Genomics Used to Improve DEpression Decisions: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Genomics Used to Improve DEpression Decisions (NCT02109939), a randomized na study. The primary outcome is Percent Change in the 17-item Hamilton Depression (HAM-D17) Score From Baseline to 8 Weeks.

2. Clinical Question

The clinical question concerns whether assignment to a testing strategy was associated with a different average change in depressive-symptom scores during the randomized comparison. It is not a comparison of the test result itself with a treatment: the intervention is the testing strategy, while treatment as usual is the comparator. The registry record also lists responder, remission, symptom-score, and follow-up outcomes. Those outcomes address related but distinct questions and should not be treated as substitutes for the prespecified primary measure.

3. Trial Design

NCT02109939, titled Genomics Used to Improve DEpression Decisions, is recorded as a randomized, triple-masked, parallel-group trial. The phase is listed as not applicable. The registry record reports enrollment of 1398 and describes participants as outpatients with major depressive disorder who had an inadequate response to at least one medication in the current episode. Assurex Health Inc. is the sponsor. The record lists the study as completed, with a start date of 2014-04 and primary completion on 2017-04-20.

The randomized comparison consisted of GeneSight Psychotropic Tested and Treatment As Usual. In the treatment-as-usual description, participants were not to see GeneSight results or know their arm assignment until after Week 12. The official title describes a 12 Week, randomized, double-blind, controlled evaluation followed by an open label 12-Week follow-up period. The record labels masking as triple; details about which people or roles were masked are not reported in the registry record.

The primary outcome was measured from baseline to the end of Week 8, within the randomized period. The recorded primary analysis population is per-protocol, and the reported analysis uses a mixed models analysis. The registry record does not report the protocol’s original sample-size justification, planned allocation ratio, detailed handling of missing observations, or whether interim analyses were planned. Enrollment and the numbers contributing to the primary estimates are different kinds of information: enrollment describes the trial, while the endpoint estimates are reported for their analysis populations.

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: Per-Protocol

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: Per Protocol Population at Baseline

6. Primary Endpoint

The registry lists the primary outcome as Percent Change in the 17-item Hamilton Depression (HAM-D17) Score From Baseline to 8 Weeks, measured over from baseline to end of Week 8.

The primary endpoint is percent change in the 17-item Hamilton Depression score from baseline to the end of Week 8. HAM-D17 is a continuous measure of depressive symptoms, with scores ranging from 0 to 50; the registry description states that lower scores are better outcomes. Percent change is defined in the record as (Week 8 score - baseline score) / (baseline score) x 100. A negative value therefore represents a decrease in the score relative to baseline, while the between-group estimate expresses the difference in mean percent change.

A continuous endpoint retains information about the measured score rather than reducing each participant to a responder/non-responder category. Comparing group means can describe the average change, while a confidence interval conveys uncertainty around the estimated difference. The registry reports least-squares means and standard errors for the primary outcome, and identifies the primary analysis population as per-protocol. Least-squares means are model-based group estimates; they are not necessarily identical to raw arithmetic averages.

Symptom scores are used as outcomes because they provide a structured way to assess change over time. They do not, by themselves, establish how participants experienced every aspect of illness or treatment. Nor does an average group difference show the distribution of individual responses. The registry record’s definition and scale support interpretation of score direction, but additional details about the clinical importance threshold for a difference are not reported in the registry record.

7. Results Posted in the Registry

• GeneSight Psychotropic Tested: mean -27.23 versus -24.39 on control; difference -2.84 (95% CI -6.29 to 0.61), standardized effect d = -0.09; z-test p = 0.107.

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.

Percent Change in the 16-item Quick Inventory of Depression Symptomology (QIDS-C16) Score From Baseline to 8 Weeks

• 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 errors.
• Result:
  • GeneSight Psychotropic Tested: mean -35.11 versus -32.90 on control; difference -2.21 (95% CI -5.44 to 1.03), standardized effect d = -0.07; z-test p = 0.182.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between GeneSight Psychotropic Tested and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Responders at Week 8 for HAM-D17

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • GeneSight Psychotropic Tested: 26.1% (146/560) versus 19.9% (121/607) on control; risk difference 6.1 percentage points (95% CI 1.3 to 11.0), risk ratio 1.31 (1.06–1.62); unadjusted two-proportion z-test p = 0.013.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for GeneSight Psychotropic Tested than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Remitters at Week 8 Defined as HAM-D17 ≤7 Each Treatment Group;

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • GeneSight Psychotropic Tested: 15.4% (86/560) versus 10.0% (61/607) on control; risk difference 5.3 percentage points (95% CI 1.5 to 9.1), risk ratio 1.53 (1.12–2.08); unadjusted two-proportion z-test p = 0.006.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for GeneSight Psychotropic Tested than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover symptom change on QIDS-C16, responder and remission classifications on HAM-D17, and other response, remission, time-to-event, and follow-up measures. The posted results available for the endpoints discussed include continuous mean changes and binary classifications. The broad pattern is not uniform across outcome types: the continuous QIDS-C16 comparison has an interval that includes zero, while the Week 8 HAM-D17 response and remission analyses have sponsor-reported odds-ratio intervals above one and p-values below 0.05.

These findings add context to the primary result but do not overturn or replace it. The primary HAM-D17 percent-change comparison has an interval including zero. Secondary outcomes can be informative because they capture other features of response, but they are not equivalent to the primary endpoint and may use different populations, models, or definitions. The record does not report the responder threshold, and the remission threshold is recorded as HAM-D17 ≤7.

There are multiple secondary outcomes, and the registry record lists 22 secondary outcomes as posted. If many hypotheses are tested, some nominally small p-values can occur by chance even when no underlying effects exist. Adjusting for multiplicity changes how results are interpreted and can affect the strength of evidence. The registry record does not report whether a multiplicity procedure was used, so individual unadjusted secondary p-values should carry less weight than a result supported by a prespecified analysis plan with clear control of multiple testing.

The secondary results are best treated as supportive or exploratory evidence unless the full analysis plan establishes their role and multiplicity handling. The record does not provide enough information to determine whether every listed secondary outcome was prespecified, how missing results were handled, or whether all planned comparisons are posted.

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

• Percent Change in the 16-item Quick Inventory of Depression Symptomology (QIDS-C16) Score From Baseline to 8 Weeks
• Percentage of Responders at Week 8 for HAM-D17
• Percentage of Responders at Week 12 for HAM-D17
• Percentage of Remitters at Week 12 Defined as HAM-D17 ≤7
• Percentage of Remitters at Week 8 Defined as HAM-D17 ≤7 Each Treatment Group;
• Time to Response/Remission of Depressive Symptoms Over 8 Weeks;
• Percent Change in the 17-item Hamilton Depression (HAM-D17) Score From Baseline to 24 Weeks
• Percentage of Responders at Week 8 for QIDS-C16

9. Statistical Interpretation

10. Statistical Methods Explained

The primary analysis is labeled Mixed Models Analysis, with the reported parameter described as Mean Difference (Net). A mixed model is a general statistical approach for analyzing outcomes while accounting for modeled structure in the data, such as repeated measurements or grouping. The registry record does not report the full model specification, the covariates included, or the covariance assumptions. Those details matter because the model determines how available observations contribute to estimated group means and their uncertainty.

What a least-squares mean represents

The primary outcome is a least-squares mean, also called a model-based mean. Unlike a simple raw average, it is estimated from a statistical model. When a model includes adjustment terms, the least-squares mean reflects those terms; the registry record does not identify which adjustment variables were used. The primary comparison is the difference between the model-based mean percent changes for the two groups. The sponsor reports this net difference as -2.836.

Standard errors and confidence intervals

A standard error describes the estimated sampling uncertainty of a statistic. It is not the spread of individual participant scores. For the primary arms, the registry reports standard errors of 1.27 and 1.22. A confidence interval describes a range of values compatible with the data and the specified statistical procedure. The reported 95% interval for the primary mean difference runs from -6.285 to 0.613. Because zero lies inside the interval, the result is compatible with no between-group difference as well as with differences in either direction within the interval’s bounds. A confidence interval is not a guarantee that the true value lies in the interval.

Interpreting p-values

The primary reported p-value is 0.107. A p-value summarizes how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the analysis assumptions were true. It does not give the probability that the null hypothesis is true, nor does it measure the size or clinical importance of the estimate. The p-value needs to be considered alongside the estimate, the confidence interval, the analysis population, and the study’s analysis plan. The registry record does not report the full prespecified statistical plan or whether any interim analyses were planned.

Continuous and binary outcomes

The primary HAM-D17 endpoint and the secondary QIDS-C16 endpoint are continuous outcomes. Their reported mean differences describe differences between group-level percent changes. The secondary responder and remitter outcomes are binary classifications: each participant either meets the recorded outcome definition or does not. For these outcomes, the registry reports odds ratios from generalized linear mixed models. An odds ratio compares the odds of the recorded outcome between groups; it is not itself a difference in percentage points. The reported odds ratio for Week 8 HAM-D17 response is 1.41, with a 95% confidence interval of 1.07 to 1.86. For Week 8 HAM-D17 remission, the reported odds ratio is 1.61, with a 95% confidence interval of 1.14 to 2.27.

For a binary outcome, the unadjusted recalculation also reports a risk difference, which expresses the difference in proportions between groups. The methods are not interchangeable: odds ratios and risk differences use different scales and can give different impressions of magnitude. The sponsor’s reported model-based result is the relevant posted analysis. Unadjusted recalculations from summary values are descriptive and do not reproduce all features of a fitted model.

General note on dispersion

General statistical explanation: over-dispersion occurs when variability in binary or count data exceeds what a model assumes. If it is not addressed appropriately, estimated standard errors and confidence intervals can be inaccurate. The registry record’s summary does not provide enough detail to assess whether over-dispersion was relevant or how it was handled in the reported models.

The recalculated continuous comparisons use an approximate pooled standard deviation derived from arm-level summary information, and are simple unadjusted comparisons. They can aid orientation, but they are not replacements for the sponsor’s mixed-model results. The record reports no discrepancy between the recalculation and the sponsor analysis, and the sponsor’s model remains the primary posted analysis. No claim about a particular model diagnostic, missing-data procedure, or covariate adjustment can be made from the reported details.

11. Confidence Intervals and Precision

The primary estimate is -2.836, and its reported 95% confidence interval extends from -6.285 to 0.613. The interval spans values on both sides of zero. This means the data are compatible with a more negative mean difference as well as with a small positive difference; the estimate alone does not capture this uncertainty. The interval’s upper bound is close to zero, but proximity to zero does not eliminate the uncertainty represented by the full range.

Precision concerns the range of values reasonably compatible with an estimate under the analysis procedure. A narrower interval indicates greater statistical precision, while a wider one indicates more uncertainty. Precision is distinct from clinical importance: even a precisely estimated difference would still require a judgment about whether its size matters, and the registry record does not state a threshold for clinical importance.

The posted primary standard errors, 1.27 for GeneSight Psychotropic Tested and 1.22 for Treatment As Usual, describe uncertainty in the respective least-squares means. They should not be interpreted as the variability of individual participant outcomes. The reported 95% interval and p-value are from the sponsor’s mixed-model analysis; the recalculated interval is an unadjusted approximation from posted summary statistics and is similar in overall interpretation.

Confidence intervals for secondary outcomes also have to be read on their own scales. An odds-ratio interval describes uncertainty about a ratio of odds, whereas a risk-difference interval describes uncertainty about a difference in proportions. Neither type of interval automatically establishes practical importance. The registry record does not provide additional information about the precision of unreported endpoints or the consequences of missing observations.

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 figures are planning illustrations based on the observed effect, not the trial’s original design calculation. For the listed continuous and binary outcomes, figures are participants per arm; time-to-event planning tables, when present, give total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect.

For each effect size shown, the 80% power column gives the planned number of participants per arm under those assumptions for a design with an 80% chance of detecting an effect of that size if it is present. The 90% power column gives the corresponding per-arm number for a 90% chance. These are prospective design illustrations. They are not observed power and do not describe the probability that the trial’s finding is true.

Rows planning for 100%, 75%, and 50% of the observed effect illustrate how the required sample size depends on the effect the design aims to detect. Smaller effects require more participants because they are harder to distinguish from random variation. For the primary outcome, the listed per-arm planning values rise as the target fraction of the observed effect gets smaller. These figures should not be mistaken for the actual number randomized or analyzed: the record reports enrollment of 1398, while the primary arm estimates are based on the per-protocol analysis population.

The registry record does not report the trial’s own original sample-size justification or its assumptions. The planning values therefore help explain what future designs might require under the stated assumptions; they do not establish whether this trial was adequately powered according to its original plan.

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 = -2.840; pooled standard deviation σ = 30.056.
2. n = 2 × (1.96 + 0.842)² × 30.056² / -2.840² = 1,758.1, rounded up to 1,759 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 <- -2.84                      # observed difference in means
sigma <- 30.0557171983803                     # 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(2109939)
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 = 1,759 (80% power), 2,354 (90% power)
 75% of observed effect: n per arm = 3,126 (80% power), 4,185 (90% power)
 50% of observed effect: n per arm = 7,033 (80% power), 9,415 (90% power)
power.t.test: 1759.1 
Simulated power at n = 1,759 per arm: 0.799 (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. Analyses Reported by the Sponsor

The sponsor’s primary analysis is a mixed models analysis reporting a mean difference (net) of -2.836, a 95% confidence interval from -6.285 to 0.613, and p = 0.107. The simple unadjusted recalculation from posted summary statistics gives a mean difference of -2.84, with an approximate 95% confidence interval from -6.292 to 0.612 and an unadjusted p-value of 0.107. The estimates and overall interpretation are close.

The recalculation is not the same procedure as the sponsor’s model. It uses arm-level summary information and does not reproduce the model specification or any adjustments in the sponsor analysis. The registry record reports no discrepancy between the posted result and the recalculation. The sponsor’s own analysis therefore remains the principal result to interpret, with the recalculation serving only as a simple check using the available summaries.

The same distinction applies to the binary secondary outcomes: unadjusted risk differences can describe the posted group percentages, while the sponsor’s reported results use generalized linear mixed models and odds ratios. Those estimates are on different scales and need not match numerically.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, time points, or treatment comparisons. Each additional statistical test creates another opportunity for a small p-value to occur by chance. A prespecified strategy can control the interpretation of a family of tests, for example by adjusting significance thresholds or defining a hierarchy of outcomes. The registry record lists multiple secondary outcomes and reports secondary analyses, but a multiplicity adjustment or testing hierarchy is not reported in the registry record. The nominal p-values for secondary findings therefore should not be interpreted as if they were necessarily adjusted for the number of analyses.

The registry record states that the primary population was per-protocol. That label alone does not explain how participants entered or were excluded from the analysis, how protocol deviations were handled, or how this population relates to all randomized participants. The record reports enrollment of 1398 and gives endpoint-specific arm counts, but a complete participant flow and reasons for differences in analysis counts are not reported in the registry record.

Missing outcome data can affect estimates if the probability of missingness is related to outcomes or differs between groups. Statistical models can use available repeated observations under assumptions, but the specific missing-data method and assumptions cannot be established from the summary. Imputation procedures, sensitivity analyses, and details of participants with unavailable Week 8 scores are not reported in the registry record.

These omissions limit independent assessment of whether the primary analysis population, missing-data handling, and multiplicity strategy were appropriate. The posted estimates remain interpretable as reported, but the record does not contain the detail needed to reconstruct the complete statistical plan or evaluate all assumptions. Interim analyses and any associated decision rules are also not reported in the registry record.

16. 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

Serious events with the highest arm-level rate: pulmonary embolism (0.1%), muscle twitching (0.1%), transient hypoglycemia (0.1%), gastrointestinal hemorrhage (0.1%), non cardiac chest pains (0.1%).

The serious adverse event summary reports 3 events among 681 participants at risk in GeneSight Psychotropic Tested and 2 among 717 at risk in Treatment As Usual. No other adverse events are listed in either group in this summary, and deaths are reported as 0 in both groups. These are counts within the registry’s stated at-risk groups; they are not the same denominators as the primary per-protocol analysis.

The listed serious-event terms each have a reported frequency of 0.1%: pulmonary embolism, muscle twitching, transient hypoglycemia, gastrointestinal hemorrhage, and non cardiac chest pains. These terms are presented as the top serious-event terms; the registry summary does not establish that any event was caused by the testing strategy. Event counts are small, and the record does not provide enough information to assess event timing, relatedness, severity details, or differences in exposure time.

Safety counts should therefore be read descriptively. The presence or absence of a listed event in this summary does not by itself establish comparative safety, and zero deaths in the recorded groups is not evidence that all possible harms were absent.

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

The primary result estimates the difference between groups in mean percent change in HAM-D17 score from baseline to the end of Week 8, among the per-protocol population. The sponsor’s estimate is negative, but its 95% confidence interval includes zero and the reported p-value is 0.107. The posted data therefore do not provide a statistically clear primary between-group difference under the reported analysis. They also do not prove that the groups were equivalent or exclude every potentially meaningful difference.

The result is specific to this trial’s population, intervention strategy, comparator, outcome definition, analysis population, and measurement time. It should not be interpreted as a direct estimate for other populations or settings without additional evidence. The study’s open-label follow-up is described in the official title, but the primary endpoint is at Week 8; the primary estimate does not summarize all follow-up outcomes.

Secondary analyses include nominally favorable findings for some Week 8 binary HAM-D17 outcomes, alongside a continuous QIDS-C16 comparison whose interval includes zero. Because these are distinct outcomes and multiplicity handling is not reported, they should not be combined into a claim that the primary result has been confirmed. The available summaries also do not establish individual-level benefit or harm, nor do they show how participants’ treatment decisions were affected.

Statistical evidence is one part of interpreting a trial. The record supports a careful description of the posted estimates and their uncertainty, but does not report the clinical-importance threshold, the full statistical plan, or enough missing-data detail to make broader claims. No treatment recommendation follows from these statistics alone.

18. Why This Trial Matters Statistically

This trial raises a statistical question about evaluating a testing strategy rather than a drug against placebo: the outcome is whether assignment to GeneSight Psychotropic testing is associated with different symptom-score change than treatment as usual. That design question highlights the importance of defining the intervention, comparator, population, and time point precisely. The registry record describes a randomized parallel-group comparison and identifies the Week 8 HAM-D17 percent change as the primary endpoint.

The results also illustrate why an estimate should be read together with its confidence interval. The primary group difference points in one direction, but the interval includes zero and the p-value is 0.107. Reporting only the point estimate would conceal uncertainty; reporting only whether a p-value passes a threshold would conceal the range of compatible effects.

The distinction between the primary continuous endpoint and secondary binary outcomes matters as well. Mean score change, response, and remission encode different information. A study can produce differing patterns across these measures without those results being contradictory. When many outcomes are assessed, multiplicity and prespecification determine how much weight isolated secondary findings deserve.

Finally, the record underscores the value of transparent reporting. The population label, model name, estimates, confidence intervals, and safety counts are available, while the full model details, missing-data approach, multiplicity plan, and original sample-size rationale are not reported in the registry record. Those omissions constrain reproducibility and the confidence with which readers can assess the analysis beyond its posted summary.

19. 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 registry summary does not report several details needed for a complete assessment of this trial. The primary analysis is identified as per-protocol, but participant flow, protocol-deviation rules, and reasons for exclusion from that population are not reported in the registry record. This limits interpretation of how the analysis population relates to the randomized groups.

The primary analysis is labeled as a mixed models analysis, but the model specification, included covariates, covariance structure, and missing-data assumptions are not reported in the registry record. The record also does not report whether the analysis was adjusted for baseline factors, how incomplete Week 8 data were handled, or whether sensitivity analyses were conducted. The confidence interval is therefore interpretable as posted, but the underlying analysis cannot be fully reconstructed from the registry summary.

The registry record does not report a multiplicity plan despite multiple secondary outcomes, nor does it state the original sample-size justification or any planned interim analyses. The responder definition is not reported in the registry record recorded for the secondary endpoint. These gaps make it difficult to determine the confirmatory status of secondary findings or assess whether the study was designed around the observed primary effect.

The primary result is measured at Week 8, whereas the official title also describes an open-label follow-up period. The primary estimate alone does not represent that later period. The safety summary provides event counts and selected terms but not event relatedness, timing, or detailed exposure context. These are limitations of the available registry record and do not, by themselves, establish that the trial’s conduct or analyses were inadequate.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Genomics Used to Improve DEpression Decisions — NCT02109939.
• Registry record last updated 2020-01-14; status last verified 2020-01.
• Results first posted 2020-01-14.
• 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.

23. Record Summary

NCT02109939 is a completed randomized, triple-masked, parallel-group trial sponsored by Assurex Health Inc. It enrolled 1398 outpatients with major depressive disorder and an inadequate response to at least one medication in the current episode. The intervention was GeneSight Psychotropic testing, compared with treatment as usual.

The primary outcome was per-protocol mean percent change in HAM-D17 from baseline to the end of Week 8. The sponsor-reported mean difference was -2.836, with a 95% confidence interval from -6.285 to 0.613 and p = 0.107. The interval includes zero, indicating uncertainty about the direction and size of the between-group difference.

Some posted secondary Week 8 binary outcomes have sponsor-reported odds-ratio intervals above one, while the continuous QIDS-C16 comparison has an interval including zero. Multiplicity handling and important statistical-plan details are not reported in the registry record, so secondary findings require caution. The serious adverse event summary reports 3 events in the tested group and 2 in treatment as usual, with no deaths reported in either group.