Exercise Training for the Treatment of Generalized Anxiety Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Exercise Training for the Treatment of Generalized Anxiety Disorder (NCT00953654), a randomized na study. The primary outcome is Generalized Anxiety Disorder (GAD) Remission as Measured by Anxiety Disorders Interview Schedule-Adult Version (ADIS-IV) Severity Ratings.

2. Clinical Question

The registry describes a randomized controlled trial comparing two forms of exercise with a waiting-list comparison condition. The posted comparisons are each exercise group against the waiting-list group; the registry does not post a direct strength-versus-endurance comparison. The clinical question is therefore about the observed remission counts in each of those exercise-versus-waiting-list comparisons, not about a ranking of the exercise approaches.

3. Trial Design

The study is registered as a randomized, parallel-group trial with single masking. Enrollment is recorded as 30. The registry does not identify which participants or study personnel were masked, so the masking label alone does not establish who was unaware of group assignment. The trial status is completed. The recorded study start is 2009-08, and primary completion is recorded as 2010-05.

Three conditions are described: resistance exercise training, aerobic exercise training, and waiting-list control. The intervention descriptions specify six weeks of training, with both exercise conditions scheduled twice weekly. Strength training used lower-body exercises and progressed from 50% to 75% of predicted one-repetition maximum. Endurance training used lower-body dynamic cycling and was matched to strength training for total work, active exercise time, and load progression. The waiting-list condition involved maintaining the current lifestyle without entering the exercise intervention, while completing outcome measures on the same time progression.

The primary outcome was GAD remission measured by ADIS-IV severity ratings at pre- and post-intervention assessments. The description states that remission was assessed from 1-16 days after the six-week intervention. The record identifies two outcome measures and one analysis entry, but it does not report the full analysis plan, the allocation procedure, or details about how outcomes were assessed beyond the named measure and timeframe.

The posted outcome counts are 4 participants with remission in the resistance exercise group, 6 in the aerobic exercise group, and 7 in the waiting-list control group. Each group has a recorded size of 10. The two posted statistical comparisons use the waiting-list group as the comparator. Details about withdrawals, adherence, protocol deviations, and the analysis population 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: Thirty women met exclusion criteria and were enrolled in the trial.

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 Generalized Anxiety Disorder (GAD) Remission as Measured by Anxiety Disorders Interview Schedule-Adult Version (ADIS-IV) Severity Ratings, measured over Pre- and post- 6 week training intervention.

The primary endpoint is the number of participants classified as having GAD remission, measured with the Anxiety Disorders Interview Schedule-Adult Version (ADIS-IV) Severity Ratings. The registry describes GAD as persistent, excessive or pathologic worry that is difficult to control, accompanied by specified symptoms, and associated with distress or functional impairment. It also states that symptoms are not caused by a substance or another disorder. The outcome is recorded as a count of participants with remission rather than as an average symptom score.

Remission is a clinically interpretable binary outcome: each participant is classified as meeting or not meeting the remission criterion. This allows the registry to report the number with the event in each group and to compare the observed event frequency between groups. As a general statistical point, binary outcomes can be summarized using measures such as a risk difference, risk ratio, or odds ratio; each expresses the group contrast on a different scale.

The outcome timeframe is pre- and post- the six-week training intervention, with remission assessed from 1-16 days after the intervention. The registry does not give the ADIS-IV threshold or operational scoring rule used to determine remission, nor does it report a continuous severity-score result. Interpretation is therefore limited to the remission counts and the posted comparisons, rather than changes in symptom severity across the full score range.

7. Results Posted in the Registry

• Resistance Exercise Training: 40.0% (4/10) versus 70.0% (7/10) on control; risk difference -30.0 percentage points (95% CI -71.6 to 11.6), risk ratio 0.57 (0.24–1.35); unadjusted two-proportion z-test p = 0.178.
• Aerobic Exercise Training: 60.0% (6/10) versus 70.0% (7/10) on control; risk difference -10.0 percentage points (95% CI -51.6 to 31.6), risk ratio 0.86 (0.45–1.64); unadjusted two-proportion z-test p = 0.639.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the risk difference describe?

As a general statistical explanation, a risk difference subtracts the event proportion in the comparator group from the event proportion in the intervention group. The event is GAD remission. A value of zero represents equal observed event proportions, a negative value indicates fewer events in the intervention group, and a positive value indicates more events in the intervention group. The registry's posted estimates are negative for both exercise-versus-waiting-list comparisons. This is a description of the observed direction, not a statement that the difference is established or causal.

The risk difference is an absolute-scale measure. It keeps attention on the difference in event frequency rather than expressing the result as a relative change. The registry also posts relative measures, including risk ratios and odds ratios, but the posted intervals for those measures are wide. The risk difference is useful for reading the size and direction of the comparison in the units of event frequency, while the relative measures answer a different question about how the groups compare proportionally.

How should the relative measures be understood?

As a general explanation, a risk ratio compares the event proportion in one group with the event proportion in another. A risk ratio of one corresponds to equal event proportions; values below or above one indicate a lower or higher event proportion in the numerator group, respectively. An odds ratio compares event odds rather than event proportions. When an event is not rare, odds and probabilities are not interchangeable, so the odds ratio should not be read as if it were a risk difference or risk ratio.

These measures do not remove uncertainty or resolve limitations in the data. Their confidence intervals are broad, consistent with the limited information in the recorded group counts. The registry posts the estimates, but the risk difference is the stated metric for the comparisons. The relative measures are additional summaries and do not constitute separate outcomes.

What do the p-values tell the reader?

The registry gives unadjusted p-values of 0.178 for resistance exercise versus waiting-list control and 0.639 for aerobic exercise versus waiting-list control. As a general statistical explanation, a p-value measures how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, the probability that a result is due to chance, or a measure of clinical importance.

The registry labels these values unadjusted. It does not report a multiplicity procedure for the comparisons or explain whether the p-values are from the sponsor's primary analysis. The recorded sample-size assumptions use a two-sided significance level of 0.05, but that planning assumption does not establish that the same threshold was used in the trial's original analysis. The trial's original statistical plan is not reported in the registry record.

Why do the confidence intervals matter alongside the estimates?

As a general statistical explanation, a confidence interval presents a range of values compatible with the observed data under the interval method and its assumptions. It communicates uncertainty around an estimate and is more informative than a point estimate alone. For both risk differences, the posted intervals include zero and extend to values on either side of zero. This means the data are compatible with no difference as well as with differences in either direction.

Interval width is relevant to interpretation: wide intervals indicate that the estimate is imprecise. In this trial, the groups are small and the outcome is binary, so the observed remission counts offer limited information about the size of a population contrast. A point estimate should therefore not be read as a precise estimate of the effect. The intervals are not guarantees that the true value lies within their limits, and they do not account for unreported issues such as missing outcomes or departures from the analysis plan.

What can randomization support?

As a general principle, random allocation is intended to make comparison groups similar on average with respect to measured and unmeasured characteristics, so differences in outcomes can be evaluated as group contrasts. Randomization does not guarantee that groups will be balanced in a small trial, nor does it resolve missing-data or measurement issues. The registry identifies the allocation as randomized but does not report the sequence-generation method, concealment procedures, or balance details. The results should therefore be read as the posted randomized-group comparisons, with the limitations of the record in view.

The comparisons are unadjusted in the posted summaries. No covariate-adjusted result is posted in the registry record. The record also does not explain whether the analysis followed assigned groups, included only participants who completed training, or used another analysis population. These omissions prevent a fuller assessment of the assumptions underlying the estimates.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data constrain the possible values of a group difference. The posted risk-difference confidence intervals for resistance exercise versus waiting-list control and for aerobic exercise versus waiting-list control are both broad. Each crosses zero and extends to positive as well as negative values. The estimates therefore do not pin down the direction or size of a population difference with precision.

The observed point estimates are negative for both comparisons, and the remission counts are 4 in the resistance group, 6 in the aerobic group, and 7 in the waiting-list group. However, a point estimate is only one value within a range of uncertainty. The intervals indicate that the data remain compatible with a range of outcomes, including no difference and contrasts in either direction. This uncertainty matters more than a simple description of which group has the larger observed count.

As a general statistical explanation, confidence intervals reflect sampling uncertainty under the assumptions of the method used to calculate them. They do not automatically include uncertainty arising from outcome misclassification, missing data, nonadherence, or differences between the trial participants and people outside the study. The registry does not report enough detail to assess those sources of uncertainty.

The small group sizes and binary outcome contribute to limited precision: each participant's remission status has a substantial effect on the observed group count. The posted intervals and p-values should be interpreted together, rather than treating a point estimate or a single p-value as a complete summary. The results offer an uncertain estimate of the remission contrast at the recorded follow-up, not a precise estimate of longer-term or broader clinical 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 sample-size figures are planning illustrations, not observed trial results. For these binary-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, and a normal approximation. The effect to detect is set as the stated fraction of the observed effect. These assumptions describe the planning illustration only; the trial's original design assumptions are not reported in the registry record.

The column for 80% power gives the estimated participants per arm for a design with an 80% chance of detecting the assumed effect under the stated assumptions, if that effect were the true effect and the model assumptions held. The column for 90% power gives the corresponding planning count for a 90% chance. These are design probabilities for hypothetical future studies, not probabilities that the trial result is true. This is not observed power.

Rows for planning at 100% of the observed effect use the full observed estimate as the target; rows for 75% and 50% use smaller target effects. Smaller effects need more participants to detect with the same planned power and significance level, which is why the requirements rise as the assumed effect becomes more modest. The number of participants per arm also differs markedly between the two posted comparisons because the observed effects used in the planning calculations differ.

The displayed values should not be treated as recommendations or as evidence that the original trial was designed with those group sizes. They are conditional illustrations based on the observed effect, which is uncertain, especially given the broad confidence intervals. The planning curve should be read in the same way: it illustrates how the required participants per arm change as the assumed effect changes under the stated planning assumptions. The registry record does not report the study's own sample-size justification.

12. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Control proportion p₀ = 0.700; treatment proportion p₁ = 0.400; average p̄ = 0.550.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7036 = 1.3790.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6708 = 0.5646.
4. n = (1.3790 + 0.5646)² / (-0.3000)² = 42.0, rounded up to 42 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.7                          # control proportion
p1 <- 0.4                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

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

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(953654)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 42 (80% power), 56 (90% power)
 75% of observed effect: n per arm = 74 (80% power), 99 (90% power)
 50% of observed effect: n per arm = 163 (80% power), 217 (90% power)
power.prop.test: 42 
Simulated power at n = 42 per arm: 0.829 (20000 replicates)

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

13. Multiplicity, Missing Data and Analysis Populations

The registry identifies two outcome measures and one analysis entry, and it posts two comparisons of exercise groups with the waiting-list control. It does not report a multiplicity adjustment or explain whether the comparisons were designated as co-primary, secondary, or exploratory. The posted p-values are unadjusted. Without a reported analysis plan, it is not possible to determine how the number of outcomes or comparisons was handled when drawing the original study conclusions.

The registry reports primary remission counts but does not provide a participant-level account of missing outcome data, withdrawals, or exclusions. The method for handling missing outcomes is not reported in the registry record. The analysis population is also not described: it is not clear from the record whether participants were analyzed according to randomized assignment, according to intervention received, or under another rule. No imputation approach or sensitivity analysis is reported.

These gaps matter because the interpretation of a randomized comparison depends on how outcomes from all assigned participants are treated. If outcome availability or inclusion in analysis differs between groups, the posted comparison may not reflect the same contrast as a complete analysis of all randomized participants. The record does not provide enough information to assess that possibility.

The registry also does not report an interim-analysis plan, stopping rules, or details of any monitoring of accumulating outcome data. Those elements are not reported in the registry record and should not be inferred from the completed status. The exact statistical model or test used for the primary analysis is not described beyond the posted comparison estimates and unadjusted p-values.

Consequently, the numerical comparisons can be described as posted, but the record does not support a full reconstruction of the sponsor's statistical analysis. The absence of reported details is not evidence that a procedure was or was not used; it means the method cannot be evaluated from this record.

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.

The registry reports zero serious adverse events affecting the participants at risk in the resistance exercise group, zero in the aerobic exercise group, and zero in the waiting-list control group. It also reports zero other adverse events affecting the participants at risk in each of those groups. The reported at-risk count is 10 in each group.

These figures are counts recorded in the registry, not proof that no event occurred outside the reporting process or that all possible harms were assessed in the same way. The record does not provide a list of serious adverse-event terms; the serious-event terms field is empty. Deaths are not reported in the registry record. The safety information should therefore be read as the adverse-event counts posted for the groups, with the limits of the registry information in mind.

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

The posted results show fewer recorded remission events in each exercise group than in the waiting-list group: 4 in the resistance exercise group, 6 in the aerobic exercise group, and 7 in the waiting-list control group. The reported risk-difference estimates are negative for both exercise-versus-waiting-list comparisons. The confidence intervals are broad and include values corresponding to no difference as well as differences in either direction, so the estimates are uncertain.

These findings do not establish that exercise caused fewer remissions, nor do they establish that the groups had the same outcome. The trial's small group counts, broad intervals, and missing information about the analysis population and missing outcomes constrain interpretation. The unadjusted p-values are 0.178 and 0.639; neither should be interpreted in isolation or as a measure of the importance of the result.

The comparison addresses remission measured shortly after a six-week intervention in sedentary women with GAD. It does not report longer-term remission, symptom-score changes, functioning, or a direct comparison between strength and endurance training. Although the intervention descriptions detail exercise frequency and progression, the registry does not post adherence results or enough information to assess how much exercise participants completed.

The results also cannot be generalized automatically to other populations, intervention schedules, or settings. They describe the outcomes recorded in this particular trial and the comparisons posted in its registry record. The data are useful for understanding the observed counts and their uncertainty, but they do not provide a definitive estimate of the effect of either exercise condition on remission.

16. Why This Trial Matters Statistically

This trial illustrates why a difference in observed counts should be interpreted together with its uncertainty. The remission counts differ across the three groups, but the posted confidence intervals for both exercise-versus-waiting-list risk differences include no difference and extend in both directions. A point estimate alone can appear more decisive than the underlying evidence warrants.

It also shows the value of stating the effect measure. A risk difference describes an absolute contrast in event frequency, while a risk ratio and odds ratio describe relative contrasts on different scales. Those measures are not interchangeable. Clear interpretation depends on knowing the event being counted, the comparison group, and the scale on which the result is expressed.

The registry record also highlights the importance of transparent reporting. Details about the original statistical plan, missing outcomes, analysis populations, and multiplicity are not reported. Without them, readers cannot fully reconstruct how the posted estimates relate to the planned analysis. Such information is particularly useful when a trial has multiple outcomes or comparisons.

Finally, the planning illustrations emphasize that required sample sizes depend on the effect a future study aims to detect. Smaller target effects require more participants under the same assumptions. Because the observed effects in a small trial are uncertain, planning based directly on those estimates should be treated cautiously. The illustrations are useful for showing that dependence, not as a substitute for a prospectively justified design.

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 recorded group sizes are small, and the primary outcome is binary. As a result, each remission outcome has a noticeable influence on the group counts and estimates. The confidence intervals are broad, so the observed direction of each comparison is uncertain. The record does not provide enough information to determine whether the outcome counts include all randomized participants or only those with available assessments.

The endpoint is assessed from 1-16 days after the six-week intervention. The registry does not report longer-term follow-up, so persistence of remission beyond that assessment period cannot be evaluated from these results. It also does not post continuous symptom-severity changes or a direct comparison between the strength and endurance groups.

Important trial details are not reported in the registry record, including the method of random sequence generation, allocation concealment, the identity of those masked, adherence, withdrawals, missing-data handling, and the analysis population. The operational threshold for ADIS-IV remission is also not stated in the posted endpoint description. These omissions limit assessment of potential bias and of the exact meaning of the remission classification.

The study population is described as sedentary women with GAD. The registry does not give further participant characteristics or eligibility details, so applicability to other populations cannot be assessed from this record alone. The completed status and recorded enrollment do not resolve these limitations. The results should be understood as a limited registry report of short-term remission counts and unadjusted comparisons.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Exercise Training for the Treatment of Generalized Anxiety Disorder — NCT00953654.
• Registry record last updated 2016-09-09; status last verified 2016-07.
• Results first posted 2010-09-29.
• 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 for NCT00953654 describes a completed randomized, parallel-group trial comparing lower-body strength training, endurance training, and a waiting-list control condition among sedentary women with generalized anxiety disorder. Enrollment is recorded as 30. The exercise conditions were scheduled twice weekly for 6 weeks.

The primary outcome was GAD remission measured using ADIS-IV Severity Ratings from 1-16 days after the intervention. The registry reports remission counts of 4 for resistance exercise training, 6 for aerobic exercise training, and 7 for waiting-list control, with a recorded group size of 10 in each arm. The posted comparisons are each exercise group versus waiting-list control; both risk-difference intervals include no difference, and the unadjusted p-values are 0.178 and 0.639.

The record does not report the original sample-size justification, full statistical analysis plan, missing-data methods, or analysis population. It reports zero serious and zero other adverse events in each group; deaths are not reported in the registry record.