Dextro-Amphetamine Versus Caffeine in Treatment-resistant OCD: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Dextro-Amphetamine Versus Caffeine in Treatment-resistant OCD (NCT00363298), a randomized na study. The primary outcome is Number of Subjects With Clinical Global Impressions Scale - Improvement (CGI-I) Score of 1 or 2.

2. Clinical Question

The trial’s stated hypothesis concerned whether dextro-amphetamine could be used as augmentation for obsessive-compulsive disorder and whether tolerance, described as loss of therapeutic effect, would develop over several weeks. The posted primary outcome is a count of participants meeting a specified improvement-score threshold. It is not a direct measurement of tolerance, and the registry record does not report a separate result for tolerance.

The outcome population is important to the question. Entry into the continuation phase depended on an initial Y-BOCS decrease after the first week of study medication. The primary result therefore describes improvement ratings among a selected group that met this entry condition; it does not describe the result for all enrolled participants. The registry record does not report how the primary result would look in a population that included those who did not enter continuation.

The comparison is randomized, double-blind, and parallel-group. However, the posted outcome counts alone do not establish a precise comparative effect. No confidence interval, p-value, or analysis estimate for the binary comparison is reported in the registry record. The result should therefore be read as the reported counts in the defined population, rather than as a conclusive estimate of a treatment difference.

3. Trial Design

This was a randomized, double-blind, parallel-group trial sponsored by Stanford University. The registry lists the phase as NA and the status as completed. Enrollment is recorded as 24. The recorded study dates are 2006-08 for start and 2008-03 for primary completion. These are registry dates, not evidence about when each participant began or completed treatment.

The intervention was dextro-amphetamine, and the comparator was described as a sham comparison using caffeine capsules identical in appearance to the dextro-amphetamine capsules. The dextro-amphetamine arm description specifies 15 mg per capsule, with a morning capsule from each of Bottles A and B. The caffeine description specifies 200 mg in Bottle A and 100 mg in Bottle B, also taken as one capsule from each bottle each morning. The registry describes double masking but does not report further details about which study personnel or participants were masked.

The posted primary outcome concerns a continuation phase. Participants had to meet the entry criterion of a >20% decrease in Y-BOCS score after 1 week of double-blind study medication. Those who entered then received a 4-week double-blind continuation phase. The primary outcome timeframe is stated as the end of week 5, except for 1 dextro-amphetamine participant rated at the end of week 2. The registry record does not explain why that participant was assessed at the earlier time.

The primary endpoint is the number of participants with a Clinical Global Impressions Scale-Improvement score of 1 or 2. The arm results are 6 of 6 for dextro-amphetamine and 7 of 7 for caffeine pills. The outcome record also lists a safety population of 12 at risk in each group. Those safety denominators differ from the primary-outcome denominators and should not be treated as interchangeable.

The registry lists 2 outcome measures, 0 posted secondary outcomes, and 0 analyses. It does not report a detailed statistical analysis plan, allocation procedure, sample-size justification, or further information about follow-up beyond the stated outcome timeframe.

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.

First Study Week

Reasons for not completing this period:

4-week Study Continuation Phase

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Subjects who met the study continuation phase entry criterion of \>20% decrease in Y-BOCS score after 1 week of double-blind study medication, and entered this 4-week double-blind continuation phase

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: 24 Subjects meeting the listed entry criteria.

6. Primary Endpoint

The registry lists the primary outcome as Number of Subjects With Clinical Global Impressions Scale - Improvement (CGI-I) Score of 1 or 2, measured over At end of week 5, except 1 d-amphetamine subject rated at end of week 2.

The primary endpoint counts participants whose Clinical Global Impressions Scale-Improvement score was 1 or 2. The registry defines score 1 as very much improved and score 2 as much improved. Other possible scale values range from 3, minimally improved, through 7, very much worse, with 4 indicating no change. The endpoint therefore classifies participants according to whether their recorded improvement rating reached either of the two specified categories.

A thresholded improvement outcome can be useful because it expresses a clinical rating in an interpretable category: whether a participant was judged very much or much improved. It is a binary endpoint for analysis, even though the underlying scale has multiple possible scores. That conversion also discards distinctions within the two selected categories and among the other possible scores. The registry record does not report the full distribution of scale scores, so the posted counts cannot show how participants outside the threshold were rated.

The population for this endpoint is restricted to participants who met the continuation-phase entry criterion of a >20% decrease in Y-BOCS score after 1 week and then entered the 4-week continuation phase. Thus, the result answers a question about a selected continuation population rather than an unrestricted comparison of everyone enrolled. The timeframe is the end of week 5, with 1 dextro-amphetamine participant rated at the end of week 2. The registry record does not explain the reason for that timing exception.

The reported counts are 6 of 6 in the dextro-amphetamine group and 7 of 7 in the caffeine-pills group. These are descriptive counts. The registry record does not report a confidence interval or a statistical test for the binary comparison, and the counts alone do not quantify the uncertainty around a treatment contrast.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a binary outcome count represent?

A binary outcome places each participant into one of two categories for analysis: meeting the stated criterion or not meeting it. Here, the criterion is a Clinical Global Impressions Scale-Improvement score of 1 or 2. The registry reports the number meeting that criterion and the denominator in each arm: 6 of 6 for dextro-amphetamine and 7 of 7 for caffeine pills. A count is easy to interpret, but by itself it does not describe the uncertainty in a population-level effect. It also leaves out the specific scores of participants who did not meet the threshold, and the record does not report the full score distribution.

Why does the outcome population matter?

The primary endpoint is not recorded for an unrestricted group defined only by enrollment. Participants first had to achieve a >20% decrease in Y-BOCS score after 1 week of double-blind medication and enter the 4-week continuation phase. A result in this selected group answers a different question from a result among all participants starting treatment. Selection on an early response can change the composition of the group whose later outcome is described. The registry record does not report outcomes for people who did not meet the entry criterion, nor does it provide an analysis that combines them with continuation-phase participants.

This distinction does not make the posted result invalid; it defines its scope. The count pertains to participants who entered continuation and had a reported primary rating. It should not be interpreted as a result for every participant randomized at the outset. The registry record does not describe a broader analysis population or state how missing primary ratings, if any, were handled.

What can the arm counts tell us?

The reported counts show how many participants in each reported denominator reached the score threshold. They are descriptive evidence about those groups. A comparative treatment effect requires a contrast between groups and an assessment of its uncertainty. The registry record does not report a confidence interval or p-value for this binary endpoint, and the comparison is marked as not estimable because values are missing or not estimable for a two-arm comparison. No inferential conclusion should be substituted for those absent results.

In general, a confidence interval expresses a range of values compatible with data and a statistical model, conveying uncertainty around an estimate. With a small number of observations, estimates of binary outcomes can be especially imprecise; observing every participant in a reported denominator meet the threshold does not remove that uncertainty. This is a general statistical point, not an additional result from this trial. The registry record contains no interval estimates from which to judge precision directly.

How should the assessment timing be read?

The stated primary timeframe is the end of week 5, except for 1 dextro-amphetamine participant rated at the end of week 2. Timing matters because the endpoint is an assessment at a specified point in treatment. An earlier rating may reflect a different point in the course of treatment than the stated endpoint. The registry record does not explain the exception or report a sensitivity analysis assessing it. There is therefore no basis to assume that the timing difference was inconsequential or that an adjustment was made.

What do the planning illustrations mean?

The planning illustrations are based on the observed effect, with an effect to detect specified as a fraction of that observed effect. The stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. Such calculations are hypothetical design exercises: they ask what sample size might be required under stated assumptions to detect a target effect. They are not observed trial results and are not observed power. The registry record does not report the trial’s own sample-size justification, so the illustrations should not be treated as evidence of the design rationale.

In general, a sample-size calculation depends on the target effect, variability or event rate, significance threshold, desired power, allocation, and analysis method. Changes in those assumptions can change a planning estimate. Because the illustrations use fractions of the observed effect, they do not independently validate that effect or show that the trial was adequately sized. Nor do they replace an estimate and uncertainty interval for the observed comparison. No additional sample-size assumptions or design justification are reported in the registry record.

Why can binary comparisons be hard to estimate?

A binary outcome contains counts of participants meeting and not meeting a criterion. Estimating a contrast requires the relevant outcome information in both groups and a defined statistical method. The record explicitly marks this two-arm binary comparison as not estimable because values are missing or not estimable. Thus, a recalculated comparison is not presented. No alternative model, adjustment set, or sponsor analysis is reported in the registry record. Since the recorded discrepancy flag is false, there is no stated conflict between a recalculation and a sponsor analysis to resolve.

Over-dispersion is a general concern when observed variation exceeds what a statistical model assumes. It is often relevant for count outcomes or clustered data, where a simple model may understate uncertainty if extra variability is present. The primary endpoint is a participant count for a binary outcome, and the registry record does not describe an over-dispersion assessment or a model-based analysis. It would be inappropriate to infer that such a method was used or needed based on these posted counts alone.

Overall, the available statistical information supports a descriptive account of the primary counts and safety counts, plus a clear description of the selected population and assessment timing. The registry record does not provide enough inferential output to quantify the comparative effect or its precision. General statistical principles can explain why those omissions matter, but they cannot supply missing trial-specific estimates.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("D-amphetamine", "Caffeine Pills")
n      <- c(6, 7)
events <- c(6, 7)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  D-amphetamine  6  6  100.0  54.1  100.0  61.0  100.0  59.0  99.6
ARM  Caffeine Pills  7  7  100.0  59.0  100.0  64.6  100.0  63.1  99.7

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 2 outcome measures, but it posts 0 secondary outcomes and 0 analyses. The record does not report a multiplicity plan, such as whether outcomes were prioritized, whether multiple tests were adjusted, or whether any subgroup comparisons were planned. Without those details, it is not possible to assess how multiple outcome questions or analyses were handled. The absence of posted secondary results should not be interpreted as evidence that no secondary outcomes were considered during the study; it means they are not reported in the registry record.

The primary endpoint is defined for participants who met the continuation entry criterion and entered the 4-week phase. The registry record does not report primary outcomes for those who did not enter continuation, or describe a full analysis population that includes them. It also does not report how missing endpoint values were handled, whether any participants lacked ratings, or whether an intention-to-treat analysis was conducted. The exception in assessment timing—1 dextro-amphetamine participant rated at the end of week 2 rather than the stated end-of-week-5 timeframe—is noted, but the record does not explain how it was handled analytically.

These omissions limit interpretation of selection and missing-data risks. No assumptions about imputation, exclusions, protocol deviations, or sensitivity analyses can be made from the posted record. The binary comparison is marked as not estimable because values are missing or not estimable for a two-arm comparison, and no p-value or confidence interval is reported. The record does not provide further detail about what information was unavailable or why the comparison could not be estimated.

12. 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 safety record lists 12 participants at risk in each group. Serious adverse events are recorded as 0 in the dextro-amphetamine group and 0 in the caffeine-pills group. Other adverse events are recorded as 5 in the dextro-amphetamine group and 4 in the caffeine-pills group. These are the registry’s reported counts, not estimates of comparative risk or evidence about whether an event was caused by study medication.

The serious-event counts and other-event counts describe different categories. The record does not list serious adverse-event terms; it also does not provide the specific terms for the other adverse events in the registry record reported for this analysis. Severity details beyond the serious-versus-other classification, event timing, duration, recurrence, and participant-level overlap are not reported in the registry record. Counts should not be interpreted as counts of unique event types or as a complete account of tolerability.

Deaths are not reported in the registry record. A missing death entry should not be read as a confirmed count of zero. The safety denominators also differ from the primary endpoint denominators, which are 6 and 7. The registry does not explain the relationship between the safety and continuation-phase populations. These safety results therefore remain a limited descriptive account rather than a comprehensive comparison of harms.

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

The registry results establish what was recorded for the primary endpoint: 6 of 6 participants in the dextro-amphetamine group and 7 of 7 participants in the caffeine-pills group had an improvement score of 1 or 2. The population was restricted to participants who met the continuation entry criterion after 1 week and entered the 4-week continuation phase. The stated assessment was at the end of week 5, with 1 dextro-amphetamine participant rated at the end of week 2.

These observations do not establish that dextro-amphetamine is effective, ineffective, superior, equivalent, or non-inferior to caffeine. The registry record does not report a statistical test or confidence interval for the comparison, and it marks the binary two-arm comparison as not estimable. Without a comparative estimate and its uncertainty, the observed counts cannot show how compatible the data are with different underlying effects. Nor can the threshold counts reveal the full distribution of improvement ratings.

The result also does not answer the stated hypothesis about tolerance over several weeks. The primary outcome is an improvement rating at a stated assessment, not a reported measure of loss of therapeutic effect over time. No separate tolerance result or repeated outcome trajectory is reported in the registry record.

The safety entries show 0 serious adverse events among 12 at risk in each group, and other adverse-event counts of 5 and 4, respectively. They do not determine causality or provide a complete description of event types, severity, timing, or clinical importance. Deaths are not reported in the registry record.

In short, the posted data describe counts for a selected continuation population and limited safety categories. They do not support a broader conclusion about treatment effects across all enrolled participants, or about the overall balance of benefits and harms. The missing inferential and analysis details leave substantial uncertainty about the comparative interpretation.

14. Why This Trial Matters Statistically

This trial illustrates why the population behind an endpoint is as important as the outcome itself. The primary result is not simply an outcome among everyone enrolled: it is among participants who first met a >20% Y-BOCS decrease criterion and entered continuation. That selection defines the question the counts answer. Readers who overlook it may mistake a conditional continuation-phase result for an estimate that applies to the entire randomized group.

It also shows the limits of interpreting raw binary counts. The reported counts are straightforward to understand, but a treatment comparison requires an estimate of the contrast and information about its uncertainty. Neither a confidence interval nor a p-value is reported, and the comparison is marked as not estimable. Descriptive similarity in the threshold counts cannot substitute for an inferential comparison or demonstrate equivalence.

The timing exception reinforces the need to examine endpoint definitions closely. One participant in the dextro-amphetamine group was rated at the end of week 2 while the stated timeframe is the end of week 5. The record notes this, but does not report how it was handled or why it occurred. Such details can affect how consistently outcomes are interpreted across participants.

Finally, separate safety and efficacy denominators underscore that different analyses may concern different populations. Serious adverse-event counts are reported for groups of 12 at risk, whereas the primary outcome denominators are 6 and 7. The registry does not explain the relationship between these groups. This trial is therefore a useful example of reading an outcome together with its population, timing, and analysis status, rather than treating a count as a complete statistical conclusion.

15. 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 several trial-specific uncertainties unresolved. The primary endpoint applies only to participants who achieved a >20% decrease in Y-BOCS score after 1 week and entered the continuation phase. Results for participants who did not meet that criterion are not reported. As a consequence, the primary counts do not describe the outcome for all enrolled participants, and the record does not provide an analysis connecting the selected continuation group to the full randomized population.

The assessment timeframe is not fully uniform: 1 dextro-amphetamine participant was rated at the end of week 2, although the stated endpoint is at the end of week 5. The reason for this exception and its analytical handling are not reported. The registry also does not provide the full distribution of Clinical Global Impressions Scale-Improvement scores, so interpretation is limited to whether the score was 1 or 2.

No confidence interval, p-value, or estimated two-arm contrast is reported for the binary endpoint. The record marks the comparison as not estimable because values are missing or not estimable. It does not explain the underlying reason or offer an alternative analysis. A sample-size justification, detailed statistical plan, missing-data approach, multiplicity procedure, interim analysis information, and analysis-population definitions are not reported in the registry record.

The safety information is also limited. Serious events are recorded as zero in both safety groups, but event terms are not listed. Other adverse events are given only as counts, without terms or detail about severity, timing, duration, or attribution. Deaths are not reported in the registry record. The relationship between the safety denominators and the primary-outcome population is not described.

These limitations do not negate the reported counts; they constrain the conclusions that can be drawn from them. In particular, the registry record does not provide enough information to assess the precision of the treatment comparison, the effects of selection and missing data, or the completeness of the safety account.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Dextro-Amphetamine Versus Caffeine in Treatment-resistant OCD — NCT00363298.
• Registry record last updated 2017-03-28; status last verified 2017-02.
• Results first posted 2017-03-28.
• 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.

19. Record Summary

The registry describes a completed, randomized, double-blind, parallel-group study sponsored by Stanford University, with enrollment recorded as 24. The intervention was dextro-amphetamine and the comparator was caffeine capsules. The primary outcome counted participants with a Clinical Global Impressions Scale-Improvement score of 1 or 2 in a continuation population selected after a >20% Y-BOCS decrease following 1 week of study medication.

The primary counts were 6 of 6 for dextro-amphetamine and 7 of 7 for caffeine pills. The stated assessment was the end of week 5, except for 1 dextro-amphetamine participant assessed at the end of week 2. The registry lists 0 posted secondary outcomes and 0 analyses, and the binary comparison is marked as not estimable. Serious adverse events were recorded as 0 in each safety group; other adverse-event counts were 5 and 4. Deaths are not reported in the registry record.

The record does not report a comparative confidence interval or p-value, a detailed statistical plan, or a sample-size justification. The results therefore support a descriptive account of the posted counts, not a precise conclusion about comparative effects.