Improving Dental Care and Oral Health in Children With Autism Spectrum Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Improving Dental Care and Oral Health in Children With Autism Spectrum Disorder (NCT03003221), a randomized na study. The primary outcome is Frequency of Tooth Brushing at Home.

2. Clinical Question

This question focuses on a reported home-care behavior rather than a direct measure of dental disease. The primary endpoint is described as the difference in brushing frequency between baseline and 6 months. That wording makes the timing of measurement central: the result concerns change over the stated interval, not simply a general report of brushing at an unspecified time.

The record describes a randomized controlled trial conducted at sites, with the toolkit as the control condition and the combined toolkit-plus-training regimen as the intervention condition. The randomized comparison is intended to assess whether the groups' reported brushing outcomes differed under the study conditions. The record does not explain how randomization was implemented or whether the outcome analysis adjusted for any particular baseline characteristics.

The primary question is narrower than the trial's broader aims, which also included oral-health outcomes and dental-office experiences. Secondary outcomes listed in the record cover brushing behavior, dentist ratings of caries and decayed teeth, observed anxiety and behavior at a dental visit, and dentist-reported behavioral compliance. Those outcomes provide context for the study's scope but do not replace the prespecified primary endpoint in interpreting its principal statistical result.

3. Trial Design

The registry identifies this as a randomized, single-masked, parallel-group trial. Randomization means treatment assignment was intended to allocate participants to study conditions by chance, while a parallel design means participants were assigned to one condition rather than crossing between conditions. The record does not identify who was masked, so the meaning and practical reach of the single-masking designation cannot be determined from this entry alone.

The control condition was the AIR-P Dental Toolkit. The parent-training condition included the same toolkit together with individual behavioral parent training. The record describes 7 core in-person sessions, including a home visit and a dental office coach, as well as 4 phone booster sessions. These details distinguish the intervention package from the toolkit-only comparison; they do not establish which component, if any, accounts for a group difference.

The registry lists enrollment as 119. Arm-level primary-outcome summaries are reported for 59 participants in the AIR-P Dental Toolkit condition and 60 in the Parent Training condition. These figures describe the denominators attached to those summaries; the record does not clarify whether they represent all randomized participants, participants with outcome data, or another analysis population.

The study is marked completed. The recorded start is 2016-12 and the primary completion is 2019-06. The phase is listed as NA. The sponsor is the University of California, Irvine. The registry entry describes the study as taking place at sites but does not give site names or the number of sites in the registry record for this account.

The trial's brief summary frames its aims as improving home dental care, oral-health outcomes, and dental-office visit experiences. The primary endpoint reported in this record is home brushing frequency, while the secondary outcomes include several caries and dental-visit measures. The record does not report the randomization sequence, allocation concealment, masking procedures, a protocol, or a statistical analysis plan. Those details therefore cannot be used to evaluate design implementation or to determine whether the posted analysis followed a prespecified 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.

Overall Study

Reasons for not completing this period:

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 Frequency of Tooth Brushing at Home, measured over Difference in brushing frequency between Baseline and 6 months.

The primary endpoint is parent-reported frequency of successful child tooth brushing at home, expressed in days per week. The description defines successful brushing as twice-daily brushing completed at home during the past week. The stated timeframe is the difference in brushing frequency between baseline and 6 months.

This measure is used to capture a home-care behavior that is directly related to the intervention's focus on family routines and parent support. A frequency measure can make a behavioral outcome interpretable in practical terms: it records how many days in the week the parent reports that the brushing criterion was met. It is not, by itself, a measure of brushing quality, duration, or dental health.

The registry reports an arm-level mean of 4.3 days per week for the AIR-P Dental Toolkit condition, with a 95% confidence interval from 3.5 to 5.1, and an arm-level mean of 5.5 days per week for the Parent Training condition, with a 95% confidence interval from 4.8 to 6.0. The record's endpoint description refers to change from baseline to 6 months, but the arm summaries do not explicitly clarify in the listed results whether those values are follow-up levels, changes, or another summary of the endpoint. The reported contrast itself is not given.

As a parent-reported outcome, the measure reflects the information collected through the questionnaire and may be affected by recall and reporting. The record does not report validation details for the questionnaire or how missing weekly responses were handled. These considerations matter when interpreting the endpoint as a measure of reported brushing behavior rather than as an objective dental examination.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a confidence interval describe?

A confidence interval expresses uncertainty around an estimated quantity under the assumptions of the method used to calculate it. In repeated use of the same method, intervals constructed this way have a defined long-run coverage interpretation. A 95% confidence interval is not a statement that there is a 95% probability that a fixed effect lies inside this particular interval. Rather, it describes the performance of the interval procedure under its assumptions.

The registry gives a 95% confidence interval for each arm's mean. Those intervals convey uncertainty around separate arm-level summaries. They do not directly answer how uncertain the between-arm contrast is. To understand the comparison, a reader would ideally have the estimated contrast and its own confidence interval, alongside a clear definition of whether the contrast represents follow-up values, change from baseline, or a model-adjusted quantity. That contrast is not reported in the registry record.

What does the p-value add?

A p-value summarizes how incompatible the observed data are with a specified null hypothesis, assuming the statistical model and its assumptions are appropriate. Smaller p-values indicate greater incompatibility with that null under those conditions. A p-value does not measure the size or practical importance of an effect. It is also not the probability that the null hypothesis is correct, nor a direct measure of the chance that the finding would replicate.

The registry reports p = 0.025 for the generalized mixed model superiority analysis. That is the inferential result recorded for the group comparison. Without the model's estimated contrast and confidence interval, the p-value alone does not identify the direction and magnitude of the modeled effect with the same clarity as a direct estimate would. The arm means are descriptive summaries and should not be used to reconstruct a model-adjusted contrast.

Why might a mixed model be used?

General statistical explanation: mixed models can represent relationships among repeated observations or observations grouped within participants, families, clinics, or other units. The model can include fixed effects for systematic comparisons and random effects to represent variation among clusters or individuals. This structure may be useful when observations are not independent, as repeated measurements from the same person commonly are not. The registry identifies a generalized mixed model but does not report its terms or the correlation structure used, so the particular role of these features cannot be confirmed for this analysis.

Generalized models can also be selected when the outcome distribution does not fit the assumptions of a conventional linear model. The exact distribution and link function depend on the outcome and analysis plan. The registry does not identify those choices. It would be inappropriate to infer them from the phrase generalized mixed model alone.

What is over-dispersion, and why can it matter?

General statistical explanation: over-dispersion occurs when observed variability is greater than the variability assumed by a statistical model. If it is not accommodated, standard errors can be too small and confidence intervals can be too narrow, making inferential evidence appear stronger than warranted. A model can address over-dispersion through an appropriate variance structure or other methods, depending on the outcome and design. The registry does not report whether over-dispersion was assessed or how variance was modeled for this endpoint; there is therefore no basis for saying whether it affected this analysis.

Why does the endpoint's time definition matter?

The endpoint is described as a difference in brushing frequency between baseline and 6 months, but the reported arm summaries are labeled only by condition and measurement unit in the record. A change score, a follow-up mean adjusted for baseline, and a repeated-measures model contrast can answer related but distinct questions. Their estimates need not be identical. Clear reporting of the analyzed time point, the contrast definition, and any baseline adjustment helps readers understand what the result represents. Those details are not reported in the registry record.

The scale is days per week, and the outcome is described as a mean. Means summarize the central level of a continuous measure, but interpretation also depends on the distribution, data collection, and handling of incomplete responses. The record does not report the distribution of individual responses, questionnaire completion patterns, or the model diagnostics needed to assess those issues.

How should planning calculations be understood?

General statistical explanation: sample-size planning calculations estimate how many participants might be needed to detect a chosen effect under stated assumptions. They are prospective illustrations, not results observed in the trial and not a measure of the trial's observed power. The assumptions listed for the planning illustration in this record include a two-sided significance level of 0.05, equal allocation, a normal approximation, and an effect defined as a stated fraction of the observed effect. For time-to-event outcomes, the listed convention is total events; for other outcomes, it is participants per arm.

Such illustrations depend on the effect chosen for planning and on the assumptions about allocation and statistical testing. A smaller target effect generally requires more information to distinguish it from random variation, while a larger target effect generally requires less, all else equal. These are general relationships, not calculations for this trial. No sample-size justification or prospective assumptions for the trial's original design are reported in the registry record, and a planning illustration based on an observed effect should not be read as evidence that the original study was designed to detect that effect.

In this record, the reported arm-level summaries and model p-value serve different purposes: the summaries describe the outcome by group, and the p-value reports the model's comparison. Neither provides the missing effect estimate and confidence interval for the contrast. Statistical interpretation is consequently limited to the information the registry actually records.

10. Analyses Reported by the Sponsor

The registry reports the sponsor's comparison as a generalized mixed model with p = 0.025. It does not report the model's estimated between-group contrast, the confidence interval for that contrast, or enough model detail to reproduce the analysis from the registry entry. The arm-level mean and confidence interval are available for each group, but the record does not give an estimable unadjusted two-arm recalculation for the continuous endpoint. The stated reason is that values were missing or not estimable for such a comparison.

Accordingly, there is no reported recalculated contrast to place beside the sponsor's model-based result. The arm summaries alone do not provide a like-for-like substitute for the mixed-model comparison, particularly when the registry does not clarify whether those summaries represent the baseline-to-6-month change or another endpoint summary. No new difference or p-value should be inferred from those separate summaries.

The record marks the discrepancy indicator as false. Thus, the stated condition that would make the sponsor's own analysis take precedence is not triggered. Even so, the sponsor's analysis is the only reported inferential comparison available in the registry entry, and its p-value should be identified as such rather than presented as an independently recalculated result. The missing contrast estimate and interval remain important reporting gaps.

11. Multiplicity, Missing Data and Analysis Populations

The registry records 15 outcome measures and 15 analyses, and 9 secondary outcomes are posted. The listed secondary outcomes include parent-reported brushing behavior, blinded dentist ratings of caries and decayed teeth, observed child anxiety and behavior during a dental-office visit, and dentist-reported behavioral compliance. Some outcome descriptions appear more than once in the record. The entry does not explain whether those repeated descriptions correspond to separate time points, analyses, or duplicated registry labels.

When a study examines multiple outcomes or conducts multiple analyses, the chance of seeing at least one apparently notable result can increase. Statistical methods can address that issue by controlling a family-wise error rate, controlling a false discovery rate, or using a prespecified hierarchy of outcomes. These approaches answer different questions and can affect how p-values are interpreted. The registry record does not report whether any multiplicity adjustment was planned or applied, nor does it explain how the primary and secondary analyses were organized.

The record identifies a primary endpoint and reports a primary analysis p-value, but it does not provide a protocol or statistical analysis plan against which the reported analyses can be checked. It is therefore not possible to determine from this entry whether the primary endpoint, analysis model, or multiplicity strategy was set before the results were known. No interim analyses or sample-size justification are reported in the registry record.

The entry also does not report the analysis population for the primary outcome, the number of participants with usable measurements at each assessment, or the amount and pattern of missing outcome data. It does not say whether missingness was handled by likelihood-based model estimation, imputation, a complete-case analysis, or another approach. These choices can affect estimates and uncertainty, especially if the availability of outcome data is related to participants' characteristics or outcomes.

The registry gives arm denominators alongside the primary outcome summaries, but does not clarify their relationship to the randomized population or to the analysis population. It also does not report adherence to the assigned intervention or withdrawals by reason. Consequently, the posted results cannot establish whether the analysis followed an intention-to-treat approach or another population definition. These omissions limit assessment of potential selection and missing-data biases.

12. 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 reports zero serious adverse events in the AIR-P Dental Toolkit condition and zero in the Parent Training condition. It also reports zero other adverse events in each condition and zero deaths in each condition. The at-risk denominators listed for the safety summaries are 59 for the AIR-P Dental Toolkit condition and 60 for the Parent Training condition.

These figures describe what was recorded in the registry for the trial's safety reporting. A zero count means no events of the listed type were recorded among the stated at-risk participants; it does not establish that no event could occur, or that the interventions have no risks. Rare events, incomplete reporting, how events were defined, the observation period, and how actively events were sought all affect what can be learned from zero recorded events.

The registry does not list serious adverse-event terms, which is consistent with the reported zero serious events. It does not provide detailed definitions of serious and other adverse events, follow-up duration for safety, or a narrative description of monitoring procedures. These figures should therefore be read as a concise registry summary, not as a comprehensive characterization of safety across settings or over longer periods.

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

The posted results support a limited statistical description: the registry reports different arm-level mean brushing frequencies and a generalized mixed model p-value of 0.025 for the superiority comparison. The reported means are 4.3 days per week for the AIR-P Dental Toolkit condition and 5.5 days per week for the Parent Training condition. Their confidence intervals describe uncertainty around the separate arm summaries. The registry does not report the estimated between-group contrast or its confidence interval.

This distinction matters because statistical evidence about a comparison does not by itself tell readers how large the difference is or how precisely it has been estimated. The p-value is not a measure of clinical importance and does not tell whether the observed group difference would matter to children or families. The record does not define a threshold for an important change in brushing frequency, so the result cannot be classified against such a threshold from the information reported.

The outcome is parent-reported frequency of successful twice-daily brushing at home during the past week, expressed in days per week. It does not directly measure brushing technique, dental plaque, cavities, or long-term oral health. Other dental and behavioral outcomes are included among the secondary outcomes, but the primary result cannot stand in for those measures. Conclusions about those outcomes require their own reported estimates and uncertainty.

The trial compared a toolkit with the same toolkit combined with parent training. A group difference would relate to the combined intervention package as studied; it would not isolate the effect of parent training apart from any other feature of that package. The record does not report a component-specific comparison or enough implementation information to determine which parts of the training were received or influential.

The result also does not establish whether effects persist beyond the stated baseline-to-6-month timeframe, apply to children beyond the studied population, or are robust to alternative analyses. The registry does not give sufficient detail about missing data, analysis populations, model specification, or multiplicity to resolve those questions. The appropriate reading is therefore bounded: the record reports a model-based statistical comparison and group-level summaries, while leaving the contrast's size and precision and several aspects of analysis and generalizability unclear.

14. Why This Trial Matters Statistically

This trial is statistically interesting because its primary outcome is a repeated behavioral measure tied to a practical home routine, while the registry describes the endpoint as change from baseline to 6 months. Interpreting that result depends on whether the reported summary represents a follow-up level, a change, or a model-based comparison over time. The entry reports arm means and confidence intervals but does not give the estimated contrast or its confidence interval, making the distinction between description and inference especially important.

The intervention comparison also illustrates a common challenge in behavioral trials: the intervention condition combines the established toolkit with a substantial parent-training program. Randomization can support a comparison of assigned conditions, but it cannot separate the effects of individual components when they are delivered together and no component comparison is reported. The trial's broader outcomes—including dental examinations and dental-office behavior—also emphasize that a change in a reported home behavior is not automatically equivalent to a change in oral health.

For statistical readers, the registry's model-based p-value is informative but incomplete without the contrast estimate, uncertainty interval, model specification, and a clear account of missing data and analysis population. Those omissions make it difficult to assess the magnitude and robustness of the comparison from this record alone. The absence of a reported multiplicity strategy also matters in a record with multiple outcome measures and analyses.

The entry further shows why registry interpretation should distinguish completed study status and posted summaries from full reporting of design and analysis methods. Randomization, masking, and a named model provide useful descriptors, but they do not answer how allocation was concealed, who was masked, how outcomes were analyzed, or how incomplete observations were treated. The result is a useful case for reading statistical evidence with attention to both the reported estimate and the reporting gaps that constrain its interpretation.

15. 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 arise specifically from what this trial's registry record does and does not report. The primary endpoint is described as a baseline-to-6-month difference, but the arm summaries are not explicitly identified as change scores or follow-up values. The model-based contrast is not reported, and neither is a confidence interval for that contrast. As a result, readers cannot directly assess the estimated size and precision of the primary comparison.

The outcome is based on parent report of brushing during the past week. The record does not report questionnaire validation, response rates, or how recall and reporting differences were addressed. The measure captures frequency under the stated definition, but does not directly characterize brushing quality or oral-health status. Blinded dentist ratings and dental-visit measures appear among secondary outcomes, yet the registry entries do not give their results.

The intervention condition combined the toolkit with individual parent training, including in-person and phone sessions. The record does not report attendance, adherence, fidelity, or variation in delivery. Consequently, it does not show how consistently the intervention was implemented or which components participants received. The record also does not describe the randomization procedure, allocation concealment, masking procedures, or who was masked, despite identifying the design as randomized and single-masked.

Important analysis details are absent: the registry does not identify the mixed model's specification, the analysis population, missing-data methods, or the relationship between the arm denominators and the full randomized group. It does not report a protocol, statistical analysis plan, interim analyses, or sample-size justification. The registry also does not state whether multiplicity was addressed across the listed outcomes and analyses.

These limitations do not negate the reported result, but they narrow what can be concluded from the registry entry. They restrict evaluation of model assumptions, selective outcome reporting, incomplete data, and the generalizability of the findings. Further details are not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Improving Dental Care and Oral Health in Children With Autism Spectrum Disorder — NCT03003221.
• Registry record last updated 2026-06-15; status last verified 2026-04.
• Results first posted 2026-06-15.
• 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.

19. Record Summary

The registry record describes a completed randomized, single-masked, parallel-group trial sponsored by the University of California, Irvine. It compared the AIR-P Dental Toolkit with the toolkit combined with individual behavioral parent training for children with autism spectrum disorder. Enrollment is listed as 119; the primary-outcome summaries are reported for 59 participants in the toolkit condition and 60 in the parent-training condition.

The primary endpoint was parent-reported frequency of successful twice-daily brushing at home, measured in days per week, with a stated timeframe of change between baseline and 6 months. The arm-level means were 4.3 and 5.5 days per week, with 95% confidence intervals of 3.5 to 5.1 and 4.8 to 6.0, respectively. The registry reports a generalized mixed model p-value of 0.025, but does not report the model's between-group estimate or confidence interval.

The record lists secondary dental and behavioral outcomes, but important details about missing data, analysis populations, multiplicity, and the statistical plan are not reported in the registry record. Safety summaries list zero serious adverse events, zero other adverse events, and zero deaths in each condition.