Cali Sin Tos Aim 2: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Cali Sin Tos Aim 2 (NCT06338462), a na na study. The primary outcome is Implementation of the mHealth Strategy to Assess Feasibility..

2. Clinical Question

The broader objective was to design and iteratively adapt a home-based mobile health and oral testing strategy for tuberculosis contact tracing in Cali, Colombia. The study was intended to assess implementation characteristics, including feasibility, acceptability, usability, and appropriateness, in preparation for a future implementation-effectiveness trial. It was not registered as a trial comparing the strategy with usual care or another intervention.

The primary endpoint focuses narrowly on whether a person engaged with the chatbot, not whether contact tracing led to a diagnosis, treatment, or other health outcome. The record does not report a numerical estimate of engagement, nor does it describe a comparator-based effect. The clinical question is therefore about an implementation measure and the feasibility of using the approach, rather than a causal question about clinical benefit.

3. Trial Design

The registry describes a completed, single-group study sponsored by Yale University. Its title is “Cali Sin Tos Aim 2,” and its official title is “Mobile Health and Oral Testing to Optimize Tuberculosis Contact Tracing in Colombia.” Tuberculosis is the listed condition. Enrollment is recorded as 269. The study start date is 2024-03-28, and primary completion is recorded as 2024-09-11. The status is completed.

The intervention is described as “Mobile Health and Oral Testing.” It combines a chatbot with oral testing and was to be adapted through three phases of participatory co-design. The brief summary describes an iterative, community-engaged process intended to optimize feasibility, acceptability, usability, and appropriateness. The study protocol includes baseline contact-tracing procedures and a nominal group technique for determining adaptations to the mobile health strategy.

The design is not a randomized comparison: allocation is recorded as NA, masking as none, and the model as single group. No control group or like-for-like comparison is reported. The registry lists phase as NA. These classifications mean that the study is principally described as an iterative implementation and adaptation effort, rather than a confirmatory efficacy comparison.

The registry record does not report a sample-size justification, a statistical analysis plan, or interim analyses. Although a planning illustration may describe assumptions for a hypothetical two-arm calculation, those assumptions are not evidence that this single-group study was designed or analyzed as a two-arm trial. No numerical results or analyses are reported in the record for the primary outcome. The completion status alone does not establish what the observed engagement rate was or whether the adaptation objectives were achieved.

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: No randomization in this trial. Single arm interventional study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Number of individuals who successfully received the chatbot. 52 did not receive the chatbot. We did not collect reasons why the chatbot was not received.

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 Implementation of the mHealth Strategy to Assess Feasibility., measured over within 14 days of invitation.

The primary outcome is titled “Implementation of the mHealth Strategy to Assess Feasibility.” Its stated timeframe is within 14 days of invitation. Feasibility is defined as the proportion of index persons with tuberculosis and household contacts at risk of tuberculosis who engaged with the chatbot, where engagement means responding to at least one chatbot message.

This endpoint represents a practical step in delivering a digital contact-tracing approach: whether the intended participants interact with the chatbot at all. It can describe reach or initial uptake, but it does not by itself establish whether a person completed contact tracing, received or understood oral testing, obtained a diagnosis, or experienced a health benefit. It is an implementation measure, not a clinical efficacy endpoint.

There is a reporting distinction within the record. The outcome parameter is described as a count of participants, while the endpoint definition is a proportion. The population note refers to individuals who successfully received the chatbot and says that 52 did not receive it, but does not state how nonreceipt was handled in calculating the defined proportion. The numerator, denominator, estimate, and uncertainty interval are not reported in the registry record. Accordingly, the endpoint’s intended meaning can be explained, but its observed value and precision cannot be assessed from the registry information.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary endpoint capture?

The endpoint is an engagement indicator: a participant is defined as engaged after responding to at least one chatbot message within 14 days of invitation. In general, a binary endpoint records whether a specified event occurred for each person. When reported as a proportion, it describes the share of the defined analysis population meeting that event definition. It does not convey how many messages were exchanged, the quality of engagement, or whether later steps in contact tracing were completed.

The registry labels the outcome parameter as a count of participants, although its description defines feasibility as a proportion. Those descriptions are related but not interchangeable as reported results: a count gives the number meeting a criterion, whereas a proportion also requires a defined denominator. The record does not state the observed count, the denominator used for the proportion, or the resulting estimate. It also does not explain how participants who did not receive the chatbot were treated in the endpoint calculation.

Why does the denominator matter?

For an implementation measure, the denominator determines what question the estimate answers. An estimate among everyone invited would reflect both delivery of the chatbot and subsequent response. An estimate among only people who received it would focus on response after delivery. These are general distinctions, not claims about which approach this study used. The registry says that 52 individuals did not receive the chatbot and that reasons were not collected, but does not specify the primary endpoint denominator or the handling of these cases.

Nonreceipt should not automatically be treated as a negative engagement outcome or as ordinary missing data. It may be relevant to reach and implementation, but the record does not state the analysis rule. Likewise, a missing response cannot be classified from the record without knowing whether delivery was successful and how outcome ascertainment was handled. Clear reporting would distinguish invitation, receipt, and response, and would state how each contributes to the endpoint.

What would an uncertainty interval contribute?

In general, a confidence interval communicates the range of values reasonably compatible with an estimate under the statistical model and its assumptions. A narrow interval indicates greater precision, while a wide interval indicates more uncertainty. For a proportion, its interpretation depends on the definition of the analysis population and the method used to calculate the interval. The registry record reports no proportion and no confidence interval, so the precision of the endpoint cannot be assessed.

A hypothesis test would address a specified statistical comparison or hypothesis. The record does not report a comparator, a null hypothesis, a test statistic, or a p-value. A significance test is not necessary to describe a feasibility proportion, and a p-value alone would not communicate its practical magnitude or precision. No such result is available in the registry record.

How does the study design shape inference?

A single-group design can describe outcomes among participants exposed to the strategy, but it does not provide a concurrent control group for estimating what would have happened under a different approach. General statistical comparisons require a defined contrast, and causal interpretation depends on design and assumptions that support that contrast. Here, allocation is recorded as NA, masking as none, and the model as single group. The record therefore does not support a randomized treatment-effect comparison.

The intervention was intended to be adapted through participatory co-design. In general, adaptation can make an approach more appropriate to its setting, but it also means that the delivered strategy may evolve during implementation. The record describes three phases of co-design, but does not report phase-specific outcome results or analyses. It is not possible to assess differences across phases or attribute any change to a particular adaptation.

What can be inferred from the analysis record?

The registry lists zero analyses. It does not report an analysis plan, adjustment variables, handling of repeated observations, missing-data methods, or interim analyses. These methods should not be inferred from the endpoint type or study objective. The registry also does not report a sample-size justification. Any planning illustration based on hypothetical assumptions is not a report of the study’s analysis or observed statistical power.

In general, statistical methods should match the endpoint and the design. A proportion estimate needs a clearly specified population and denominator; uncertainty assessment needs a stated method; and a comparative effect needs a defined comparator. Without the reported result and analysis details, the available information supports a description of intended measurement, but not reconstruction of a numerical result or assessment of its statistical precision.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Mobile Health and Oral Testing")
n      <- c(217)
events <- c(112)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Mobile Health and Oral Testing  217  112  51.6  44.8  58.4  45.0  58.2  45.0  58.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures, but the details of the other outcome are not reported in the registry record. It lists zero analyses. The record does not describe a multiplicity strategy, such as prioritizing outcomes for formal hypothesis testing or adjusting for multiple comparisons. Since no analysis results or tests are reported, it is not possible to assess whether multiplicity affected an inferential conclusion.

The population note reports that 52 individuals did not receive the chatbot and that reasons for nonreceipt were not collected. It does not state whether those individuals were excluded from the primary endpoint, counted as not engaged, or handled in another way. No other missing-data rules are reported. Nonreceipt is a delivery issue that may be important to implementation; it should not be conflated with a missing response among people who received the chatbot unless the analysis definition supports that treatment.

The registry does not define the analysis population beyond describing individuals who successfully received the chatbot, nor does it state how that description relates to the endpoint’s defined population of index persons and household contacts at risk. The numerator, denominator, and missing-outcome count are not reported. The extent and potential direction of missing-data bias therefore cannot be evaluated from this information.

The record does not report subgroup analyses, adjusted models, interim analyses, or a prespecified statistical analysis plan. These omissions do not establish that such methods were absent from all study procedures; they mean that they are not reported in the registry record. Without the endpoint analysis population and results, readers cannot determine how many observations contributed to the primary estimate or whether the stated feasibility question was analyzed as planned.

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

The study addresses an implementation question: whether intended users respond to a chatbot used as part of a home-based tuberculosis contact-tracing strategy. The endpoint definition makes the event of interest clear at a basic level: a response to at least one chatbot message within 14 days of invitation. That definition helps distinguish initial engagement from broader outcomes such as completion of contact investigation or clinical benefit.

The registry does not report the observed engagement result. It therefore cannot establish from the posted information whether engagement was common or uncommon, whether the strategy met a feasibility target, or how precise an estimate would be. The enrollment of 269 and the report that 52 individuals did not receive the chatbot do not determine the endpoint result. No numerator, denominator, or uncertainty estimate is recorded.

The study also does not establish a comparative effect. It is registered as single group, with no control comparison. Any descriptive experience with the approach would be specific to the participants, setting, and implementation process recorded for the study; it would not by itself show what would have occurred under another contact-tracing strategy. The planned co-design and adaptation process is central to the study’s purpose, but phase-specific outcomes and the final adapted strategy are not described in the results information.

The record describes preparation for a future, appropriately powered implementation-effectiveness trial. That stated aim should not be confused with evidence that the present study was powered for comparative effectiveness. The record does not report a sample-size justification or analysis plan. Likewise, the zero adverse-event and death entries are registry counts, not a complete account of how safety was monitored or ascertained.

In short, the record explains what feasibility was intended to mean and how the intervention was to be developed. It does not provide the numerical results needed to judge observed engagement, its uncertainty, or comparative performance. No treatment recommendation or conclusion about clinical benefit follows from the registry information.

13. Why This Trial Matters Statistically

Implementation research asks whether an approach can be delivered and used in its intended setting, not only whether it changes a clinical outcome under controlled conditions. This study’s focus on adapting a chatbot and oral testing for tuberculosis contact tracing makes the distinction important. A response to a chatbot is a meaningful implementation step, but it is not equivalent to completing contact tracing or achieving a health outcome.

The endpoint also illustrates why operational definitions matter. “Engagement” is given a specific event definition: responding to at least one message within 14 days of invitation. Yet the record labels the measure as a participant count while describing feasibility as a proportion. Reporting the numerator, denominator, and treatment of people who did not receive the chatbot would connect the operational definition to a reproducible estimate.

The study’s participatory co-design approach is another statistically relevant feature. Adaptation is part of the intervention-development process, so results would need to be interpreted in relation to the version of the strategy and the stage of implementation. The registry describes three co-design phases but does not report phase-specific outcomes. A future evaluation could benefit from clearly linking each measured outcome to the intervention version and implementation stage.

Finally, the single-group design aligns with a descriptive and developmental purpose but cannot provide a controlled estimate of comparative effect. The stated intention to prepare for a future implementation-effectiveness trial separates feasibility work from confirmatory evaluation. The registry record conveys that distinction, while the lack of posted results limits assessment of what the feasibility work found.

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 interpreting outcomes is that the registry record contains no numerical primary result. It reports no engagement estimate, numerator, denominator, or uncertainty interval. As a result, the central feasibility question cannot be answered quantitatively from the posted information.

The endpoint’s population and analysis denominator are not fully specified. Its definition refers to index persons with tuberculosis and household contacts at risk, while the population note describes individuals who successfully received the chatbot and reports that 52 did not receive it. The record does not explain how these descriptions relate or how nonreceipt was handled. Reasons for nonreceipt were not collected, limiting interpretation of delivery barriers and the contribution of nonreceipt to implementation performance.

The measure is described both as a participant count and as a proportion. Although a proportion is the stated feasibility construct, the registry does not report the denominator needed to interpret it. It also does not report how engagement was ascertained or whether all eligible people had outcome data.

The single-group design has no reported comparator, and masking is recorded as none. This design cannot support a like-for-like comparison of the mobile health and oral testing strategy with another approach. The intervention was intended to be adapted through co-design, but the record does not provide phase-specific results or detail how outcome measurement related to adaptations.

Further limitations in the posted information include the absence of a reported statistical analysis plan, sample-size justification, missing-data approach, and details for the other registered outcome measure. The record lists zero analyses. These gaps restrict assessment of precision, bias, and the relationship between the stated aims and the results. They should be understood as limitations of the information reported in the registry, not proof that no additional study documentation exists.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Cali Sin Tos Aim 2 — NCT06338462.
• Registry record last updated 2025-10-09; status last verified 2025-09.
• Results first posted 2025-08-28.
• 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

NCT06338462, “Cali Sin Tos Aim 2,” is a completed, single-group study sponsored by Yale University. Its official title concerns mobile health and oral testing to optimize tuberculosis contact tracing in Colombia. Enrollment is recorded as 269; the study start date is 2024-03-28, and primary completion is 2024-09-11. Phase is listed as NA, allocation as NA, and masking as none.

The intervention combines a chatbot and oral testing, with an iterative participatory co-design process intended to adapt the approach. The primary feasibility endpoint concerns chatbot response within 14 days of invitation. The record states that 52 individuals did not receive the chatbot and that reasons were not collected.

The study is intended to inform a future implementation-effectiveness trial. The registry lists two outcome measures and zero analyses, but does not report the primary outcome result, its denominator, or an uncertainty estimate. It records zero serious adverse events, other adverse events, and deaths for the intervention group.