Antidepressant Medication Treatment for Depression in Individuals With Chronic Heart Failure: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Antidepressant Medication Treatment for Depression in Individuals With Chronic Heart Failure (NCT00078286), a randomized phase 2, phase 3 study. The primary outcome is Symptoms of Depression in Congestive Heart Failure Patients With Clinical Depression After Treatment With Sertraline or Placebo.

2. Clinical Question

The clinical question concerns a continuous symptom measure rather than a yes-or-no outcome. The registry record describes the study more broadly as examining antidepressant treatment in relation to rates of death and disease in depressed people with chronic heart failure. However, the posted primary outcome is change in depression symptoms at week 12. A secondary outcome concerns cardiac events and morbidity or mortality, including rehospitalization, after treatment with sertraline or placebo.

The posted group means describe improvement in both groups: a negative change score means the week-12 score was lower than the baseline score, and lower scores indicate fewer depressive symptoms. The sertraline group mean change was -7.1, while the placebo group mean change was -6.8. The difference between those reported means is -0.3, with the sertraline mean numerically more negative. The sponsor's own analysis reports a difference in mean change score of -0.3 but gives p = 0.89. Because a recalculation from summary statistics yields a conflicting inferential result, the sponsor's analysis takes precedence for interpretation.

This comparison addresses the recorded outcome and assessment time only. It does not, by itself, answer whether either group experienced differences in longer-term heart-failure outcomes, mortality, or other health outcomes. Those results are not reported in the registry record.

3. Trial Design

The registry record identifies NCT00078286 as a completed, randomized, parallel-group trial in Phase 2 and Phase 3. It lists an enrollment of 469 and identifies Duke University as the sponsor. The study began in 2003-11, and primary completion was recorded in 2008-03. The record describes quadruple masking. It does not specify which parties were masked.

Participants were assigned to sertraline or placebo. Both interventions were described as being taken for 12 weeks. The primary outcome was depression symptom change from baseline to week 12, measured using the Hamilton Depression Rating Scale (HDRS). The analysis population is described as intent to treat, and the primary-outcome description says random coefficient modeling was used. The sponsor's posted analysis is labeled “Mixed Models Analysis.”

The group results list 234 participants for sertraline and 235 for placebo. The mean change scores were -7.1 and -6.8, respectively, and the reported dispersion is identified as standard deviation. The registry record does not describe the randomization sequence, allocation concealment, masking procedures, treatment adherence, or the handling of participants who did not complete the assessment. Those details are not reported in the registry record.

The registry lists a secondary outcome addressing cardiac events and morbidity or mortality, including rehospitalization. It records two outcome measures and two analyses, but only one secondary outcome is described in the posted information. The detailed results available for interpretation focus on the week-12 depression symptom outcome. The registry's original sample-size justification and design assumptions are not reported in the registry record; planning illustrations based on the observed effect are separate from the trial's original design.

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: Analysis was based on intent to treat (ITT), using random coefficient modeling.

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 Symptoms of Depression in Congestive Heart Failure Patients With Clinical Depression After Treatment With Sertraline or Placebo, measured over Measured at Week 12.

The primary endpoint is change in depression symptom score between baseline and week 12, measured with the 17-item Hamilton Depression Rating Scale. The scale is administered by a rater and assesses depression severity. Its total score ranges from 0, described as not at all depressed, to 52, described as most severely depressed. The endpoint is a continuous measure, and the posted results report group mean change and standard deviation.

Change is defined as the week-12 HDRS score minus the baseline HDRS score. Accordingly, a negative change means the score decreased over the 12-week period, which the registry describes as improvement in depressive symptoms. In the posted results, both group means are negative: -7.1 for sertraline and -6.8 for placebo. The sertraline mean is numerically more negative, while the difference in mean change is reported as -0.3.

Using a continuous symptom score retains information about the amount of change, rather than reducing each participant's outcome to a category such as improved or not improved. In general, a mean summarizes the average observed change across a group, while its spread describes variation among participants. The endpoint is relevant to the study's depression-treatment question, but it does not directly measure heart-failure events, rehospitalization, or mortality. The registry record lists a cardiac-event and morbidity/mortality outcome separately.

7. Results Posted in the Registry

• Sertraline: mean -7.10 versus -6.80 on control; difference -0.30 (95% CI -0.39 to -0.21), standardized effect d = -0.60; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant's HDRS change score is numerical, and the posted arm summaries report means and standard deviations. The sponsor describes its comparison as a difference in mean change score, analyzed with mixed models. The primary-outcome description separately says that the intent-to-treat analysis used random coefficient modeling.

What does a mean change represent?

A change score subtracts the baseline score from the follow-up score. Here, follow-up is week 12. The registry defines negative values as improvement because the follow-up HDRS score is lower than baseline. A group mean change summarizes the average change, not the change experienced by every participant. The reported means are -7.1 for sertraline and -6.8 for placebo; the posted difference in mean change is -0.3.

In general, a standard deviation describes how dispersed individual measurements are around a group mean. The arm summaries label the dispersion as standard deviation and report 0.5 in each group. The record does not provide individual-level observations or further detail about the calculation of these summaries. A small reported spread, taken alone, does not explain how the sponsor modeled repeated measurements or any missing observations.

What does a mixed model contribute?

In general, mixed models can account for the structure of repeated measurements and model variation among participants. Random coefficient modeling is one approach that can allow patterns of change to vary across individuals. The registry names these methods but does not report the full model formula, covariance structure, covariates, estimation method, or assumptions about missing data. Those choices matter for reproducing an analysis, so the method label alone is not enough to reconstruct the sponsor's result.

The registry's sponsor analysis is labeled a mixed-model analysis and reports p = 0.89 for a difference in mean change of -0.3. The recalculation from the arm-level summaries is an unadjusted comparison, not a reproduction of the mixed model. It does not use the model's participant-level data or any unreported modeling choices. This distinction is important because the recalculation gives p < 0.001, while the sponsor's result is p = 0.89. The discrepancy is not resolved by the summary statistics, and the sponsor's analysis takes precedence.

How should the p-values be understood?

In general, a p-value describes how compatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The p-value from the unadjusted recalculation and the sponsor's p-value come from analyses that do not agree; they should not be read as interchangeable evidence.

The sponsor analysis is categorized in the registry as non-inferiority or equivalence. In general, such a conclusion depends on a prespecified margin defining what difference would be considered acceptably small, together with an appropriate inferential comparison. The margin and sponsor confidence interval are not reported in the registry record. The p-value alone therefore cannot establish that the groups met a particular equivalence or non-inferiority criterion.

What is the role of the intent-to-treat description?

In general, an intent-to-treat analysis aims to analyze participants according to their randomized groups, preserving the comparison created by random assignment. The primary-outcome description says the analysis was based on intent to treat. It does not state how missing week-12 scores were handled, whether all enrolled participants contributed to the reported model, or whether any additional analysis populations were used. Those details are not reported in the registry record.

The available methods information therefore identifies the broad outcome, population, and model labels, but not enough detail to reproduce the sponsor's analysis or assess every assumption. The conflict with the unadjusted recalculation makes it especially important to distinguish the sponsor's mixed-model result from a simple calculation using group summaries.

10. Confidence Intervals and Precision

A confidence interval gives a range of values compatible with an estimate under the specified statistical method and its assumptions. In general, a narrower interval indicates greater statistical precision, while a wider interval indicates more uncertainty about the estimated effect. The interval does not show the range of individual participant responses and does not establish clinical importance by itself.

The unadjusted recalculation from posted group summaries gives a confidence interval that lies below zero for the difference in mean change. That interval is tied to the recalculation, not the sponsor's mixed-model analysis. Because the registry flags a discrepancy, the sponsor's own analysis takes precedence; the recalculated interval should not be substituted for an interval from that analysis.

The sponsor reports a mean difference of -0.3 and p = 0.89, but its confidence-interval bounds are not reported in the registry record. The record also does not report the non-inferiority or equivalence margin. Consequently, the precision of the sponsor's estimate cannot be assessed from a sponsor confidence interval in the posted record, and the p-value does not provide the missing interval bounds.

Precision also needs to be interpreted in relation to the endpoint. The observed group means summarize average HDRS change, while the reported standard deviations describe spread in the arm summaries. Neither feature supplies a threshold for clinically meaningful change. No such threshold is reported in the registry record. The data support describing the reported estimates and the disagreement between analyses, but they do not support a claim about clinical importance based on a margin or precision criterion that is absent from the record.

11. Planning a Similar Trial: Sample Size

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

The planning figures are illustrative calculations based on the observed effect, not the trial's original design assumptions. The listed assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is set as a stated fraction of the observed effect. The figures for this continuous-outcome comparison are participants per arm, not total events.

The 80% power column gives the planned number of participants per arm for a design that would have an 80% chance of detecting the specified effect under the listed assumptions, if that effect were the true one. The 90% power column gives the corresponding number for a 90% chance. These are design probabilities conditional on the assumed effect and statistical model; they are not probabilities that the trial result is true.

For planning at 100% of the observed effect, the figures are 44 participants per arm for 80% power and 59 for 90% power. At 75% of the observed effect, they are 78 and 104 participants per arm. At 50% of the observed effect, they are 175 and 234 participants per arm. The pattern illustrates the general principle that smaller effects require more participants to detect, holding the other planning assumptions fixed.

This is not observed power. The figures do not describe the trial's actual enrollment plan, and the registry record does not report the original sample-size justification or design assumptions. They are planning illustrations calculated from the observed effect, so they should not be used to infer what power the completed trial had for its observed result.

12. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Observed difference Δ = -0.300; pooled standard deviation σ = 0.500.
2. n = 2 × (1.96 + 0.842)² × 0.500² / -0.300² = 43.6, rounded up to 44 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.3                      # observed difference in means
sigma <- 0.5                     # 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(78286)
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 = 44 (80% power), 59 (90% power)
 75% of observed effect: n per arm = 78 (80% power), 104 (90% power)
 50% of observed effect: n per arm = 175 (80% power), 234 (90% power)
power.t.test: 44.6 
Simulated power at n = 44 per arm: 0.797 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The sponsor's own result is a mixed-model analysis of the difference in mean change score. It reports -0.3 and p = 0.89 and is categorized in the registry as non-inferiority or equivalence. The sponsor's confidence interval and the margin used for that classification are not reported in the registry record.

A separate unadjusted recalculation from the posted group summaries produces the same rounded difference, -0.3, but a different p-value, p < 0.001, and an interval that lies below zero. This is a discrepancy between the reported sponsor analysis and a summary-statistic recalculation. The recalculation does not reproduce the sponsor's mixed model, which may use information or modeling choices not contained in the arm summaries; the specific reason for the difference is not reported in the registry record.

Because the discrepancy is flagged, the sponsor's analysis takes precedence. The recalculated significance result is useful as a check on the posted summary values, but it should not replace the sponsor's result or be presented as the trial's primary finding. The available record does not contain enough model detail to resolve the disagreement independently.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures and two analyses, while the posted information describes one secondary outcome. The primary outcome is week-12 change in HDRS score. The secondary outcome described in the record concerns cardiac events and morbidity or mortality, including rehospitalization. Results for that secondary outcome are not reported in the registry record.

Multiplicity arises when several outcomes, time points, or statistical comparisons are tested. In general, conducting more tests can increase the chance of at least one apparently unusual result under a set of null hypotheses. Adjustment methods or a prespecified hierarchy can address this issue, but the registry record does not report whether multiplicity was controlled, how outcomes were prioritized, or whether all planned analyses are posted. No adjustment should be assumed.

The primary outcome description states that the analysis was based on intent to treat, using random coefficient modeling. The sponsor's analysis is labeled mixed models. These are the available descriptions; the registry record does not report participant-level analysis details, the model specification, covariates, or the handling of missing week-12 measurements. Attrition, reasons for missing data, and the number contributing to the modeled outcome are not reported in the registry record.

In general, missing outcome data can affect an estimate when the chance of missingness is related to participant characteristics or outcome values. Model-based analyses rely on assumptions about the observed data and missingness. Without information about completeness and the sponsor's specific missing-data approach, those assumptions cannot be assessed from the posted record. The record also does not describe additional analysis populations or sensitivity analyses.

These omissions limit independent evaluation of the analysis and of how robust the sponsor's result is to alternative assumptions. They do not establish that missingness or multiplicity caused the discrepancy with the unadjusted recalculation; the reason for that discrepancy is not reported in the registry record.

15. 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 0.80% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cardiovascular (18.7%), noncardiovascular and nonpsychiatric (7.3%), psychiatric (0.9%).

The safety summaries list 234 participants at risk in the sertraline group and 235 in the placebo group. Serious adverse events are recorded for 57 sertraline participants and 63 placebo participants. The “other” adverse-event counts are 22 and 23, respectively. These are counts, not percentages calculated from the groups, and no additional calculation is needed to read them.

The registry also lists leading serious-event categories with percentages: cardiovascular, 18.7%; noncardiovascular and nonpsychiatric, 7.3%; and psychiatric, 0.9%. The record does not specify in this summary whether these category percentages use participants, events, or another denominator, so they should be read as posted rather than converted into counts or compared as rates.

These descriptive figures do not establish whether an event was caused by treatment, nor do they account for differences in exposure time or clinical context. Death counts are not reported in the registry record. The safety information should be interpreted separately from the primary depression symptom analysis and does not provide results for the listed cardiac-event or morbidity/mortality outcome.

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

The posted primary outcome describes mean depression-symptom change from baseline to week 12. Both groups had negative mean changes: -7.1 for sertraline and -6.8 for placebo. By the registry's definition, negative change represents improvement in HDRS symptoms over the period. The reported mean difference is -0.3, with the sertraline group numerically more negative.

The inferential evidence is not internally consistent across the sponsor analysis and the recalculation from group summaries. The recalculation gives p < 0.001, while the sponsor's mixed-model result is p = 0.89. The sponsor's own analysis takes precedence because the record flags a discrepancy. Its classification as non-inferiority or equivalence cannot be fully evaluated without the margin and confidence interval, neither of which is reported in the registry record.

The mean change is a group summary, not a description of every participant's experience. The registry record does not provide a threshold for a clinically meaningful difference, individual response distributions, or enough model detail to independently reproduce the sponsor's analysis. It is therefore not possible to infer clinical importance solely from the mean difference or the recalculated interval.

The trial's broader description refers to death and disease rates, and the secondary outcome concerns cardiac events and morbidity or mortality, including rehospitalization. Results for that secondary outcome and death counts are not reported in the registry record. The posted depression endpoint does not answer whether the groups differed in those outcomes, or in outcomes beyond the recorded week-12 assessment. The findings describe the reported statistical results, not medical advice or a treatment recommendation.

17. Why This Trial Matters Statistically

This trial illustrates why the analysis method matters when interpreting a posted treatment comparison. The arm summaries give mean changes and standard deviations, allowing an unadjusted recalculation. The sponsor's reported result comes from a mixed-model analysis. Although both report a mean difference of -0.3, their p-values differ sharply: the recalculation gives p < 0.001 and the sponsor reports p = 0.89. A single effect estimate does not guarantee that different analytical approaches yield the same uncertainty or inferential result.

The record also illustrates the distinction between statistical and clinical interpretation. A p-value addresses compatibility with a statistical model, while clinical importance depends on the meaning of the outcome and an appropriate threshold. No clinically meaningful HDRS difference is reported in the registry record. The sponsor's non-inferiority or equivalence label also cannot be fully appraised without its margin and confidence interval.

Finally, a trial can have a clear primary symptom endpoint while leaving important questions unresolved in the posted record. The registry identifies a secondary cardiac-event and morbidity/mortality outcome, but its results are not reported. Details of missing-data handling, multiplicity, and the original design assumptions are also absent. Together, these points show why statistical interpretation depends not only on an estimate, but also on the analysis specification, uncertainty, outcome context, and completeness of reporting.

18. 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 posted information does not report the sponsor's confidence interval or the margin underlying the non-inferiority or equivalence classification. This limits evaluation of that classification. The disagreement between the sponsor's mixed-model result and the unadjusted recalculation cannot be resolved because the registry does not provide enough detail to reproduce the sponsor's model.

Missing-data handling, the number of participants contributing to the week-12 analysis, and reasons for any missing measurements are not reported in the registry record. The record identifies an intent-to-treat analysis and random coefficient modeling, but does not give the model specification or sensitivity analyses. Multiplicity procedures are also not reported.

The record describes a cardiac-event and morbidity/mortality secondary outcome, but its results are not reported in the registry record. Death counts are not reported. The posted safety summary does not identify the denominator for its serious-event category percentages or establish whether events were attributed to treatment.

The registry lists an enrollment of 469, while the group summaries give arm sizes of 234 and 235; the record does not explain whether these counts represent all enrolled participants or the analysis set. Details of eligibility beyond the stated population, adherence, and follow-up beyond the 12-week treatment period are not reported in the registry record. These limitations constrain interpretation of the posted summaries and do not, by themselves, show that any particular bias occurred.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Antidepressant Medication Treatment for Depression in Individuals With Chronic Heart Failure — NCT00078286.
• Registry record last updated 2013-02-27; status last verified 2013-01.
• Results first posted 2013-02-27.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

NCT00078286 is a completed randomized, parallel-group Phase 2 and Phase 3 trial sponsored by Duke University. It compared sertraline with placebo, with both interventions taken for 12 weeks, in people with depression and chronic heart failure. The record lists enrollment of 469 and describes quadruple masking.

The primary outcome was change in HDRS score from baseline to week 12. Mean changes were -7.1 for sertraline and -6.8 for placebo, with a reported difference of -0.3. The sponsor's mixed-model analysis reports p = 0.89; an unadjusted recalculation gives p < 0.001. Because the registry flags a discrepancy, the sponsor's own analysis takes precedence.

The record does not report the sponsor confidence interval or equivalence margin, the original design assumptions, detailed missing-data methods, or results for the listed secondary cardiac-event and morbidity/mortality outcome.