Duloxetine Treatment of Major Depression and Chronic Low Back Pain For Older Adults: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Duloxetine Treatment of Major Depression and Chronic Low Back Pain For Older Adults (NCT00696293), a na phase 4 study. The primary outcome is Change in Montgomery Asberg Depression Rating Scale(MADRS) Score From Baseline and 12 Weeks.

2. Clinical Question

The study concerns older adults with major depressive disorder and chronic low back pain. Its registered primary outcome is change in depression severity, measured with the Montgomery Asberg Depression Rating Scale (MADRS), from baseline to 12 weeks. The study also describes a combined response criterion based on depression and back-pain improvement, and a planned investigation of whether changes in the two conditions are correlated.

The registry record describes an initial treatment step with duloxetine plus clinical management. It also discusses a second step in which duloxetine plus problem-solving therapy for depression and pain (PST-DP) would be compared with duloxetine plus clinical management. However, only one arm is listed in the posted record, and the registered design is single-group. No results for the proposed second-step comparison are reported in the registry record.

Thus, the record describes a clinical question about depression severity during treatment, alongside questions about combined depression and pain response. It does not provide a randomized estimate comparing duloxetine with another treatment or with no treatment. In the absence of the primary outcome value and a concurrent comparison group, the posted record cannot establish the size of change attributable to treatment.

3. Trial Design

This is a phase 4 study sponsored by the University of Pittsburgh. The registry lists a start of 2007-05, primary completion at 2010-03, completed status, and enrollment of 30. The listed intervention is duloxetine, and the only arm description is duloxetine plus clinical management. The registry identifies the model as single-group, allocation as NA, and masking as none. In practical terms, the posted design does not describe random assignment, masking, or a concurrent control arm.

The brief summary describes hypotheses involving two treatment steps. In the first step, participants receiving duloxetine plus clinical management were to be evaluated for a response involving both depression and back pain. The proposed response definition was a MADRS score at or below 9 together with at least a 30% improvement in back pain. The summary refers to an initial 60 subjects for that hypothesis, while the registry lists enrollment of 30. The record does not explain this difference.

The brief summary also describes a second-step comparison of duloxetine plus PST-DP with duloxetine plus clinical management, with response assessed during the second 8 weeks. Yet the posted record lists only the duloxetine plus clinical management arm and identifies the trial as single-group. It does not report assignment procedures, results for a second arm, or an analysis comparing the proposed treatments. The record therefore does not support treating the second-step hypothesis as a completed comparative result.

The primary outcome is described as change in MADRS score from baseline to 12 weeks. Although the outcome metadata label the parameter as a mean with standard deviation, the population description refers to a description of median change. This inconsistency is not resolved in the registry record, and no numerical outcome value or dispersion is posted. The study's sample-size rationale and detailed statistical analysis plan are not reported in the registry record. The planned response definition and other hypotheses describe aims, but do not substitute for reported outcome estimates or a full analysis plan.

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

Analysis population for the primary outcome, as defined in the registry: description of median change

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 Montgomery Asberg Depression Rating Scale(MADRS) Score From Baseline and 12 Weeks, measured over baseline and 12 weeks.

The primary endpoint is change in Montgomery Asberg Depression Rating Scale score from baseline to 12 weeks. The MADRS is a measure of depression severity, with a theoretical scale range of 0-60; lower scores represent better outcome. The registry describes a larger reduction between baseline and 12 weeks as representing a better outcome.

Change in a continuous symptom score can retain more information about the degree of improvement or worsening than a response category alone. A response definition classifies participants according to whether they meet a prespecified threshold, whereas a change score can describe the extent of movement across the scale. These are general measurement considerations, not findings from this study.

The record's endpoint metadata are not fully consistent: the parameter is labeled as a mean and the dispersion as standard deviation, but the population description says “description of median change.” The numerical result, including any summary value and variability, is not reported in the registry record. Therefore, the direction and magnitude of the observed MADRS change cannot be assessed from the posted results, and the record does not permit a judgment about the precision of that estimate.

The brief summary also sets out a combined response criterion involving MADRS and back-pain improvement. That criterion is distinct from the registered primary endpoint of MADRS change. In addition, the back-pain scale is described differently across the two treatment-step hypotheses: the first refers to a 20-point numeric rating scale, while the second refers to a 2-point numeric rating scale. The record does not clarify this difference. Neither the combined response result nor the correlation between depression and pain improvement is reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The methods information in the registry record is sparse. The outcome is continuous, the listed endpoint is change in MADRS score, and the outcome metadata name a mean and standard deviation. However, the population description refers to median change. No numerical estimate, dispersion value, confidence interval, test statistic, or p-value is reported in the registry record. The methods below explain how such quantities are generally interpreted; they do not describe analyses confirmed to have been conducted in this study.

What a change score represents

In general, a change score expresses how a measurement at follow-up differs from its baseline measurement for a participant or for a group. For a depression scale on which lower values represent better outcome, a reduction corresponds to improvement on the scale's direction. A group summary can be reported as a mean change or a median change. The mean is sensitive to all observed values, including unusually large or small ones; the median describes the middle of the ordered observations and is less affected by extreme values. These are general distinctions. The registry's mixed labeling of mean and median means the intended summary cannot be identified unambiguously from the record.

A standard deviation describes the spread of individual values around a mean. It is not the same as the uncertainty in the estimated group mean. A confidence interval, when reported, expresses a range of values compatible with the data and the statistical model at a stated confidence level. A narrow interval generally indicates greater precision than a wide interval, but precision alone does not remove bias or limitations in the design. No confidence interval or standard deviation value for the primary outcome is reported in the registry record.

What a one-group design can establish

General statistical interpretation distinguishes a before-and-after description from a controlled comparison. A one-group analysis can estimate change over time among the observed participants, provided the relevant measurements and analysis are reported. It lacks a concurrent control group that would help separate treatment-associated change from other time-related influences. Random assignment is one way to balance known and unknown factors between groups on average, while masking can reduce some expectation-related influences on care or measurement. The registry lists neither a randomized allocation nor masking for this study.

These design features do not make a within-group description meaningless. They determine what the estimate can answer. Without a reported estimate, the registry record does not permit even a descriptive conclusion about the amount of MADRS change. Without a suitable comparator, a reported change would not by itself quantify the causal effect of duloxetine plus clinical management relative to another option.

Combined response and multiple outcomes

The brief summary defines response using both a depression threshold and a minimum improvement in back pain. A combined endpoint can require participants to meet both criteria, making the result dependent on each component's measurement and definition. It answers a different question from mean or median change in MADRS alone. The record describes different numeric rating scale ranges in its two step hypotheses, referring to a 20-point scale for the first and a 2-point scale for the second. Since the discrepancy is not explained, the back-pain component cannot be interpreted consistently from the description alone.

The hypotheses also span depression, pain, their correlation, and exploratory outcomes such as disability, sleep, functioning or quality of life, caregiver burden or depression, and analgesic use. When many outcomes or hypotheses are tested, general statistical practice considers multiplicity: the chance of at least one apparently notable result can increase as the number of tests grows. Procedures for controlling or describing that issue depend on the analysis plan and scientific aim. Multiplicity procedures are not reported in the registry record, and no results for these hypotheses are posted.

Planning illustrations and power

General power calculations assess how likely a planned study is to detect an effect of a specified size under stated assumptions. A planning sample size is not an observed result and does not show that the study had a particular probability of detecting an effect after the data are collected. The record describes planning illustrations based on a fraction of an observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. It also characterizes the figures as participants per arm for non-time-to-event outcomes. These are planning assumptions, not evidence that this single-group study used that design.

The registry record does not state its own sample-size justification. It also does not report a numerical observed effect for the primary endpoint, so the stated basis for the planning illustrations cannot be evaluated using the posted outcome results. Any planning figure should be separated from the actual enrollment of 30 and from any observed treatment effect. It is not observed power.

What cannot be reconstructed

Because numerical outcome results and analysis details are absent, no unadjusted recalculation can be checked against sponsor analyses. The registry lists no analyses. Details such as how missing measurements were handled, which participants entered each analysis, how any departures from the intended protocol were managed, and whether covariates were included are not reported in the registry record. General methods cannot fill those gaps: the actual analysis and its assumptions must be documented before its estimate or uncertainty can be evaluated.

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 Plus Clinical Management")
n   <- c(28)
m   <- c(-11.7)
sp  <- c(6.5)
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 Plus Clinical Management  28  -11.700  6.500  -14.220  -9.180

11. Multiplicity, Missing Data and Analysis Populations

The brief summary sets out several related hypotheses: response during the first treatment step, a proposed treatment comparison during the second step, correlation between depression and pain improvement, and exploratory outcomes covering disability, sleep, functioning or quality of life, caregiver burden or depression, and analgesic use. This range of outcomes raises a general multiplicity question. If multiple hypotheses are tested, the chance of at least one apparently unusual result can increase. A prespecified hierarchy, adjustment procedure, or clear distinction between confirmatory and exploratory analyses can help readers understand how results should be interpreted. Such procedures are not reported in the registry record.

The registry record does not report how missing outcome measurements were handled. It also does not describe an analysis population, such as whether the primary analysis included all enrolled participants or only those with observed follow-up measurements. The listed enrollment is 30, but the record does not give the number with primary endpoint measurements or explain exclusions. These details matter because missingness and exclusion can affect both the estimate and the population to which it applies.

The record contains no numerical primary outcome result, and it lists no analyses. It is therefore not possible to determine whether all planned outcomes were analyzed, whether additional analyses were undertaken, or whether any adjustments were made for repeated measurements or multiple endpoints. The study-specific statistical analysis plan, missing-data methods, and population definitions are not reported in the registry record. The multiple-step hypotheses do not establish that these methods were specified or used.

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.

For the listed duloxetine plus clinical management group, the registry records 30 participants at risk for serious adverse events and 0 affected. For other adverse events, it also records 30 participants at risk and 0 affected. These are the posted counts for the group and should be read as registry entries, not as a comprehensive description of every possible safety outcome.

The deaths field is null, so deaths are not reported in the registry record; a null entry should not be read as a reported count of zero. The list of serious adverse-event terms is empty. The record does not explain event-collection procedures, the observation period for adverse events, or whether events were assessed using additional definitions. No comparison group is listed, so the adverse-event counts cannot serve as a comparative estimate of relative safety.

Zero affected participants in the posted count does not establish that the risk of an event is zero. In general, observed event counts describe what was recorded in the participants under the specified reporting process; they do not guarantee that an uncommon event could not occur or that all events were captured. The record does not provide enough detail to assess completeness or precision beyond the stated counts.

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

The registry record describes a completed, open-label, single-group study with enrollment of 30 and a primary outcome focused on change in MADRS score from baseline to 12 weeks. It specifies that lower MADRS values represent better depression severity scores. These details identify the clinical measurement and the intended time point, but the primary outcome's numerical result is not reported in the registry record.

Without that result, there is no basis in the posted results for describing the size, direction, or precision of the observed change. The record also does not provide the summary statistic's value or an estimate of variability. Its labeling is internally unclear about whether the intended change summary is a mean or a median. Accordingly, the primary endpoint cannot be quantitatively interpreted from the registry record.

Even if a within-group change estimate were available, the single-group design would describe change among participants rather than isolate the effect of treatment against a concurrent control. The registry lists no randomized comparison and no masking. General explanations for change over time can include treatment exposure, natural symptom variation, expectations, and other care; the record does not establish the contribution of any one explanation.

The second-step comparison and other hypotheses in the brief summary are not accompanied by posted results. The record lists one arm, no analyses, and no comparator. It therefore does not support conclusions about whether PST-DP outperformed clinical management, whether depression improvement tracked pain improvement, or whether exploratory outcomes differed.

The safety entries report 0 serious events and 0 other events among 30 participants at risk in the listed group, while deaths are not reported in the registry record. These figures should be read as the posted counts, not as proof of no risk or as evidence from a comparison. Overall, the record documents study intent and selected design and safety fields, but leaves the central efficacy result and many analysis details unreported.

14. Why This Trial Matters Statistically

Statistically, this record illustrates why a clearly specified endpoint and a reported estimate are both needed to interpret a study. The MADRS endpoint is identifiable, its scale direction is described, and its intended assessment spans baseline and 12 weeks. Yet the numerical change and its uncertainty are absent. Without them, readers cannot assess the magnitude or precision of the outcome, regardless of how clearly the clinical question is stated.

The design details are also important. The registry identifies a single-group, unmasked study with no allocation and one listed treatment arm. Those features distinguish a description of outcomes during treatment from a randomized comparative estimate. The brief summary's proposed second-step comparison does not change the posted design or supply comparative results. Recognizing this distinction prevents a hypothesis from being mistaken for an observed treatment contrast.

There are further statistical interpretation issues in the record: the primary outcome metadata refer both to a mean and to a median change; the back-pain rating scale is described with different ranges across treatment steps; and the first-step hypothesis refers to 60 subjects while enrollment is listed as 30. These are reporting ambiguities rather than evidence of any particular analytic choice. Their presence underscores the value of consistency between outcome definitions, design descriptions, and results.

Finally, the proposed hypotheses cover multiple outcomes, while the registry record does not describe multiplicity procedures, missing-data handling, analysis populations, or the study's own sample-size rationale. These omissions limit assessment of the planned inferential framework. A completed status and posted adverse-event counts do not replace outcome estimates and analysis methods. The record is therefore useful for identifying the study's stated aims and design, while its statistical conclusions remain constrained by the lack of posted results.

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 most consequential limitation for interpreting this specific record is that the primary MADRS change result is not reported in the registry record. No numerical estimate, uncertainty measure, or test result is available. The record's outcome description also labels the parameter as a mean and the dispersion as a standard deviation, while its population description refers to median change. This makes the intended summary statistic unclear.

The registered design is single-group, allocation is NA, and masking is none. Only duloxetine plus clinical management is listed as an arm. This design does not provide a concurrent treatment comparison in the posted record, so a change over time, if reported, would not by itself establish a treatment-attributable effect. The second-step comparison described in the brief summary is not accompanied by a listed second arm or reported analysis.

There are inconsistencies or gaps in the study description. The first-step hypothesis refers to an initial 60 subjects, whereas recorded enrollment is 30; no reconciliation is reported. The brief summary describes the back-pain scale as a 20-point numeric rating scale for the first step and a 2-point numeric rating scale for the second. The record does not explain whether this reflects a correction, different measurement, or another reason, so no interpretation should be imposed.

Analysis populations, missing-data handling, multiplicity procedures, and the study's sample-size justification are not reported in the registry record. The registry lists no analyses. The short safety entries do not state event-collection methods, and deaths are not reported. These limitations concern what the registry documents; they do not establish what may have been reported elsewhere, which cannot be assessed from this record alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Duloxetine Treatment of Major Depression and Chronic Low Back Pain For Older Adults — NCT00696293.
• Registry record last updated 2017-02-01; status last verified 2017-01.
• Results first posted 2016-05-26.
• 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 describes a completed phase 4, open-label, single-group study sponsored by the University of Pittsburgh, with enrollment of 30. The listed treatment arm is duloxetine plus clinical management. The study focuses on older adults with major depressive disorder and chronic low back pain, and its primary endpoint is change in MADRS score from baseline to 12 weeks.

The brief summary describes additional response, comparison, correlation, and exploratory hypotheses. However, only one arm is listed, no analyses are posted, and numerical results for the primary outcome and the other hypotheses are not reported in the registry record. The record also leaves unresolved differences in the endpoint summary description, the back-pain scale description, and the stated initial sample size. Safety entries list 0 serious and 0 other affected participants among 30 at risk; deaths are not reported.