Cranial Electric Stimulation to Modify Suicide Risk Factors in Psychiatric Inpatients.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Cranial Electric Stimulation to Modify Suicide Risk Factors in Psychiatric Inpatients. (NCT02846740), a randomized na study. The primary outcome is Change From Baseline in the Modifiable Suicide Risk Factors (MSRF's) Global Score.

2. Clinical Question

This was described as a pilot study, with randomized allocation and a parallel-group design. The registry records 30 participants enrolled, with 15 in each arm in the posted primary outcome data. The central statistical question is therefore how to interpret the observed difference and its uncertainty in a small trial with incomplete repeated assessments, rather than whether the results alone settle the effectiveness of CES.

3. Trial Design

The registry describes a randomized, quadruple-masked, parallel-group trial comparing adjunctive cranial electric stimulation with a sham control. The active intervention used the Alpha-Stim® device at 100µA for one hour daily, five to seven days per week, alongside usual treatment. The sham device did not emit electricity, was designed to appear identical to the active device, and had its current intensity preset and locked by the manufacturer. The registry says the other procedures were the same for both groups.

The enrollment was 30. The posted primary outcome data include 15 participants in the adjunctive CES arm and 15 in the sham control arm. The registry describes the study as a pilot. It records a start date of 2017-03-01, a primary completion date of 2018-01-22, and a status of terminated. The reason for termination is not reported in the registry record.

Rating scales were planned at baseline, twice a week, and at the end of the study. The primary outcome was a global score combining depression and anxiety scales. Although the registry outcome timeframe reads “Baseline, 22 months,” the record also explains that short hospital stays limited repeat measurements. The actual assessment timing and the analysis set used for each posted comparison are not reported in the registry record.

The registry lists length of stay as a secondary outcome. It also describes agitation and sleep quality measures, but says those scales could not be used meaningfully because of short stays. The original sample-size rationale, allocation details beyond randomized assignment, and the trial’s planned statistical analysis methods are not reported in the registry record.

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: Many patients were discharged within a few days of admission, some left in two days, which left us with just baseline and several missing scales.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in the Modifiable Suicide Risk Factors (MSRF's) Global Score, measured over Baseline, 22 months,.

The primary endpoint was titled change from baseline in the Modifiable Suicide Risk Factors global score. It was intended to combine measures of depression and anxiety, both treated in the registry as modifiable suicide risk factors. The component measures were the Montgomery-Åsberg Depression Rating Scale and the Hamilton Anxiety Rating Scale. Their summed score had a theoretical range of 0–116, with higher scores indicating a worse outcome.

The registry describes severity thresholds for the individual scales: the depression scale ranges from 7 for recovered to 44 for very severe, while the anxiety scale ranges from 0 to 56, with higher values indicating greater severity. The global score was intended to capture both domains together, rather than either scale alone. The registry says agitation and sleep quality scales were not included in the analysis because short hospital stays prevented meaningful use of those measures.

A combined score can provide a single summary of the targeted risk factors, but it also makes the result dependent on the chosen components and scoring rules. In this record, the displayed endpoint title refers to change from baseline, while the posted arm means are not explicitly identified as change scores or as scores at a named follow-up visit. That measurement detail is not reported in the registry record. The distinction matters because a difference in change scores and a difference in follow-up scores do not necessarily represent the same statistical quantity.

7. Results Posted in the Registry

• Adjunctive CES: mean 35.33 versus 36.50 on control; difference -1.17 (95% CI -1.91 to -0.43), standardized effect d = -1.14; z-test p = 0.002.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Length of Stay

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Adjunctive CES: mean 11.00 versus 12.00 on control; difference -1.00 (95% CI -1.48 to -0.52), standardized effect d = -1.50; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Adjunctive CES than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Length of stay is the only posted secondary endpoint. Its summary-statistic comparison points in the same direction as the primary score contrast: the mean value was lower in the adjunctive CES group than in the sham control group. The secondary endpoint therefore does not show a directional pattern that conflicts with the primary result. It measures a different aspect of the inpatient course, however, and does not directly verify a change in the combined depression and anxiety score.

The registry lists one secondary outcome, but the record does not report a multiplicity adjustment or a prespecified hierarchy for interpreting secondary outcomes. When several outcomes or analyses are considered, unadjusted results can overstate the strength of evidence because each comparison creates another opportunity for a chance finding. The weight of a secondary result is therefore lower when its inferential role and adjustment plan are not documented, particularly in a small pilot study.

The registry lists agitation and sleep quality measures among the planned assessment instruments, but says the Agitated Behavior Scale and Pittsburgh Sleep Quality Index could not be used meaningfully because of short hospital stays. They were not included in the analysis. Thus, the secondary pattern is limited to the posted length-of-stay endpoint, rather than a consistent set of independently analyzed outcomes covering every intended risk factor. Formal analyses for the secondary endpoint are not reported in the registry record.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Length of Stay

9. Statistical Interpretation

10. Statistical Methods Explained

The posted primary outcome is continuous: it is a score on a scale, and the registry reports group means and standard deviations. A mean describes the average value within a group, while a standard deviation describes how spread out individual values are around that mean. These summaries are useful for describing the groups, but they do not show the distribution of each participant’s measurements or the pattern of repeated measurements over time.

The comparison is a mean difference, with the adjunctive CES arm compared against the sham control arm. The posted estimate is -1.17. Its sign depends on the order used for the comparison; a negative estimate means the CES group’s reported mean was lower than the sham group’s mean. Since the registry defines higher global scores as worse, lower scores point in a favorable direction on the scale, but this interpretation depends on whether the recorded values are change scores or measurements at a specified time. That point is not clarified in the registry record.

General statistical explanation: a confidence interval describes a range of values compatible with the data and the assumptions of the analysis. It reflects uncertainty about the estimated contrast, rather than a range containing a known proportion of individual participant outcomes. For the primary mean difference, the interval is below zero, from -1.907 to -0.433. The interval indicates the estimate is more precise than a result whose interval spans a wide range of directions, but it does not address missing data, measurement ambiguity, or bias in the trial conduct.

The unadjusted p-value for the primary comparison is 0.002. General statistical explanation: a p-value measures how incompatible the observed result is with a specified null model, assuming that model and the analysis conditions are appropriate. It is not the probability that the null hypothesis is true, and it does not quantify the size or clinical importance of an effect. Here it is described as unadjusted because the posted value does not account for multiple comparisons.

The record also gives a standardized effect estimate, d = -1.136, based on the posted summaries. General statistical explanation: a standardized mean difference expresses a contrast in units of pooled standard deviation, which can help describe magnitude on a scale-independent basis. It is especially sensitive to the variability estimate in small samples, so it should not be treated as a stable prediction for a future study. It also does not resolve whether the reported values are change scores or follow-up scores.

The primary comparison has a reported standard deviation of 1.02 for the adjunctive CES group and 1.04 for the sham group. These are the spread measures attached to the posted arm summaries. The registry does not report individual observations, baseline-to-follow-up correlations, or a model for repeated measurements. It is therefore not possible to assess from the record whether a longitudinal model, baseline adjustment, or another approach was used in the trial’s own analysis. The original statistical analysis plan is not reported in the registry record.

General statistical explanation: repeated measures from the same participant are usually correlated. An analysis that uses several measurements per person needs to account for that dependence; treating each measurement as independent can misstate uncertainty. The planned assessments were at baseline, twice a week, and at the end of the study, but short inpatient stays led to missing scales. The registry does not say how the available repeated observations were incorporated or whether the posted comparison uses only one measurement per participant.

General statistical explanation: missing outcomes can affect both precision and interpretation. If participants with missing measurements differ systematically from those with observed measurements, a complete-case comparison may not represent all randomized participants. Other approaches can use available measurements under explicit assumptions, but the assumptions and method need to be stated. The registry notes short stays and missing scales but does not report an imputation method, missingness model, or analysis population. No particular missing-data approach should be inferred.

Over-dispersion is a concern when observed variability is greater than a statistical model expects. General statistical explanation: if a model understates variability, its standard errors may be too small and its confidence intervals too narrow. The primary summaries do not report an over-dispersion assessment, and the registry does not describe a model-based analysis for the primary endpoint. There is no basis in the record to claim that over-dispersion was tested or that it affected the comparison.

The comparison values are recalculated from posted summary statistics and are unadjusted. The record does not list a sponsor-reported analysis for the primary endpoint. The trial’s own inferential method, multiplicity strategy, interim analysis plans, and prespecified covariate adjustments are not reported in the registry record. Those omissions limit what can be concluded about whether the displayed comparison matches the analysis originally intended by the investigators.

11. Confidence Intervals and Precision

The primary comparison is a mean difference of -1.17, with a 95% confidence interval from -1.907 to -0.433. The interval lies below zero, so the posted summary-statistic comparison is consistent with a lower mean in the adjunctive CES group than in the sham control group. The interval communicates uncertainty around the mean difference; it does not show the range of responses among individual participants or establish the likely result in a different population.

The interval’s endpoints are relatively close to the estimate compared with intervals that span both directions, but its practical interpretation depends on the scale and on whether the posted values represent change from baseline or scores at a particular assessment. The registry defines higher scores as worse, but does not clarify the measurement status of the arm means. The clinical importance of the interval cannot be judged from the registry record because a threshold for a meaningful difference is not reported.

Precision is also affected by the limited number of participants in the posted comparison: each arm includes 15. The registry reports that several scale measurements were missing after short hospital stays. The posted summary does not tell readers how many repeated assessments were available per participant or how missing values contributed to the comparison. The confidence interval reflects the calculation’s statistical uncertainty under its assumptions, not uncertainty from every possible source of bias or incomplete follow-up.

The unadjusted p-value is 0.002, but that value does not make the interval narrower or remove concerns about outcome selection, multiple comparisons, or the analysis population. The endpoint is a composite of depression and anxiety ratings, while the registry says agitation and sleep quality were excluded from meaningful analysis. Taken together, the interval gives information about the precision of the reported contrast, while the unreported analysis details and incomplete assessments constrain broader interpretation.

12. 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 sample-size entries are planning illustrations expressed as participants per arm, not total events. They apply to continuous outcomes and use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The assumed effect is a stated fraction of the observed effect. These are planning assumptions for the illustration; the original design assumptions for the trial are not reported in the registry record.

The 80% power column gives the estimated participants per arm for a future study to have an 80% chance of detecting the assumed effect under those planning assumptions, if that effect and the model assumptions are correct. The 90% column uses a higher target chance of detection and therefore generally calls for more participants per arm. For the primary comparison, the entries range from 13 per arm when planning for 100% of the observed effect to 49 per arm when planning for 50% of the observed effect at 80% power. At 90% power, the corresponding entries range from 17 to 66 per arm.

The rows planning for 75% and 50% of the observed effect illustrate that smaller assumed effects need more participants to detect. The displayed numbers are not observed power, and they are not the number enrolled in this trial. They are sensitivity illustrations based on the observed contrast, which may be unstable in a small pilot study. A future study’s actual requirements would depend on its chosen outcome, expected variability, missingness, design, and analysis assumptions; those choices should not be inferred from these illustrative entries.

13. 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 = -1.170; pooled standard deviation σ = 1.030.
2. n = 2 × (1.96 + 0.842)² × 1.030² / -1.170² = 12.2, rounded up to 13 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 <- -1.17                      # observed difference in means
sigma <- 1.03004854254545                     # 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(2846740)
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 = 13 (80% power), 17 (90% power)
 75% of observed effect: n per arm = 22 (80% power), 29 (90% power)
 50% of observed effect: n per arm = 49 (80% power), 66 (90% power)
power.t.test: 13.2 
Simulated power at n = 13 per arm: 0.793 (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.

14. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a multiplicity strategy for the primary and secondary outcomes, or for repeated assessments over time. It also does not report whether the investigators prespecified one primary time point, a hierarchy of outcomes, or any adjustment to control error across multiple comparisons. The posted p-values are unadjusted. If several outcomes, time points, or analyses were examined, an unadjusted result can be less persuasive than a result interpreted under a clearly specified testing plan.

Missing data are a material limitation. The registry states that many patients were discharged within a few days, some left in two days, and that this left baseline data and several missing scales. It further notes that short stays prevented meaningful use of the agitation and sleep quality instruments. The extent of missingness at each assessment, reasons for missing data beyond short stays, and differences between participants with and without follow-up measurements are not reported in the registry record.

The analysis population is also not reported. The record does not state whether the posted comparisons include all randomized participants, only those with available outcome data, or another defined group. It does not report an intention-to-treat or per-protocol analysis, an imputation method, or sensitivity analyses addressing missing outcomes. The randomized assignment is noted, but the consequences of missing follow-up data for preserving the randomized comparison cannot be assessed from the posted information.

The registry does not report an interim analysis plan, stopping rules, or the reason the study was terminated. The original sample-size justification and the protocol’s statistical analysis plan are also not reported in the registry record. These are distinct uncertainties: the displayed planning illustrations describe hypothetical future sample sizes based on observed effects, not the study’s original enrollment rationale or conduct.

15. 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.

The posted adverse event counts list zero serious adverse events, zero other adverse events, and zero deaths in both arms. Each group has 15 participants at risk in the registry’s safety summaries. No serious adverse event terms are listed.

These entries mean that no events of those types were recorded in the posted safety data for the listed groups. They do not establish that no adverse events occurred outside the definitions or collection period used, or that the intervention has no safety risks. The registry does not report detailed event definitions, surveillance methods, exposure duration for each participant, or a narrative explaining how safety events were collected. The small group sizes also limit how much can be learned from zero recorded events. The counts should therefore be read as descriptive trial-record information, not as a complete safety profile.

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

The primary posted comparison shows a lower mean global score in the adjunctive CES arm than in the sham control arm, with a mean difference of -1.17 and an unadjusted p-value of 0.002. The confidence interval lies below zero. Since the score is defined so that higher values indicate worse outcomes, the direction is consistent with a lower score in the active-device group. However, the registry does not clarify whether the arm means are changes from baseline or measurements at a named follow-up time, which limits the interpretation of that direction.

This result does not establish that CES reduces suicide risk. The endpoint is a combined depression and anxiety score intended to capture modifiable risk factors, not a direct count of suicide attempts or deaths. The registry describes a pilot trial and reports missing repeated measures associated with short hospital stays. It does not provide the planned analysis, analysis population, missing-data method, or a stated threshold for a clinically meaningful score difference.

The result also does not establish that the observed contrast would recur in another sample, in a different care setting, or over a longer observation period. The registry record does not report outcome follow-up beyond its stated timeframe, and the posted summary does not allow assessment of individual trajectories. Statistical evidence for a difference in a specific comparison should not be confused with certainty about its size, durability, cause, or relevance to outcomes beyond the rating scales.

The most defensible reading is narrow: the posted summaries show a difference between the two groups on the recorded global score, in the direction of a lower score for adjunctive CES. The uncertainty interval and p-value describe that comparison under its calculation assumptions. Missingness, unclear measurement timing, and unreported analysis details mean the record does not resolve how broadly the result should be interpreted.

17. Why This Trial Matters Statistically

This trial raises several important statistical issues common in small clinical studies. First, randomized assignment and a sham comparator support a direct group comparison, but interpretation still depends on knowing what was measured and when. The endpoint title refers to change from baseline, while the posted arm means are not explicitly identified as change scores or scores at a particular visit. That ambiguity affects what the reported contrast represents.

Second, the inpatient setting created a practical challenge for repeated measurement. The registry says many patients left within a few days and several scale assessments were missing. That makes the choice of analysis population and missing-data method especially relevant, yet neither is described in the record. In a repeated-measures study, missingness can affect both the amount of information available and the comparability of the groups.

Third, this is a pilot with a modest enrollment, and its observed effect is also used as the basis for planning illustrations. Such illustrations are useful for showing how assumptions about effect size influence future sample-size estimates, but an effect observed in a small study can be uncertain. Planning for a smaller effect requires more participants, which is why the fractions of the observed effect lead to different illustrative requirements.

Finally, the registry record illustrates the distinction between statistical evidence and evidential completeness. A confidence interval that excludes zero and a small unadjusted p-value describe one numerical comparison, but do not substitute for a prespecified analysis plan, clear measurement timing, complete follow-up, or appropriate consideration of multiple outcomes. Those details are essential for judging reliability and deciding how much weight the observed result should carry.

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.

Several limitations are specific to this trial record. The registry says participants were often discharged within a few days, with some leaving in two days. This curtailed repeated scale collection and left multiple assessments missing for some patients. The record does not give counts of missing observations by arm or assessment, so the completeness of the longitudinal data cannot be evaluated.

The primary endpoint combines depression and anxiety rating scales, but the posted arm means are not clearly identified as change-from-baseline values or scores from a particular follow-up. The timeframe field reads “Baseline, 22 months,” while the trial description emphasizes brief inpatient stays. The timing and meaning of the reported endpoint values are therefore not fully clear in the registry record.

The registry says agitation and sleep quality scales could not be used meaningfully and were excluded from analysis. This narrows the assessed set of intended risk factors. It also means the global outcome represents the two included scales, rather than all the instruments mentioned in the study description.

The sample was small, with 15 participants per arm in the posted primary and secondary summaries. The original sample-size justification, analysis population, missing-data handling, multiplicity approach, and reason for termination are not reported in the registry record. The result should be interpreted with these specific gaps in mind, rather than as a complete account of the trial’s planned and conducted statistical analyses.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Cranial Electric Stimulation to Modify Suicide Risk Factors in Psychiatric Inpatients. — NCT02846740.
• Registry record last updated 2019-12-26; status last verified 2019-12.
• Results first posted 2019-12-26.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

The registry describes a randomized, quadruple-masked, parallel-group pilot comparing adjunctive Alpha-Stim® cranial electric stimulation with a sham control in psychiatric inpatients. It records 30 participants enrolled and a terminated study status. The primary endpoint was a global score combining depression and anxiety ratings; the record defines higher scores as worse. Posted primary data include 15 participants per arm and a mean difference of -1.17, with a 95% confidence interval from -1.907 to -0.433 and an unadjusted p-value of 0.002.

Length of stay is the only posted secondary outcome. The registry reports zero serious adverse events, zero other adverse events, and zero deaths in each arm. It also notes missing scale assessments associated with short hospital stays and says agitation and sleep quality scales could not be used meaningfully. Measurement timing, missing-data methods, the analysis population, multiplicity procedures, and the original sample-size rationale are not reported in the registry record.