Medication Treatment for Depression in Nursing Home Residents: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Medication Treatment for Depression in Nursing Home Residents (NCT00076622), a registered phase not stated study. The primary outcome is Geriatric Depression Scale (GDS) Score.

2. Clinical Question

The study question concerns people whose depressive symptoms had already resolved, rather than treatment of current depressive symptoms. It asks whether outcomes differ in the context of continuing or stopping medication, while also attending to adverse effects. The record includes randomized groups and groups based on participant preference. Those categories should not be treated as interchangeable: preference-based groups are not a randomized comparison, and the registry record cautions that the arms are not a like-for-like control comparison. The primary outcome is named and its measurement time is stated, but the registry record does not report an estimated group difference, uncertainty interval, or test of that difference. Thus, the stated question can be described, but the posted information does not establish a comparative result for it.

3. Trial Design

The study is titled “Medication Treatment for Depression in Nursing Home Residents,” with the official title “Drug Treatment of Depression in the Nursing Home Aged.” The sponsor is the University of Pennsylvania. The registry lists depression as the condition, a start date of 2003-11, primary completion in 2009-07, a completed status, and enrollment of 94.

The study description says it examines therapeutic and adverse effects of continuing versus discontinuing antidepressant medication in nursing home residents who had no more than a single depression episode and no longer had depressive symptoms. The intervention descriptions distinguish antidepressant medication from no antidepressant medication. Arm descriptions further separate participants randomized to continue or discontinue from participants whose participation was based on a preference to continue or discontinue.

This distinction matters for interpretation. A randomized comparison is intended to support comparison according to assigned groups, subject to the actual design and analysis. Preference-based groups are formed differently and may differ for reasons related to participants' choices. The registry record explicitly warns that allocation is not stated and that the arms are not a like-for-like control comparison. It therefore does not support treating all four groups as one randomized contrast. Further details of sequence generation, concealment, masking, treatment duration, adherence, and eligibility criteria beyond the brief description are not reported in the registry record. The registry identifies a primary outcome and time point but does not report outcome analyses.

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 Geriatric Depression Scale (GDS) Score, measured over Measured at Month 12.

The primary endpoint is the Geriatric Depression Scale (GDS) score, measured at Month 12. The record describes the GDS as a measure of the presence and severity of depressive symptoms in older adults. Its scale ranges from zero, meaning no depression symptoms, to thirty, meaning severe depression symptoms. The reported endpoint type is a mean, with standard deviation as the dispersion measure, and the unit is units on a scale.

A continuous score retains gradations across the scale rather than reducing each participant to a simple present-or-absent category. In general, a group mean describes the average score, while a standard deviation describes how dispersed individual scores are around that average. These summaries can help characterize both the central tendency and variability of scores, but they do not by themselves describe the full distribution or the uncertainty in a comparison between groups.

The registry identifies the endpoint and measurement time but does not report Month 12 means, standard deviations, a treatment contrast, or a confidence interval. It also does not report the analysis population or handling of missing GDS measurements. Consequently, the endpoint's intended meaning is clear, but its comparative result cannot be interpreted from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies the GDS outcome as a continuous measure summarized by a mean and standard deviation. It does not report the statistical methods used for a primary-outcome comparison. The concepts below are general explanations of how such methods are commonly interpreted, not claims about methods applied in this study.

What does a mean summarize?

For a continuous measure, the mean is the arithmetic average of the observed scores in a group. It gives a compact description of the group's central level, but can be affected by unusually high or low observations. The standard deviation describes how much individual scores vary around the mean. A standard deviation is not the uncertainty around the mean and is not a measure of the treatment contrast. To interpret group summaries comparatively, readers need the estimated difference and a measure of its uncertainty, along with the analysis population and the number of observations contributing to each summary.

The GDS is described as having a bounded range from zero to thirty. The registry does not report the score distributions, so it is not possible to assess whether scores were concentrated near either end of the scale, whether there were notable outliers, or whether a mean adequately represents the observed pattern. Such considerations can affect the suitability and interpretation of statistical models for continuous outcomes.

What would a confidence interval add?

In general, a confidence interval expresses the precision of an estimate under a specified statistical procedure and its assumptions. A relatively narrow interval indicates greater precision than a wide interval, but neither width alone nor whether the interval includes a particular value answers every question about importance. The interval should be interpreted with the estimated effect, the measurement scale, the study design, and the clinical context. The registry record does not report a confidence interval or an estimated primary-outcome contrast.

A p-value, when reported, summarizes how incompatible the observed data are with a specified null model, subject to assumptions. It is not the probability that the null hypothesis is true and does not measure the size or practical importance of an effect. The registry record does not report a p-value for the primary endpoint. Its absence does not establish that the groups were the same or that a comparison was not made outside the registry; it means the registry record does not present that result.

How should assignment groups be distinguished?

General statistical interpretation depends on how groups are formed. Random assignment can, in expectation, balance measured and unmeasured characteristics across assigned groups, although the design and implementation details still matter. An analysis according to assigned group is often used to preserve the benefits of randomization. The registry describes randomized continuation and discontinuation groups, but also describes preference-based continuation and discontinuation groups. It states that allocation is not stated and that the arms are not a like-for-like control comparison. The registry does not describe the analysis population or how those categories were used in any outcome analysis.

Preference-based group comparisons can be confounded: characteristics associated with a participant's choice may also be related to subsequent outcomes. In general, statistical adjustment can address measured differences under appropriate assumptions, but cannot guarantee removal of unmeasured confounding. No adjustment method is reported in the registry record. Accordingly, the preference groups should be described separately and not assumed to provide the same evidential basis as randomized groups.

What do the planning assumptions mean?

The record describes a planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It says the effect to detect is a stated fraction of the observed effect. For time-to-event calculations, the illustration concerns total events; for other calculations, it concerns participants per arm. These are planning assumptions, not a record of the trial's original design justification. The registry does not report the trial's own sample-size justification.

In general, sample-size planning asks how much information a future study would need to detect a specified effect under chosen assumptions. An illustration based on an observed effect is sensitive to the estimate used: an uncertain or unusually large estimate can make a planning result unstable. Planning values do not indicate the probability that the completed study found a result, and they are not observed power. The primary-outcome estimate needed to evaluate such an illustration is not reported in the registry record. The planning assumptions therefore cannot replace an outcome analysis or be treated as evidence about the study's findings.

Other design and analysis details—including interim analyses, model specification, covariate adjustment, and assumptions about repeated measurements—are not reported in the registry record. Their absence limits assessment of how the endpoint was analyzed. The general explanations above clarify statistical ideas; they do not identify which procedures were actually used.

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("Randomized Arm-continue", "Randomized Arm-discontinue", "Non Randomized/Preference Arm-continue", "Non-Randomized Preference-discontinue")
n   <- c(23, 13, 40, 18)
m   <- c(5.11, 5.03, 4.59, 4.64)
sp  <- c(0.86, 1.15, 0.49, 0.75)
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  Randomized Arm-continue  23  5.110  0.860  4.738  5.482
CARM  Randomized Arm-discontinue  13  5.030  1.150  4.335  5.725
CARM  Non Randomized/Preference Arm-continue  40  4.590  0.490  4.433  4.747
CARM  Non-Randomized Preference-discontinue  18  4.640  0.750  4.267  5.013

11. Multiplicity, Missing Data and Analysis Populations

The registry lists three outcome measures and identifies GDS score at Month 12 as the primary outcome. Cognitive Function (MMSE) is listed as a secondary outcome. The remaining outcome measure is not identified in the available record, and no outcome analyses are listed. The registry record does not report a multiplicity plan describing how inference across outcomes or multiple comparisons was handled.

In general, examining several outcomes or contrasts can increase the chance of obtaining at least one apparently notable result by chance. Statistical approaches to multiplicity may adjust inference or designate a primary comparison in advance. Whether an adjustment is needed and how it should be implemented depend on the study's analysis plan and the set of tests. Neither the testing strategy nor any adjustment is reported, so no claim about control of multiplicity is possible.

The record also does not report missing-outcome counts, reasons for missingness, imputation, or sensitivity analyses. In general, missing data can affect an estimate when the observed participants differ in relevant ways from those without measurements. The direction and magnitude of any such effect cannot be inferred without information about the missingness pattern and analysis assumptions.

Analysis populations are also not reported. The registry does not say whether the primary outcome was analyzed by randomized assignment, among participants with observed scores, or using another population definition. This is particularly relevant because randomized and preference-based groups are both described, while the record warns that the arms are not a like-for-like comparison. Without the analysis population and missing-data approach, the precision and potential biases of any unreported outcome estimate cannot be assessed.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: syncope (23.1%), unstable angina with aspiration pneumonia (11.1%), breast cancer (7.7%), fall (7.7%), bronchitis (7.7%).

The registry reports serious and other adverse-event counts separately for four groups: randomized continuation, randomized discontinuation, preference-based continuation, and preference-based discontinuation. For serious adverse events, the recorded affected and at-risk counts are 8 and 23 for randomized continuation, 4 and 13 for randomized discontinuation, 14 and 40 for preference-based continuation, and 5 and 18 for preference-based discontinuation. The record also lists one death in each group. Other adverse-event counts are recorded as 21 of 23, 12 of 13, 33 of 40, and 17 of 18, respectively.

These are registry counts, not a complete comparative safety analysis. The record does not report event timing, follow-up duration by group, definitions, severity details beyond the serious-event category, or whether events were considered related to medication. It lists serious-event terms and percentages including syncope at 23.1%, unstable angina with aspiration pneumonia at 11.1%, breast cancer at 7.7%, fall at 7.7%, bronchitis at 7.7%, and congestive heart failure at 5.6%. The denominator for each listed percentage is not stated in the registry record.

Because the groups differ in how they were formed, these figures should not be read as a like-for-like estimate of comparative risk. They describe recorded counts; they do not by themselves establish that medication caused or prevented an event.

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

The study record establishes that investigators set out to examine continuation versus discontinuation of antidepressant medication among nursing home residents described as having no more than a single depression episode and no longer having depressive symptoms. It identifies the GDS score at Month 12 as the primary endpoint and includes safety-event reporting. It also records randomized and preference-based groups, making the distinction between those group types central to interpretation.

What the registry does not establish is the result of the primary comparison. No Month 12 GDS group summaries, effect estimate, confidence interval, or p-value are reported, and the registry lists zero analyses. Accordingly, the posted registry information cannot show the direction, size, or precision of any difference in GDS scores. It also cannot support a conclusion that scores were equivalent: absence of a reported result is not evidence that a comparison found no difference.

Likewise, the safety counts are not enough to establish a comparative treatment effect. The registry provides counts across groups but does not report the details needed to judge comparability, event timing, or attribution. In particular, preference-based groups should not be interpreted as equivalent to randomized groups, and the record itself cautions that the arms are not a like-for-like control comparison.

In general, a clinical trial result is interpreted by considering the estimated effect, its precision, the outcome's meaning, the analysis population, and the design. Here, the outcome definition and some group descriptions are available, but the primary results and key analysis details are not reported in the registry record. The defensible conclusion is therefore about reporting: the registry documents the study question and selected safety counts, but does not provide the primary-outcome analysis needed to determine what the study found on its principal measure.

14. Why This Trial Matters Statistically

This record illustrates an important distinction in clinical-trial reporting: a named primary endpoint is not itself a reported result. The GDS score and Month 12 measurement time define what the study intended to assess, while the absence of means, a treatment contrast, and uncertainty estimates prevents statistical interpretation of the endpoint. Clear outcome definitions are useful, but they do not answer the comparative question without analysis results.

The record also shows why group formation matters. Randomized groups and preference-based groups may appear within the same study record, but they do not automatically support the same inference. Randomization and participant choice create different bases for comparison. The registry's warning that allocation is not stated and the arms are not a like-for-like control comparison is therefore important context rather than a minor labeling issue.

Finally, safety counts and planning assumptions serve different purposes from an efficacy estimate. Counts can describe what was recorded, while an analysis is needed to support a comparative interpretation. Planning illustrations concern hypothetical future information requirements under stated assumptions; they are not trial findings or observed power. Keeping these distinctions separate helps prevent descriptive data, design labels, or planning calculations from being mistaken for evidence of an outcome effect.

The statistical lesson is broader than this individual record: readers need estimates, uncertainty, analysis populations, and transparent methods to judge what a study can support. Where those elements are absent, careful interpretation means stating the limits rather than filling gaps with assumptions.

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.

Several limitations specific to this record constrain interpretation. The most consequential is that the primary GDS outcome is identified at Month 12, but no outcome summaries or analyses are reported. Enrollment is recorded as 94, yet the number with Month 12 measurements and the number included in any analysis are not reported. The registry also does not describe missing-data handling or analysis populations.

The group structure includes both randomized and preference-based participation. The record states that allocation is not stated and that the arms are not a like-for-like control comparison. It does not report how the groups were combined or contrasted in an analysis, or whether adjustment was made for differences between them. This prevents a clear account of the intended comparative estimand and limits any inference from the descriptive group counts.

Other methodological details are also not reported in the registry record: the randomization procedure, masking, treatment adherence, follow-up completeness, statistical model, covariate adjustment, interim analyses, and multiplicity strategy. The registry lists three outcome measures and a secondary cognitive-function measure, but does not provide analyses for them. The trial's own sample-size justification is not reported. The planning assumptions described in the record should not be mistaken for that justification or for observed power.

For safety, the record lists event counts and selected serious-event terms, but does not state event definitions, timing, exposure duration, attribution, or the denominator behind each listed percentage. The counts therefore provide limited descriptive information and do not resolve comparative safety questions. These reporting gaps do not show that methods or analyses were absent from the study; they mean that such details are not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Medication Treatment for Depression in Nursing Home Residents — NCT00076622.
• Registry record last updated 2018-01-31; status last verified 2018-01.
• Results first posted 2018-01-31.
• 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 record describes a completed study sponsored by the University of Pennsylvania, with a start date of 2003-11, primary completion in 2009-07, and enrollment of 94. It concerns continuation versus discontinuation of antidepressant medication among nursing home residents whose depressive symptoms had resolved and who had no more than a single depression episode.

The primary endpoint is GDS score at Month 12, reported as a continuous measure intended to be summarized by a mean and standard deviation. Cognitive Function (MMSE) is listed as a secondary outcome. The registry lists zero analyses and does not report primary-outcome results. It includes serious and other adverse-event counts for randomized and preference-based groups, while cautioning that the arms are not a like-for-like control comparison.

Thus, the registry record documents the study question, endpoint, group descriptions, and selected safety information, but does not report the analysis needed to interpret the primary comparison.