Safer Use of Antipsychotics in Youth: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safer Use of Antipsychotics in Youth (NCT03448575), a randomized na study. The primary outcome is Percent of Youths With Antipsychotic Orders at 6 Months.

2. Clinical Question

3. Trial Design

The registry describes a randomized, quadruple-masked, parallel-group study. It is marked completed, and its phase is listed as NA. Kaiser Permanente is the sponsor. The registry dates the study start to 2018-03-29 and primary completion to 2020-06-30. Enrollment is listed as 733.

The comparison was between a provider medication alert alone and a broader intervention combining an alert with child psychiatrist review and enhanced behavioral health access. In the intervention arm, the alert could prompt a prescriber to keep or remove an antipsychotic order and/or order study services. The arm description includes behavioral health navigation, expedited psychotherapy access, and a virtual consult with a child and adolescent psychiatrist. It also describes passive case review by the study child and adolescent psychiatrist for all, although the registry excerpt does not complete that description.

The study summary calls this a practical clinical trial and says the aim was to recruit 800 eligible patients in 4 health systems. That aim is distinct from the registry's enrollment figure; the record does not explain the relationship between the aim and enrollment. The actual allocation ratio, methods used to generate or conceal the random sequence, and details of the masking procedure are not reported in the registry record. The planning illustration assumes equal allocation, but that assumption does not establish the allocation used in the trial.

The primary outcome concerns orders placed within the health system during the 180 day period following the index date. Several secondary outcomes address medication use, healthcare visits, safety assessments, treatment-plan changes, and behavioral health services. The record lists 12 outcome measures and 10 secondary outcomes, but it does not report a single statistical analysis for every listed outcome.

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: We define "started" as the number of eligible participants for whom a study medication alert fired in Epic when the provider ordered an antipsychotic. A blinded study team member assessed for eligibility according to study criteria and marked as eligible or ineligible.

Overall Study

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 Percent of Youths With Antipsychotic Orders at 6 Months, measured over 180 day period following index date.

The primary endpoint is the percentage of youths with antipsychotic orders at 6 months. The record defines the measure as medication orders placed within the health system during the 180 day period following the index date. It is therefore an order-based measure, not a direct measure of whether medication was taken, how consistently it was taken, or whether an order was placed outside the health system.

This endpoint is relevant to the trial’s intervention because both arms used an electronic alert triggered when a prescriber initiated an antipsychotic prescription for an eligible patient. The intervention added clinical review and offers of behavioral health services. Tracking orders in the subsequent period can describe prescribing activity within the system after that index event. It does not, on its own, identify why an order was placed or establish whether a change in orders resulted from any particular component of the intervention.

The posted arm results are 54.6% in the control arm and 47.6% in the intervention arm. The registry record does not provide an estimable two-arm comparison for this primary count-type outcome. A confidence interval, p-value, or adjusted effect estimate for the primary comparison is therefore not reported in the registry record. The percentages can be read as descriptive arm results, but they do not provide the precision of an inferential comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The posted results include descriptive arm summaries and recalculated unadjusted comparisons for three secondary outcomes. The registry record does not report an estimable two-arm comparison for the primary count-type outcome. These are different levels of statistical information: an arm percentage or mean describes the observed summary in that arm, while a contrast estimates how the arms differ and, when accompanied by a confidence interval, conveys uncertainty about that estimate.

What does a confidence interval express?

In general, a confidence interval gives a range of values compatible with the data and the method used to construct it. A narrower interval indicates greater precision under that method; a wider interval indicates greater uncertainty. For a difference, an interval that includes zero means that the data are compatible with no difference as well as with differences in either direction. It does not prove that the true difference is zero. The confidence intervals reported for the secondary recalculations include zero, so those estimates do not resolve whether the underlying arm contrast differs from zero.

How should the mean differences be read?

For days' supply and emergency department or urgent care visit frequency, the arm summaries are means with standard deviations. The recalculated contrast is intervention minus control. A negative mean difference indicates a lower observed mean in the intervention arm for the relevant outcome. The estimated mean differences are -7.5 days for days' supply, with a 95% confidence interval from -16.37 to 1.375, and -0.07 visits for emergency department or urgent care frequency, with a 95% confidence interval from -0.213 to 0.073. Their unadjusted p-values are 0.098 and 0.337. These comparisons use posted summary statistics and are not reported sponsor analyses.

A standard deviation describes variation among observations within an arm; it is not the uncertainty interval for the arm mean or the between-arm difference. For count-like outcomes, the distribution may be asymmetric or contain many low values. In general, over-dispersion means that variability exceeds what a chosen statistical model expects. Accounting for it can matter when fitting count models, but the registry record does not identify a model used for these outcomes or report whether over-dispersion was assessed. The recalculated mean comparisons should not be mistaken for a model-based analysis of the underlying count distribution.

How should the risk difference be read?

For change to the psychotropic medication treatment plan, the posted arm results are percentages and the recalculated contrast is a risk difference. The estimate is -0.015, with a 95% confidence interval from -0.087 to 0.056 and an unadjusted p-value of 0.674. The estimate is expressed as a proportion, while the arm summaries are expressed as percentages. The interval includes zero and permits effects in either direction. The registry also posts relative measures for this outcome, but neither a relative measure nor a single p-value replaces consideration of the interval and the endpoint’s definition.

In general, a p-value measures how incompatible the observed data are with a specified null hypothesis under the assumptions of the test. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of an effect. These secondary p-values are explicitly unadjusted. The registry record does not identify the tests or models used by the sponsor, and reported analyses are not listed.

What do the planning illustrations mean?

The sample-size illustrations are planning calculations based on the observed effects, not observed power. Their stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, a normal approximation, and an effect to detect expressed as a fraction of the observed effect. For outcomes other than time-to-event outcomes, the figures represent participants per arm. The illustrations show how required sample sizes can increase when a smaller effect is selected for planning. They do not show that the study was designed around the observed effects, nor do they supply a justification for the trial’s original enrollment target. The record does not report that justification.

Because these calculations use observed effects, they are sensitive to the uncertainty and possible instability of those effects. A planning figure is not evidence that a study has, or lacks, adequate power after results are known. It is an illustration of hypothetical design requirements under the stated assumptions. The record does not report interim analyses or a prespecified statistical analysis plan, so neither should be inferred from these calculations.

What is missing from the primary comparison?

The primary outcome is designated as a count-type endpoint, but it is presented as arm percentages and the registry record identifies values as missing or not estimable for a two-arm comparison. It therefore supplies no primary confidence interval or p-value to interpret. The arm summaries remain informative as descriptions, but they cannot answer how precisely the between-arm difference is estimated. The registry record does not explain why a comparison was not estimable, identify a primary analysis model, or report an alternative adjusted estimate.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists 12 outcome measures and 10 secondary outcomes. It does not report a multiplicity adjustment strategy, a hierarchy for testing outcomes, or a plan for designating secondary results as confirmatory or exploratory. Without that information, the secondary p-values should be read as unadjusted comparisons rather than as results protected against the chance of false-positive findings across multiple tests. The record does not report an analysis plan that would clarify how the outcomes were prioritized.

Missing-data methods and the handling of incomplete outcome measurements are not reported in the registry record. The outcome summaries include arm-specific counts for some measures, but the record does not explain whether participants with unavailable outcome data were excluded, how missing values were handled, or whether any sensitivity analyses were conducted. The listed enrollment and outcome denominators do not, by themselves, establish the completeness of follow-up or the reasons for any difference between them.

The analysis populations are not reported in the registry record. In particular, it does not state whether the analyses followed an intention-to-treat principle, used a per-protocol population, or applied another definition. The record also does not report how protocol deviations, crossover, or withdrawals were handled. These details can affect interpretation of a randomized comparison, so they cannot be inferred from the randomization label alone.

The registry lists no reported sponsor analyses, and the primary comparison is not estimable in the posted results. The recalculated secondary comparisons are unadjusted and are based on posted summary statistics. The record does not report the statistical models, covariate adjustments, or procedures used by the sponsor for the primary or secondary outcomes. These omissions limit assessment of how the analyses were specified and how robust the estimates are to alternative assumptions.

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 0% reporting threshold.

The registry safety summary lists zero serious adverse events and zero deaths in each arm. It also lists zero other adverse events in each arm. The at-risk counts are 385 for the control arm and 348 for the intervention arm for both the serious and other adverse event summaries.

These figures describe the events recorded in the registry summary; they do not establish that no adverse events occurred outside the definitions, collection procedures, or reporting period used for the study. The registry record does not provide serious adverse event terms, and no top serious-event terms are listed. It also does not report the methods used to collect, classify, or verify adverse events in the information presented.

A zero count in a registry summary should therefore be read as a reported count, not as proof of zero risk. The figures do not show how completely events were ascertained or whether the safety results were compared using a statistical analysis. No risk difference, confidence interval, or p-value for adverse events is reported in the registry record. These descriptive figures are not a basis for conclusions beyond what the recorded counts state.

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

The posted primary results describe the percentage of youths with antipsychotic orders placed within the health system during the 180 day period following the index date. They show arm-specific percentages, but the registry record does not provide an estimable two-arm comparison for this endpoint. The primary result therefore does not include a reported confidence interval or p-value that would quantify the uncertainty in the contrast.

The result does not establish whether an order was filled or whether a medication was taken. It is limited to orders placed within the health system and does not capture orders outside it. Nor does the primary outcome isolate the contribution of the electronic alert, psychiatrist review, behavioral health navigation, or bridging therapy. The intervention combined several components, and the posted primary endpoint is an outcome measure rather than a component-specific evaluation.

The secondary estimates add descriptive information about medication supply, visits, and treatment-plan changes. Their unadjusted comparisons have uncertainty intervals that include zero, and they do not provide a substitute for the missing primary contrast. The registry does not report a multiplicity strategy or sponsor analyses that would support stronger claims from the secondary results. Absence of a clear difference in these recalculations does not demonstrate equivalence or prove that the arms had identical outcomes.

The trial’s randomized design is relevant to how a comparison could be interpreted, but design labels alone do not resolve limitations in the posted results. The registry record does not report the primary analysis model, analysis population, missing-data procedures, or sample-size justification. Consequently, the descriptive percentages and recalculated secondary estimates should be interpreted as the statistics available in the registry, not as a complete account of every analysis conducted or as a definitive estimate of the intervention’s effects.

13. Why This Trial Matters Statistically

This trial raises a useful statistical question about how to evaluate a multicomponent clinical intervention using routine prescribing data. Its primary endpoint is tied to medication orders recorded within the health system, while its intervention includes an electronic alert, clinical review, and offers of behavioral health services. The contrast therefore concerns two care strategies, not a single isolated feature.

The record also illustrates why descriptive arm results and an inferential comparison are not interchangeable. The primary percentages describe each arm, but the record does not provide an estimable comparison or its uncertainty. Without that contrast, readers cannot assess the precision of the primary between-arm result from the posted statistics. A randomized design does not compensate for an unavailable primary estimate.

The secondary outcomes show another important issue: different endpoints can capture different parts of care, and their estimates may be uncertain even when they point in a similar direction. Unadjusted testing across multiple outcomes can make isolated findings difficult to interpret. A clear analysis plan, transparent population definitions, and reporting of missing-data methods would help readers assess how much weight to place on both primary and secondary results.

Finally, the sample-size illustrations are distinct from the trial’s original design rationale. They use observed effects and explicit assumptions to explore hypothetical planning requirements; they do not reveal what the investigators planned in advance or establish observed power. This distinction is important when interpreting completed trials: design assumptions, observed estimates, and post-result planning calculations answer different statistical questions.

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 primary limitation in interpreting the posted result is that the registry reports arm percentages but no estimable two-arm comparison for the primary endpoint. The reason the comparison is unavailable is not reported in the registry record. As a result, the precision of the primary contrast cannot be evaluated using a confidence interval or p-value from the posted results.

The endpoint counts orders placed within the health system, rather than medication use across all settings. The registry record does not report whether orders outside the system were captured, or whether an order corresponded to medication initiation, dispensing, or adherence. These distinctions constrain what the endpoint can represent. The record also does not report additional detail about the primary-outcome population.

The study is described as randomized and quadruple-masked, but the sequence-generation and allocation-concealment methods, masking procedures, and actual allocation ratio are not reported in the registry record. The planning illustration’s equal-allocation assumption should not be read as documentation of trial allocation. The record does not report the analysis population, missing-data procedures, protocol-deviation handling, or the primary statistical model.

Numerical secondary comparisons are unadjusted recalculations from summary statistics, and sponsor-reported analyses are not listed. A multiplicity strategy is not reported in the registry record. These factors limit how strongly secondary findings can be interpreted, particularly alongside the number and variety of listed outcomes. The sample-size illustrations are based on observed effects and stated planning assumptions, not the trial’s original design justification. The record does not report that justification, interim analyses, or sensitivity analyses.

Further details about eligibility, participant characteristics, outcome ascertainment, and the incomplete intervention description are not reported in the registry record. These gaps limit assessment of how broadly the results may apply and how completely the intervention was delivered. They do not change the descriptive values reported, but they define the boundaries of what can be concluded from the registry record.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Safer Use of Antipsychotics in Youth — NCT03448575.
• Registry record last updated 2024-10-31; status last verified 2024-10.
• Results first posted 2024-10-31.
• 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 NCT03448575, “Safer Use of Antipsychotics in Youth,” as a completed, randomized, quadruple-masked, parallel-group study sponsored by Kaiser Permanente. Enrollment is listed as 733. The study compared a provider medication alert alone with an alert plus child psychiatrist review and enhanced behavioral health access.

The primary outcome was the percentage of youths with antipsychotic orders placed within the health system during the 180 day period following the index date. The record posts arm percentages but no estimable two-arm comparison for that outcome. Three secondary outcomes have numerical arm comparisons in the posted results; those estimates are unadjusted recalculations with confidence intervals that include zero.

The registry record does not report sponsor analyses, the primary statistical model, analysis populations, missing-data methods, or a multiplicity strategy. Safety summaries list zero serious adverse events, zero other adverse events, and zero deaths in both arms. Planning illustrations are hypothetical calculations based on observed effects, not observed power or the trial’s original design justification.