Duloxetine Treatment in Elderly With Dysthymia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Duloxetine Treatment in Elderly With Dysthymia (NCT01852383), a na phase 4 study. The primary outcome is Change in Hamilton Rating Scale for Depression (HAM-D, 24-item) From 0 Weeks to 12 Weeks..

2. Clinical Question

3. Trial Design

This was a Phase 4, open treatment trial of duloxetine in elderly patients with dysthymic disorder. The registry identifies a single-group model, no masking, and allocation as NA. These features describe a treatment study without a randomized or masked comparison group; the registered design does not include a concurrent control arm.

The record lists an enrollment of 30. The study start is recorded as 2006-01, primary completion as 2010-09, and status as completed. New York State Psychiatric Institute is listed as the sponsor. These registry entries establish the recorded administrative and design context, but they do not supply outcome results.

Before treatment, the arm description required a minimum 1-week psychotropic medication washout, with a 3-week washout for fluoxetine and monoamine oxidase inhibitors. The dose schedule then describes duloxetine at 20 mg daily for the first week, 30 mg daily for the second week, and 60 mg daily for another 4 weeks. The remaining dosing description is cut off in the registry record, so the full treatment schedule cannot be reconstructed from that text.

The primary assessment was the change in HAM-D score from Week 0 to Week 12. The record also names change in Cornell Dysthymia Rating Scale scores from Week 0 to Week 12 as a secondary outcome. It lists 4 outcome measures and 4 analyses, but does not provide the full outcome and analysis details in the record summarized by the posted results.

Because treatment was not compared with a separate group, changes over time cannot be separated from other influences using a randomized between-group contrast. An open design also means treatment and outcome assessment were not masked as part of the registered design. Details about assessor masking, adherence, concomitant care, and the planned statistical analysis 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

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 Hamilton Rating Scale for Depression (HAM-D, 24-item) From 0 Weeks to 12 Weeks., measured over Screen (0) and 12 weeks.

The primary endpoint is change in the 24-item Hamilton Rating Scale for Depression (HAM-D) between Week 0 and Week 12. A research rater completed the scale at each visit. The instrument is described as ranging from 0-50, with lower scores or decreasing scores representing lower depressive symptom severity and better outcome, and higher or increasing scores representing greater severity and worse outcome.

The registry defines change by subtracting the Week 12 score from the Week 0 score. With that direction, a positive change indicates that the recorded score was lower at Week 12 than at Week 0; a negative change indicates a higher Week 12 score. This sign convention is important when reading a reported mean because change scores can be defined in either direction. The measure's scale and scoring direction help give the change a clinical interpretation, but the registry record does not report the mean change or its standard deviation.

A repeated symptom rating can describe how scores differ between the start and end of treatment among those included in an analysis. It does not, on its own, establish that treatment caused the observed change. Without a concurrent comparator, time trends, expectations, changes in other care, and other factors cannot be distinguished from a treatment effect by a between-group comparison. The specific rationale for choosing HAM-D as the primary measure is not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies a continuous primary outcome summarized by a mean, with standard deviation as the dispersion measure. It does not report the outcome values or the statistical analysis plan. The following concepts are general explanations of how these methods are commonly read, not descriptions of analyses confirmed for this trial.

What does a mean change describe?

A mean change is an average of participants' individual change scores, using the direction specified for the endpoint. Here, the score at Week 12 is subtracted from the Week 0 score. A positive mean would therefore indicate a lower average HAM-D score at the later assessment, while a negative mean would indicate a higher average score. An average does not show whether every participant changed in the same direction. Different individual patterns can produce the same mean.

The mean is most informative alongside a measure of variation and a clear account of who contributed data. The standard deviation describes how spread out individual observations are around the mean. It is not the same as the uncertainty in the estimated mean: uncertainty also depends on how many observations were analyzed and on the analysis method. Neither the standard deviation nor the analyzed sample size for the primary endpoint is reported in the registry record.

What would a confidence interval add?

General: A confidence interval expresses the uncertainty associated with an estimated quantity under a specified statistical procedure and its assumptions. For a mean change, it would give a range of values compatible with the observed data at the stated confidence level. A narrow interval would indicate greater precision than a wide interval, but precision alone does not establish clinical importance or causation. The registry record does not report a confidence interval or the confidence level used for the primary outcome.

A confidence interval must be interpreted in relation to the question and design. In this single-group study, an interval around the mean change would concern the estimated average change in the analyzed participants under the analysis assumptions. It would not create a comparison with an untreated group and would not remove the limitations of an open, uncontrolled design. The interval's width and location cannot be evaluated because the estimate and its uncertainty are not reported.

Why does the design matter for the estimate?

General: Random assignment to concurrent groups can, in expectation, balance both measured and unmeasured influences on outcomes. A single-group design does not offer that comparison. A before-and-after estimate may be affected by changes that occur over time even without the intervention, by participants' expectations, or by differences in who has measurements at each visit. These considerations do not prove that any particular bias occurred; they describe why a within-group change has a narrower interpretation than a randomized treatment contrast.

Open treatment means that masking was absent in the registered design. Knowledge of treatment can affect participant-reported or rater-assessed outcomes through expectations or assessment behavior. HAM-D was completed by a research rater, but further details about assessor procedures or masking are not reported. The design information alone cannot determine whether measurement bias occurred.

How should the planning assumptions be read?

General: A sample-size calculation is a planning exercise that asks how much information a future study might need to detect a specified effect under chosen assumptions. It is not an observed result, and it is not the same as observed power. The registry's planning assumptions describe a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the target effect expressed as a fraction of the observed effect. These are planning assumptions, not evidence that this single-group trial used a two-arm design.

The sample-size assumptions are explicitly described as an illustration computed from the observed effect, rather than as the trial's own design. The primary observed effect needed to interpret such an illustration is not reported in the registry record. Consequently, the illustration cannot be independently assessed from the endpoint information available. Even when a planning estimate is available, its sample sizes are prospective planning quantities, not numbers of participants observed in a trial and not proof that a study had adequate power.

General: Over-dispersion means that variability is greater than a model expects. When present, a model that understates variability can make uncertainty appear too small and statistical evidence too strong. The registry specifies a continuous endpoint and standard deviation, but does not report a model assessment or indicate that over-dispersion was relevant to the analysis. It should not be assumed to have occurred in this trial.

What analysis details remain unknown?

The record does not report whether the primary comparison used paired observations, a model-based approach, or another method; how covariates were handled; whether any assumptions were checked; or whether any sensitivity analyses were conducted. Those details matter because methods can use incomplete repeated measurements differently and can rely on different assumptions. No particular method should be attributed to the investigators based solely on the endpoint description.

These distinctions separate the endpoint's definition from the analysis used to estimate it. The registry supplies the scoring direction and identifies a mean and standard deviation, but does not provide the primary estimate, uncertainty interval, test result, analysis population, or complete statistical plan. A quantitative conclusion about the endpoint is therefore not available from the posted results.

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("Duloxetine")
n   <- c(30)
m   <- c(8)
sp  <- c(6.1)
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  Duloxetine  30  8.000  6.100  5.722  10.278

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 4 outcome measures and 4 analyses. It identifies HAM-D change from Week 0 to Week 12 as the primary outcome and names Cornell Dysthymia Rating Scale change over the same period as a secondary outcome. The complete descriptions of the remaining listed measures and analyses are not reported in the registry record. A multiplicity strategy—such as a hierarchy of tests or an adjustment for considering several outcomes—is also not reported.

General: When several outcomes or statistical tests are examined, the chance of at least one apparently unusual result can increase if each is interpreted without accounting for the others. A prespecified primary outcome can help focus interpretation, but it does not reveal whether any adjustment was planned or applied. Because outcome results and the analysis plan are unavailable, the impact of multiplicity cannot be assessed in this trial.

The record does not report how missing Week 12 scores were handled, whether participants who discontinued treatment remained in an analysis, or whether a particular analysis population was defined. It also does not give the number with endpoint measurements or the number completing treatment. Enrollment of 30 should therefore not be read as the primary analysis denominator.

General: Missing data can affect an estimate when the likelihood of having an observed measurement is related to outcomes or participant characteristics. Complete-case analyses, imputation, and model-based approaches address missingness differently and depend on different assumptions. No method can be attributed to this trial from the registry record. Without missing-data counts, reasons for missingness, and analysis details, the effect of missing observations on the endpoint is not assessable.

12. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports that 0 participants were affected in the serious adverse event category among 30 at risk in the Duloxetine group. In the other adverse event category, 7 participants were affected among 30 at risk. These are the recorded counts for the listed categories; they are not efficacy results and do not establish a causal relationship between treatment and an event.

The serious adverse event count should be read as a registry-reported count for this group, not as proof that serious harms are impossible or that none occurred outside the way the record defines and captures them. The registry record does not report a follow-up window for these safety counts or further details about event collection and attribution. No serious adverse event terms are listed. The death count is not reported in the registry record.

General: Adverse event counts describe events or affected participants according to the registry's reporting convention. They do not, by themselves, establish comparative risk. A control group and clear information about event definitions, observation periods, and ascertainment would be needed for a comparative safety interpretation. This single-group record does not provide a like-for-like control comparison.

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

The registry record establishes that an open, single-group duloxetine trial was conducted in elderly patients with dysthymic disorder, with HAM-D change from Week 0 to Week 12 designated as the primary outcome. It also gives the endpoint's scale, scoring direction, and intended summary as a mean with standard deviation. Those details describe the question the study set out to address.

The posted results do not include the primary mean change, its standard deviation, a confidence interval, or a test result. It is therefore not possible to say what change was observed or how precisely it was estimated. The trial's completed status does not supply this missing information. The secondary Cornell outcome is named, but its result is not reported in the registry record.

Even if a within-group score change were available, the design would not establish how much of that change was caused by treatment. With no concurrent comparator, the study cannot distinguish a treatment-related change from time effects, expectations, concurrent care, or other influences. This is a limitation of what a single-group design can establish, not a claim that any specific alternative explanation occurred.

The record also does not report the analysis population, endpoint completion, missing-data handling, or full statistical plan. These gaps prevent assessment of whether the analyzed participants represent the enrolled group or whether analytic choices could affect the estimate. The serious adverse event count and other adverse event count are descriptive registry entries; without a control group and fuller ascertainment details, they do not support comparative safety conclusions.

Accordingly, the record supports a description of the study design, endpoint, and selected adverse event counts. It does not support a conclusion about the size or precision of the primary outcome, statistical significance, or a causal treatment effect. It also does not support a comparison with another treatment or with no treatment.

14. Why This Trial Matters Statistically

This trial is statistically notable for the distinction between measuring change and estimating a treatment effect. The registry defines a before-and-after symptom measure over a 12-week period, but its single-group design has no concurrent control condition. A change score can summarize how measurements differ over time; by itself, it cannot identify what would have happened to the same population under an alternative condition.

The endpoint's sign convention also matters. The score at Week 12 is subtracted from the score at Week 0, so positive values point toward a lower score at follow-up. Clear interpretation depends on keeping that direction in mind. The registry gives the scoring framework but does not report the mean change or standard deviation needed to interpret the actual result.

There is a second statistical issue: the gap between a registry's outcome definitions and its outcome results. The record names a primary measure and a secondary measure and lists 4 outcome measures and 4 analyses, yet the numerical primary result and full analysis details are not reported. This makes it impossible to assess magnitude, precision, multiplicity, or the influence of missing observations. It also illustrates why an outcome label or a completed study status is not a substitute for a numerical estimate and its uncertainty.

Finally, planning calculations must not be confused with evidence from participants. The stated planning assumptions describe an illustrative two-arm calculation based on an observed effect; they do not change the registered single-group design or show observed power. The primary effect needed to judge that illustration is not reported. Taken together, these points show why design, endpoint definition, outcome reporting, and analysis assumptions each matter to statistical interpretation.

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 main trial-specific limitation is the absence of a concurrent comparison group. Because the registry records a single-group open treatment design, any within-group change would not isolate the effect of duloxetine from other influences over time. The lack of masking may also matter for a rater-completed symptom measure, although the registry does not report whether assessors were otherwise blinded or how assessments were standardized.

The primary endpoint result is not reported. The registry describes the mean change and standard deviation but gives neither value, and it does not report a confidence interval or test result. This prevents evaluation of the direction actually observed, the magnitude of change, and its statistical precision. The secondary Cornell outcome is identified but its results are not reported either.

Enrollment is listed as 30, but the registry does not state the number with Week 12 data, the number completing treatment, or the analysis population. Missing-data methods, reasons for missing observations, protocol deviations, adherence, and sensitivity analyses are not reported. These gaps prevent assessment of how the observed data, if available, would represent the enrolled group.

The registry lists 4 outcome measures and 4 analyses without full details of all measures, the statistical plan, or a multiplicity approach. The dosing description is also incomplete after the stated schedule through another 4 weeks. These reporting limitations do not establish that the study lacked such procedures; they mean the relevant details are not reported in the registry record. The planning assumptions do not substitute for the trial's own statistical design information or its outcome estimates.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Duloxetine Treatment in Elderly With Dysthymia — NCT01852383.
• Registry record last updated 2014-05-01; status last verified 2014-02.
• Results first posted 2014-05-01.
• 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

This completed Phase 4 study is registered as an open, single-group duloxetine treatment trial in elderly patients with dysthymic disorder. The registry lists 30 enrolled participants and records the study start as 2006-01 and primary completion as 2010-09. The primary endpoint was mean change in 24-item HAM-D score from Week 0 to Week 12, defined as the Week 0 score minus the Week 12 score.

The registry record does not report the primary endpoint estimate, its standard deviation, or a confidence interval. It names Cornell Dysthymia Rating Scale change as a secondary outcome, but does not report its result. Analysis populations, missing-data methods, and a multiplicity strategy are not reported in the registry record.

For safety, the registry lists 0 participants affected by serious adverse events and 7 affected in the other adverse event category, each among 30 at risk. These are descriptive counts without a control comparison. The design and available reporting do not permit a causal estimate of treatment effect or a comparative safety assessment.