Treatment of Psychotic Major Depression With Mifepristone: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Treatment of Psychotic Major Depression With Mifepristone (NCT00867360), a randomized phase 3 study. The primary outcome is Change in Psychotic Symptoms Subscale (PSS) of the Brief Psychiatric Rating Scale (BPRS).

2. Clinical Question

The registry record situates the trial within a broader investigation of the hypothalamic-pituitary-adrenal and dopamine axis in psychotic depression. It also says the study was conducted in conjunction with an observational study of clinical and biological characteristics of psychotic depression. A secondary outcome concerns percentage change in mean evening cortisol before and after Florinef treatment. The record names this outcome but does not post a result for it. Therefore, the clinical question is described more fully than the comparative statistical evidence available in the record.

3. Trial Design

The registry classifies this as a phase 3, randomized, triple-masked study sponsored by Stanford University. The listed intervention groups are mifepristone and placebo, and the arm descriptions specify mifepristone for 8 days or placebo rather than mifepristone. The registry also labels the study model single group. That classification sits alongside the listed two-arm interventions and randomized allocation; the record does not explain how those design labels relate to one another.

The study start is recorded as 2005-08, and primary completion as 2009-05. Its status is terminated. The enrollment figure in the registry is 10. For the primary continuous outcome, the posted group summaries report 3 participants for mifepristone and 5 for placebo. The registry does not explain the relationship between the enrollment figure and the outcome-group counts, nor does it describe the disposition of participants.

The primary outcome is a mean change score from baseline to day 9. Other outcome measures are also registered: the record lists 3 outcome measures and names one secondary outcome, a percentage change in mean evening cortisol after treatment with mifepristone or placebo. However, the analysis count is 0, and the cortisol outcome does not have a posted result in the record. The study description says the research examined hormones and mood and thinking, but details of the protocol, timing of other measurements, and analysis procedures are not reported in the registry record.

The short treatment duration and endpoint timing define what the primary outcome can address: symptom-scale change across the interval from day 1 to day 9. They do not, on their own, establish longer-term outcomes. The record does not report follow-up beyond the primary endpoint timeframe, an allocation sequence, masking procedures, or a statistical analysis plan. Those omissions limit how much can be inferred about conduct and analysis beyond the design labels and arm descriptions that are registered.

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

Reasons for not completing this period:

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 Psychotic Symptoms Subscale (PSS) of the Brief Psychiatric Rating Scale (BPRS), measured over baseline to day 9.

The primary endpoint is change in the Psychotic Symptoms Subscale (PSS) of the Brief Psychiatric Rating Scale (BPRS), expressed in units on a scale. The BPRS rates psychiatric symptoms, with each item scored from 1 to 7 and 1 meaning not present. The PSS sums 4 items representing positive psychotic symptoms. The record describes adjusting the sum by subtracting 4, yielding a range from 0 to 24; higher scores indicate greater levels of positive psychotic symptoms.

The endpoint is defined as the PSS total on day 1 minus the PSS total on day 9. Under that direction, zero indicates no change, and positive values indicate a decrease in psychotic symptoms. This definition matters because the numerical sign is the reverse of a simple day 9 minus day 1 difference: a positive change score is interpreted as symptom reduction, not worsening. The measure is a mean continuous outcome, with standard deviation as its reported measure of spread.

A symptom subscale can provide a structured way to quantify change in a defined domain of psychiatric symptoms. The posted summaries give a mean change of 0.0 units for mifepristone and 1.0 units for placebo. These group means describe the registered change-score measure; they are not themselves a formal estimate of the contrast between groups. The registry does not report confidence intervals or a formal two-arm statistical test for this outcome. The record’s endpoint definition supports interpreting the direction of each group mean, but not a conclusion about comparative efficacy.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous measure summarized with a mean and standard deviation. A mean describes the arithmetic center of the recorded change scores, while a standard deviation describes how dispersed those scores are around the mean. These summaries answer different questions: the mean describes average change within a group, and the standard deviation describes variation among participants. Neither quantity alone gives the uncertainty around a comparison between randomized groups.

What does the change score mean?

The registry defines change as day 1 PSS total minus day 9 PSS total. A positive value therefore means the later score was lower than the earlier score, consistent with a decrease in positive psychotic symptoms. A value of zero means the two scores were the same. The group means, 0.0 for mifepristone and 1.0 for placebo, are means of this specifically oriented change score. They should not be reinterpreted using the more common convention of follow-up minus baseline without reversing the sign.

The score is based on 4 BPRS items. The individual items range from 1 to 7, and subtracting 4 from the sum produces a PSS range of 0 to 24. Higher PSS scores indicate greater positive psychotic symptoms. The scale description explains the direction and range of the endpoint, but does not establish that the score is normally distributed or that its change scores meet any particular statistical assumptions.

Why is a treatment contrast different from group means?

A randomized trial’s comparative question is generally about the difference between outcomes under the intervention and comparator, not merely whether either group’s mean is above or below zero. A direct contrast estimates how the groups differ, subject to the analysis method and uncertainty. The registry posts the group means and standard deviations but does not report a formal two-arm analysis for this continuous endpoint. The recalculated-comparison pathway is not available because values are missing or not estimable for a two-arm comparison. No contrast, confidence interval, or p-value should therefore be inferred from the descriptive summaries.

In general, a confidence interval gives a range of values compatible with the data and the statistical model, while reflecting sampling uncertainty. A narrow interval indicates greater precision than a wide interval, all else equal. The registry record does not report a confidence interval for the primary comparison, so it is not possible to assess the precision of a treatment contrast from the posted outcome summaries. A p-value, if reported, would address compatibility with a specified null model; it would not by itself measure the size or clinical importance of an effect. No p-value is reported for the primary endpoint.

What does the standard deviation tell the reader?

The standard deviations are 2.6 in the mifepristone group and 0.625 in the placebo group. They describe dispersion in the recorded scores, not the uncertainty of the means or the difference between means. With small outcome groups, a few individual observations can have substantial influence on a group mean and its spread. Individual data are not reported, so the shapes of the score distributions and the presence of influential observations cannot be checked.

In general, over-dispersion means that observed variability is greater than a model assumes. If a method underestimates variability, its uncertainty estimates can be too small and its intervals too narrow. This is a general statistical consideration, not a finding established by the registry summaries. The posted standard deviations alone are not enough to diagnose over-dispersion or select an appropriate model.

How should planning illustrations be read?

The stated planning assumptions describe an illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. The target effect is defined as the stated fraction of the observed effect. These are assumptions for planning, not information about the original trial’s sample-size justification. That justification is not reported in the registry record.

In general, a sample-size calculation estimates the number of participants needed under specified assumptions about the target effect, variability, significance level, and power. It is conditional on those assumptions and is not a measure of the probability that a completed trial’s result is true. A planning illustration based on an observed effect can be especially uncertain when the observed groups are small, because the effect estimate itself may be imprecise. Any calculated participant counts should be read as planning illustrations rather than observed power or a retrospective verdict on the trial.

The registry does not report the statistical analysis plan, the method used to estimate variability for planning, any adjustment for baseline score, or any handling of repeated measurements. Consequently, the assumptions stated for planning do not document the method used to analyze the primary endpoint. The registry also does not report interim analyses, stopping rules, or a prespecified strategy for multiplicity. It would be inappropriate to infer any of these from the phase, masking, or allocation labels.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Mifepristone", "Placebo")
n   <- c(3, 5)
m   <- c(0, 1)
sp  <- c(2.6, 0.625)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Mifepristone  3  0.000  2.600  -6.459  6.459
CARM  Placebo  5  1.000  0.625  0.224  1.776

11. Multiplicity, Missing Data and Analysis Populations

The record lists 3 outcome measures and names one secondary outcome: percentage change in mean evening cortisol before and after treatment with mifepristone or placebo. It does not post a result for that secondary measure. The analysis count is 0. The registry therefore does not provide a set of reported analyses from which to determine how many outcomes were tested, how outcomes were prioritized, or whether any adjustment for multiple comparisons was used. A multiplicity plan is not reported in the registry record.

The primary endpoint summaries include 3 participants in the mifepristone group and 5 in the placebo group. The record does not state whether these are all randomized participants with available outcomes, identify reasons for absent measurements, or explain how missing values were handled. It does not report an intention-to-treat population, a per-protocol population, or another analysis population definition. The relationship between the registry enrollment figure of 10 and the outcome-group counts is not explained.

In general, missing outcome data can affect both the precision and interpretation of an estimate. The direction and extent of possible bias depend on why outcomes are missing and on the assumptions used to address them. Those details are unavailable here, so no particular missing-data mechanism or correction can be assumed. Similarly, a registered outcome without a posted result cannot be treated as evidence of either a positive or a negative finding. The record’s limited analysis reporting constrains assessment of selective reporting and the robustness of the primary result.

12. 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

Serious events with the highest arm-level rate: psychaitric decompensation requiring hospitalization (16.7%).

The serious adverse event figures are descriptive counts, not a comparative safety analysis. In the mifepristone group, the record lists 0 serious adverse events among 4 participants at risk. In the placebo group, it lists 1 among 6 participants at risk. The record identifies the serious event term as “Psychaitric decompensation requiring hospitalization” and reports 16.7% for that term. The spelling is reproduced as recorded.

The at-risk counts differ from the primary outcome group counts: the primary summaries include 3 participants for mifepristone and 5 for placebo. The registry does not explain the difference between these denominators. No deaths are recorded in either group, and the count of other adverse events is 0 in each group. These small descriptive counts do not establish whether the groups differ in safety, and the record does not report a statistical comparison, confidence interval, event follow-up details, or a broader safety analysis. A zero count in a small group does not demonstrate that an event cannot occur.

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

The posted primary outcome data show mean PSS change scores of 0.0 units for mifepristone and 1.0 units for placebo, measured from day 1 to day 9 using day 1 minus day 9. Given that definition, the mifepristone group mean corresponds to no average change, while the placebo group mean corresponds to a decrease in psychotic symptoms. The accompanying standard deviations are 2.6 and 0.625, respectively. These are descriptive results for the recorded outcome groups.

They do not establish that the treatments differ. The registry does not report a formal two-arm analysis, a confidence interval, or a p-value for the continuous primary endpoint. The group means alone cannot show the precision of a contrast, account for potential baseline imbalance, or determine the influence of individual observations. Nor does the registry explain the difference between the enrollment figure and the outcome-group counts or report missing-data handling.

The findings also do not establish longer-term effects. The primary measure covers baseline to day 9, and the arm description specifies an 8-day treatment period. The record does not report follow-up beyond that endpoint timeframe. Although the study description concerns hormones as well as mood and thinking, the named secondary cortisol outcome has no posted result. A registry outcome label or study objective is not itself a result.

The record is marked terminated, but does not state why. Its randomized and triple-masked labels are accompanied by a single-group model classification, which is not explained. These features, along with the lack of a reported statistical plan and analysis, leave important questions about conduct and inference unresolved. The registered record supports a narrow description of the posted means and safety counts; it does not support a definitive comparative conclusion about treatment efficacy or safety.

14. Why This Trial Matters Statistically

This trial is statistically notable because the registry combines design labels that are difficult to reconcile from the record alone. Allocation is classified as randomized and masking as triple, and two intervention arms are described, yet the model is labeled single group. The record does not explain the classification. For readers assessing a comparative question, this makes the design documentation less complete than the arm descriptions might suggest.

The primary endpoint is clearly defined in direction and scale, which is essential for interpreting its sign. The score uses day 1 minus day 9, so positive values indicate a symptom decrease. The posted group means can be understood correctly only with that orientation in mind. At the same time, the absence of a formal between-group analysis means that the descriptive means cannot answer the central comparative question with known precision.

The record also illustrates why outcome counts and analysis populations matter. The primary summaries use groups of 3 and 5 participants, while safety denominators are 4 and 6. The registry enrollment is 10, but it does not explain how these figures relate or describe participant flow. This prevents a clear reconstruction of which participants contributed to each outcome and whether missing measurements could affect the results.

Finally, the registry lists multiple outcomes but gives no analyses and no result for its named secondary cortisol measure. Without a stated multiplicity strategy or complete outcome reporting, readers cannot assess the planned hierarchy of evidence. These are not merely technical details: they determine how confidently a result can be interpreted, how uncertainty should be communicated, and whether the registered clinical question is answered by the posted statistical record.

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 primary outcome groups are small, with 3 participants in the mifepristone summary and 5 in the placebo summary. Estimates from small groups can be sensitive to individual observations. The registry does not post individual scores, so that sensitivity cannot be examined. It also does not provide confidence intervals for the group means or a treatment contrast, leaving the precision of the comparative result unreported.

The record gives no formal analysis of the continuous primary endpoint, no p-value, and no statistical analysis plan. It does not state whether the comparison was adjusted for baseline PSS score or identify any other covariates or model assumptions. No interim-analysis or stopping-rule information is reported. The record’s terminated status is not accompanied by a reason, so its effect on study completion or reporting cannot be assessed.

There is also incomplete information about participant accounting. The registry lists enrollment of 10, while primary outcome group summaries include 3 and 5 participants and safety denominators are 4 and 6. These figures should not be combined into a total or treated as interchangeable: the record does not explain the differences. It does not report participant flow, analysis-population definitions, missing-data reasons, or methods for handling missing outcomes.

Finally, the registry lists 3 outcome measures, names one secondary outcome, and reports 0 analyses. It provides no result for the named cortisol outcome and no multiplicity strategy. The single-group model label appears alongside randomized allocation and two listed arms, without explanation. These reporting gaps make it difficult to evaluate the intended comparison, completeness of outcome reporting, and reliability of the primary result beyond the descriptive summaries.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Treatment of Psychotic Major Depression With Mifepristone — NCT00867360.
• Registry record last updated 2017-04-04; status last verified 2017-02.
• Results first posted 2017-04-04.
• 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 record describes a phase 3 study of mifepristone versus placebo for psychotic depression, sponsored by Stanford University. It labels the study randomized and triple-masked, lists a single-group model, and records status as terminated. The primary outcome is mean change in the PSS of the BPRS from baseline to day 9, defined as day 1 minus day 9.

The posted primary group summaries are 0.0 units for mifepristone, with standard deviation 2.6 and 3 participants, and 1.0 units for placebo, with standard deviation 0.625 and 5 participants. The record reports 0 analyses and no formal two-arm result for this continuous endpoint. It also names a secondary cortisol outcome without a posted result. Safety summaries list 0 serious adverse events among 4 at risk in the mifepristone group and 1 among 6 at risk in the placebo group; no deaths are recorded. The record does not report a statistical analysis plan, missing-data methods, or the reason for termination.