Study to Evaluate TV-46000 as Maintenance Treatment in Adult and Adolescent Participants With Schizophrenia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study to Evaluate TV-46000 as Maintenance Treatment in Adult and Adolescent Participants With Schizophrenia (NCT03503318), a randomized phase 3 study. The primary outcome is Time to Impending Relapse (Number of Participants With Impending Relapse) for Intent-to-treat [ITT] Analysis Set.

2. Clinical Question

The clinical question is about maintenance treatment and relapse-related events, not simply symptom change at a single visit. The registry describes a randomized, quadruple-masked, parallel-group study and identifies efficacy, safety, and tolerability as its aims. Relapse included specified symptom worsening, hospitalization for worsening psychotic symptoms, severe suicidal status or worsening, or violent behavior resulting in clinically significant harm or property damage. These criteria make the endpoint a composite: an event could arise through more than one clinical route. The registry record does not report the exact criteria used to determine clinical stability or eligibility beyond the description in the study summary.

3. Trial Design

This was a phase 3, randomized, quadruple-masked, parallel-group, placebo-controlled study of risperidone extended-release injectable suspension TV-46000. The registry lists an enrollment of 544 and records the study as completed. The start date was 2018-04-27, and the primary completion date was 2020-09-30. Teva Branded Pharmaceutical Products R&D, Inc. is listed as the sponsor.

The registry describes three treatment groups: placebo, TV-46000 q1m, and TV-46000 q2m. The q1m group received a subcutaneous injection at baseline and every 4 weeks thereafter. The q2m group received TV-46000 at baseline and every 8 weeks thereafter, with placebo injections at the intervening schedule described in the arm record. Placebo recipients received matching injections at baseline and every 4 weeks thereafter. The arm descriptions include maximum doses described as comparable to oral risperidone doses, but the registry record excerpt does not contain the complete descriptions of treatment continuation or all dose details.

The primary outcome was assessed from randomization up to 108 weeks. Its stated analysis population was the intent-to-treat set of adults randomized to double-blind maintenance treatment, regardless of whether they received treatment. The posted results include a separate extended intent-to-treat result for time to impending relapse. Secondary outcomes also include week 24 impending relapse, stability and remission at endpoint, an adolescent time-to-relapse outcome, and changes in drug attitudes and schizophrenia quality of life. The registry record does not report the study's original sample-size justification, detailed randomization procedures, or the full statistical analysis plan.

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.

Before assignment: Participants were randomized to receive TV-46000 once monthly (q1m) subcutaneous (SC) injections, TV-46000 once every 2 months (q2m) SC injections, or placebo q1m SC injections in a 1:1:1 ratio. Open-label oral risperidone (2 to 5 mg/day) was used to stabilize participants to the treatments (the dose was based on clinical judgment) before randomization.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat (ITT) analysis set included adult participants randomized to the double-blind maintenance treatment, regardless if they had received treatment or not.

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: Extended intent-to-treat (eITT) analysis set included all participants (adults and adolescents) randomized to the double-blind maintenance stage treatment, regardless if they had received treatment or not.

6. Primary Endpoint

The registry lists the primary outcome as Time to Impending Relapse (Number of Participants With Impending Relapse) for Intent-to-treat [ITT] Analysis Set, measured over From randomization up to 108 weeks.

The primary endpoint counts participants with impending relapse and also frames the outcome as time to that event. The assessment window was from randomization up to 108 weeks. Relapse was defined using one or more specified events: a Clinical Global Impression-Improvement (CGI-I) score of at least 5 together with specified worsening on Positive and Negative Syndrome Scale (PANSS) items; hospitalization due to worsening psychotic symptoms; specified severe suicidal status or worsening on the Clinical Global Impression-Severity of Suicidality (CGI-SS); or violent behavior resulting in clinically significant self-injury, injury to another person, or property damage.

This endpoint is used to assess whether participants remain free of the defined relapse event over follow-up. Counting participants with an event gives a direct description of event occurrence, while time-to-event analysis also uses information about when events occurred. General statistical point: time-to-event methods can account for follow-up duration and censoring, when those are handled according to a suitable analysis plan. The registry record does not report the censoring rules or detailed time-to-event model assumptions. The composite definition also matters to interpretation: the count combines events with different clinical features, and the posted total does not show how much each criterion contributed to the overall result.

7. Results Posted in the Registry

• TV-46000 q1m: 7.1% (13/183) versus 29.3% (53/181) on control; risk difference -22.2 percentage points (95% CI -29.8 to -14.6), risk ratio 0.24 (0.14–0.43); unadjusted two-proportion z-test p = <0.001.
• TV-46000 q2m: 12.8% (23/179) versus 29.3% (53/181) on control; risk difference -16.4 percentage points (95% CI -24.7 to -8.2), risk ratio 0.44 (0.28–0.68); unadjusted two-proportion z-test p = <0.001.

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.

Time to Impending Relapse (Number of Participants With Impending Relapse) for Extended ITT [eITT] Analysis Set

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • TV-46000 q1m: 7.1% (13/183) versus 29.3% (53/181) on control; risk difference -22.2 percentage points (95% CI -29.8 to -14.6), risk ratio 0.24 (0.14–0.43); unadjusted two-proportion z-test p = <0.001.
  • TV-46000 q2m: 13.3% (24/180) versus 29.3% (53/181) on control; risk difference -15.9 percentage points (95% CI -24.2 to -7.7), risk ratio 0.46 (0.29–0.70); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for TV-46000 q1m than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for TV-46000 q2m than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Proportion of Participants With Impending Relapse

• 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:
  • TV-46000 q1m: 0.0% (0/183) versus 0.6% (1/181) on control; risk difference -0.6 percentage points (95% CI -1.6 to 0.5); unadjusted two-proportion z-test p = 0.314.
  • TV-46000 q2m: 0.0% (0/179) versus 0.6% (1/181) on control; risk difference -0.6 percentage points (95% CI -1.6 to 0.5); unadjusted two-proportion z-test p = 0.319.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between TV-46000 q1m and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between TV-46000 q2m and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants Who Maintain Stability at the Endpoint

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • TV-46000 q1m: 86.9% (159/183) versus 60.8% (110/181) on control; risk difference 26.1 percentage points (95% CI 17.5 to 34.7), risk ratio 1.43 (1.26–1.63); unadjusted two-proportion z-test p = <0.001.
  • TV-46000 q2m: 79.9% (143/179) versus 60.8% (110/181) on control; risk difference 19.1 percentage points (95% CI 9.9 to 28.3), risk ratio 1.31 (1.14–1.51); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for TV-46000 q1m than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for TV-46000 q2m than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover several related dimensions: impending relapse in an extended intent-to-treat analysis, a week 24 relapse proportion, stability and remission at endpoint, adolescent time to impending relapse, and changes in drug attitudes and schizophrenia quality of life. The posted results available in the record include selected outcomes, while the registry lists 17 outcome measures and 16 posted secondary outcomes. Taken together, the secondary material broadens the description beyond the primary adult intent-to-treat time-to-event endpoint.

The overall pattern in the posted secondary results is generally aligned with the primary direction for the extended intent-to-treat relapse count and endpoint stability: the arm comparisons are not all measuring the same outcome or time frame, but they provide additional context about relapse-related events and maintenance status. The week 24 impending-relapse comparisons are a separate, fixed-time measure, and their recalculated intervals include values on both sides of no difference. The outcome titles also show that some secondary measures concern adolescents or patient-reported scales, which should not be treated as equivalent to the primary adult endpoint.

These secondary findings do not independently establish the primary result, and they should be interpreted with caution. The registry record does not report whether secondary comparisons were adjusted for multiplicity, how missing observations were handled, or whether a hierarchy of secondary hypotheses was specified. General statistical point: testing several outcomes increases the opportunity for chance findings unless the analysis plan accounts for multiplicity. Unadjusted secondary results therefore carry less inferential weight than a prespecified primary analysis. The secondary outcomes are best read as supportive descriptive evidence, with attention to their distinct populations, definitions, and assessment times rather than as a single combined confirmation.

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

• Time to Impending Relapse (Number of Participants With Impending Relapse) for Extended ITT [eITT] Analysis Set
• Proportion of Participants With Impending Relapse
• Number of Participants Who Maintain Stability at the Endpoint
• Number of Participants Achieving Remission at the Endpoint
• Observed Rate of Impending Relapse (Number of Participants With Impending Relapse) at the Endpoint
• Time to Impending Relapse (Number of Participants With Impending Relapse) in the Adolescent Participants
• Change From Baseline in Drug Attitudes Inventory 10-item Version (DAI-10) Total Score at the Endpoint and End of Treatment
• Change From Baseline in Schizophrenia Quality of Life Scale (SQLS) Total Score at the Endpoint and End of Treatment

9. Statistical Interpretation

10. Statistical Methods Explained

The results combine event counts, a sponsor-reported time-to-event analysis, and unadjusted calculations based on posted summary counts. Each method describes a different aspect of the data. Understanding those distinctions helps prevent a count, a risk difference, a hazard ratio, and a p value from being treated as interchangeable.

What does a participant count show?

A count records how many participants met the endpoint definition during the stated assessment period. The primary counts are 53.0 for placebo, 13.0 for q1m, and 23.0 for q2m. A count alone does not communicate the time at which events occurred or the duration of observation for each participant. The comparison of counts is therefore most informative when read alongside the group denominators and the time-to-event analysis. It also depends on the event definition: this endpoint combines several clinical criteria rather than representing one single type of event.

What is a risk difference?

A risk difference subtracts one group's event proportion from the comparison group's event proportion, using the direction stated for the comparison. The recalculated risk differences are -0.222 for q1m versus placebo and -0.164 for q2m versus placebo. Their negative signs indicate that the event proportion was lower in the TV-46000 group in these posted data. The confidence intervals quantify uncertainty around these estimates. These recalculations are unadjusted comparisons derived from the posted counts and denominators; they do not reproduce the sponsor's full analysis plan or account for any adjustments not available in the registry record.

General statistical point: a risk difference is an absolute contrast on the proportion scale. It can be more directly interpretable than a relative measure, but it is tied to the follow-up period and population used in its calculation. When follow-up varies or censoring is relevant, a simple event proportion may not use all the timing information in the same way as a time-to-event analysis.

What does a hazard ratio describe?

The sponsor reports hazard ratios from log-rank analyses: 0.200 for q1m versus placebo and 0.375 for q2m versus placebo. A hazard ratio compares the rate of experiencing the event over time between groups, within the time-to-event framework. It is not the same as a ratio of final event proportions, nor does it tell how much earlier or later any individual event occurred. The sponsor's reported confidence intervals and p values belong to those analyses. The registry record does not report the detailed model specification or assumptions used to produce the hazard ratios.

General statistical point: time-to-event methods are useful when event timing matters and when some participants may not have an event before observation ends. The interpretation depends on how censoring and follow-up are handled. Those details are not reported in the registry record, so the posted hazard ratios should be understood as the sponsor's reported analyses rather than independently reconstructed estimates.

How should confidence intervals and p values be read?

A confidence interval gives a range of values compatible with the observed data under the method used, with the stated confidence level. It reflects statistical uncertainty in an estimate; it does not show the range of individual responses. A p value describes how unusual data at least as incompatible with a null hypothesis would be under that hypothesis and the analysis assumptions. It is not the probability that the null hypothesis is true, and it does not measure clinical importance. The sponsor reports p < 0.001 for both primary comparisons, while the unadjusted recalculations report p < 0.001 for both.

The record includes multiple outcomes and comparisons. A p value from an unadjusted calculation does not account for the fact that several hypotheses may be examined. General statistical point: when many comparisons are made, the chance of at least one small p value under a collection of null hypotheses can increase. Whether and how this trial addressed that issue is not reported in the registry record. Consequently, the primary result and the secondary results should not be assigned the same inferential weight solely because their p values are small.

11. Confidence Intervals and Precision

The primary sponsor-reported time-to-event estimates have confidence intervals that exclude 1: 0.109 to 0.367 for q1m versus placebo, and 0.227 to 0.618 for q2m versus placebo. The corresponding reported p values are p < 0.001. These intervals convey uncertainty around the hazard-ratio estimates under the sponsor's analysis. They do not capture every possible source of uncertainty, such as limitations in endpoint definition, missing-data handling, or generalizability.

The unadjusted recalculated risk-difference intervals are -0.298 to -0.146 for q1m versus placebo and -0.247 to -0.082 for q2m versus placebo. They describe uncertainty around differences in event proportions calculated from the posted counts. The registry also posts 95% confidence intervals for the week 24 proportion outcome: 0.205 to 0.347 for placebo, 0.03 to 0.109 for q1m, and 0.065 to 0.165 for q2m. Those are group-specific intervals for that secondary measure, not intervals for the primary hazard ratios.

General statistical point: a narrower interval indicates greater precision for the estimate, while a wider interval indicates more uncertainty. Precision is influenced by event frequency, sample size, follow-up, and the analysis method. Confidence intervals do not mean that the true value has a particular probability of lying inside the calculated interval. Their formal interpretation depends on repeated use of the same method under its assumptions. The intervals also do not by themselves establish clinical relevance; that judgment requires a meaningful context for the effect and the endpoint.

The posted estimates are rounded, so the displayed endpoints of an interval need not reproduce every calculation exactly if used as inputs. The record does not report confidence intervals for every secondary outcome, nor does it provide a full account of uncertainty from missingness or other analysis choices. The strength of a conclusion should therefore reflect both the numerical precision and the limits of the registry detail.

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 planning figures are illustrations based on observed effects, not the trial's original design assumptions. For the binary participant-count comparisons, the figures are participants per arm. The 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. The trial's own sample-size justification is not reported in the registry record.

For example, planning for 100% of the observed q1m primary effect gives 47 participants per arm for 80% power and 62 participants per arm for 90% power. Planning for 75% of the observed effect gives 93 participants per arm for 80% power and 124 participants per arm for 90% power. For q2m, the corresponding figures are 96 and 128 participants per arm when planning for 100% of the observed effect. These are illustrations, not recommendations or evidence about the trial's actual design.

In plain terms, the 80% power column gives the estimated number of participants per arm for a design with an 80% chance of detecting the specified effect under the listed assumptions, if that effect is real. The 90% column describes the corresponding target for a 90% chance. Smaller effects need more participants to detect under otherwise unchanged assumptions. The curve should be read as a visual expression of that general relationship between the effect size being targeted and the required sample size; no additional assumptions beyond those stated should be inferred from it. These figures are not observed power.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.293; treatment proportion p₁ = 0.071; average p̄ = 0.182.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5456 = 1.0693.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5226 = 0.4398.
4. n = (1.0693 + 0.4398)² / (-0.2218)² = 46.3, rounded up to 47 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
p0 <- 0.292817679558011                          # control proportion
p1 <- 0.0710382513661202                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(3503318)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 47 (80% power), 62 (90% power)
 75% of observed effect: n per arm = 93 (80% power), 124 (90% power)
 50% of observed effect: n per arm = 230 (80% power), 308 (90% power)
power.prop.test: 46.3 
Simulated power at n = 47 per arm: 0.836 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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 registry reports sponsor analyses using the log-rank method for the primary time-to-impending-relapse comparisons. The hazard ratio is 0.200 for TV-46000 q1m versus placebo, with a 95% confidence interval from 0.109 to 0.367 and p < 0.001. For TV-46000 q2m versus placebo, the hazard ratio is 0.375, with a 95% confidence interval from 0.227 to 0.618 and p < 0.001.

The recalculated results are simple unadjusted comparisons of the posted event counts and denominators, expressed as risk differences. They are not a reanalysis of individual participant times or a reproduction of the sponsor's log-rank method. The two approaches use different statistical summaries, so their estimates are not expected to be numerically identical. The direction of the recalculated risk differences is consistent with the sponsor-reported comparisons: both are negative when expressed as the TV-46000 group minus placebo event proportion.

There is no flagged discrepancy between the posted counts and the sponsor analyses in the registry record. For interpretation of the primary endpoint, the sponsor's own time-to-event analysis takes precedence over the unadjusted recalculation, because it is the reported analysis for that endpoint. The registry record does not provide enough detail to reproduce that analysis independently or assess every modeling and censoring choice.

15. Multiplicity, Missing Data and Analysis Populations

The registry record identifies the primary outcome and reports multiple secondary outcomes and comparisons. It does not report a multiplicity-control procedure, a testing hierarchy, or a complete statistical analysis plan. Without that information, the unadjusted p values for comparisons should not be read as if they necessarily preserve a trial-wide error rate across all tested hypotheses. This is particularly relevant to secondary outcomes, where several domains and assessment times are listed.

The registry describes the primary intent-to-treat population as adults randomized to double-blind maintenance treatment, regardless of whether they received treatment. A separate extended intent-to-treat analysis set is named for a secondary relapse outcome, but its detailed membership rules are not reported in the registry record. The primary population is therefore not the same as a description of all enrolled participants, and the adult primary analysis should not be assumed to include adolescent participants.

Missing-data procedures are not reported in the registry record. The record also does not describe how participants who stopped treatment, withdrew, or had incomplete follow-up contributed to the primary time-to-event analysis, nor does it specify imputation or sensitivity analyses. General statistical point: missingness can affect estimates when the probability of an observed outcome is related to participant characteristics or outcomes. The impact depends on the amount, reasons, and handling of missing data, details that cannot be assessed from the posted summary alone.

The registry does not report interim analyses, stopping rules, or the original sample-size justification. These omissions do not establish that such procedures were absent; they mean the registry record does not report them. Consequently, the available record supports interpretation of the posted endpoint estimates, but not a complete audit of the trial's prespecification, multiplicity strategy, or missing-data robustness.

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

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: schizophrenia (2.2%), pelvic abscess (1.4%), psychotic disorder (1.1%), psychotic symptom (1.1%), suicidal ideation (1.1%).

The registry lists serious adverse event figures for 179 participants at risk in placebo, 183 in TV-46000 q1m, and 180 in TV-46000 q2m. The corresponding numbers affected are 14, 8, and 10. Deaths recorded in the groups are 1, 0, and 4, respectively. These counts describe events recorded in the registry and are not, on their own, an adjusted comparison of safety between groups.

Other adverse events are listed for 179 participants at risk in placebo, 183 in q1m, and 180 in q2m; the corresponding affected counts are 35, 49, and 48. The registry lists serious-event terms and percentages including Schizophrenia at 2.2%, Pelvic abscess at 1.4%, Psychotic disorder at 1.1%, Psychotic symptom at 1.1%, Suicidal ideation at 1.1%, and Myocardial infarction at 0.6%. The excerpt does not clearly assign those listed terms to specific treatment groups.

These figures should be read with their stated at-risk groups and the study context. They do not provide event timing, duration, severity detail beyond the serious-event classification, or causal attribution. Comparative safety interpretation is limited because the registry record does not report those details or a complete safety analysis plan.

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

The posted primary results show fewer recorded impending-relapse events in each TV-46000 group than in the placebo group during the stated follow-up, and the sponsor's time-to-event analyses report hazard ratios below 1 for both comparisons. The unadjusted risk-difference recalculations point in the same direction. Together, these statistics describe an observed between-group difference in this randomized trial under its endpoint definition and analysis populations.

They do not establish that every participant experienced the same outcome, nor do they predict an individual participant's course. The endpoint combines multiple event types, and the aggregate counts do not identify which component accounted for each event. A hazard ratio is a relative time-to-event measure, not a direct estimate of an individual's probability of relapse. Confidence intervals describe uncertainty around group estimates rather than the distribution of individual outcomes.

The findings also do not answer every question about treatment. The registry record does not report the complete missing-data approach, full analysis plan, multiplicity handling, or detailed assumptions for the sponsor's time-to-event analysis. The primary population is defined as adults in the intent-to-treat set; results should not automatically be generalized to adolescents or populations outside the study eligibility criteria. The record lists additional safety events, but those counts alone do not establish causation or provide a comprehensive benefit-risk assessment.

General statistical point: statistical evidence and clinical importance are related but distinct. A small p value does not quantify the importance of an effect, and a confidence interval does not resolve whether an effect is meaningful to patients. The registry results can be interpreted as group-level evidence about the defined outcomes and follow-up, with conclusions bounded by the populations, endpoints, and analysis details actually reported.

18. Why This Trial Matters Statistically

This trial illustrates why the choice of endpoint and analysis matters in clinical research. Impending relapse was defined as a composite of symptom worsening, hospitalization, suicidality criteria, and violent behavior with clinically significant consequences. That definition makes the outcome relevant to maintenance follow-up, while also meaning that the overall event count can combine clinically different experiences.

It also provides a useful contrast between time-to-event analysis and a straightforward comparison of event proportions. The sponsor's log-rank analyses incorporate the timing framework and report hazard ratios, while recalculation from the registry counts gives risk differences. The estimates are related in direction but describe different statistical quantities. Readers benefit from distinguishing the sponsor's reported analysis from unadjusted summary-statistic calculations rather than treating them as interchangeable.

The registry includes adult primary results and secondary outcomes spanning extended intent-to-treat relapse, endpoint stability and remission, adolescent outcomes, and patient-reported measures. This breadth shows why multiplicity and population definitions matter: a collection of outcomes can answer complementary questions, but not every result has the same inferential status. The available record does not report the full prespecified analysis strategy or missing-data procedures, limiting assessment of how the set of findings was managed.

Finally, the safety counts underscore that efficacy and safety summaries require separate interpretation. Event counts need their denominators, classification, follow-up context, and group attribution. The registry record is useful for understanding the reported trial outcomes, while its omissions set clear boundaries on what can be concluded from summary results alone.

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.

Several limitations specific to this record constrain interpretation. First, the primary outcome is a composite that can be met through different clinical events. The registry provides the definition but not a component-by-component breakdown of the primary events, so the overall count cannot show which criteria drove the group differences.

Second, the primary analysis population is adult intent-to-treat participants, although the study summary describes adults and adolescents as eligible and lists adolescent outcomes among secondary measures. The primary result should not be taken as a direct estimate for adolescents. The detailed eligibility criteria and the definition of clinical stability are not reported in the registry record beyond the brief summary.

Third, the registry gives a timeframe up to 108 weeks, but the excerpt does not describe follow-up completeness, censoring rules, withdrawals, or missing-data methods. These details are especially relevant to interpreting time-to-event results. The sponsor's log-rank hazard ratios are reported, but the full model specification and assumptions are not reported in the registry record.

Fourth, multiple outcomes and comparisons are listed without a reported multiplicity strategy or complete analysis hierarchy. Secondary results may therefore be less conclusive than the primary result. The registry also does not report the original sample-size justification, interim analysis procedures, or detailed protocol deviations. Finally, the serious-event summary does not provide enough information to determine event timing, causation, or group attribution for the listed serious-event terms. These are limitations of the registry information available, not proof that the underlying study did not address these issues.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study to Evaluate TV-46000 as Maintenance Treatment in Adult and Adolescent Participants With Schizophrenia — NCT03503318.
• Registry record last updated 2023-03-10; status last verified 2023-02.
• Results first posted 2021-12-09.
• 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

NCT03503318 is a completed phase 3 randomized, quadruple-masked, parallel-group, placebo-controlled study sponsored by Teva Branded Pharmaceutical Products R&D, Inc. The study evaluated subcutaneous TV-46000 on monthly and every-8-weeks schedules as maintenance treatment for schizophrenia. Enrollment was 544; the registry lists a start date of 2018-04-27 and a primary completion date of 2020-09-30.

The primary endpoint counted adult intent-to-treat participants with impending relapse from randomization up to 108 weeks. The sponsor reported log-rank hazard ratios of 0.200 for q1m versus placebo and 0.375 for q2m versus placebo, with p < 0.001 for each comparison. Unadjusted risk-difference recalculations from posted counts were negative for both comparisons. The registry also lists secondary relapse, stability, remission, adolescent, and patient-reported outcomes, as well as adverse-event counts.

Interpretation is limited by registry entries not reported on missing-data handling, multiplicity, the full analysis plan, and time-to-event assumptions. The posted results describe group-level outcomes under the registry's endpoint definitions and should be read within those limits.