judging the precision of the estimated difference. The re-calculated comparison uses the posted group summaries and an approximate pooled standard deviation, so it is an unadjusted calculation from summaries rather than a reproduction of the sponsor's full repeated-measures model.

What does a p-value say?

In general, a p-value measures how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the statistical model were true. It does not measure the probability that a treatment has no effect, the probability that the result occurred by chance, or the size or importance of an effect. The sponsor reports p < 0.001 for the primary comparison. That finding is evidence against a zero treatment contrast under the reported model, but it must be read alongside the estimated difference, its confidence interval, the outcome definition, and the analysis population.

The sponsor's model-based analysis and the simple recalculation answer related but not identical statistical questions. MMRM adjusts for the listed covariates and repeated measurements; the summary-based recalculation does not recreate that adjustment. The estimates agree closely at the displayed precision, and the registry does not flag a discrepancy. The sponsor's model result is therefore the direct reported analysis for the endpoint; the recalculation offers a check based on posted summary statistics, not a replacement for the model.

Why can over-dispersion matter?

For general statistical context, binary outcomes are often modeled under assumptions about the variability of event counts. Over-dispersion means that the observed variability is greater than the model assumes. If it is not accounted for, standard errors can be too small and confidence intervals or p-values can overstate precision. This general issue is not a claim that over-dispersion occurred in this trial. The registry record does not describe an over-dispersion assessment or a specific model for the secondary response outcome.

For continuous outcomes, the analysis still depends on how well the model represents the repeated measurements and their uncertainty. The registry record does not report residual diagnostics, sensitivity analyses, or detailed handling of protocol deviations. Readers should distinguish the stated primary model and result from analytic choices that are not documented in the registry record.

11. Confidence Intervals and Precision

The sponsor reports a 95% confidence interval from -4.03 to -1.27 for the primary least-squares mean difference of -2.65. In general, a confidence interval describes a range of values compatible with the data and the statistical model at the stated confidence level. It communicates uncertainty around the estimate; it is not a range containing a fixed proportion of individual participant outcomes and does not guarantee that the true effect lies within those particular limits.

The interval lies entirely below zero, consistent with a larger mean reduction in the duloxetine group under the outcome's coding. Its width also shows that the exact size of the group difference is not known with certainty. Values nearer either limit remain compatible with the model-based analysis. The interval does not determine whether the difference is clinically important, since a threshold for clinically meaningful change is not reported in the registry record.

A recalculation from the posted summary statistics gives a 95% interval from -4.039 to -1.261. This is a simple, unadjusted re-calculation and not the sponsor's MMRM result. The sponsor's reported interval is based on the registered mixed-model analysis and should be used when describing that analysis. The small differences in displayed limits are consistent with the distinction between the methods and their reported rounding; the registry does not flag a discrepancy.

Precision is specific to this endpoint, time point, model, and analysis population. It does not describe uncertainty for every secondary outcome, nor does it address possible bias from missing data or site exclusion. The registry record does not report sensitivity analyses that would show how the interval changes under alternative assumptions.

12. 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 illustrative calculations based on the observed effect, not a reconstruction of the trial's original design. For the continuous and binary outcomes in these calculations, the figures represent participants per arm, not total events. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The calculations consider detecting the stated fraction of the observed effect.

The 80% power column gives the approximate number of participants per arm that the planning calculation says would be needed for an 80% chance of detecting the specified effect under those assumptions, if that effect and model assumptions were correct. The 90% power column describes the corresponding planning target for a 90% chance. These are design calculations, not guarantees about what a future study will find. Smaller effects require more participants to detect under the same assumptions, which is why the planning figures increase when the assumed effect is reduced from the full observed effect to a smaller fraction.

These figures are not observed power. Observed power is not a useful substitute for interpreting the estimated effect and its uncertainty after a study; the confidence interval is more informative about precision. The registry record does not report the assumptions or calculation used in the trial's original sample-size justification, so the illustrative planning values should not be taken as evidence about how the study was designed or whether it had adequate planned power.

13. 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 = -2.650; pooled standard deviation σ = 5.800.
2. n = 2 × (1.96 + 0.842)² × 5.800² / -2.650² = 75.2, rounded up to 76 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 <- -2.65                      # observed difference in means
sigma <- 5.79984196051214                     # 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(1226511)
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 = 76 (80% power), 101 (90% power)
 75% of observed effect: n per arm = 134 (80% power), 179 (90% power)
 50% of observed effect: n per arm = 301 (80% power), 403 (90% power)
power.t.test: 76.2 
Simulated power at n = 76 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.

14. Analyses Reported by the Sponsor

The sponsor's primary analysis is a mixed-model analysis of repeated measurements. It reports an LS mean difference of -2.65, a 95% confidence interval from -4.03 to -1.27, and p < 0.001. The summary-based recalculation also estimates a difference of -2.65 and gives an interval from -4.039 to -1.261. The registry does not flag a discrepancy between the calculations.

The two intervals are not identical because the recalculation is a simple, unadjusted calculation from posted summary statistics, whereas the sponsor's estimate comes from its specified repeated-measures model. The sponsor's reported model analysis is the result to use when describing the trial's primary analysis; the recalculation is a supplementary check and does not reproduce the full model. The difference in interval limits should not be read as an independent trial result.

The sponsor's model includes baseline and other listed covariates, along with visit and interaction terms. Those adjustments and the use of repeated observations are not represented by a calculation based only on arm summaries. Thus, agreement in the estimated contrast does not make the two approaches methodologically interchangeable. The registry record does not provide enough model detail to independently reproduce the sponsor's analysis fully.

15. Multiplicity, Missing Data and Analysis Populations

The registry record identifies a primary endpoint and several secondary outcomes, but a formal multiplicity strategy is not reported in the registry record. Multiplicity concerns arise when statistical tests are conducted across several outcomes, visits, or comparisons: the more results examined, the greater the opportunity for apparently notable findings to occur by chance. In general, prespecifying a testing sequence or adjusting significance levels can help control that risk. Neither a testing hierarchy nor an adjustment procedure is reported for the secondary outcomes.

The primary result should therefore be distinguished from the secondary findings. Its status as the designated primary endpoint gives it a different role in the trial's statistical interpretation, while unadjusted secondary results remain less conclusive. The registry record also does not report whether other outcomes or analyses were examined but not posted, so the full scope of all analyses cannot be established from this record.

The primary analysis population is described as randomized participants who had a baseline and at least 1 post-baseline PARS severity score during acute treatment, excluding 9 participants from 1 site because of major quality issues. The registry record does not describe the missing-data pattern, the reasons for missing outcome measurements, or sensitivity analyses for alternative missing-data assumptions. MMRM is a repeated-measures method, but its use does not remove the need to consider assumptions about missingness.

The record does not report whether the primary analysis used an intention-to-treat population as a formal analysis set, how protocol deviations were handled, or whether other analysis populations were used. It also does not report how missing secondary outcomes were handled. These omissions limit assessment of how robust results are to analytic choices; they do not, on their own, establish that the reported analysis is biased.

16. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: suicidal ideation (14.3%), adenoiditis (1.0%), tonsillitis (1.0%), acute psychosis (0.9%), bipolar disorder (0.9%).

The registry lists serious adverse events by treatment group and period. In the acute-treatment groups, serious adverse events affected 1 of 135 participants at risk in the duloxetine group and 0 of 137 at risk in the placebo group. During the extension, the record lists 2 of 104 in the duloxetine/duloxetine-extension group and 2 of 106 in the placebo/duloxetine-extension group. In the taper groups, it lists 1 of 97 and 1 of 7, respectively. These are counts of participants with serious adverse events, not estimates of treatment effect.

The registry also lists serious-event terms with percentages, including suicidal ideation at 14.3%, but the denominator and context for those percentages are not reported in the registry record. The percentages should not be read as rates for a particular randomized arm without that information. Death counts are not reported in the registry record.

Serious adverse event counts do not establish whether an event was caused by treatment. The record does not provide a causal assessment in the registry record, and the small group counts do not support precise comparisons. The different treatment periods and group definitions should be kept separate rather than combined into a single safety comparison.

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

The primary result describes a statistically distinguishable difference in average change on the specified PARS severity score over the 10-week acute-treatment period. The estimated change was more negative in the duloxetine group than in the placebo group, and the sponsor's confidence interval lies below zero. This is evidence about a group-level outcome under the reported analysis, not a prediction of an individual participant's experience.

The result does not show that every participant improved, that the improvement was equally large across participants, or that the observed average difference meets a threshold of clinical importance. The registry record does not report a minimum clinically important difference for the primary measure. Nor does the result establish how durable the difference is after the 10-week endpoint, even though an extension is described. A later endpoint requires its own reported results and analysis.

The finding also does not by itself establish a complete benefit-risk assessment. Safety outcomes are recorded separately and include different treatment periods and group definitions. Their counts require cautious interpretation and do not establish causation. Secondary outcomes add context but, without a reported multiplicity strategy and complete numerical results, they should not be treated as multiple independent confirmations.

Finally, the finding applies to the measured population, outcome, and model. The registry record describes an analysis population with baseline and post-baseline measurements and excludes participants from a site with major quality issues. Eligibility details and sensitivity analyses are not reported in the registry record. The result should therefore be understood as a finding from this trial's registered comparison, not automatically generalized to every child or adolescent with GAD or to other treatment settings.

18. Why This Trial Matters Statistically

This trial illustrates why a primary endpoint, analysis population, and model need to be read together. A change score may appear straightforward, but the reported estimate is a model-adjusted least-squares mean difference derived from repeated measurements, not simply an unadjusted subtraction of two raw group averages. Knowing the model terms helps explain what the sponsor's estimate represents and why a summary-based recalculation is only a check rather than a full reproduction.

It also shows the value of reporting uncertainty alongside a p-value. The primary estimate is accompanied by a confidence interval, which communicates a range of plausible effects under the model and makes the degree of precision visible. Statistical evidence against a zero difference and the practical importance of the difference are separate considerations; the registry does not report a threshold for clinically meaningful change on this measure.

The registry lists numerous secondary outcomes, making multiplicity a practical interpretive issue. A consistent direction across related outcomes can add context, but unadjusted secondary findings do not all carry the same evidential weight as a designated primary result. Missing-data assumptions and the stated site exclusion also matter when judging what population the analysis represents.

Finally, the study record makes clear that efficacy and safety evidence come from different outcome summaries and treatment periods. Keeping those distinctions intact is essential: counts of serious events are not an efficacy comparison, and an acute-treatment efficacy endpoint does not describe the later extension. These principles apply broadly to the statistical reading of clinical trials.

19. 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 registry record leaves important details unavailable for independent assessment. It does not report the original sample-size justification, its assumptions, or an interim-analysis plan. The illustrative planning calculations based on observed effects cannot fill that gap. Detailed randomization procedures and the identities of the parties masked are also not reported in the registry record.

The primary analysis excluded 9 participants from 1 site because of major quality issues. The record does not describe the nature of those issues, the timing or rationale for the exclusion, or sensitivity analyses that include or otherwise address that site. The analysis population also required baseline and post-baseline measurements. Missing-data amounts, patterns, reasons, and sensitivity analyses are not reported in the registry record.

Although the MMRM adjustment variables are specified, the full model details are not: covariance structure, diagnostic checks, and further estimation details are not reported in the registry record. The record does not report a clinically important difference threshold for the primary score. This limits interpretation of the magnitude beyond its position on the scale and its statistical uncertainty.

The registry lists 13 secondary outcomes, but numerical findings for all those outcomes are not reported in the registry record. A multiplicity strategy is also not reported, so secondary results have to be interpreted cautiously. Safety figures span acute treatment, extension, and taper groups; these are not one uniform comparison, and the causal relationship of events to treatment is not established by the counts alone. The percentages attached to listed serious-event terms have no denominator or group context reported in the registry record.

Finally, the primary outcome concerns a 10-week endpoint. The record describes an 18-week extension but the primary result does not establish longer-term persistence. Specific age eligibility limits are not reported in the registry record, which constrains assessment of how broadly the pediatric findings apply.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study in Pediatric Participants With Generalized Anxiety Disorder — NCT01226511.
• Registry record last updated 2014-03-05; status last verified 2014-01.
• Results first posted 2014-03-05.
• 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.

23. Record Summary

The registry record describes a completed randomized, parallel-group Phase 3 study of duloxetine versus placebo in children and adolescents with GAD. The blinded acute-treatment period lasted 10 weeks, with flexible once-daily duloxetine dosing of 30-120 mg and an 18-week extension described after the blinded period. The record lists 281 participants enrolled.

The primary outcome was change from baseline to the 10-week endpoint in the GAD-related PARS severity score. The sponsor's MMRM analysis reported a least-squares mean difference of -2.65, with a 95% confidence interval from -4.03 to -1.27 and p < 0.001. The primary analysis population required baseline and at least 1 post-baseline score and excluded 9 participants from 1 site for major quality issues.

The record also lists secondary efficacy, functioning, and safety outcomes. A formal multiplicity strategy, the original sample-size assumptions, detailed missing-data sensitivity analyses, and numerical results for all registered secondary outcomes are not reported in the registry record.