A Study of ALKS 3831 in Subjects With Schizophrenia and Alcohol Use Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of ALKS 3831 in Subjects With Schizophrenia and Alcohol Use Disorder (NCT02161718), a randomized phase 2 study. The primary outcome is Number of Participants With Independent Adjudication Committee (IAC) Adjudicated Event of Exacerbation of Disease (EEDS).

2. Clinical Question

The clinical question is whether the two assigned regimens differed in the occurrence of events meeting the trial’s definition of disease exacerbation among the efficacy-analysis population. That definition included several kinds of events, ranging from hospitalization or specified symptom-score changes to rescue medication, emergency visits, and discontinuation for stated reasons. The endpoint therefore represents a composite clinical and treatment-course outcome, rather than a measure of one symptom alone.

The trial also listed secondary outcomes concerning exacerbation events and change in World Health Organization drinking-risk level. The available registry information does not give a single comparative result for those outcomes either. No conclusion about the direction or size of treatment differences follows from the trial question alone.

3. Trial Design

The registry describes a randomized, parallel-group Phase 2 study of ALKS 3831 in people with schizophrenia and alcohol use disorder. Enrollment was 300. The named randomized arms were samidorphan plus olanzapine (ALKS 3831) and placebo plus olanzapine. The brief summary calls the study double-blind, while the registry’s masking classification is quadruple. The record does not explain how those descriptions relate to one another.

The study status is completed. The recorded start was 2014-05, and primary completion was 2017-01-11. The listed primary endpoint covered up to 15 months and assessed participants with an Independent Adjudication Committee-adjudicated event of exacerbation of disease. The primary endpoint description places events within the double-blind period.

The primary efficacy population was defined as randomized participants who received at least one dose of study drug after randomization. This definition is important: it is not simply every person randomized, because randomized participants who did not receive a post-randomization dose are not included in the stated efficacy population.

The adverse-event record also lists an Open- Label ALKS 3831 group, separately from the two named randomized arms. Its presence should not be treated as evidence that it was part of the primary randomized comparison. The registry record does not provide enough detail to describe the timing or relationship of that group to the randomized phase.

Further design details—including the exact allocation ratio, eligibility criteria, treatment duration by participant, randomization procedure, and protocol-level analysis plan—are not reported in the registry record. The enrollment figure alone does not establish the planned allocation or the number contributing to each efficacy analysis.

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: After screening, subjects began a 4-week open-label olanzapine dosing period followed by a 2- week open-label ALKS 3831 period. Subjects who did not tolerate ALKS 3831 during the second open-label period were discontinued from the study prior to randomization. A total of 300 subjects were enrolled in the study, and 255 completed the open-label olanzapine period. 234 subjects were randomized following the open-label ALKS 3831 period, however, 5 subjects discontinued prior to receiving treatment.

Open-Label ALKS 3831 Period

Reasons for not completing this period:

Post- Randomization Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Efficacy analyses were carried out using the intent to treat (ITT) population, i.e. all randomized subjects who received at least one dose of study drug post-randomization

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 Number of Participants With Independent Adjudication Committee (IAC) Adjudicated Event of Exacerbation of Disease (EEDS), measured over Up to 15 months.

The primary endpoint was the number of participants with an Independent Adjudication Committee-adjudicated event of exacerbation of disease, abbreviated EEDS in the registry. It is a participant count: a participant meeting at least one qualifying event criterion contributes to the outcome. The endpoint was assessed during the double-blind period, with the recorded timeframe stated as up to 15 months.

The definition combined several types of events considered related to worsening of disease symptoms and confirmed by the committee. These included hospitalization, specified increases in PANSS total score, qualifying PANSS item scores, self-injury or aggressive behavior, suicidal or homicidal ideation, initiation of rescue medication, an emergency-room visit, specified forms of withdrawal or discontinuation, and arrest or incarceration related to the underlying disease. The criteria also included stated numerical thresholds for score changes and item scores.

Composite endpoints can capture different ways a clinically important worsening may present, rather than relying on a single symptom or event type. That breadth can be useful when events are uncommon or varied, but it also means that the total count does not by itself reveal which component contributed most. Component-specific results are not reported in the registry record.

The endpoint’s committee adjudication indicates that event classification involved review against the stated definition. It does not, by itself, tell readers how consistently events were identified, how many were judged eligible, or whether event adjudication was blinded; those details are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Several statistical ideas help clarify what the registry information can and cannot establish. The primary outcome is a count of participants meeting an event definition, not a continuous score or a time-to-event measure as described in the record. A comparison of such outcomes would ordinarily require an explicit contrast between the randomized groups and an estimate of uncertainty. The registry record does not report one single primary comparison, so no such estimate can be interpreted.

What does an effect estimate represent?

In general, an effect estimate summarizes the contrast between groups for a chosen outcome and estimand. For a participant-count outcome, the analysis might report group-specific counts and a measure of how those counts compare. The meaning of that estimate depends on the outcome definition, analysis population, handling of missing observations, and statistical model. Those details cannot be reconstructed from the endpoint title alone. The registry record does not report an effect estimate for the primary contrast.

A confidence interval is a range of values compatible with the data under a specified statistical procedure. In general, a narrower interval indicates greater precision, while a wider interval indicates more uncertainty. It does not show the probability that a particular treatment effect is true. Because no primary effect estimate or confidence interval is reported in the registry record, the precision of the randomized primary comparison cannot be assessed.

Why does the composite definition matter?

The primary endpoint combines multiple qualifying events into a participant-level outcome. This can capture a broad range of worsening-related events, but it may also combine events that differ in their nature and frequency. An overall result would need to be read alongside the component definitions and, where available, component-specific findings. Without component counts, the contribution of each criterion to any total is unknown. The adjudication requirement is part of the endpoint definition, but the record does not describe adjudication procedures in further detail.

How should the analysis population be read?

The registry says efficacy analyses used an intent-to-treat population defined as all randomized participants who received at least one dose of study drug post-randomization. In general, intent-to-treat approaches seek to preserve the benefits of random assignment by analyzing participants according to their assigned groups. The particular definition in this record has a post-randomization dosing condition, so it does not include randomized participants who received no post-randomization dose. The number excluded for that reason is not reported in the registry record.

For a binary participant-level event outcome, complete and consistent follow-up is important. Missing outcome information can affect group comparisons, particularly if the reasons for missingness differ between groups or relate to the outcome. General approaches include explicit assumptions and sensitivity analyses, but no missing-data method or sensitivity analysis is reported in the registry record. It would therefore be inappropriate to presume that missing outcomes were absent or handled in a particular way.

What can planning illustrations say?

General sample-size calculations estimate how many participants may be needed to detect a chosen effect under specified assumptions. The registry record describes planning illustrations based on a fraction of the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. These are planning illustrations, not observed power and not evidence that the trial was designed using those assumptions. The trial’s own sample-size justification is not reported in the registry record.

More generally, a planning calculation depends on the assumed effect size, event frequency or outcome variability, statistical test, allocation, and desired power. Changing an assumption can change the calculated sample size. A calculation based on an observed effect is particularly dependent on that effect estimate and should not be read as a retrospective probability that the trial would detect a true effect. No numerical planning result is stated in the registry information described for this analysis, so no sample-size figure can be interpreted.

Why do multiplicity and model details matter?

When a trial examines multiple outcomes, time points, or analyses, the chance of at least one apparently notable result can increase if many tests are interpreted without adjustment. General multiplicity strategies include prespecifying a hierarchy or adjusting inference across a defined family of tests. The registry record lists three outcome measures and three analyses, but a multiplicity procedure, testing hierarchy, and confirmatory statistical plan are not reported. The primary endpoint itself is described as having several categories or time points and is not reduced to one comparison in the posted result.

Model choice also affects interpretation. A binary outcome can be analyzed in different ways, and different models may target different effect measures or make different assumptions. The registry record does not specify the model used for the primary endpoint or provide diagnostics. It is therefore not possible to characterize model adequacy or reproduce the sponsor’s analysis from the available information.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists three outcome measures and three analyses. That indicates more than one outcome or analysis was recorded, but it does not reveal how the statistical testing was organized. A prespecified hierarchy, family of hypotheses, multiplicity adjustment, or division between confirmatory and exploratory analyses is not reported in the registry record.

In general, multiplicity matters because considering several outcomes, time points, or comparisons can increase the chance of a nominally unusual result by chance alone. A testing strategy can control that issue, but no such strategy should be assumed in the absence of a report. The primary result is not reduced to one comparison in the posted record, so there is no single primary contrast to interpret as a definitive statistical test.

The efficacy population is defined as randomized participants who received at least one dose of study drug post-randomization. The registry does not report the number of randomized participants excluded by this condition, the number with missing primary outcome data, reasons for missingness, or methods used to address missing observations. Imputation methods and sensitivity analyses are also not reported in the registry record.

These omissions limit assessment of whether the analysis preserved the randomized comparison and how robust any result might be to incomplete outcome information. They do not establish that missing data or multiplicity problems occurred; rather, the necessary details are not reported in the registry record. The statistical analysis plan and sample-size justification are also not reported, so the intended primary testing procedure cannot be verified from the registry information.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 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 (3.4%), suicidal ideation (2.6%), disturbance in social behavior (1.7%), electrocardiogram abnormal (0.9%), chronic obstructive pulmonary disease (0.9%).

The registry reports serious adverse-event counts and other adverse-event counts for three labeled groups. For Open- Label ALKS 3831, 255 participants were at risk, 4 were affected by a serious adverse event, and 0 were affected by other adverse events; deaths were 0. For Olanzapine + Placebo, 117 were at risk, 12 were affected by a serious adverse event, 19 by other adverse events, and deaths were 1. For ALKS 3831, 112 were at risk, 7 were affected by a serious adverse event, 26 by other adverse events, and deaths were 1.

These are descriptive counts. The Open- Label ALKS 3831 group has a different label from the two randomized groups, and the registry record does not explain its relationship to the randomized comparison. The counts should therefore be read within their listed groups rather than treated as a direct randomized comparison without additional design information.

The registry also lists serious-event terms and percentages: Schizophrenia at 3.4%, Suicidal ideation at 2.6%, Disturbance in social behavior at 1.7%, Electrocardiogram abnormal at 0.9%, Chronic obstructive pulmonary disease at 0.9%, and Transient ischaemic attack at 0.9%. These term percentages do not by themselves establish that an event was caused by a study drug, or that the groups differ statistically. The registry record does not report a formal comparative safety analysis or further detail on event attribution.

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

The trial was designed to compare samidorphan plus olanzapine with placebo plus olanzapine in a randomized Phase 2 study involving people with schizophrenia and alcohol use disorder. Its primary endpoint counted participants who had an adjudicated event of exacerbation of disease according to a composite definition. The registry record also identifies a specified efficacy population and reports descriptive adverse-event counts.

What the posted primary information does not establish is the direction, size, or precision of the primary treatment contrast. No single comparative result, confidence interval, or p-value for that endpoint is reported in the registry record. The trial’s completion and enrollment of 300 are not evidence of a particular treatment effect. Without the comparative estimate, it is not possible to determine from this record whether the primary outcome differed between randomized arms or how uncertain such a difference might be.

The endpoint is broader than any one symptom or event. It included hospitalization, score-based criteria, specified behavior or ideation, rescue medication, emergency visits, certain forms of discontinuation, and other listed events. A total participant count would not reveal which components contributed to the outcome unless component-specific results were also available. Those results are not reported in the registry record.

The safety counts also have limits. They are reported for groups with different labels, including an open-label group, and do not constitute a formal treatment comparison. Event counts and listed terms should not be interpreted as proof of causation, a relative risk, or a complete account of safety. The registry record does not report the statistical procedures for comparing safety outcomes.

Accordingly, the record can support a careful description of the design, endpoint, and recorded safety observations, but it cannot support a quantitative conclusion about comparative efficacy. It also cannot establish equivalence, benefit, lack of benefit, or a causal interpretation for individual adverse events. Those conclusions would require results and methodological details not reported in the registry record.

13. Why This Trial Matters Statistically

This trial raises a statistically important problem: the central endpoint is clinically broad, yet the registry record does not give a single randomized comparison for it. A study can define a meaningful outcome and complete enrollment while still leaving readers unable to assess the estimated difference or its precision from the posted information. The endpoint description and a comparative result answer different questions; the former does not substitute for the latter.

The composite endpoint also illustrates a trade-off in outcome design. Combining several event types may capture diverse forms of worsening, but it can make an overall result difficult to interpret without component-level reporting. A total count would not show whether the result reflected hospitalizations, symptom-score thresholds, rescue medication, or another criterion. For statistical interpretation, transparent reporting of both the composite and its components can clarify what a group contrast represents.

The defined efficacy population deserves attention as well. It is called intent-to-treat, but includes randomized participants only if they received at least one post-randomization dose. Reporting the number excluded and their outcome status would help readers understand how the analysis relates to randomization and whether the population differs from all randomized participants.

Finally, the registry’s descriptions of masking differ in wording: the brief summary says double-blind, whereas the classification is quadruple. Without further explanation, readers cannot determine how the labels were operationalized. Along with unreported multiplicity procedures, missing-data handling, and sample-size justification, this illustrates why complete statistical documentation matters. Transparent methods and results enable readers to distinguish a defined research question from evidence that answers it.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The main limitation for interpreting efficacy is the absence of a single reported comparison for the primary endpoint. The record states that the primary outcome has several categories or time points and is not reduced to one comparison. It therefore does not give a primary arm-specific estimate or its uncertainty in a form that supports a quantitative interpretation.

The composite endpoint includes multiple kinds of events, but component-specific results are not reported in the registry record. The adjudication process is identified, yet details about how events were collected, reviewed, or classified are not reported. These omissions limit assessment of the endpoint’s components and implementation.

The registry’s efficacy-population definition requires at least one post-randomization dose. Counts of randomized participants who did not meet that condition, as well as the extent and handling of missing outcome data, are not reported. The statistical model, analysis assumptions, sensitivity analyses, and multiplicity approach are not reported either. The registry record also does not report the trial’s own sample-size justification or a protocol-level statistical plan.

Masking is described as double-blind in the brief summary and classified as quadruple in the registry. The record does not explain the difference. The adverse-event record includes an open-label group as well as the two named randomized groups, but the relationship between these groups is not explained in the information available for this analysis. Safety counts therefore require careful, group-specific reading.

These are limitations in the information available for interpretation, not evidence that particular procedures were absent or that particular biases occurred. The registry record does not provide enough detail to resolve them.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Study of ALKS 3831 in Subjects With Schizophrenia and Alcohol Use Disorder — NCT02161718.
• Registry record last updated 2021-10-08; status last verified 2021-09.
• Results first posted 2021-10-08.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

The registry record describes a completed randomized, parallel-group Phase 2 study sponsored by Alkermes, Inc., in people with schizophrenia and alcohol use disorder. Enrollment was 300. The named randomized comparison was samidorphan plus olanzapine (ALKS 3831) versus placebo plus olanzapine. The record gives a start of 2014-05 and primary completion of 2017-01-11.

The primary outcome was the number of participants with an Independent Adjudication Committee-adjudicated event of exacerbation of disease during the double-blind period, with a timeframe of up to 15 months. Efficacy analyses were defined for randomized participants who received at least one post-randomization dose. A single comparative primary result is not reported in the registry record.

The registry also lists secondary outcomes concerning exacerbation events and change in World Health Organization drinking-risk level, plus descriptive safety counts and serious-event terms. Details of missing-data handling, multiplicity, statistical modeling, and the trial’s sample-size justification are not reported.