Focal Electrically-Administered Seizure Therapy (FEAST) Studies at Two Enrolling Sites to Further Test and Refine the Treatment: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Focal Electrically-Administered Seizure Therapy (FEAST) Studies at Two Enrolling Sites to Further Test and Refine the Treatment (NCT02535572), a non-randomized phase 2, phase 3 study. The primary outcome is Percentage Change in Depressive Symptoms as Assessed by Hamilton Rating Scale for Depression.

2. Clinical Question

3. Trial Design

This completed study is registered as a Phase 2 and Phase 3 investigation with a parallel-group model. It enrolled 48 participants and compared FEAST with right unilateral ultrabrief ECT, abbreviated RUL UB in the registry. The sponsor was the Medical University of South Carolina. The study began in 2015-08, and primary completion is recorded as 2019-06-10.

The design was nonrandomized and unmasked. These features matter when interpreting any between-arm contrast: participants were not assigned by random allocation, and patients or study personnel were not masked to treatment. The registry does not report the assignment procedure, whether the arms were recruited at the same time, or whether particular participant characteristics influenced treatment selection. Those details should not be inferred from the arm names or descriptions.

The two treatment groups were FEAST and standard-of-care RUL UB. The record describes a study intended to further evaluate safety, efficacy, and potential mechanisms of action, while refining the FEAST treatment. It also describes treatment modifications, including selective changes to electrode geometry to decrease interelectrode resistance and a revised titration schedule using a standard 800 ma ultrabrief pulse. The registry does not give enough detail to reconstruct how those refinements were applied to each participant or whether all participants received an identical treatment version.

The study record lists four outcome measures and identifies the primary measure as percentage change in depressive symptoms on the HRSD from baseline to the end of the acute course, typically after 4 weeks. Number of adverse events is listed as a secondary outcome. Final arm-level results for the primary endpoint are not reported in the registry record. The adverse-event counts are available, but they do not substitute for results on depressive symptoms.

The registry summary also discusses preliminary open-label experience preceding the completed trial, including results for the first 17 patients and further participants studied afterward. That earlier account is useful context about development of the treatment, but it should not be treated as the final, arm-by-arm result of this 48-participant study.

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.

Population: The participants in baseline analysis were those that maintained inclusion criteria throughout the duration of the study. 2 participants were not counted in the baseline analysis population due to exclusionary circumstances.

6. Primary Endpoint

The registry lists the primary outcome as Percentage Change in Depressive Symptoms as Assessed by Hamilton Rating Scale for Depression, measured over From Baseline to end of acute course (typically after 4 weeks).

The primary endpoint is percentage change in depressive symptoms, measured with the 24-item Hamilton Rating Scale for Depression (HRSD24), from baseline to the end of the acute course. The stated time frame is typically after 4 weeks. The outcome is a continuous measure, and the registry identifies the reported summary as a mean with standard deviation.

The registry describes the HRSD as having a total range of 0 to 72 and notes that lower scores represent less depression, from mild symptoms to no depression. It gives categories of 0-7 for no depression, 7-17 for mild depression, 17-24 for moderate depression, and 25 and higher for severe depression. It also states that a score of 21 or higher was the clinical-depression inclusion criterion. These descriptions provide context for the scale, but do not indicate the actual scores or percentage changes in the completed trial.

Change from baseline is used to describe how an individual’s measured symptom score shifts over the treatment course. Reporting a group mean can describe the average change, while the standard deviation describes variability among observed values. Neither statistic is reported for the completed trial’s primary outcome in the registry record. The formula and sign convention used to calculate percentage change are also not reported in the registry record. Accordingly, the endpoint definition alone does not establish the size or direction of the observed arm-specific changes.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary endpoint summarize?

The endpoint is percentage change in depressive symptoms measured by the HRSD24 from baseline to the end of the acute course, typically after 4 weeks. The registry identifies the measure as continuous and specifies a mean and standard deviation. In general, a mean describes the center of a set of measurements, while a standard deviation describes how dispersed the measurements are around that mean. A group mean alone does not show whether individual participants had similar changes or widely varying changes.

The HRSD is a symptom-rating scale, and the registry states that lower scores represent milder symptoms or no depression. The record does not report the exact formula used to calculate percentage change or how the sign of change was coded. It also does not report the final mean and standard deviation for either arm. Without those values, the endpoint definition cannot be converted into a numerical result.

What would a confidence interval add?

In general, a confidence interval expresses uncertainty around an estimated quantity, such as a mean change or a difference between groups. Its width depends on the variability of the data, the amount of information available, and the analysis method. A narrower interval indicates greater precision under the model; a wider interval indicates less precision. The interval is not a guarantee that the true value lies within its endpoints, nor does it describe the range of individual participant outcomes.

No confidence interval for the primary endpoint is reported in the registry record. As a result, precision cannot be judged from the primary outcome information. The absence of a reported interval is not evidence that the estimate would be either precise or imprecise.

How does the design affect a group comparison?

Random allocation, in general, helps balance both measured and unmeasured baseline characteristics between treatment groups, on average. This study is explicitly identified as nonrandomized. Thus, any difference between its arms could reflect treatment, differences in who entered each arm, or other factors connected with treatment assignment. The registry does not report the assignment procedure or an adjustment strategy, so the extent to which measured differences were addressed is unknown.

Masking is recorded as absent. In general, knowledge of treatment assignment can affect participants’ reports, clinicians’ assessments, treatment decisions, or the collection of outcomes. The registry does not state whether the HRSD assessor was independent or masked in practice. It would be inappropriate to assume an assessment procedure that is not described.

The parallel-group designation means the record describes separate treatment groups rather than a crossover design. The precise timing and sequence of enrollment, the number of treatment sessions received by participants in each arm, and the handling of treatment changes are not reported in the registry record. The brief summary gives some context about FEAST refinements, but does not fully specify their implementation for the completed study.

What do the sample-size illustrations mean?

The record describes planning illustrations based on an observed effect, a specified fraction of that effect, equal allocation, a two-sided significance level of 0.05, and a normal approximation. In general, such illustrations estimate how many participants per arm might be needed to detect a chosen effect under stated assumptions. They are planning scenarios, not evidence about how many participants actually showed a response, and not observed power.

The numerical planning outputs and the specific effect fractions are not reported in the registry record. The stated assumptions should not be mistaken for the trial’s own sample-size justification: that justification is not reported in the registry record. Nor do these illustrations establish that the completed study had adequate precision or was designed around the final observed treatment contrast.

In general, a planning calculation is sensitive to the assumed effect size and variability. If the target effect is smaller, more information is generally needed to distinguish it from random variation under the same design assumptions. If variability is greater than assumed, the planned sample may yield less precision than expected. These are general statistical relationships, not claims about what occurred in this trial.

Why does over-dispersion matter for event counts?

In general, a simple count model may assume that the mean and variance of event counts are linked in a particular way. Over-dispersion means observed variability is greater than that model expects. If it is ignored, standard errors may be too small and intervals or tests may appear more certain than warranted. The registry provides adverse-event counts, but does not report a count-model analysis, an over-dispersion assessment, or inferential comparisons of event rates. No such analysis should be inferred from the descriptive listings.

Likewise, a count of participants with at least one event is not the same as a count of all events. The registry’s safety entries give numbers at risk and numbers affected, but do not describe recurrent-event totals, exposure time, or an event-rate model. The serious-event counts should therefore be read as listed participant counts, without deriving rates or comparative estimates.

What can be said about the statistical analysis?

The registry record does not report the final statistical analysis plan, the model used for the primary endpoint, covariate adjustments, a prespecified contrast, or a primary endpoint p-value. The number of analyses listed is 0. This absence limits the ability to evaluate the analytic choices or to reproduce an inferential result from the registry record. It does not establish that no analysis was performed outside the information recorded there.

General statistical interpretation distinguishes estimation from testing. Estimation focuses on the size and uncertainty of an effect; hypothesis testing evaluates data under a specified null model. Neither is a substitute for design quality. In a nonrandomized, unmasked study, an estimated difference can remain vulnerable to confounding and measurement influences, even if a model-based test were statistically significant. Since the primary contrast and its uncertainty are not reported, those features cannot be assessed for this study.

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("FEAST", "RUL UB")
n   <- c(20, 19)
m   <- c(60.83, 61.81)
sp  <- c(18.36, 20.8)
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  FEAST  20  60.830  18.360  52.237  69.423
CARM  RUL UB  19  61.810  20.800  51.785  71.835

11. Multiplicity, Missing Data and Analysis Populations

The record lists four outcome measures, including the primary measure of percentage change in depressive symptoms and the secondary outcome of number of adverse events. It does not report a multiplicity plan describing how the chance of false-positive findings across multiple outcomes or analyses was handled. The number of analyses listed is 0, and no final primary-endpoint comparison is reported in the registry record.

In general, multiplicity arises when several outcomes, time points, subgroups, or statistical tests are examined. The more opportunities there are to identify a nominally notable result, the more important it is to distinguish prespecified primary questions from exploratory analyses. Methods for addressing multiplicity can include allocating the overall error rate across tests or using a hierarchy of outcomes. The registry does not state that any particular approach was used.

Missing-data methods and the amount or pattern of missing primary outcome data are not reported in the registry record. The analysis population is also not specified. It is therefore unknown whether the primary analysis included all enrolled participants, only participants with outcome measurements, or another defined group. The safety denominators do not establish the primary endpoint analysis population.

In general, missingness can affect interpretation when the probability of having an observed outcome is related to symptoms, treatment, or other participant characteristics. Analyses may rely on assumptions about those relationships, but the assumptions and any sensitivity analyses for this trial are not reported in the registry record. This limits assessment of how robust a primary result would be to missing observations, even apart from the fact that the primary result itself is not posted.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: suicide attempt (5.3%), spontaneous abortion (5.3%).

The registry lists serious adverse events for 0 of 20 participants at risk in the FEAST group and 2 of 19 in the RUL UB group. It lists other adverse events for 17 of 20 and 17 of 19, respectively. Deaths are listed as 0 in both groups. These are the counts recorded for the groups, not evidence by themselves of a causal treatment difference.

The record also names Suicide Attempt and Spontaneous Abortion among the top serious terms, each with 5.3%. The group attribution and event counts for these terms are not reported in the registry record, so they should not be assigned to an arm or interpreted as arm-specific rates. The registry does not report event timing, duration, severity criteria, relatedness assessments, or exposure-adjusted rates.

In general, safety counts require context. A small count can be important to describe, but it does not by itself establish a comparative risk. Here, allocation was nonrandomized and the groups are not a like-for-like control comparison. The listed denominators also differ from total enrollment; their relationship to the primary analysis population is not reported in the registry record. The figures are best read as descriptive safety entries, not as proof that one intervention caused more or fewer events.

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

The study record establishes that a completed, open-label, nonrandomized investigation enrolled 48 participants and compared FEAST with standard-of-care RUL UB in a parallel-group design. It identifies a clinically relevant measurement strategy: change in depressive symptoms assessed with the HRSD from baseline to the end of the acute course, typically after 4 weeks. It also records adverse-event counts and describes treatment refinements made during the development work.

Those details do not establish the magnitude of the primary endpoint result. Final mean percentage changes and their variability by arm are not reported in the registry record. No primary between-arm estimate, confidence interval, or inferential test is reported. Therefore, the record does not allow a reader to determine how large or precise the primary outcome contrast was.

The preliminary results described in the brief summary concern earlier open-label experience. They should not be treated as a final comparison of the two groups in the completed study. Similarly, enrollment and completion status indicate the scale and administrative status of the trial, not whether a treatment difference was observed or how uncertain it was.

General interpretation of a treatment comparison requires attention to both the observed result and the study design. Nonrandomized allocation can leave differences between groups that influence outcomes, and unmasked assessment can introduce measurement influences. These are design considerations, not proof that bias occurred in a particular direction. The registry does not report enough information to assess these issues in detail or determine whether statistical adjustment was used.

The safety listings show the numbers of participants affected as recorded, but do not identify a causal effect. The small number of serious events and the absence of a randomized comparison limit what can be inferred from arm counts alone. Overall, the registry record documents the research question and selected descriptive safety information, while leaving the completed trial’s primary efficacy result and its statistical precision unreported.

14. Why This Trial Matters Statistically

This trial is statistically notable for the contrast between a clearly defined primary measurement and limited posted outcome information. The record specifies percentage change in HRSD scores, a time frame, and a mean with standard deviation as the summary. Yet it does not report the final arm-level results for that endpoint. This illustrates why registration and outcome definitions are not substitutes for posted estimates and uncertainty measures.

The study also highlights the interpretive limits of a nonrandomized comparison. A parallel-group structure can make treatment groups appear directly comparable, but the allocation method matters: the record states that allocation was nonrandomized and masking was absent. In general, these features make it important to consider selection and measurement influences when interpreting a contrast. They do not, without further evidence, establish the direction or magnitude of bias.

Another statistical issue is the distinction between a planning illustration and evidence generated by the completed trial. The registry describes sample-size illustrations based on an observed effect and stated assumptions, but planning scenarios do not reveal the precision achieved in the actual study. They are not observed power and cannot replace the estimate and confidence interval for the clinical outcome.

The safety information further illustrates the difference between description and inference. Counts can communicate what was recorded, but do not establish comparative risk in a nonrandomized study, particularly when the record does not report event timing, exposure-adjusted analysis, or inferential methods. Together, these features make the record useful for understanding design and measurement, while underscoring the need for transparent outcome reporting before a quantitative treatment comparison can be evaluated.

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 limitation for statistical interpretation is that final arm-level results for the primary outcome are not reported in the registry record. The mean percentage change, standard deviation, between-arm estimate, confidence interval, and p-value are not reported in the registry record. Without them, neither the size of the observed contrast nor its precision can be assessed.

Nonrandomized allocation and no masking are recorded design features. The allocation procedure, baseline comparability of the arms, any adjustment for baseline characteristics, and the HRSD assessment procedures are not reported in the registry record. These gaps prevent a detailed assessment of how selection and measurement could have influenced the comparison.

The registry lists 48 enrolled participants, while the adverse-event entries give risk sets of 20 for FEAST and 19 for RUL UB. The relationship between total enrollment, these safety denominators, and the primary endpoint analysis population is not reported in the registry record. The number and pattern of missing primary outcome observations are also not reported in the registry record.

The primary endpoint formula and sign convention, the final analysis plan, any interim analyses, handling of multiplicity, and sensitivity analyses are not reported in the registry record. The record also does not give enough detail to assess the consistency of treatment delivery after the described FEAST refinements. These limitations constrain interpretation; they do not establish that any particular analytic or treatment-delivery problem occurred.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Focal Electrically-Administered Seizure Therapy (FEAST) Studies at Two Enrolling Sites to Further Test and Refine the Treatment — NCT02535572.
• Registry record last updated 2021-02-04; status last verified 2021-02.
• Results first posted 2021-02-04.
• 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

NCT02535572 is a completed Phase 2 and Phase 3, nonrandomized, unmasked, parallel-group study with enrollment of 48. It compared FEAST with standard-of-care right unilateral ultrabrief ECT. The sponsor was the Medical University of South Carolina; the study began in 2015-08, with primary completion recorded as 2019-06-10.

The primary outcome was percentage change in depressive symptoms on the HRSD from baseline to the end of the acute course, typically after 4 weeks. The registry record does not report final arm-level results for this outcome. Safety listings report serious adverse events for 0 of 20 participants at risk in FEAST and 2 of 19 in RUL UB; other adverse events for 17 in each listed group; and 0 deaths in each group. These counts are descriptive and do not establish a causal treatment difference.