Evaluation of TAK-058 and Ondansetron on P50 Auditory Gating in Participants With Stable Schizophrenia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Evaluation of TAK-058 and Ondansetron on P50 Auditory Gating in Participants With Stable Schizophrenia (NCT02614586), a randomized phase 1 study. The primary outcome is Change From Baseline in P50 Ratio S2/S1 at Central (Cz) Electrode Following Administration of TAK-058.

2. Clinical Question

3. Trial Design

The registry describes a Phase 1, randomized, quadruple-masked crossover study sponsored by Takeda. Enrollment is recorded as 11, and the study status is terminated. The recorded start is 2015-12 and the primary completion is 2016-04. These dates describe the registry record; they do not establish that the planned primary outcome was measured or analyzed.

A crossover design assigns a participant to receive more than one study condition in a sequence, allowing a participant’s measurements under different conditions to be compared. Here, the record lists sequences involving TAK-058, ondansetron, and their matching placebos. The sequence descriptions include oral administration and, where specified, a 1-week washout period between intervention periods. The record does not give a complete explanation of the sequence-generation procedure, the rationale for the washout, or the intended analysis of period and sequence effects.

Randomization and quadruple masking are recorded, but the registry does not report which groups of people were masked or how masking was maintained. It also does not report a target enrollment or the rationale used to select the recorded enrollment. The planned primary endpoint refers to Part 2, with baseline and repeated post-dose assessment in each period. The registry explicitly says that Part 2 was not initiated because it was not possible to calculate an intraclass correlation coefficient given the magnitude of within-participant variability relative to between-participant variability. The reason the overall study was terminated is not reported in the registry record beyond that statement about Part 2.

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 with a diagnosis of schizophrenia were enrolled in this 2-part study comprised of Part-1, screening and Part-2, treatment phase. Part-2 was not initiated because in Part-1 (sequential step 2), the P50 auditory gating (stimulus\[S\]2/S1) ratio for alpha and beta frequency bands demonstrated high variability and the study was terminated.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Part 2 was not initiated because it was not possible to calculate an ICC due to the magnitude of the intrasubject variability in relation to the intersubject variability.

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: Schizophrenia Set included all participants who were enrolled and completed the Screening Phase (Part-1) of the study.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in P50 Ratio S2/S1 at Central (Cz) Electrode Following Administration of TAK-058, measured over Part 2: Day 1 pre-dose and at multiple time points (up to 2 hours) post-dose in each period..

The planned primary endpoint was change from baseline in the P50 ratio S2/S1 at the central Cz electrode following administration of TAK-058. The ratio compares the amplitude of the test P50 wave, S2, with the amplitude of the conditioning P50 wave, S1. The registry defines S1 as the conditioning wave with the most positive peak in the stated latency window after the conditioning stimulus. S2 is the test wave whose positive peak is closest in latency to the conditioning P50. For both waves, amplitude is the difference between the positive peak and the preceding negative trough.

As a general explanation, paired auditory stimuli can be used to examine sensory gating: the response to the second stimulus is considered in relation to the response to the first. The S2/S1 ratio therefore provides a way to summarize the relative size of those responses. It is a physiological measurement, not itself a direct measure of a clinical symptom or day-to-day functioning. The registry describes P50 as a pharmacodynamic marker, meaning a measurement intended to reflect a biological response to an intervention.

The planned assessment was before dosing and at multiple time points up to 2 hours after dosing in each period. The endpoint’s unit, summary parameter, and measure of dispersion are not reported in the registry record. Since Part 2 was not initiated, the record does not provide observed endpoint values or a change-from-baseline estimate.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a randomized crossover study with quadruple masking. Those design labels indicate how the study was intended to allocate and conceal treatment conditions, but they do not substitute for a statistical analysis plan. The registry record does not report the detailed randomization method, the identities of the masked groups, or the planned model for the endpoint. The following explanations are general; they do not imply that a particular method was used in this study.

What does crossover mean for a comparison?

In a crossover design, participants are intended to receive multiple conditions in an assigned order. Each participant can then contribute measurements under more than one condition. A comparison may use within-participant contrasts, which can reduce the impact of stable differences between people. However, the validity of such a comparison depends on issues including treatment sequence, period effects, carryover, measurement reliability, and complete data across conditions. The registry lists several treatment sequences and a washout period in the descriptions, but it does not report how period, sequence, or carryover effects were to be handled statistically.

Masking is intended to limit the influence of expectations or knowledge of assigned treatment on study conduct and assessment. “Quadruple” is the masking label recorded in the registry. The specific roles kept unaware of treatment assignment are not reported in the registry record, so it is not possible to describe precisely how masking operated.

What does the P50 ratio summarize?

The endpoint is the ratio of the test response, S2, to the conditioning response, S1, measured at Cz. The registry describes paired auditory clicks and defines how the positive peaks and preceding negative troughs are used to obtain each wave’s amplitude. In general, a ratio expresses one quantity relative to another; here, the two quantities come from responses to the test and conditioning stimuli. A ratio can be affected by changes in either its numerator or denominator, so interpreting it as a measure of gating depends on the measurement procedure and the planned analysis.

The endpoint was to be assessed at baseline and at multiple times after dosing, up to 2 hours, in each period. Repeated measurements can describe how a response changes over time, but they also create analysis choices: for example, whether to compare a prespecified time point, summarize several time points, or model the repeated measurements together. The registry does not report which approach was planned. It does not report the endpoint’s unit, summary parameter, or dispersion measure either.

Why does the variability statement matter?

The registry states that Part 2 was not initiated because an intraclass correlation coefficient could not be calculated in view of the magnitude of within-participant variability relative to between-participant variability. In general, an intraclass correlation coefficient describes how strongly measurements from the same participant resemble one another relative to variation across participants. It can inform understanding of measurement reliability or the structure of repeated data, depending on the intended use and calculation.

When measurements vary substantially within a participant, repeated measurements may be less consistent. That can make it difficult to distinguish stable differences between participants from fluctuations within participants. For a crossover study, within-participant comparisons are central to the design, so the amount and structure of this variability can matter. The registry does not say what calculation was attempted, what data were used, or what threshold or decision rule governed whether Part 2 could proceed. It would therefore be unwarranted to infer a specific reliability estimate or a particular protocol rule.

What can be concluded from the posted analyses?

The registry reports zero analysis entries. Without an estimate and its uncertainty, there is no basis to assess the size, direction, or precision of a treatment contrast. A confidence interval, in general, describes a range of values compatible with the data and the statistical procedure used; it is not available for the primary question. A p-value, in general, summarizes how unusual data at least as extreme as those observed would be under a specified null model. No primary-outcome p-value is reported.

Planning calculations are different from analyses of observed outcomes. The sample-size assumptions described for this page characterize planning illustrations: two-sided significance level 0.05, equal allocation between two arms, a normal approximation, and an effect expressed as a stated fraction of an observed effect. Such illustrations are not observed power. They do not demonstrate that the crossover study enrolled enough participants, and they cannot replace an outcome estimate. Since this record does not report the primary outcome analysis, the planning illustrations should not be mistaken for evidence about treatment effects or precision.

Finally, the registry does not report an analysis population, missing-data method, multiplicity adjustment, or statistical model. These omissions matter because analysis decisions can affect which participants and measurements contribute to an estimate and how uncertainty is quantified. They remain unreported rather than recoverable from the design labels or endpoint description.

10. Multiplicity, Missing Data and Analysis Populations

The registry does not report how multiplicity was to be handled. The planned endpoint involved multiple post-dose measurement times in each period, and the study included more than one active intervention and placebo conditions. Those features can create multiple possible comparisons or summaries. As a general statistical concern, examining many time points or contrasts can increase the chance of findings that appear notable by chance unless the analysis plan addresses the selection and interpretation of comparisons. The record does not identify a primary time point, a hierarchy of comparisons, or an adjustment strategy. It would be inappropriate to assume that any particular correction was planned.

Missing-data handling is also not reported in the registry record. Because Part 2 was not initiated, the record supplies no basis for describing which planned endpoint observations were missing, whether measurements were collected in another part, or how any incomplete sequences would have been treated. The registry does not report an imputation method, a sensitivity analysis, or assumptions about missingness.

The analysis population is not reported. In particular, the record does not say whether the intended analysis would include all randomized participants, only participants with measurements in each condition, or another defined group. Although enrollment is recorded as 11, that figure does not by itself identify the population for an endpoint analysis.

The registry lists a secondary outcome concerning participants with at least one treatment-emergent adverse event, but no analysis result is posted for it in the information. The adverse-event entry shown is for Part 2, with zero participants at risk and zero affected. Given that Part 2 was not initiated, those entries should not be interpreted as a complete safety analysis of all enrolled participants or as proof that no events occurred.

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

This record describes a planned pharmacodynamic comparison, not an interpretable estimate of treatment effect. It documents a randomized, quadruple-masked crossover design, lists the planned P50 S2/S1 measurement, and records that Part 2 was not initiated. It does not report primary-outcome values, an effect estimate, a confidence interval, or a p-value. The posted information therefore does not establish whether P50 gating changed after TAK-058 or how any change compared with placebo.

The absence of an analyzed result is not evidence that the treatments had no effect. It is also not evidence that they had an effect. Without outcome measurements and an analysis, the direction and magnitude of any difference remain unknown. Similarly, the study’s recorded enrollment cannot be treated as the number of participants with usable endpoint data, since the registry does not report that count.

The reason given for not initiating Part 2 concerns the inability to calculate an intraclass correlation coefficient in the presence of substantial within-participant variability relative to between-participant variability. This explains the stated obstacle but does not quantify measurement reliability or show how a treatment comparison would have turned out. The registry does not provide the attempted calculation or its underlying data.

The secondary adverse-event outcome and the Part 2 safety group do not resolve the missing efficacy analysis. Zero participants at risk in that group means the posted group summary cannot serve as a safety estimate for the full enrolled study population. No broader adverse-event findings are reported.

Accordingly, the most defensible reading is limited: the registry records an intended study and an uninitiated primary-outcome part, with no analysis results available for the planned treatment question. The tables, plots, and planning material should be interpreted as summaries or illustrations, not as substitutes for observed results. No clinical recommendation follows from this record.

12. Why This Trial Matters Statistically

The study is statistically informative as an example of how a trial’s design can be clear while its planned inferential question remains unanswered. The registry identifies a crossover structure, randomization, quadruple masking, treatment sequences, and a physiological endpoint. It also records a specific reason Part 2 was not initiated: the stated difficulty calculating an intraclass correlation coefficient given the relative magnitude of within-participant and between-participant variability.

That variability issue is relevant to the logic of repeated-measures designs. In general, crossover comparisons rely on measurements within the same participant, so the consistency and structure of repeated observations matter. If within-participant measurements vary considerably, the planned comparison may be harder to characterize or interpret. The registry does not provide enough detail to assess the magnitude of the problem or judge what alternative design or analysis might have been possible, but it makes clear that measurement properties affected the study’s progression.

The record also illustrates why a planned endpoint must be distinguished from an observed endpoint. Naming the P50 ratio and specifying a measurement schedule do not provide an estimate of treatment response. Without analyzed observations, there is no treatment contrast to assess for magnitude or precision. Reporting that limitation directly is important: a terminated study with no primary analysis should not be summarized as a positive or negative finding.

Finally, the case highlights the need to interpret enrollment and safety summaries in context. Enrollment is recorded, yet the number contributing to the uninitiated Part 2 endpoint is not reported. The Part 2 adverse-event group has no participants at risk, so it does not summarize the experience of all enrolled participants. These distinctions help readers separate design information, study status, and actual analysis results.

13. 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 interpretation is that Part 2, which contained the planned primary endpoint, was not initiated. The registry therefore provides no observed P50 ratio result for the stated treatment question. It does not report an effect estimate, uncertainty interval, p-value, or endpoint-level participant count. This prevents assessment of the direction, magnitude, and precision of any treatment contrast.

The record gives a reason for not initiating Part 2, but the information is limited. It states that an intraclass correlation coefficient could not be calculated because within-participant variability was large relative to between-participant variability. It does not report the data or calculation behind that statement, the intended role of the coefficient, or the decision criteria used. The broad study status is terminated, but the record does not provide a separate overall termination explanation.

Other design and analysis details are not reported in the registry record. These include the detailed randomization procedure, which roles were masked, the planned statistical model, the prespecified time-point or summary for the primary comparison, the analysis population, and handling of missing data and multiplicity. The endpoint’s unit and dispersion measure are also not reported. The sequence descriptions provide some administration details but do not fully explain how period, sequence, or potential carryover effects were to be addressed.

Enrollment of 11 does not establish how many participants were randomized, completed a sequence, or had usable measurements. The adverse-event summary pertains to a Part 2 group with zero participants at risk and cannot be used as a complete account of safety for the enrolled population. Together, these gaps limit both efficacy and safety interpretation; they do not support conclusions about whether the interventions did or did not affect the measured outcome.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Evaluation of TAK-058 and Ondansetron on P50 Auditory Gating in Participants With Stable Schizophrenia — NCT02614586.
• Registry record last updated 2017-05-01; status last verified 2017-03.
• Results first posted 2017-05-01.
• 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.

17. Record Summary

The registry describes a randomized, quadruple-masked, Phase 1 crossover study of TAK-058 and ondansetron in participants with stable schizophrenia. Enrollment is recorded as 11, and the status is terminated. The planned primary endpoint was change from baseline in the P50 S2/S1 ratio at the Cz electrode, measured before dosing and at multiple post-dose times up to 2 hours in each period.

Part 2 was not initiated because an intraclass correlation coefficient could not be calculated in view of the magnitude of within-participant variability relative to between-participant variability. The registry reports no analysis entries and no primary endpoint results. It lists a secondary treatment-emergent adverse-event outcome, but the Part 2 safety group has zero participants at risk and does not provide a full-population safety result.

The record supports describing the intended design and the reason stated for not initiating Part 2. It does not support an estimate of treatment effect, a conclusion about the P50 outcome, or a complete safety interpretation.