Predictors of Response to Cognitive Remediation in Schizophrenia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Predictors of Response to Cognitive Remediation in Schizophrenia (NCT00333970), a randomized na study. The primary outcome is Cognitive Performance.

2. Clinical Question

3. Trial Design

The registry describes a randomized, parallel-group trial with no masking. It lists schizophrenia as the condition, cognitive remediation as a behavioral intervention, and treatment as usual as the other condition. The study is marked completed, with enrollment of 103. The listed phase is NA. The sponsor is the US Department of Veterans Affairs. The record gives a start of 2004-08 and primary completion of 2013-10.

The trial’s stated purpose was to identify variables that predict improvement on a cognitive training task battery and to determine the proportion of patients whose performance returns to normal following cognitive training. The single posted primary outcome is cognitive performance, specifically change in verbal memory score from baseline to 2 months. The primary-outcome population is described as individuals who were randomized and completed 2-month assessments.

For that outcome, the result groups are labeled Arm 1 and Arm 2. Arm 1 is the control in the posted comparison, and Arm 2 is the comparison arm. The record separately names cognitive remediation and treatment as usual, but does not connect those named conditions to the numbered result-arm labels. Consequently, the numerical comparison can be described by arm label, but its direction cannot confidently be assigned to a particular named condition from the posted result alone.

The registry record does not report the trial’s original sample-size justification, detailed allocation procedure, or protocol-level statistical plan. Masking is listed as none, but the record does not provide further details about how outcomes were assessed or whether any protections against assessment bias were used. The number of registered primary outcome measures is 1, and no secondary outcomes are posted in the record. These registry entries describe the available design and results; they do not fill in protocol information that is not reported.

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: 103 participants were enrolled. Of those, 75 were found eligible, completed baseline assessments and were then randomized to either Arm 1 (n=50) or Arm 2 (n=25). The remaining 28 participants were either discontinued from the study by the PI or chose to withdraw prior to randomization.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: individuals who were randomized and completed 2 month assessments

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: individuals randomized to Arm 1 or Arm 2

6. Primary Endpoint

The registry lists the primary outcome as Cognitive Performance, measured over baseline and 2 months.

The primary endpoint measures change in verbal memory from baseline to the end of the active phase, at 2 months. The measure is the total score across trials 1–5 of the California Verbal Learning Test-II. Its scale ranges from 0–80, with higher scores indicating better performance. The registry reports the endpoint as a mean, with standard error as the measure of dispersion.

A change score addresses how much the measured performance differs between the two assessment times, rather than reporting only a score at one time. A group mean describes the average change among the participants included in that group’s posted result. It does not show how much individual participants varied, whether every participant changed in the same direction, or how many reached a particular level of performance.

Verbal memory is relevant to the stated cognitive-training question because the study concerns cognitive performance and uses a cognitive task battery. The selected measure provides a defined task and scoring range for the registered primary outcome. The registry does not explain why this specific measure was chosen over other possible cognitive measures, nor does it report a threshold defining return to normal. Therefore, the posted endpoint can address average change in this verbal-memory score, but it cannot by itself answer the registry’s separate question about the proportion returning to normal performance.

7. Results Posted in the Registry

• Arm 2: mean -2.30 versus 4.97 on control; difference -7.27 (95% CI -12.00 to -2.54), standardized effect d = -0.81; z-test p = 0.003.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports a continuous outcome as a mean change score. A mean is the average of the measured values within a group. Here, the values represent change in verbal memory score between baseline and 2 months. A positive mean indicates an average increase in score over that interval, and a negative mean indicates an average decrease. Since higher scores on the measure represent better performance, the sign of a change score describes the direction of average score change. It does not imply that every participant changed in that direction.

What does the mean difference compare?

The reported contrast is Arm 2 minus Arm 1. Arm 1 is the designated control, and the comparison estimate is -7.27 units. This is the difference between the posted group means, as reported in the registry. The estimate is in the original measurement units, which makes it interpretable relative to the verbal-memory scale. The result labels do not identify which numbered arm corresponds to each named condition, so the contrast should remain expressed as Arm 2 compared with Arm 1.

The registry also reports a standardized mean difference of -0.812. Standardizing expresses a contrast relative to a measure of variability, allowing effects measured on different scales to be described in a common unit. It is a statistical summary, not a measure of how many score points changed. The registry reports a pooled standard deviation of 8.95 for this calculation. No additional interpretation about clinical importance follows from the standardized estimate alone.

What do the standard errors describe?

The arm-level results report standard errors of 1.38 for Arm 1 and 1.98 for Arm 2. A standard error describes the estimated uncertainty in a group mean, reflecting how precisely that mean is estimated. It is not the same as the standard deviation, which describes the spread of individual observations. The registry labels the dispersion measure as standard error, and the comparison summary also reports the pooled standard deviation. Those quantities answer different statistical questions and should not be interchanged.

In general, a standard error tends to be smaller when a mean is estimated more precisely, though it also depends on the variability and amount of data. The reported group counts are 39 for Arm 1 and 23 for Arm 2. The registry does not give participant-level observations, so the distribution of scores and the presence of unusual values cannot be assessed from these summaries alone.

How should the p-value be understood?

The unadjusted p-value is 0.003. In general, a p-value describes 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, nor does it measure the size or importance of an effect. The registry does not report an adjusted p-value or a detailed analysis plan, so the posted value should be treated as the unadjusted result recorded for this comparison.

A confidence interval and a p-value are related but convey different information. The interval gives an estimated range for the contrast and communicates its precision, while the p-value summarizes evidence against a null value under a statistical model. The posted interval for the mean difference extends from -12.0 to -2.54. The interval is expressed in the original score-change units, and its range communicates uncertainty about the size of the arm contrast.

What assumptions are not documented?

The registry does not report whether the primary comparison adjusted for baseline score or other participant characteristics, whether any transformations or diagnostic checks were used, or how departures from model assumptions were handled. It does not provide a detailed protocol for missing data, the exact analysis population beyond the description of randomized participants who completed 2-month assessments, or an explanation of how the standard errors and confidence interval were constructed. These details are not reported in the registry record and should not be inferred from the summary statistics.

General statistical considerations remain relevant. A mean comparison is most informative when the observations and analysis model support its assumptions. Differences in variability, skewed score distributions, or influential observations can affect estimates and uncertainty. Randomization can support a fair group comparison when implemented and analyzed as planned, but it does not automatically resolve missing outcome data or unclear reporting. Since no participant-level information or detailed analytic methods are posted, these aspects cannot be evaluated from the registry summary.

The registry records a randomized parallel design and no masking. General methodological guidance is that unmasked designs can leave expectations or knowledge of assignment open to influencing behavior, assessment, or reporting, depending on how the outcome is collected. The registry does not describe assessment procedures, so the practical implications for this verbal-memory measure cannot be determined from the record. Likewise, the record does not report interim analyses or a multiplicity strategy. No unreported design or analysis procedure should be assumed.

10. Confidence Intervals and Precision

The estimated mean difference is -7.27 units for Arm 2 compared with Arm 1. Its 95% confidence interval ranges from -12.0 to -2.54. The interval expresses uncertainty in the estimated contrast: values across that range are compatible with the data under the assumptions of the statistical procedure used. The interval is in score-change units, so it retains the scale of the verbal-memory measure. It is not a range containing 95% of individual participants’ changes.

The interval’s endpoints are both negative, indicating that the reported estimates in the interval favor a lower mean change in Arm 2 than in Arm 1. The interval also shows that the precise size of the contrast is uncertain: the data are compatible with more than one magnitude of difference. Its interpretation remains limited by the unreported mapping between numbered arms and named conditions, as well as by the lack of detail about the analysis and participants without completed assessments.

The registry reports standard errors of 1.38 and 1.98 for the means in Arm 1 and Arm 2, respectively. These quantities reflect uncertainty in each estimated mean; they do not describe how individual participants’ change scores are distributed. The result groups contain 39 and 23 participants. With group-level summaries alone, it is not possible to inspect the score distribution, identify influential observations, or assess whether individual responses were consistent.

The unadjusted p-value of 0.003 provides another summary of the comparison under its statistical model, but it does not replace the interval or explain the practical importance of the score difference. The registry does not report a threshold for a meaningful change or a normal-performance cutoff. Therefore, the precision of the group comparison can be discussed statistically, but the clinical meaning of a particular value within the interval is not established by the posted information.

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 planning figures are expressed as participants per arm, not as total events. They are illustrations based on the observed effect, not the trial’s original sample-size design. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The planning effect is set to the stated fraction of the observed effect. The trial’s own original design assumptions are not reported in the registry record.

For planning at 100% of the observed effect, the illustration gives 24 participants per arm for 80% power and 32 participants per arm for 90% power. At 75% of the observed effect, it gives 43 participants per arm for 80% power and 57 participants per arm for 90% power. At 50% of the observed effect, it gives 96 participants per arm for 80% power and 128 participants per arm for 90% power. These are hypothetical planning figures under the stated assumptions, not descriptions of the enrolled or analyzed group sizes.

In plain language, the 80% power column represents a planning target with an 80% chance of detecting the specified effect under the assumed model if that effect is the true one. The 90% column uses a higher chance of detection as the target. The planning display should be read by relating the assumed effect fraction to the required participants per arm: smaller effects need more participants because they are harder to distinguish from random variation. Any curve accompanying these figures is a planning visualization of that relationship, not a record of the trial’s actual results. These values are not observed power.

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 = -7.270; pooled standard deviation σ = 8.950.
2. n = 2 × (1.96 + 0.842)² × 8.950² / -7.270² = 23.8, rounded up to 24 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 <- -7.27                      # observed difference in means
sigma <- 8.94990055810677                     # 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(333970)
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 = 24 (80% power), 32 (90% power)
 75% of observed effect: n per arm = 43 (80% power), 57 (90% power)
 50% of observed effect: n per arm = 96 (80% power), 128 (90% power)
power.t.test: 24.8 
Simulated power at n = 24 per arm: 0.788 (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. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure, no posted secondary outcomes, and 0 analyses. Although one primary outcome is identified, the record does not report a multiplicity strategy or explain whether other outcomes, time points, or subgroup comparisons were considered in the study. Without that information, the effect of multiple testing on interpretation cannot be assessed. No adjustment for multiplicity should be assumed.

The primary outcome population is described as individuals who were randomized and completed 2-month assessments. The result groups contain 39 participants in Arm 1 and 23 in Arm 2, while the enrollment entry is 103. The registry does not report the flow of participants between enrollment, randomization, and completion, or reasons assessments may be missing. It also does not state how missing outcomes were handled. The observed comparison therefore describes the posted endpoint population; the potential influence of missing data is not reported in the registry record.

The record does not specify whether the analysis followed an intention-to-treat approach, used a modified analysis population, or applied another rule for including participants. It does not report imputation procedures, sensitivity analyses for missing outcomes, or an analysis of participants who did not complete the 2-month assessment. These omissions make it impossible to determine from the registry summary how robust the result is to alternative assumptions about missing outcomes.

The trial’s stated purpose includes assessing predictors of response and the proportion of patients returning to normal performance. No predictor analyses or normalization results are posted. The single reported outcome comparison cannot stand in for those unreported analyses. Details about subgroup selection, predictor models, or definitions of normal performance are not reported in the registry 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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: shortness of breath (4.0%), pulled chest muscle (4.0%), worsening of psychiatric symptoms (2.0%), substance abuse (2.0%), renal failure (2.0%).

The registry reports serious adverse event counts separately by result-arm label. In Arm 1, 50 participants are listed as at risk and 3 as affected. In Arm 2, 25 are listed as at risk and 2 as affected. These are counts associated with the registry’s serious-event reporting and should be interpreted with the listed at-risk groups in mind. The record does not provide a detailed definition of the event-collection period or clarify whether each participant could contribute more than one event.

The other adverse-event entries list 50 participants at risk in Arm 1 and 4 affected, and 50 at risk in Arm 2 and 5 affected. The at-risk counts differ between the serious-event and other-event entries, so the categories should not be combined or treated as if they share the same denominator. Death counts are not reported in the registry record.

The registry also lists serious-event terms with percentages, including shortness of breath and pulled chest muscle at 4.0%, and worsening of psychiatric symptoms, substance abuse, and renal failure at 2.0% each. The record does not assign these terms to a particular result arm in the posted information. These summaries do not establish causation or provide a complete account of comparative safety.

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

The posted result describes an average change in verbal-memory score over the 2-month assessment interval for participants included in the primary-outcome results. Arm 1 has a reported mean change of 4.97, and Arm 2 has a mean change of -2.3. The mean difference for Arm 2 compared with Arm 1 is -7.27, with a 95% confidence interval from -12.0 to -2.54. The unadjusted p-value is 0.003. These values describe the registered comparison, not the experience of every participant.

The numbered-arm labels cannot be connected confidently to cognitive remediation and treatment as usual from the posted results. Thus, the direction of the difference can be stated for Arm 2 relative to Arm 1, but not attributed to either named condition. The registry record also does not give a threshold for meaningful change or define a score that constitutes return to normal performance. The numerical result therefore does not answer whether participants returned to normal functioning.

The record’s stated purpose includes finding predictors of improvement. No predictor analyses are posted, so the results do not identify characteristics associated with response. Nor do the results establish the proportion of participants whose performance returned to normal. The endpoint is a group mean change on one verbal-memory measure, and the registry lists no secondary outcome results. Conclusions about other aspects of cognition or outcomes are not supported by this result.

The comparison also does not reveal individual variation, the distribution of changes, or the number of participants who improved or declined. The posted outcome population consists of randomized individuals who completed 2-month assessments, and the registry does not report how missing assessments may affect interpretation. The statistics can be read as an estimate and its uncertainty for the recorded group comparison, but they do not establish why the difference occurred or how it relates to each participant’s course.

16. Why This Trial Matters Statistically

This study illustrates how a trial’s stated research aim and its posted result can address different questions. The stated aim includes identifying predictors of cognitive-training response and estimating the proportion of patients returning to normal performance. The registered primary result instead reports mean change in one verbal-memory score, and no predictor or normalization analyses are posted. Recognizing this distinction prevents a group-average outcome from being treated as an answer to questions about individual response or predictors.

The registry provides several complementary statistical summaries: arm means, standard errors, a mean difference, a confidence interval, an unadjusted p-value, and a standardized mean difference. Each describes a different aspect of the result. The group means show average change in the original units; the contrast expresses the between-arm comparison; the interval communicates uncertainty; and the standardized estimate expresses the contrast relative to variability. None of these summaries supplies the missing participant-level distribution or establishes clinical importance without a meaningful-change threshold.

The result also highlights why clear arm labels matter. The registry names cognitive remediation and treatment as usual but posts the outcome groups as Arm 1 and Arm 2. Arm 1 is identified as the control, yet the mapping of each numbered arm to the named condition is not reported. Statistical interpretation should retain the comparison as labeled rather than infer a condition-specific direction.

Finally, the planning illustrations show how sample-size requirements depend on the effect assumed for a future study. They are based on the observed effect and stated planning assumptions, not the trial’s original design. Separating such illustrations from the study’s actual analysis helps avoid confusing a planning calculation with evidence about the trial’s observed precision or design choices.

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 posted outcome results are limited to the primary verbal-memory measure and to the participants described as having completed 2-month assessments. The enrollment entry is 103, while the arm-level outcome results include 39 in Arm 1 and 23 in Arm 2. The registry does not provide participant flow, reasons for missing assessments, or a missing-data analysis, so the relationship between enrollment and the analyzed outcome groups cannot be explained from this record alone.

The arm-to-condition mapping is unclear in the posted results. Although the record names cognitive remediation and treatment as usual, the numerical comparison uses Arm 1 and Arm 2, with Arm 1 designated as control. The record does not say which numbered arm received each named condition. This limits interpretation of the estimate as a comparison of the named study conditions.

The registry does not post predictor analyses or results on the proportion returning to normal performance, despite those questions appearing in the study purpose. It also does not define a normal-performance threshold or a meaningful-change threshold. No secondary outcome results are posted. The mean change comparison therefore cannot be extended to these unreported questions.

Further limitations are that the registry record does not report the original sample-size justification, detailed statistical plan, multiplicity approach, or analysis population rules beyond the completion-based outcome description. It gives no participant-level data or detailed account of outcome assessment procedures. The study is listed as having no masking, but the possible effects on assessment cannot be evaluated from the available description. These gaps limit independent appraisal of the analysis and the broader interpretation of the result.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Predictors of Response to Cognitive Remediation in Schizophrenia — NCT00333970.
• Registry record last updated 2014-10-09; status last verified 2014-10.
• Results first posted 2014-10-09.
• 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

NCT00333970, titled Predictors of Response to Cognitive Remediation in Schizophrenia, is listed as a completed randomized, parallel-group study with no masking and enrollment of 103. The condition is schizophrenia. The record names cognitive remediation and treatment as usual, and identifies the sponsor as the US Department of Veterans Affairs.

The primary outcome is mean change in verbal memory from baseline to 2 months, measured by the total across trials 1–5 of the California Verbal Learning Test-II. The posted comparison is Arm 2 versus control Arm 1: the mean difference is -7.27, with a 95% confidence interval from -12.0 to -2.54 and an unadjusted p-value of 0.003. The registry does not report the mapping of the numbered result arms to the named conditions.

The registry lists one primary outcome measure, 0 analyses, and no secondary outcome results. Predictor findings, the proportion returning to normal performance, the original sample-size justification, and details of missing-data handling are not reported in the registry record.