Does The Addition Of Divalproex Sodium ER To An Atypical Antipsychotic Drug (APD) Improve Cognition And Psychopathology In Outpatients With Schizophrenia (SCH) Or Schizoaffective Disorder (SAD)?: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Does The Addition Of Divalproex Sodium ER To An Atypical Antipsychotic Drug (APD) Improve Cognition And Psychopathology In Outpatients With Schizophrenia (SCH) Or Schizoaffective Disorder (SAD)? (NCT00306475), a randomized phase 4 study. The primary outcome is Change in Cognition and Psychopathology Disorders as Measured on the MATRICS Consensus Cognitive Battery (MCCB).

2. Clinical Question

3. Trial Design

This was a Phase 4, randomized, parallel-group trial with quadruple masking. The registry lists Northwestern University as the sponsor and an enrollment of 60 participants. The study is marked completed. The registry records a start of 2006-03 and a primary completion of 2008-04.

The two reported groups were participants receiving the investigational product and participants receiving placebo. The intervention listing identifies divalproex sodium extended-release tablets and placebo, while the study objective describes adding the assigned treatment to an atypical antipsychotic other than clozapine. The registry does not specify which atypical antipsychotics participants used, or the dose and duration of those background treatments. The study objective describes a six-week randomized trial, and the primary outcome timeframe is six weeks.

Randomization means assignment to the comparison groups was intended to occur by chance, which can reduce systematic differences between groups at the outset. A parallel design assigns participants to one group rather than having each participant receive both study treatments. Quadruple masking indicates that masking was reported, but the identities of the masked parties are not reported in the registry record. Details of the randomization procedure, allocation concealment, and masking procedures are not reported in the registry record. The arm descriptions contain no additional information about treatment delivery.

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.

Overall Study

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Change in Cognition and Psychopathology Disorders as Measured on the MATRICS Consensus Cognitive Battery (MCCB), measured over six weeks.

The primary endpoint is titled change in cognition and psychopathology disorders as measured on the MATRICS Consensus Cognitive Battery (MCCB), with results expressed in units on a scale. The registry describes the MCCB as a measure of cognitive function and says it provides an overall composite score expressed as a T-score. It also describes the scale’s reference mean as 50 and its standard deviation as 10, and notes that there is no fixed minimum or maximum because the score is a T-score.

The MCCB is used in psychiatric research to assess cognitive function and track changes over time. A standardized cognitive measure can make outcomes comparable across participants and can capture change that may not be evident from a broad symptom description alone. The registry’s study objective also refers to psychopathology, including positive, negative, and mood symptoms, but the posted primary endpoint description does not specify separate symptom scales.

The reported arm means are 14.1 and 14.3, while the primary endpoint title refers to change. The registry does not clarify in the result description how these units relate to the MCCB composite T-score or whether they represent a particular score transformation. Interpretation should therefore remain with the posted values and comparison rather than assigning them a different scale meaning.

7. Results Posted in the Registry

• Participants Which Received IP: mean 14.10 versus 14.30 on control; difference -0.20 (95% CI -1.59 to 1.19), standardized effect d = -0.07; z-test p = 0.778.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean difference represent?

For a continuous outcome, a group mean is the average of participants’ recorded values. A mean difference compares the two group means, with the direction depending on which group is subtracted from which. In the posted comparison, the estimate is -0.2, with the divalproex sodium group first and placebo as the reference. The negative value therefore corresponds to a lower reported mean in the divalproex sodium group. A mean difference remains in the units of the outcome; it is not a percentage change.

The two reported means are 14.1 for participants receiving the investigational product and 14.3 for participants receiving placebo. The reported standard deviations are 2.8 and 2.7, respectively. A standard deviation describes spread among individual measurements around a group mean. It is distinct from the uncertainty in the mean difference, which depends on both within-group variation and the number of participants contributing data.

What does the standardized effect describe?

The standardized effect estimate is -0.073. Standardizing places the mean difference on a scale relative to outcome variability, rather than leaving it in the original outcome units. This can help readers describe an effect’s relative magnitude when comparing measures with different units. It does not make an uncertain result precise, and it does not replace the original-unit estimate or its confidence interval. The registry’s outcome description refers to MCCB T-scores, but does not explain how the posted means and standardized estimate relate to a transformed or component score.

How should the p-value be understood?

The unadjusted p-value is 0.778. General statistical interpretation: a p-value measures how incompatible the observed data are with a specified null model, assuming that model and the test’s assumptions hold. It is not the probability that the null hypothesis is true, and it does not quantify the probability that the result occurred by chance. A high p-value does not establish equality or equivalence. The registry does not describe the test used to obtain this p-value or provide a prespecified statistical analysis plan, so the particular modeling and testing choices cannot be assessed from the record.

Why does the confidence interval matter?

A confidence interval conveys uncertainty around an estimated effect. General explanation: intervals that are narrower indicate greater precision under the method used; wider intervals reflect less precision. The interval for this comparison runs from -1.592 to 1.192 and includes zero. It therefore includes values corresponding to a lower mean in either group and a value of no difference. The confidence level and exact construction method are not reported in the registry record, so neither should be inferred from the interval alone.

What can and cannot be inferred from the summary statistics?

The posted comparison is unadjusted, and the record gives group means, standard deviations, participant counts, a mean difference, a confidence interval, and an unadjusted p-value. These summaries support a basic comparison of the group averages. They do not reveal whether the analysis adjusted for baseline cognition, site, or other participant characteristics. They also do not show whether all randomized participants had outcome data or whether the summaries use a particular analysis population.

General explanation: adjustment can account for prespecified covariates and may improve precision or address chance imbalance, but it also depends on a suitable model and a transparent analysis plan. The record does not describe an adjusted analysis, so no adjusted interpretation should be attributed to this result. The trial lists one primary outcome measure and no posted analyses; specific methods for handling multiple outcome domains within the MCCB are not reported.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when variation in observed data is greater than a statistical model expects. When a model does not accommodate that extra variability, standard errors can be too small and confidence intervals too narrow, making uncertainty appear less than it is. This concept is especially relevant to count or event outcomes, where a simple model may not describe participant-to-participant variation well. The posted primary endpoint is continuous, and the registry does not describe an over-dispersion assessment or a model for this outcome. There is therefore no basis to claim that over-dispersion was present or that a particular correction was applied.

What is not documented?

The registry record does not report the original sample-size justification, the detailed statistical analysis plan, the method used to construct the confidence interval, the analysis population, missing-data procedures, or any interim analyses. It also reports no posted analyses beyond the primary result. These details matter because they can affect how an estimate is calculated, how uncertainty is quantified, and how fully the results represent the randomized groups. Their absence is a limit on interpretation, not evidence that a particular method was or was not used.

10. Confidence Intervals and Precision

The estimated mean difference is -0.2 units, and its confidence interval runs from -1.592 to 1.192. The interval includes zero and extends to values on either side of it. This means the point estimate should not be treated as a precise determination of the direction or size of the group difference. The estimate is close to zero, but the interval also allows differences larger in magnitude than the point estimate in either direction.

General explanation: a confidence interval represents a range of effect values compatible with the data under the statistical method used. Its width reflects both the amount of information in the data and the variability of the outcome. A confidence interval is not a guarantee that the true effect lies within the displayed limits, and it is not a range containing a specified proportion of individual participant scores.

The group summaries include 30 participants per arm, with standard deviations of 2.8 and 2.7. These describe the number of participants and the observed spread, but the registry does not report the confidence interval’s confidence level or construction method. It also does not explain whether the endpoint values are transformed MCCB scores or another reported quantity. These gaps restrict a fuller evaluation of precision and clinical scale interpretation.

The p-value of 0.778 is consistent with substantial uncertainty about the difference, but the confidence interval is more informative about the range of estimates compatible with the data. Neither statistic demonstrates equivalence. Establishing equivalence would require an appropriate design and a prespecified margin; such a margin and an equivalence analysis are not reported in the registry record.

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 participants per arm, not total events. The registry’s planning illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers effects equal to 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. These are planning assumptions based on the observed effect, not the trial’s original design assumptions, which are not reported in the registry record.

For each assumed effect, the 80% power column gives the estimated number of participants per arm for a design intended to have an 80% chance of detecting an effect of that size under the stated assumptions. The 90% power column gives the corresponding per-arm planning number for a higher target chance of detection. These are design illustrations, not counts of participants in the completed trial.

The illustration lists 2,969 participants per arm for 80% power and 3,975 for 90% power when planning for 100% of the observed effect. For 75% of the observed effect, it lists 5,278 per arm for 80% power and 7,066 for 90% power. For 50% of the observed effect, it lists 11,876 per arm for 80% power and 15,898 for 90% power. The pattern reflects a general principle: smaller effects are harder to distinguish from random variation and therefore require more participants, all else equal.

This is not observed power. The calculations are illustrations based on the observed effect and stated planning assumptions; they do not describe the trial’s original sample-size rationale or establish what the completed trial could detect under its actual analysis plan.

12. How the Sample Size Is Derived

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

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.200; pooled standard deviation σ = 2.750.
2. n = 2 × (1.96 + 0.842)² × 2.750² / -0.200² = 2,968.8, rounded up to 2,969 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.200000000000001                      # observed difference in means
sigma <- 2.75045450789501                     # 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(306475)
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 = 2,969 (80% power), 3,975 (90% power)
 75% of observed effect: n per arm = 5,278 (80% power), 7,066 (90% power)
 50% of observed effect: n per arm = 11,876 (80% power), 15,898 (90% power)
power.t.test: 2969.8 
Simulated power at n = 2,969 per arm: 0.803 (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

Multiplicity arises when investigators evaluate more than one outcome, time point, subgroup, or statistical comparison. General explanation: testing many hypotheses can increase the chance that at least one result appears statistically notable by chance, so a plan may specify a primary comparison or adjustments to control that risk. Whether and how such procedures are needed depends on the planned analyses.

The registry lists one primary outcome measure, described as change in cognition and psychopathology assessed with the MCCB. The study objective refers to cognition and psychopathology, including positive, negative, and mood symptoms, but no secondary outcome results are posted. The record does not state whether multiple MCCB domains, symptom outcomes, or other comparisons were included in a prespecified analysis plan. Multiplicity procedures are not reported in the registry record.

The registry lists zero posted analyses. The primary mean comparison and its unadjusted p-value are available, but no further model results or secondary analyses are reported. The term unadjusted indicates that the reported p-value is not presented as adjusted for multiple comparisons; the record does not provide enough detail to determine whether any other statistical adjustment was planned or carried out.

Missing-data handling and analysis populations are also not reported in the registry record. The total enrollment and arm counts are available, but those counts alone do not establish how many participants contributed primary outcome data or whether every randomized participant was included in the analysis. The record does not describe an intention-to-treat, per-protocol, or other analysis population, and it does not state how missing outcome measurements were handled.

These omissions do not prove that missing data or multiplicity affected the result. They limit the ability to assess the analysis’s completeness, reproducibility, and sensitivity to alternative assumptions. A fuller interpretation would require the prespecified statistical plan and the numbers analyzed for the endpoint, but those details are not reported in the registry record.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

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: any situation which makes the recurrence of severe psychotic symptoms, which is harmful to patients (0.0%), severe suicidal ideation or suicidal attempt (0.0%).

The registry reports serious adverse events in the divalproex sodium group for 30 participants at risk, with zero affected, and in the placebo group for 30 participants at risk, also with zero affected. The number of deaths is reported as zero in both groups. The reported other adverse-event counts likewise list 30 participants at risk and zero affected in each group.

The registry identifies two serious-event terms: situations that make recurrence of severe psychotic symptoms harmful to patients, and severe suicidal ideation or suicidal attempt. Each is listed as 0.0%. These figures describe what was recorded in the registry; they do not establish that no adverse experiences of any kind occurred, because the posted summary reports no affected participants in the specified categories and does not provide further event narratives.

Safety counts should be read alongside the number at risk and the reporting scope. The record does not provide detailed definitions, event collection procedures, exposure time by group, or additional safety analyses. Zero reported events in these categories should not be interpreted as proof that either treatment has no safety risks.

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

The primary result is a mean difference of -0.2 units between the divalproex sodium and placebo groups at the six-week outcome assessment, with a confidence interval from -1.592 to 1.192 and an unadjusted p-value of 0.778. The estimate is close to zero, but the interval includes zero and differences in either direction. The posted data therefore do not establish a clear group difference, and they also do not establish that the groups are equivalent.

The result concerns the outcome as reported in the registry, not every possible cognitive or psychopathology measure. The registry describes the MCCB as a cognitive measure and its study objective includes psychopathology, but it does not post separate secondary analyses. The endpoint description and reported result units are not fully reconciled in the record, so the estimate should not be translated into a different scale or broader clinical meaning.

The result also does not answer whether a different duration, dose, background antipsychotic, or patient group would have a different outcome. The registry does not report enough detail about medication regimens or participant characteristics to assess those questions. Nor do the posted group averages explain individual responses; averages can conceal variation among participants.

The reported safety figures show zero affected participants for serious and other adverse events in both groups, and zero deaths. Those registry counts are limited to the recorded categories and do not prove an absence of risk. Taken together, the posted results describe a small estimated mean difference with substantial uncertainty and limited supporting analysis. They should be understood as the record’s six-week statistical comparison, not as medical advice or a definitive conclusion about treatment effects.

16. Why This Trial Matters Statistically

This trial illustrates why an effect estimate and its uncertainty should be read together. The mean difference is -0.2, but the confidence interval extends from -1.592 to 1.192. A point estimate alone may look close to no difference; the interval shows that the data remain compatible with effects in either direction. The p-value of 0.778 does not resolve that uncertainty or demonstrate equivalence.

The small enrollment, with 30 participants in each arm, is relevant to precision. The planning illustration translates the observed effect into much larger per-arm requirements under its stated assumptions. That illustration is not a reconstruction of the trial’s design rationale, and it is not observed power. It demonstrates how sample-size requirements depend strongly on the effect size chosen for planning.

The record also highlights the importance of clear endpoint definitions. The primary title refers to change measured by the MCCB, the description explains T-scores, and the posted means are 14.1 and 14.3 units. Without an explanation of how those result values relate to the described scale, readers should avoid assigning them a more specific interpretation than the registry supports.

Finally, a result is easier to evaluate when the analysis plan, missing-data approach, analysis population, and secondary findings are reported. Those details are absent or not posted in this record. Their absence does not invalidate the reported comparison, but it limits assessment of how robustly the analysis addresses the study question.

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 registry record gives limited detail about the protocol and analysis. It does not report the original sample-size justification, the full statistical analysis plan, interim analyses, missing-data procedures, or the analysis population. Although the total enrollment and group counts are listed, the number of participants contributing to the primary result is not separately stated.

The endpoint description leaves a measurement-interpretation question. It describes an MCCB composite T-score, while the reported group means are 14.1 and 14.3 units. The record does not clarify how those values relate to the described score or whether they represent a transformed measure. The title identifies the endpoint as change, but the result description does not provide further detail about baseline values or the calculation of change.

Other details that would help interpret the clinical comparison are not reported in the registry record, including the specific atypical antipsychotics used, dose information, adherence, detailed eligibility criteria, baseline characteristics, and the masking procedures. The posted results include no secondary outcome analyses, so the broad study objective cannot be assessed across separately reported cognition and psychopathology outcomes.

The safety summaries list no affected participants in the reported categories, but do not provide detailed event descriptions or collection procedures. These limitations define what can be concluded from the registry entry; they do not establish that unreported procedures or analyses were not performed.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Does The Addition Of Divalproex Sodium ER To An Atypical Antipsychotic Drug (APD) Improve Cognition And Psychopathology In Outpatients With Schizophrenia (SCH) Or Schizoaffective Disorder (SAD)? — NCT00306475.
• Registry record last updated 2025-06-27; status last verified 2022-04.
• Results first posted 2025-06-27.
• 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

This completed Phase 4 randomized, quadruple-masked, parallel-group trial enrolled 60 outpatients with schizophrenia or schizoaffective disorder. It compared divalproex sodium extended-release tablets added to an atypical antipsychotic with placebo added to an atypical antipsychotic over six weeks.

The registry reports 30 participants per arm. The primary MCCB-related result was a mean of 14.1 units in the divalproex sodium group and 14.3 units in the placebo group, with a mean difference of -0.2, confidence interval from -1.592 to 1.192, and unadjusted p-value of 0.778. No secondary analyses are posted. Serious and other adverse-event summaries report zero affected participants in both groups and zero deaths.

The estimate is uncertain, and the record does not clarify how the reported outcome units relate to its MCCB T-score description. Details of the analysis plan, missing-data handling, and original sample-size rationale are not reported in the registry record.