A Study of Olanzapine and Fluoxetine for Treatment-resistant Depression: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Olanzapine and Fluoxetine for Treatment-resistant Depression (NCT01687478), a randomized phase 3 study. The primary outcome is Mean Change From Baseline to 8 Week Endpoint in Montgomery-Äsberg Depression Rating Scale (MADRS).

2. Clinical Question

The registry describes a comparison of olanzapine plus fluoxetine with placebo plus fluoxetine for treatment-resistant depression in Chinese participants. The primary question was whether the average change in depressive-symptom severity, measured with the MADRS from baseline to the 8-week endpoint, differed between the treatment groups. The trial also recorded secondary measures of symptoms, functioning, movement-related effects, response and remission, along with adverse events.

3. Trial Design

This was a phase 3, randomized, parallel-group trial with quadruple masking. The registry lists an enrollment of 176 and identifies treatment-resistant depression as the condition. The study began in 2012-09, with primary completion recorded as 2015-11. Its status is terminated; the reason for termination is not reported in the registry record.

The intervention group received olanzapine with fluoxetine, while the comparator group received a matching placebo with fluoxetine. Both arms used daily oral treatment for 8 weeks, with the fluoxetine dose starting at 20 mg and potentially titrated to 40 mg or 50 mg. Olanzapine started at 5 mg and could be titrated to 10 mg or 15 mg. The registry does not give the actual dose distributions or describe how dose titration was decided.

The primary efficacy outcome was mean change in MADRS total score at the 8-week endpoint. The listed analysis population included randomized participants with a baseline score and at least one post-baseline score. The reported group sizes for this outcome were 87 for olanzapine plus fluoxetine and 86 for placebo plus fluoxetine. These analysis counts are not the same as the registry’s enrollment figure, and the reason for any difference is not reported in the registry record.

The registry lists eight secondary outcomes, including symptom severity, functioning, movement scales, response and remission. It also records serious and other adverse events. Details on the randomization procedure, allocation concealment, the reason the study was terminated, and the trial’s original sample-size justification are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants in the full analysis set (FAS) population: all randomized participants who had a baseline and at least one post-baseline MADRS total score measurement.

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: All randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change From Baseline to 8 Week Endpoint in Montgomery-Äsberg Depression Rating Scale (MADRS), measured over Baseline, 8 Weeks.

The primary endpoint was mean change from baseline to the 8-week endpoint in the Montgomery-Äsberg Depression Rating Scale (MADRS) total score. The scale sums ten items, and the total score can range from 0 to 60. A higher total score indicates greater depressive-symptom severity. In this setting, a negative change means the measured score was lower at follow-up than at baseline.

The registry reports least-squares mean changes and standard errors estimated with mixed model repeated measures (MMRM). The endpoint therefore compares modeled average changes between groups rather than simply presenting an unadjusted difference between raw follow-up scores. The registry describes treatment, pooled investigator, visit, and the baseline-by-treatment and visit terms as factors in that model.

MADRS is used as a symptom-severity outcome, while the group contrast addresses whether the average change differed between the assigned regimens. A mean difference does not describe every participant’s experience, and a change in a scale score alone does not establish the clinical importance of that change. The registry record does not report a threshold for interpreting the primary mean difference as clinically important.

7. Results Posted in the Registry

• Olanzapine + Fluoxetine: mean -15.92 versus -14.30 on control; difference -1.62 (95% CI -5.06 to 1.82), standardized effect d = -0.14; z-test p = 0.356.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Mean Change From Baseline to 8 Week Endpoint in Clinical Global Impressions-Severity of Depression (CGI-S) Scale

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Olanzapine + Fluoxetine: mean -1.43 versus -1.63 on control; difference 0.20 (95% CI -0.23 to 0.63), standardized effect d = 0.14; z-test p = 0.362.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Olanzapine + Fluoxetine and the control. The interval gives the range of effects that remain compatible with the data.

Mean Change From Baseline to 8 Week Endpoint in the Simpson-Angus Scale (SAS)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Olanzapine + Fluoxetine: mean 0.03 versus -0.14 on control; difference 0.17 (95% CI -0.34 to 0.68), standardized effect d = 0.10; z-test p = 0.512.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Olanzapine + Fluoxetine and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants Who Achieve a Response Based on a ≥50% Reduction From Baseline in MADRS Total Score

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Olanzapine + Fluoxetine: 65.5% (57/87) versus 61.6% (53/86) on control; risk difference 3.9 percentage points (95% CI -10.4 to 18.2), risk ratio 1.06 (0.85–1.33); unadjusted two-proportion z-test p = 0.595.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Olanzapine + Fluoxetine and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes span symptom ratings, movement-related scales, functioning, response, and remission. Among the secondary comparisons with numerical results in the analysis record, the estimates are not clearly separated from their null values: the reported recalculated intervals include no difference for the CGI-S comparison, the SAS comparison, and the MADRS response comparison. Taken together, these results do not supply a clear secondary-outcome signal that changes the broad reading of the primary result.

The secondary outcomes provide different views of the trial, but they are not interchangeable. Symptom scales measure related constructs, movement scales address different potential effects, and response or remission classifications use thresholds rather than average score change. Agreement or disagreement across such outcomes can be informative, but interpretation depends on how they were specified and analyzed.

The registry lists multiple secondary outcomes, and the posted unadjusted comparisons do not account for the multiplicity of examining several outcomes. General explanation: when many hypotheses are tested, the chance of at least one apparently unusual result can increase even if there are no true effects. Without a reported multiplicity procedure, individual unadjusted secondary findings should carry less weight than a prespecified primary analysis. The registry does not report a multiplicity adjustment or a hierarchy for interpreting the secondary outcomes.

The registry lists secondary measures beyond the comparisons with numerical estimates described in the analysis record. Their presence in the outcome list should not be mistaken for a fully interpretable set of results: the relevant estimates, uncertainty, and analysis details are not reported for every listed outcome in the record summarized by the posted analysis. No claim about the direction or size of those unreported comparisons can be made from the outcome names alone.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Mean Change From Baseline to 8 Week Endpoint in Clinical Global Impressions-Severity of Depression (CGI-S) Scale
• Mean Change From Baseline to 8 Week Endpoint in the Simpson-Angus Scale (SAS)
• Mean Change From Baseline to 8 Week Endpoint in the Short-Form 36 Health Survey (SF-36)
• Mean Change From Baseline to 8 Week Endpoint in the Sheehan Disability Scale (SDS)
• Percentage of Participants Who Achieve a Response Based on a ≥50% Reduction From Baseline in MADRS Total Score
• Percentage of Participants Who Achieve Remission Based on MADRS Total Score ≤10 at 8 Weeks
• Mean Change From Baseline to 8 Week Endpoint in the Barnes Akathisia Scale (BAS)
• Mean Change From Baseline to 8 Week Endpoint in the Abnormal Involuntary Movement Scale (AIMS)

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports least-squares mean change for the primary outcome, with standard error as the measure of uncertainty around each group estimate. It says these estimates came from mixed model repeated measures (MMRM), with factors for treatment, pooled investigator, visit, and the baseline-by-treatment and visit terms. The registry record does not report further model details such as the covariance structure, degrees-of-freedom method, or handling of missing values.

What does a mean change represent?

A mean change summarizes the average within-group movement in the outcome from baseline to follow-up. Here the scale runs from 0 to 60, and higher scores indicate greater symptom severity, so a negative mean change corresponds to a decrease in the measured score. The group mean is not a description of each participant, and it can conceal variation in individual changes.

A between-group mean difference compares those average changes. The recalculated difference of -1.62 is calculated in the direction of olanzapine plus fluoxetine minus placebo plus fluoxetine. A negative value therefore indicates a numerically greater decrease in the first group’s score. The sign does not, on its own, establish that the difference is meaningful or reliable.

Why use a repeated-measures model?

General explanation: repeated-measures models can use measurements collected at multiple visits and account for the fact that measurements from the same participant are related. They can estimate group patterns over follow-up while including baseline and other specified factors. Their results depend on the model specification and assumptions about the observed data. The registry identifies MMRM and some model factors, but the full specification and assumptions are not reported in the registry record.

The posted group means are least-squares means, which are model-based estimates rather than necessarily raw averages. The recalculation uses the posted group summaries and does not reproduce the model from participant-level records. The primary comparison’s pooled standard deviation is marked as approximate in the recalculation. This is another reason to distinguish the recalculated summary-statistics result from the trial’s reported MMRM result.

What does the confidence interval add?

General explanation: a confidence interval describes the range of values reasonably compatible with the data and the method used, under that method’s assumptions. It communicates precision as well as direction. An interval that includes zero for a mean difference includes no between-group difference among its compatible values. It can also include values in either direction, so it does not demonstrate equivalence.

For the recalculated primary mean difference, the interval runs from -5.059 to 1.819. Its inclusion of zero aligns with the unadjusted recalculated p-value of 0.356: these are two ways of conveying that the summary-statistics comparison does not give clear evidence against a zero difference under the corresponding unadjusted test. Neither statistic measures the probability that a treatment is effective.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when observed variability is greater than a model assumes. When that happens, standard errors can be too small if the extra variability is ignored, which can make intervals narrower and tests appear more decisive than warranted. The primary endpoint is continuous, and the registry reports standard errors from MMRM; it does not report that over-dispersion was assessed or identified. The concept is relevant to interpreting statistical models generally, but it should not be taken as a claim about this trial.

What is the role of the recalculated test?

The unadjusted recalculation provides a consistent, transparent comparison based on the posted group summaries. It is not the trial’s full adjusted model and does not account for unreported details of the analysis. The registry’s own estimate and confidence interval should be read as the reported trial result; the recalculated values offer a separate summary-level perspective. The registry does not report the primary analysis p-value, multiplicity procedure, interim analysis plan, or complete statistical analysis plan.

11. Confidence Intervals and Precision

The primary recalculated mean difference is -1.62, with a 95% confidence interval from -5.059 to 1.819. The interval includes zero and extends to values in both directions. It therefore leaves uncertainty about whether the group contrast is negative or positive, as well as about its magnitude. The confidence interval is more informative than the point estimate alone because it shows the range of estimates compatible with the recalculation’s method and posted summaries.

The registry’s reported analysis gives the same mean difference, -1.62, with a 95% confidence interval from -5.04 to 1.80. The registry does not report a p-value for its own analysis. The recalculated interval is based on summary statistics and an approximate standard deviation, whereas the trial’s reported primary endpoint analysis used MMRM. The two intervals should therefore be understood in the context of their different bases, rather than treated as if both came from the same fitted model.

Neither interval establishes that the groups are equivalent. A confidence interval containing zero means that a zero difference is among the values compatible with the analysis; it does not show that all clinically important differences have been excluded. The registry record does not report a clinical-importance threshold for the MADRS mean difference, so the clinical meaning of the interval’s endpoints cannot be determined from the record alone.

General explanation: precision concerns how narrowly an estimate is located, while statistical significance concerns a decision under a specified testing procedure. A result can be uncertain even when its point estimate has a particular direction. Here, the range crossing zero is a central feature of the uncertainty, not evidence that the groups have exactly the same outcome.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are participants per arm for the continuous and binary outcomes represented in the planning display; they are not total event counts. The planning illustration assumes two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its starting point and presents planning for 100%, 75%, and 50% of that observed effect. These are illustrations based on the posted result, not the trial’s original design assumptions; the original sample-size justification is not reported in the registry record.

The 80% power column indicates the illustrative number of participants per arm under the stated assumptions for a design intended to have an 80% chance of detecting the specified effect if that effect were present. The 90% power column expresses the analogous planning target for a 90% chance under those assumptions. Power is a design property under an assumed effect and analysis, not a probability that the posted result is true.

Each row changes the effect size used for planning. Planning for 75% or 50% of the observed effect means targeting a smaller effect than the full observed estimate. Smaller effects are harder to distinguish from random variation, so the illustration requires more participants per arm as the assumed effect becomes smaller.

These figures are not observed power. They should not be used to infer that the completed trial was adequately or inadequately powered without the original design assumptions, the trial’s planned analysis, and other relevant design information. The registry record does not report those original planning details.

13. 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 = -1.620; pooled standard deviation σ = 11.535.
2. n = 2 × (1.96 + 0.842)² × 11.535² / -1.620² = 795.8, rounded up to 796 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 <- -1.62                      # observed difference in means
sigma <- 11.5345359884061                     # 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(1687478)
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 = 796 (80% power), 1,066 (90% power)
 75% of observed effect: n per arm = 1,415 (80% power), 1,894 (90% power)
 50% of observed effect: n per arm = 3,184 (80% power), 4,262 (90% power)
power.t.test: 796.8 
Simulated power at n = 796 per arm: 0.801 (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.

14. Analyses Reported by the Sponsor

The registry’s own primary analysis reports a mean difference of -1.62 and a 95% confidence interval from -5.04 to 1.80. It does not report a p-value for that analysis. This is the trial’s reported result and should be distinguished from the separate recalculation based on group summary statistics.

The recalculation gives the same point estimate, -1.62, with a 95% confidence interval from -5.059 to 1.819 and an unadjusted p-value of 0.356. Its interval is derived from posted summaries rather than from the full MMRM fit. The registry identifies the primary outcome estimates as least-squares means from MMRM, while the recalculation uses an approximate standard deviation. The recalculated p-value is consequently not the sponsor’s analysis p-value.

Both intervals include zero, so the broad interpretation of the primary contrast is similar: neither result establishes a clear difference between groups on the posted primary endpoint. The numerical bounds are not identical, and the registry does not provide enough model detail to reconcile their precise differences. The recalculation is a supplementary summary-level check, not a replacement for the registry’s reported analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary MADRS outcome and lists multiple secondary outcomes. It reports an MMRM approach for the primary outcome, but the complete statistical analysis plan is not reported in the registry record. In particular, the record does not report a multiplicity adjustment, a testing hierarchy across primary and secondary outcomes, or whether any secondary hypotheses were intended to be confirmatory.

General explanation: multiplicity arises when several hypotheses or outcomes are examined. Without an adjustment or an explicit hierarchy, the chance of finding at least one apparently unusual result can rise. This does not make every unadjusted result invalid, but it means isolated secondary findings are less persuasive than results interpreted under a prespecified plan. The secondary comparisons in the analysis record are unadjusted recalculations and should be read with that limitation.

The primary analysis population is defined as the full analysis set: all randomized participants who had a baseline MADRS total score and at least one post-baseline MADRS measurement. This definition excludes randomized participants without the required measurements from the stated primary population. The reported primary outcome group sizes are 87 and 86. The registry does not report how missing outcome measurements were handled in the MMRM, whether additional imputation was used, or the results of sensitivity analyses for missing data.

General explanation: missing-data methods rely on assumptions about why measurements are unavailable and how missing outcomes relate to observed information. Different assumptions can affect estimates and uncertainty. The record’s population definition and model label do not, by themselves, establish which missing-data assumptions were used or whether those assumptions were examined.

The record also does not report interim analyses, stopping rules, or the reason for termination. Those omissions limit assessment of whether the reported analysis followed the intended plan. They do not demonstrate that the analysis was inappropriate; they identify information that is not available in the registry record.

16. 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: suicide attempt (1.1%), restlessness (1.1%).

The registry reports serious adverse events among 1 participant in each group: the olanzapine-plus-fluoxetine group had 1 serious event among 88 participants at risk, and the placebo-plus-fluoxetine group had 1 among 87 participants at risk. The record also lists suicide attempt and restlessness as top serious-event terms, each at 1.1%. It does not assign those terms to a particular group in the posted summary, so they should not be attributed to one arm.

Other adverse events were recorded for 55 participants among 88 at risk in the olanzapine-plus-fluoxetine group and 45 among 87 at risk in the placebo-plus-fluoxetine group. These are descriptive registry counts, not evidence that a particular event was caused by a treatment. The registry does not provide the detail needed to evaluate event severity beyond the serious-event classification, timing, duration, or individual clinical context.

Deaths are not reported in the adverse-event record. The trial’s safety counts should be interpreted alongside its 8-week treatment period and the registry’s reported population at risk, not as a complete assessment of longer-term safety.

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

The primary result describes an estimated difference in average MADRS change at the 8-week endpoint. The point estimate is numerically in the direction of a larger decrease in the olanzapine-plus-fluoxetine group, but both the sponsor-reported and recalculated confidence intervals include zero. The posted results therefore do not establish a clear between-group difference on this measure.

This does not show that the treatments are equivalent, nor does it prove that there is no difference. The interval remains compatible with estimates in either direction, and the registry does not report a clinical-importance threshold that would show which values matter to participants. The result is a comparison of group averages, not a prediction for an individual participant.

The finding is specific to the population, treatment regimens, outcome definition, and 8-week endpoint described in the registry. It does not establish the comparison at other time points, in other populations, or under different dose patterns. The registry record also does not report the reason the study was terminated or the original design assumptions, both of which limit contextual interpretation.

Safety information is also descriptive. The posted serious and other adverse-event counts do not establish whether events were caused by treatment, and the record does not report deaths. Secondary outcomes offer additional measures, but unadjusted comparisons across multiple outcomes carry added uncertainty and should not override the primary result in the absence of a reported analysis hierarchy.

Overall, the posted data support a cautious reading: a numerical difference was estimated, but uncertainty around that estimate is substantial enough that the registry results do not clearly distinguish the groups on the primary endpoint.

18. Why This Trial Matters Statistically

This trial illustrates why a point estimate should be read together with its uncertainty. The primary mean difference has a definite numerical direction, but the confidence interval includes zero and extends to estimates on either side. Reporting only the direction or magnitude of the point estimate would omit an important part of the statistical evidence.

It also illustrates the distinction between a trial’s reported model and a summary-statistics recalculation. The registry identifies an MMRM analysis of least-squares means, while the recalculated comparison is built from posted group summaries and an approximate standard deviation. Such a recalculation can make the comparison easier to inspect, but it cannot reproduce all features of the fitted model. The two analyses should be labeled and interpreted according to how they were obtained.

The trial also has a broad outcome set, including symptom, functioning, movement, response, remission, and safety measures. Multiple outcomes can help describe different aspects of a study, but unadjusted testing across them creates a multiplicity concern. Without a reported hierarchy or adjustment, secondary results are more appropriately treated as supporting or exploratory evidence rather than as a collection of independent confirmations.

Finally, the registry record shows why population definitions and missing-data details matter. The primary analysis population required baseline and post-baseline measurements, but the record does not describe the missing-data method or sensitivity analyses. Those details are essential to judge how robust the model-based estimate is to incomplete follow-up. The information available supports a careful account of what was estimated while leaving these design and analysis questions unresolved.

19. 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 study status is terminated, but the reason for termination is not reported in the registry record. The record also does not give the original sample-size justification or the assumptions used to plan the trial, so the planning illustration based on the observed effect cannot establish whether the original design was adequate for its intended question.

The primary outcome was measured at the 8-week endpoint, and the registry describes short-term treatment. The reported result therefore does not address longer-term outcomes. The actual dose distributions and the process used for titration are not reported, limiting interpretation of how closely the observed treatment groups followed the permitted dosing ranges.

The primary analysis population required a baseline and at least one post-baseline MADRS measurement. The registry record does not describe missing-data handling, imputation, sensitivity analyses, or the extent and reasons for missing measurements. It also does not report the full MMRM specification, including its covariance structure and other technical details. These omissions constrain independent assessment of model robustness.

The registry lists multiple secondary outcomes, but the analysis record does not provide numerical estimates for every listed outcome. No multiplicity adjustment or outcome hierarchy is reported. Safety figures are also limited to posted counts and terms; deaths are not reported, and the record does not provide enough detail to infer causality or longer-term risk.

Finally, the posted summary cannot show participant-level distributions or explain whether average changes were similar across participants. The available group-level estimates are useful for describing the reported comparison, but they cannot answer questions requiring individual data or unreported analyses.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Olanzapine and Fluoxetine for Treatment-resistant Depression — NCT01687478.
• Registry record last updated 2019-10-09; status last verified 2019-09.
• Results first posted 2017-08-24.
• 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.

23. Record Summary

This phase 3 randomized, quadruple-masked, parallel-group study compared olanzapine plus fluoxetine with placebo plus fluoxetine for treatment-resistant depression in Chinese participants. The registry records enrollment of 176, a start date of 2012-09, primary completion in 2015-11, and status as terminated. The reason for termination is not reported in the registry record.

The primary outcome was mean change in MADRS total score at the 8-week endpoint. The registry reports least-squares mean changes of -15.92 and -14.3, with a reported mean difference of -1.62 and a 95% confidence interval from -5.04 to 1.80. The recalculated interval also includes zero. The posted result does not establish a clear between-group difference on the primary outcome.

The record lists secondary symptom, functioning, movement, response, and remission outcomes, as well as adverse events. Important details about missing-data methods, multiplicity, original sample-size planning, and the reason for termination are not reported in the registry record.