Tailoring Treatment Targets for Early Autism Intervention in Africa: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Tailoring Treatment Targets for Early Autism Intervention in Africa (NCT04068688), a non-randomized na study. The primary outcome is Supported Joint Engagement Measured by the Joint Engagement Rating Inventory (JERI).

2. Clinical Question

The study’s broader rationale concerns whether behaviors targeted by Naturalistic Developmental Behavioral Interventions are cross-culturally valid, and whether a digital outcome measure can capture treatment response. Its COVID-19 adaptation also aimed to develop and implement telehealth coaching and gather implementation and preliminary effectiveness data using mixed methods. Those aims describe the research questions, not findings: the registry record does not report numerical outcome results or conclusions about intervention effects.

3. Trial Design

This is a completed, non-randomized, parallel study sponsored by Duke University. The registry record gives an enrollment of 73 and lists the study start as 2019-09-01 and primary completion as 2021-08-26. The phase is listed as not applicable. Masking is listed as none. These design features matter because participants were not randomly allocated and the listed groups do not form a like-for-like control comparison.

The groups span caregiver-child dyads with ASD receiving different delivery formats, children with typical development, children with developmental delay, early childhood development practitioners and school supervisors, and ESDM supervisors. The ASD delivery groups include in-person intervention only, telehealth intervention only, and in-person intervention followed by telehealth. These categories reflect differing participant types and delivery experiences, rather than interchangeable arms assigned under random allocation.

The primary outcome population is described as participants who completed the study in the ASD, typical-development, and developmental-delay groups. However, the registry record says data were collected only at baseline for the typical-development and developmental-delay groups. Therefore, the stated follow-up comparison is not available across all three groups as described. The record does not state the number with paired baseline and follow-up observations.

The registry lists 22 outcome measures and 22 analyses, alongside secondary measures concerning telehealth acceptability, appropriateness and feasibility, as well as fidelity of intervention delivery. The presence of these measures does not establish which analyses were completed, how they were conducted, or what they found. Detailed allocation procedures, protocol deviations, recruitment by group, and the statistical analysis plan are not reported in the registry record.

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: One participant from the Children with Typical Development group was deemed ineligible post-enrollment.

Only the first 6 of 10 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants who completed the study. Only applicable to Caregiver-Child Dyads With ASD, Children with typical development, and Children with developmental delay groups. Data only collected at baseline for Children with Typical Development and Children with Developmental Delay.

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.

Only the first 6 of 11 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Supported Joint Engagement Measured by the Joint Engagement Rating Inventory (JERI), measured over Baseline and Follow-up (after 12 intervention sessions, up to 3 months).

The primary endpoint is supported joint engagement measured by the Joint Engagement Rating Inventory (JERI). It concerns child engagement states during interaction. The instrument contains eighteen 7-point Likert scale items intended to characterize different aspects of joint engagement. The description says the items span engagement behaviors likely to be observed in typically developing children aged 18 to 30 months and children of similar ages or older with developmental difficulties, including ASD.

On the described rating scale, 1 indicates a very low rating and no time spent in the engagement state, 4 is the midpoint for moderate time and quality, and 7 indicates a very high rating and high-quality engagement. These anchors help explain what a rating represents; they do not establish that ratings are equally spaced or that a particular change has clinical importance.

The endpoint is measured at baseline and at follow-up after 12 intervention sessions, up to 3 months. The registry specifies the median and full range as the summary format. The endpoint is used to describe engagement, a behavior relevant to the study’s rationale about early language-development interventions. The registry record does not report the observed medians, ranges, changes, or uncertainty estimates, nor does it explain how item ratings are combined into the reported score. Typical-development and developmental-delay groups had baseline data only according to the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the median and full range describe?

The primary outcome is summarized using a median and full range. The median identifies the central observation when values are ordered. The full range describes the lowest and highest reported observations. These summaries can be useful when a score distribution is skewed or when the extremes are important to describe. They do not show how observations are distributed between the endpoints, and the range is sensitive to extreme values. Because the registry record does not include the actual medians and ranges, it is not possible to characterize the observed score distributions.

The JERI ratings use a 1-to-7 scale with stated anchors at 1, 4, and 7. Although the registry describes a score on a scale, it does not report how the eighteen items are combined into the endpoint or whether item-level scores are also analyzed. A median of a composite score and a median of an individual item would have different interpretations. The scoring details are not reported in the registry record, so the precise construction of the summary score cannot be assessed.

What would a baseline-to-follow-up contrast mean?

In general, a within-participant change compares an individual’s measurement at baseline with that same individual’s measurement later. This differs from comparing group summaries at separate visits, which may reflect changes in who was measured as well as changes in scores. The registry record names baseline and follow-up after 12 intervention sessions, up to 3 months, but does not report the analysis used to relate measurements across time or the number of participants with both measurements. It also does not report whether follow-up timing varied within the stated window.

The typical-development and developmental-delay groups are described as having baseline data only. Their baseline scores could describe the groups at that assessment, but cannot show change over the intervention period from the information in the record. No follow-up comparison for these groups should be inferred.

Why does non-random allocation matter?

In general, random allocation helps make groups comparable, on average, with respect to measured and unmeasured characteristics at the start of a study. Here, allocation is listed as non-randomized. Differences among the groups may therefore reflect who entered each group, their circumstances, delivery format, or other factors, as well as any intervention-related change. The registry explicitly cautions that the arms are not a like-for-like control comparison. Statistical adjustment, even if used, would not automatically remove differences that were not measured or modeled; whether adjustment was used is not reported in the registry record.

The parallel model indicates distinct groups rather than a reported crossover design. The registry does not provide sufficient detail about assignment rules, group comparability, or the planned statistical model to support a more specific description of the analysis. Masking is listed as none. In general, lack of masking can affect how behavioral ratings are collected or interpreted, particularly when observers know a participant’s group or assessment occasion. The record does not say who rated the JERI or whether the raters knew group assignment.

How should precision and planning be distinguished?

In general, a confidence interval describes a range of values compatible with data under a specified statistical model and reflects uncertainty in an estimate. A point estimate alone does not communicate that uncertainty. The registry record does not report confidence intervals for the primary outcome, so the precision of any estimated difference cannot be evaluated from these registry entries.

The registry describes a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It states that the effect to detect is a stated fraction of the observed effect. Such calculations are planning illustrations, not observed power and not the trial’s own design justification. The numerical planning outputs and fractions are not reported in the registry record. They cannot establish that the study was designed to detect a particular clinically meaningful effect.

In general, over-dispersion means that observed variability is greater than a model assumes. If present, a model that underestimates variability can make uncertainty appear too small. No distributional diagnostic or over-dispersion assessment is reported in the registry record, and the outcome summaries alone do not show which model, if any, was used.

The record lists 22 outcome measures and 22 analyses. When many outcomes or analyses are considered, the chance of at least one apparently unusual result can increase if multiple tests are conducted without an appropriate plan. General approaches include defining a primary analysis in advance and using methods that account for multiple comparisons where suitable. The registry does not report a multiplicity strategy, the exact analysis corresponding to each outcome, or which analyses were prespecified. The number of listed outcomes and analyses does not itself establish how many statistical tests were performed.

10. Multiplicity, Missing Data and Analysis Populations

The record identifies a primary outcome and lists 22 outcome measures and 22 analyses. It also names secondary measures of telehealth acceptability, appropriateness, feasibility, and fidelity of intervention delivery. However, the registry record does not describe a multiplicity adjustment, a hierarchy of secondary outcomes, or a rule for interpreting results across the measures. Without such information and the numerical results, it is not possible to determine how the study handled the risk of chance findings across multiple analyses.

The primary outcome population is defined as participants who completed the study in the ASD, typical-development, and developmental-delay groups. The typical-development and developmental-delay groups had baseline data only. The number of completers, the number assessed at each time point, the reasons for missing observations, and the handling of missing outcome data are not reported in the registry record. It is therefore unclear whether the primary outcome summaries include all enrolled participants, only participants with complete measurements, or another analysis set.

In general, missing data can affect both the estimate and its uncertainty, especially when the likelihood of missingness is related to participant characteristics or outcomes. A complete-case summary may describe those with observed measurements rather than the entire enrolled population. Other methods may make different assumptions. The registry record does not identify an imputation method, sensitivity analysis, or assumption about why data may be missing.

The record refers to participants who completed the study but does not define a separate intention-to-treat population, per-protocol population, or safety population beyond the group-specific adverse-event listings. The analysis populations and exclusions are not reported in the registry record. Accordingly, the enrollment figure should not be treated as the denominator for every outcome or analysis.

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 safety listings report serious adverse events, other adverse events, and deaths separately for caregivers, children, typical-development children, developmental-delay children, early childhood development practitioners and school supervisors, and ESDM supervisors. Each listed group has zero reported serious adverse events, zero reported other adverse events, and zero deaths. The numbers at risk differ by group and are displayed separately for caregivers and children in the ASD delivery groups.

These counts mean that no events of the listed types are recorded for the corresponding groups in the registry record. They do not establish that no event occurred outside the reporting process, and they do not show whether an intervention caused an event. Zero reported events also do not demonstrate that the risk is zero; the number of people at risk and the extent of observation affect how much can be learned from an absence of recorded events.

The safety groups include distinct intervention histories and participant roles. Their counts should be read within each listed group, not combined into a single rate or used as a direct comparison between intervention formats. The registry record does not provide event narratives, definitions of the reporting window, or an explanation of relatedness assessments. No serious adverse-event terms are listed.

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

The record establishes that a completed, non-randomized study evaluated early autism intervention approaches in a South African setting and included an in-person intervention, a telehealth adaptation, and several participant groups. It identifies supported joint engagement measured with the JERI as the primary outcome and describes when that outcome was intended to be measured. It also lists implementation-related secondary outcomes and safety counts.

It does not establish the size, direction, or precision of a change in supported joint engagement, because numerical primary-outcome results are not reported in the registry record. It also does not establish whether telehealth and in-person delivery differed, whether any observed group differences were attributable to intervention, or whether intervention targets were cross-culturally valid. Those conclusions require results and analyses that are not reported in the registry record.

Because allocation was non-randomized and the arms are not described as like-for-like controls, comparisons between groups would be vulnerable to differences in participant characteristics and other factors. The typical-development and developmental-delay groups had baseline data only, so their role in the outcome description is not evidence of change over time. The record also does not supply completion counts or missing-data methods to clarify which participants contributed to each analysis.

The JERI’s rating anchors explain the meaning assigned to low, midpoint, and high ratings, but the record does not report the score construction or evidence of measurement performance in this study. The stated research aims are therefore distinct from demonstrated findings. Similarly, zero recorded adverse events in the listed groups is a report of observed counts, not proof of no risk.

The defensible reading is limited to the registered design, participant categories, outcome definitions, and event counts. Estimates of intervention effect, comparative effectiveness, measurement validity, and the precision of results cannot be recovered from the information reported.

13. Why This Trial Matters Statistically

This study raises a statistical question that is important for behavioral intervention research: whether intervention targets and outcome measures behave meaningfully across cultural settings. A measure can be clearly described and have interpretable rating anchors without its cross-cultural validity being established. Evidence about validity generally depends on how the measure performs in the populations and contexts where it is used. The registry record states this as a motivation, but does not report validation results.

The combination of in-person and telehealth delivery also highlights the need to distinguish implementation questions from effectiveness questions. Acceptability, appropriateness, feasibility, and fidelity concern different aspects of delivering an intervention. They do not substitute for a comparative estimate of change in the primary behavioral outcome. The registry lists these implementation measures, but their numerical results are not reported in the registry record.

Statistically, the non-randomized design makes careful description of group composition and analysis populations especially important. The registry’s warning that the arms are not like-for-like controls helps prevent a simple group contrast from being mistaken for a causal treatment effect. The baseline-only data for the typical-development and developmental-delay groups also define a limit on the comparisons those groups can support.

The record further illustrates why outcome definitions alone are not enough to assess evidence. Readers need observed values, denominators, uncertainty estimates, and transparent missing-data and multiplicity methods to understand the strength and limits of findings. Those details are not reported in the registry record. The study’s aims are relevant to culturally responsive measurement and remote intervention delivery, while the available record supports a description of the questions and methods rather than conclusions about results.

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 principal limitation for interpretation is that numerical primary and secondary outcome results are not reported in the registry record. There are no reported JERI medians or ranges, group-specific changes, confidence intervals, or inferential test results. The stated endpoint and its rating anchors therefore cannot be connected to an observed effect or an estimate of precision.

The non-randomized allocation and explicit warning that arms are not like-for-like controls limit causal interpretation. The listed groups differ in participant type and intervention delivery, and the record does not describe assignment rules or provide baseline comparisons that would show their comparability. Masking is listed as none, while the identity and masking status of JERI raters are not reported in the registry record.

Outcome data were collected only at baseline for the typical-development and developmental-delay groups, according to the record. Their measurements cannot support a reported change over time. The number of participants completing follow-up in the ASD groups, the timing of each assessment, and the handling of missing measurements are also not reported in the registry record.

The scoring details for the JERI endpoint are incomplete: the record describes eighteen items and the scale anchors, but does not state how the score analyzed as the primary endpoint is constructed. The statistical analysis plan, adjustment for multiple outcomes, and analysis populations are not reported in the registry record. The registry lists many outcomes and analyses, but does not provide enough detail to assess how results were selected or interpreted.

Finally, the safety listings record no events in the named groups, but the registry does not provide event narratives, reporting-window details, or relatedness assessments. The absence of reported events should not be interpreted as proof of absence of risk.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Tailoring Treatment Targets for Early Autism Intervention in Africa — NCT04068688.
• Registry record last updated 2022-12-02; status last verified 2022-11.
• Results first posted 2022-12-02.
• 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

NCT04068688 is a completed, non-randomized, parallel study sponsored by Duke University, with enrollment of 73 and no masking. The study began on 2019-09-01 and had primary completion on 2021-08-26. It concerns early autism intervention, including ESDM and telehealth coaching, in South Africa.

The primary outcome is supported joint engagement measured using the JERI at baseline and follow-up after 12 intervention sessions, up to 3 months. The registry describes median and full-range summaries. Typical-development and developmental-delay groups had baseline data only. Numerical outcome results and detailed statistical methods are not reported in the registry record.

The registry lists 22 outcome measures and 22 analyses, including secondary implementation measures. Its safety listings report zero serious adverse events, zero other adverse events, and zero deaths for each listed group. The record cautions that the non-randomized arms are not a like-for-like control comparison.