Low Frequency TMS for Depression in Epilepsy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Low Frequency TMS for Depression in Epilepsy (NCT03105700), a na na study. The primary outcome is Change in Seizure Frequency Expressed as the Average Number of Seizures Experienced by All Participants and Recorded at Specified Time Points Throughout the Study..

2. Clinical Question

3. Trial Design

This completed study, registered as NCT03105700, evaluated accelerated low-frequency TMS in patients with epilepsy and medication-resistant depression. Dartmouth-Hitchcock Medical Center is listed as the sponsor. The study began on 2017-04-01, and the primary completion date is 2021-11-15. The registry status is completed.

The design is recorded as single group, with enrollment of 15. Allocation is listed as NA, and masking is listed as none. The intervention group is described as receiving low-frequency TMS across three consecutive days. There is no like-for-like control arm, so the design does not support a randomized between-group treatment contrast. Measurements at different times can describe the course of outcomes among participants, but changes over time alone cannot separate an intervention effect from other changes occurring during follow-up.

The registry describes in-person follow-up visits after one month and again after six months. The primary endpoint also lists baseline and 1-week post-treatment measurement points, as well as 1-month and 6-month post-treatment follow-up. Seizure information was reported directly by participants during check-ins with research staff. The registry does not report the detailed visit procedures, adherence to the schedule, or a statistical analysis plan.

The study is listed as phase NA. Its stated purpose was to determine whether low-frequency TMS is safe and feasible for depressive symptoms in patients with epilepsy. Depression severity is listed as a secondary outcome, but numerical secondary-outcome findings are not reported in the registry record. The design is therefore best understood as a single-group safety and feasibility investigation, rather than a controlled efficacy trial.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Subjects 3 and14 were lost to 6-month follow-up for unknown reasons. Subjects 9 and 13 were lost to 6-month follow-up due to COVID 19.

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: Patients that were enrolled.

6. Primary Endpoint

The registry lists the primary outcome as Change in Seizure Frequency Expressed as the Average Number of Seizures Experienced by All Participants and Recorded at Specified Time Points Throughout the Study., measured over Baseline, 1-week post-treatment, 1-month post-treatment, 6-month post-treatment follow-up.

The primary endpoint is change in seizure frequency, expressed as the average number of seizures experienced by participants. The listed measurement points are baseline, 1-week post-treatment, 1-month post-treatment, and 6-month post-treatment follow-up. Participants reported seizures directly during check-ins with research staff. The registry lists the outcome’s dispersion as the full range.

Seizure frequency is relevant to the study’s stated safety question because the registry defines the hypothesized serious safety concern in terms of an increase in average seizures across participants. The endpoint therefore tracks an outcome of particular interest during and after TMS exposure, rather than measuring depressive symptoms as the primary outcome. Depression severity is identified separately as a secondary outcome.

In general, an average describes the central level of observations, while a range describes the spread from the lowest to highest recorded values. A range can make variation visible, but it does not show how observations are distributed between its endpoints and can be strongly affected by extreme observations. Repeated measurements can also help describe whether a measure changes over follow-up, but interpretation depends on the observed values, the number of measurements available at each timepoint, and how missing measurements are handled.

The registry record supplies the endpoint definition and timepoints, but not the actual average seizure counts or ranges at those timepoints. It therefore does not permit a numerical assessment of the direction or size of change in the primary endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a mean outcome with full-range dispersion, measured repeatedly in a single group. It does not report the statistical analysis method used for the primary endpoint. The explanations below are general statistical context, not a description of analyses documented for this trial.

What does an average seizure count describe?

In general, an average combines observations into a measure of central tendency. It can be a concise way to describe the group’s seizure frequency at a timepoint, but it does not show whether most participants had similar values or whether a small number of high values influenced the average. That distinction matters for count outcomes, where observations may vary substantially among participants.

Seizure frequency is a count, and counts are often non-negative whole-number observations. In general, simple methods that assume normally distributed outcomes may not represent count data well, particularly when counts are highly variable or include many low values. A model for count outcomes can account for the relationship between the average and variability. When variability exceeds what a basic count model expects, over-dispersion is a concern. Ignoring over-dispersion can make uncertainty appear smaller than it is. The registry does not report whether any count model, over-dispersion assessment, or alternative analysis was used.

Why report the full range?

In general, the full range gives the lowest and highest observations and can reveal that the data include values far apart. It is easy to interpret, but it does not describe the distribution between those values. It may also change substantially because of one unusually low or high observation. A range is therefore useful context, but it is not a substitute for participant-level distributions or an uncertainty interval around an estimated average. The registry record does not report the numerical ranges or additional dispersion measures for the seizure endpoint.

How should repeated measurements be understood?

The endpoint is measured at baseline and multiple post-treatment timepoints. In general, measurements from the same participant are related: a participant’s observations over time are not independent in the way observations from unrelated people might be. An analysis of repeated observations should account for that within-participant relationship if it estimates changes across time or summarizes a trajectory. The specific statistical method, any adjustment for baseline, and any contrasts among timepoints are not reported in the registry record.

Timepoint-specific summaries can describe the observed data at each visit, while a model can estimate change using all available observations under its assumptions. Neither approach eliminates the need to understand missing measurements. A summary at a later visit may describe only the participants with data at that timepoint, depending on the analysis population and missing-data procedure. Those details are not reported for the primary endpoint.

What can a confidence interval contribute?

In general, a confidence interval expresses uncertainty around an estimated quantity under the assumptions of the analysis. A narrow interval indicates greater precision than a wide interval for that estimate, while an interval that spans a range of plausible values signals uncertainty about the estimate’s magnitude. An interval is not a guarantee that the true value lies within its bounds, and it does not correct biases caused by study design, measurement, or missing data. No confidence intervals for seizure-frequency results are reported in the registry record.

A hypothesis test and a confidence interval answer related but different questions. A test assesses compatibility of the observed data with a specified null hypothesis under a statistical model. An interval gives a range of values compatible with the data and model at a stated confidence level. Neither can be interpreted for the primary endpoint without posted estimates and analysis results. The registry does not report a primary endpoint p-value, test statistic, or interval.

How do the planning illustrations differ from trial results?

In general, a sample-size calculation estimates the number of participants or events needed under chosen design assumptions. The registry describes planning illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is described as a stated fraction of the observed effect. For time-to-event calculations, the stated quantity is total events; for other calculations, it is participants per arm.

Those assumptions describe planning illustrations, not the design of this single-group trial. In particular, equal allocation between two arms is not the allocation structure recorded for this study. A planning calculation based on an observed effect is not an observed power analysis and does not establish the reliability of the effect estimate. The registry record does not report numerical sample-size illustration results, and it does not report the trial’s own sample-size justification.

More broadly, a statistical method must be interpreted in light of the study design and available data. A method can quantify uncertainty under its assumptions, but it cannot create a control group, supply missing outcome observations, or resolve an unreported analysis plan. Since the registry does not provide primary outcome values or an analysis record, statistical conclusions about the magnitude and precision of seizure-frequency change cannot be reproduced from the posted information.

10. Multiplicity, Missing Data and Analysis Populations

The primary endpoint has several listed measurement times, and depression severity is named as a secondary outcome. Considering multiple outcomes or timepoints can raise multiplicity concerns: in general, testing many hypotheses can increase the chance of finding an apparently notable result by chance if no adjustment or prespecified hierarchy is used. The registry record does not report a multiplicity strategy, a hierarchy among timepoints, or which comparisons were planned for formal testing. It also does not report an analysis plan that would establish whether the listed outcomes were tested or only summarized descriptively.

Missing follow-up is relevant to interpretation. The registry states that subjects 3 and 14 were lost to 6-month follow-up for unknown reasons, and subjects 9 and 13 were lost to 6-month follow-up due to COVID 19. It does not report the amount of primary endpoint data available at each timepoint, the reasons for any other missing observations, or how missing data were handled. In general, missingness can affect estimates when the availability of an observation is related to the outcome or to participant characteristics. A complete-case analysis, imputation procedure, or repeated-measures model would each require assumptions; the record does not say which, if any, was used.

The registry does not define an analysis population for the primary endpoint, such as a population based on enrollment, treatment receipt, or availability of measurements. The record also reports no analysis entries. Consequently, the population included in any primary endpoint estimate and the treatment of missing observations are not reported in the registry record. These gaps limit assessment of both the completeness of follow-up and the robustness of any outcome estimate.

11. 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 non-serious adverse event terms

The adverse-event summary lists 12 participants at risk for serious adverse events, with 0 serious adverse events and 0 deaths. It also lists 12 participants at risk for other adverse events, with other adverse events recorded for 12. The record does not provide the terms or descriptions for those other events in this summary.

These counts are descriptive. They report what the registry’s adverse-event summary records, but they do not indicate the type, timing, severity, or duration of the other adverse events. The record also does not report an event-by-event relationship assessment or a comparison with an untreated or alternative-treatment group.

In general, a count of zero observed serious events does not prove that the underlying risk is zero. Interpretation depends on the number of people observed, duration and completeness of follow-up, how events were defined and collected, and the precision of the estimate. A single-group study also cannot make a randomized comparison of event rates. The serious-event counts should therefore be read alongside the missing numerical seizure-frequency results and the lack of a control group, not as a complete resolution of the study’s safety question.

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

The record establishes that a completed single-group study examined accelerated low-frequency TMS in patients with epilepsy and medication-resistant depression. It specifies a three-consecutive-day intervention schedule, follow-up timing, and a primary endpoint based on average seizure frequency. It also records adverse-event counts. These details clarify what the study set out to examine and what was recorded in its registry summary.

The record does not report the numerical seizure-frequency results at baseline or the post-treatment timepoints. It therefore does not show whether the average changed, how large any change was, how variable the observations were, or how precisely a change was estimated. Depression severity is listed as a secondary outcome, but its numerical results are also not reported in the registry record. Feasibility measures are not reported either. The study’s stated aims should not be mistaken for findings.

The serious adverse-event summary reports 0 serious events among 12 at risk and 0 deaths, while other adverse events are recorded among 12 at risk. Those counts do not establish the absence of risk, and they do not tell readers what the other events were. The record’s reported losses to 6-month follow-up are also relevant when considering longer-term observations.

Because the trial has no concurrent comparison group, observed changes over time, even if numerical results were available, would not by themselves establish that TMS caused those changes. The registry record also lacks a documented analysis method, analysis population, and missing-data procedure. The appropriate statistical conclusion is limited: the study design and endpoint are described, but the posted information does not permit a quantitative conclusion about the primary outcome or the intervention’s effect.

This interpretation is not medical advice. It is a distinction between the study’s objectives, the outcomes listed, and the results actually reported in the registry.

13. Why This Trial Matters Statistically

This trial raises statistical issues common to early or exploratory studies that measure safety-related outcomes over time. Its primary endpoint is a repeated count outcome, collected from participant reports at specified timepoints. Count distributions can be uneven, and a group average may conceal substantial participant-to-participant variation. Reporting a range offers some context about spread, but without numerical values, additional distributional summaries, and uncertainty estimates, readers cannot judge the observed pattern or its precision.

The single-group design is also central. Within-person follow-up can describe change after an intervention, but a concurrent comparator is needed for a direct randomized contrast. Without one, change over time cannot be attributed to TMS from the design alone. This matters even when the study’s objective is safety and feasibility: clear descriptive reporting and transparent limitations remain necessary for readers to assess what was observed.

Missing follow-up, especially at the 6-month timepoint, underscores why analyses should report the number of observations available at each visit and explain how incomplete data are handled. The registry identifies reasons for some losses, but does not describe the analysis population or missing-data method. Those choices can affect both the estimated change and its uncertainty.

Finally, a registry record can describe objectives and outcome definitions without containing the numerical findings needed for statistical interpretation. In this case, the primary endpoint values and analysis results are not reported in the registry record. That distinction prevents the study’s safety hypothesis from being mistaken for an observed conclusion and keeps planning assumptions separate from trial evidence.

14. 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 are specific to interpreting this study from its registry record. The primary endpoint is defined as average seizure frequency with full-range dispersion, but the record does not include the numerical averages or ranges at any listed timepoint. There is consequently no reported estimate, interval, or test result from which to judge the direction, size, or precision of change.

The study is single group, with no like-for-like control comparison. Allocation is listed as NA and masking as none. Therefore, any time-related pattern would not constitute a randomized comparison, and the design alone could not distinguish an intervention-related pattern from other changes over follow-up.

The record describes loss to 6-month follow-up for subjects 3 and 14 for unknown reasons and for subjects 9 and 13 due to COVID 19. It does not report primary endpoint completeness at each timepoint, the handling of missing values, or the analysis population. The extent to which missing follow-up affects the posted outcome summaries cannot be assessed.

The adverse-event summary reports counts, but the other adverse-event terms are not listed, and the record does not provide event timing or further descriptions. The endpoint description also does not supply numerical seizure data. Depression severity is named as a secondary outcome, but its results are not reported in the registry record. The statistical analysis plan, multiplicity approach, and trial-specific sample-size justification are also not reported. These omissions limit assessment of the study’s findings, not merely their precision.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Low Frequency TMS for Depression in Epilepsy — NCT03105700.
• Registry record last updated 2023-03-21; status last verified 2023-03.
• Results first posted 2023-03-21.
• 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.

18. Record Summary

NCT03105700 is a completed, single-group study titled “Low Frequency TMS for Depression in Epilepsy,” sponsored by Dartmouth-Hitchcock Medical Center. It recorded enrollment of 15 and evaluated low-frequency TMS administered over three consecutive days to patients with epilepsy and medication-resistant depression. Allocation is listed as NA, masking as none, and phase as NA.

The primary endpoint is average seizure frequency, with a full range listed as the measure of dispersion. Timepoints are baseline, 1-week post-treatment, 1-month post-treatment, and 6-month post-treatment follow-up. Depression severity is listed as a secondary outcome. Numerical primary and secondary outcome findings are not reported in the registry record.

The adverse-event summary lists 0 serious adverse events among 12 at risk, 0 deaths, and other adverse events among 12 at risk. The record reports losses to 6-month follow-up for subjects 3 and 14 for unknown reasons and subjects 9 and 13 due to COVID 19. It reports no analysis entries.