This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Tracking Depression Symptoms With a Health Chatbot (NCT03990389), a randomized na study. The primary outcome is Longitudinal Change in Depression Severity.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03990389 |
| Phase | NA |
| Condition | Depression, Postpartum, Depression, Unipolar |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 31 |
| Arms in the primary outcome | n/a |
| Lead sponsor | University of Chicago |
| Results posted | Yes |
2. Clinical Question
Population
The study concerns people with postpartum or unipolar depression who were screened for eligibility. Screening included the Edinburgh Postnatal Depression Scale (EPDS), the Rapid Estimate of Adult Literacy in Medicine (REALM-R), and additional questions. The registry record does not report detailed eligibility criteria or participant characteristics, so the enrolled population cannot be described more precisely from this information. It also describes a provider subject cohort, but does not report how that cohort relates to the participant groups in the primary outcome analysis.
Intervention
The intervention is chatbot care added to usual care. The study summary describes a chatbot intended to support measurement-based care through Facebook Messenger. The intervention description and the arm description indicate that participants in chatbot care also received monthly reminder phone calls from a research coordinator. The registry record does not report how often participants used the chatbot or whether they completed the planned surveys.
Comparator
The comparator is usual care. Participants in this group were also to receive monthly reminder phone calls from a research coordinator. The registry description of usual care does not provide a full account of the care participants received outside those calls, so the comparator should not be interpreted as a single standardized treatment.
Primary question
The primary question concerns longitudinal change in depression severity, measured with the EPDS over a 3 months timeframe. The registry describes the primary outcome as a score on a scale and identifies missing values at month 1, month 2, and month 3. It does not report an estimated difference between groups, a change estimate for either group, a confidence interval, or a significance test. Consequently, the record identifies the intended outcome but does not establish what the posted primary outcome results show.
3. Trial Design
The study is described as randomized, with a parallel-group model and no masking. Its planned comparison is usual care against usual care with additional chatbot care. The registry record lists 31 participants as the enrollment and marks the study completed. It identifies University of Chicago as the sponsor. The record lists the study start as 2019-02-20 and primary completion as 2020-02-29.
The group descriptions name a Usual Care Group and a Chabot Care Group. The spelling varies between the intervention description and the group label; the study summary calls the added intervention chatbot care. The usual-care description mentions screening and monthly reminder calls. The chatbot-care description mentions the same screening and calls, plus an email with a survey link, but the description ends before fully explaining the survey process. These descriptions indicate the intended contrast, but do not establish how consistently either arm received its planned activities.
The record also lists a Provider Subject Cohort and says that 20 provider subjects from each study site will be enrolled. The number of study sites and the relationship of this cohort to the participant enrollment and primary outcome analysis are not reported in the registry record. The listed participant enrollment should therefore not be combined with the provider cohort description or treated as a count for any particular analysis population.
The registry record does not report the randomization method, allocation concealment, the actual allocation ratio, or procedures for handling deviations from assigned care. It does not report a detailed statistical analysis plan or a sample-size justification for the trial. The assumptions described for planning illustrations are not a statement of the trial’s original design rationale. These gaps limit how fully the design and conduct can be reconstructed from the registry entry alone.
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: Excluded Patient Subjects (n = 35) * EPDS \< 12 (n = 12) * Spanish speaking only (n = 2) * No Facebook account (n=1) * Consent not returned (n=18) * Refused (n = 2) No data was collected for the providers as no results related to the providers were reported.
Overall Study
| Stage | No Intervention: Usual Care Group | Experimental: Chabot Care Group |
|---|---|---|
| STARTED | 15 | 16 |
| COMPLETED | 12 | 16 |
| NOT COMPLETED | 3 | 0 |
Reasons for not completing this period:
| Reason | No Intervention: Usual Care Group | Experimental: Chabot Care Group |
|---|---|---|
| Lost to Follow-up | 3 | 0 |
Analysis population for the primary outcome, as defined in the registry: missing values at month 1, month 2, and month 3
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.
| Characteristic | Unit | No Intervention: Usual Care Group | Experimental: Chabot Care Group | Total |
|---|---|---|---|---|
| Participants analysed | 15 | 16 | 31 | |
| Age, Continuous | years | 31.3 (SD 4.4) | 33.1 (SD 2.7) | 32.2 (SD 3.7) |
| Sex: Female, Male: Female | Participants | 15 | 16 | 31 |
| Sex: Female, Male: Male | Participants | 0 | 0 | 0 |
| Race and Ethnicity Not Collected | Participants | — | — | 0 |
6. Primary Endpoint
The registry lists the primary outcome as Longitudinal Change in Depression Severity, measured over 3 months.
| Element | Registry entry |
|---|---|
| Result type | Mean |
| Measure of spread | Standard Error |
| Unit | score on a scale |
| Endpoint classification | Continuous Endpoint |
The primary endpoint is longitudinal change in depression severity, assessed with the Edinburgh Postnatal Depression Scale. The registry describes the outcome as a mean score on a scale, with a 3 months timeframe. The EPDS range is 0-30, and the registry notes that risk of depression goes up with an increasing score. Thus, the measure records symptom severity on a scale; it is not itself a diagnosis or a direct measure of every aspect of a participant’s experience.
Repeated symptom measurements can be used to describe how scores change over follow-up and whether the patterns differ between groups. This is the general rationale for a longitudinal endpoint: it can represent change over time rather than relying on a single follow-up measurement. However, the registry record does not describe the specific visits included in the primary analysis, the baseline measurement’s role, or the statistical model used to estimate change. It identifies missing values at month 1, month 2, and month 3, but does not report how those missing observations were handled.
The registry record does not report a numerical primary outcome estimate, group-specific means, uncertainty intervals, or a test of the between-group contrast. The outcome’s definition can therefore be explained, but the magnitude and precision of any treatment-group difference cannot be interpreted from the record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record establishes the study’s intended comparison and primary outcome, but it does not report a numerical result for longitudinal change in EPDS score. There is no posted group contrast, change estimate, confidence interval, or primary-outcome p-value in the registry record recorded for this analysis. Without those quantities, it is not possible to describe the direction, magnitude, or statistical precision of the primary comparison. The study’s completion status and enrollment count do not substitute for an outcome estimate.
The primary outcome is described as longitudinal, and the record notes missing values at month 1, month 2, and month 3. A longitudinal analysis can use repeated measurements to assess patterns over time, but the interpretation depends on the analysis population, model, assumptions, and missing-data approach. Those details are not reported in the registry record. It is therefore not possible to determine whether the intended contrast was an overall average change, a difference at a particular visit, or another summary of the trajectories.
The registry lists secondary outcomes for depression severity, side-effect burden, maternal function, maternal confidence, and medication adherence. It also lists 6 outcome measures and 5 analyses. The record does not provide the secondary outcome estimates or explain how the listed analyses correspond to the outcomes. No conclusions about those outcomes can be drawn from their names alone.
The safety record lists 0 serious adverse events and 0 other adverse events in the Usual Care Group, with 15 at risk in each category. It lists 0 serious adverse events and 0 other adverse events in the Chabot Care Group, with 16 at risk in each category. Deaths are listed as 0 in both groups. These are the recorded event counts, not proof that no events occurred outside the registry’s definitions, observation period, or reporting process. The registry does not report further detail about event collection or follow-up.
General statistical interpretation: an estimate describes the size and direction of a contrast, while a confidence interval describes the range of values compatible with the data under the analysis assumptions. A p-value alone would not describe effect size or clinical importance. In this case, the relevant primary estimates and uncertainty measures are not reported in the registry record, so neither statistical significance nor the practical importance of the group difference can be assessed.
9. Statistical Methods Explained
Randomized parallel-group design means participants are assigned to one of the study groups and followed in those groups. Randomization is intended, in general, to make treatment groups comparable on average at assignment. The registry identifies the study as randomized, but does not report the sequence-generation method, concealment procedure, or actual allocation ratio. The arm-specific risk counts in the safety record are 15 and 16; they describe the safety denominators recorded there and do not by themselves give the analysis population for the primary outcome.
What does a longitudinal outcome represent?
A longitudinal outcome consists of repeated measurements over time. General statistical practice can represent such data by estimating change within each group, comparing groups at follow-up, or modeling the pattern across visits. These approaches answer related but not identical questions. A group-by-time contrast, for example, addresses whether the trajectory differs between groups, while a single-visit contrast addresses the difference at that visit. The registry calls the primary outcome longitudinal change in depression severity and gives a 3 months timeframe, but does not report which estimand or model was used. The primary comparison therefore cannot be reconstructed from the outcome label alone.
The registry describes the endpoint as a mean score and its dispersion as a standard error. A mean is a summary of the score distribution. A standard error describes uncertainty in an estimated mean, rather than the spread of individual scores. General statistical interpretation should distinguish a standard error from a standard deviation: the former concerns the precision of an estimate, while the latter describes variability among observations. No endpoint means or standard errors are reported in the registry record recorded for this analysis.
How should uncertainty be interpreted?
In general, a confidence interval combines an estimate with a range expressing uncertainty under the model and assumptions used. Its width depends on factors such as the amount of information and the variability in the data. A narrow interval indicates more precise estimation than a wide interval, but neither width nor location alone determines whether a difference matters to patients or clinicians. The primary outcome’s confidence interval is not reported in the registry record, so precision cannot be assessed. A p-value, if reported, would address compatibility with a specified null hypothesis under the analysis assumptions; it would not measure the size or clinical importance of an effect. No primary-outcome p-value is reported.
The record identifies missing values at month 1, month 2, and month 3. General statistical methods for repeated measurements may use all available observations, impute missing values, or restrict analysis to participants with specified measurements. Each choice relies on assumptions and can affect the estimate. The registry record does not report the amount or pattern of missingness, the method used to address it, or sensitivity analyses. The phrase identifying missing values does not disclose how many participants had missing measurements or why they were missing.
What do the planning assumptions mean?
The record describes sample-size assumptions for planning illustrations: two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. The illustrations are based on fractions of an observed effect. These assumptions describe hypothetical planning scenarios, not the trial’s own design justification, and do not supply observed treatment effects. Any sample-size figure associated with such a scenario is a planning illustration, not observed power and not evidence that the study had a particular probability of detecting an effect. The registry does not report a trial-specific sample-size justification.
General power calculations depend on a target effect, variability, significance level, allocation, and design assumptions. If an effect is chosen from an observed estimate, the resulting calculation is conditional on that estimate and should not be read as an independent finding. In this study, the primary effect estimate is not reported in the registry record, so the underlying effect cannot be assessed from the registry entries. No new effect estimate or sample-size calculation should be inferred from the enrollment count.
What is not available about the analysis?
The record lists 6 outcome measures and 5 analyses, but it does not describe the statistical model, covariate adjustment, analysis population, handling of protocol deviations, or the mapping between each outcome and analysis. The number of outcome measures and analyses does not, by itself, establish which comparisons were prespecified or how results were selected for reporting. The registry also lists secondary outcomes, but provides no numerical results for them in the registry record recorded for this analysis.
General methods principle: when many outcomes or comparisons are tested, the chance of at least one false-positive finding can rise if testing is not accounted for. Multiplicity procedures can control that risk, but their appropriateness depends on the scientific questions and analysis plan. The record does not report whether any adjustment was planned or applied. It is therefore not possible to characterize the inferential approach as adjusted or unadjusted.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists 6 outcome measures and 5 analyses. It names secondary outcomes for change in depression severity, side-effect burden, maternal function, maternal confidence, and medication adherence. The primary outcome is longitudinal change in depression severity. The record does not explain how the listed outcome measures and analyses correspond to one another, which comparisons were designated primary or secondary in the analysis, or whether any outcomes were measured at multiple visits and treated as separate tests.
General statistical context: when researchers test multiple outcomes or contrasts, the chance of at least one apparently unusual result can increase. A prespecified testing hierarchy or multiplicity adjustment can address this issue, but the choice depends on the intended questions and analysis plan. Whether the study used such a strategy is not reported in the registry record. No adjustment should be assumed from the number of outcomes or analyses alone.
The primary outcome description identifies missing values at month 1, month 2, and month 3. It does not report the number of missing observations, the reasons for missingness, or how missing values were handled. The analysis population is also not reported in the registry record. In general, missing-data methods depend on assumptions about why values are missing; those assumptions can affect the interpretation of longitudinal results. Without the method and supporting details, the possible influence of missing data cannot be evaluated.
The record does not report whether the primary analysis followed assignment regardless of intervention use, required a minimum level of chatbot participation, or used another population definition. It also does not report sensitivity analyses. Thus, neither the treatment-policy interpretation nor the robustness of the outcome analysis can be established from the registry entry.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| No Intervention: Usual Care Group | 0 | 15 | 0.0% |
| Experimental: Chabot Care Group | 0 | 16 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| No Intervention: Usual Care Group | 0 of 15 |
| Experimental: Chabot Care Group | 0 of 16 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| No Intervention: Usual Care Group | 0 of 15 |
| Experimental: Chabot Care Group | 0 of 16 |
The registry safety record lists 0 serious adverse events, 0 other adverse events, and 0 deaths for the Usual Care Group. The recorded at-risk count is 15 for serious and other adverse events. For the Chabot Care Group, the record lists 0 serious adverse events, 0 other adverse events, and 0 deaths, with 16 at risk for serious and other adverse events. No serious-event terms are listed.
These entries describe the counts reported in the registry, not a complete account of how safety information was collected. A zero count means that no events were recorded in that category in the registry entry; it does not establish that participants experienced no symptoms or events of any kind. The record does not report the safety follow-up period, event definitions, collection procedures, or whether event ascertainment was active or passive. It also does not provide detail about non-serious events beyond the zero count.
These figures should be read separately from the primary outcome. They do not indicate whether depression scores changed, whether participants used the chatbot, or whether the study was sufficiently informative to rule out uncommon events. The registry record does not report enough information to assess the completeness or precision of the safety comparison.
12. What the Result Does, and Does Not, Mean
The study addresses a practical measurement question: whether adding chatbot-supported symptom tracking to usual care can be evaluated as a way to support measurement-based care for depression. Its randomized, parallel-group design provides a framework for comparing usual care with usual care plus chatbot care. The registry identifies an EPDS-based longitudinal primary outcome and lists secondary outcomes relevant to symptoms, side-effect burden, maternal function, maternal confidence, and medication adherence.
However, the registry record does not report numerical results for the primary outcome or the secondary outcomes. It does not provide estimates of change, between-group contrasts, uncertainty intervals, or significance tests. Therefore, the record cannot establish the direction or size of a difference between groups, how precise any estimate was, or whether the data were compatible with effects in either direction. A completed study and a stated enrollment count do not answer those questions.
The safety entries report no serious adverse events, other adverse events, or deaths in either listed group. Those counts are useful as a description of what the registry records, but they do not show that no event occurred under all possible definitions or observation conditions. The reporting details needed to assess that limitation are not reported in the registry record.
General interpretation: a statistically detectable difference, if one were reported, would not automatically establish practical importance. Conversely, a non-significant result would not establish that two approaches are equivalent. Such judgments require effect estimates, uncertainty, the outcome’s meaning, and an appropriate threshold for importance. The registry entries available for this study do not support either conclusion. They describe the research question and design, but not enough outcome evidence to determine what the intervention comparison found.
13. Why This Trial Matters Statistically
This trial is statistically relevant because it tests a measurement-support approach rather than defining the question only as a comparison of two named treatments. The planned contrast is usual care with additional chatbot care against usual care. Its primary outcome focuses on repeated depression severity scores, making the choice of longitudinal estimand, analysis model, and missing-data strategy central to interpretation.
The registry also highlights a common challenge in studies using digital tools: assignment to an intervention does not establish how much participants used it or how consistently it was delivered. The record describes chatbot care and survey-related communication, but does not report usage, completion, or engagement results. Those details would help distinguish the effect of being assigned the additional support from the effect among participants who actually used it. They are not reported in the registry record.
The endpoint is a score with a defined range and direction, while the study lists several additional outcome domains. Interpreting multiple outcomes requires clarity about which question is primary, how repeated measurements are summarized, and whether multiplicity was addressed. The registry lists outcome and analysis counts but does not report the analysis plan or multiplicity strategy. This makes transparent reporting especially important for evaluating the evidence.
Finally, the registry illustrates why design information and results must be read separately. Randomization and completion describe aspects of the study, but they do not reveal the treatment contrast’s estimated size or precision. Without primary and secondary estimates, statistical interpretation remains limited to the design and outcome definitions. The record therefore identifies a meaningful methodological question while leaving the central empirical comparison unresolved.
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 main limitation for interpretation is that the registry record does not report primary or secondary outcome estimates. There are no group-specific EPDS summaries, change estimates, between-group contrasts, confidence intervals, or p-values in the registry record recorded for this analysis. The planned endpoint can be described, but its observed result and precision cannot be evaluated.
The population is also incompletely characterized. The record lists postpartum and unipolar depression, screening with EPDS and REALM-R, and an enrollment of 31, but does not report detailed eligibility criteria or participant characteristics. It does not clarify the relationship between the participant groups and the provider subject cohort. The description of provider enrollment refers to each study site, while the number of sites is not reported in the registry record.
Intervention delivery is not fully described. The study summary identifies Facebook Messenger, while the arm description mentions an email link to surveys and ends before fully detailing the process. Actual chatbot use, survey completion, and intervention fidelity are not reported. The usual-care description also does not characterize care beyond the stated study activities.
For the primary analysis, the registry does not report the statistical model, analysis population, missing-data method, covariate adjustment, or sensitivity analyses. It identifies missing values at month 1, month 2, and month 3 without giving their number or pattern. The relationship between the listed outcome measures and analyses, and any multiplicity strategy, is also not reported. These gaps prevent assessment of the assumptions behind any estimated effect.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Tracking Depression Symptoms With a Health Chatbot — NCT03990389.
- Registry record last updated 2022-05-12; status last verified 2022-04.
- Results first posted 2022-05-12.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
NCT03990389 is a completed randomized, unmasked, parallel-group study sponsored by University of Chicago. It lists enrollment of 31 and compares usual care with usual care plus chatbot care. The primary outcome is longitudinal change in EPDS score over a 3 months timeframe; the scale range is 0-30, with increasing scores associated with increased risk of depression.
The registry lists secondary outcomes concerning depression severity, side-effect burden, maternal function, maternal confidence, and medication adherence. It records 0 serious adverse events, 0 other adverse events, and 0 deaths in each group, with 15 at risk in the usual-care group and 16 in the chatbot-care group for serious and other adverse events.
Numerical efficacy results, analysis details, missing-data procedures, and multiplicity methods are not reported in the registry record. The study’s design and intended measurement question can be described, but the primary comparison’s direction, magnitude, and precision cannot be assessed from these registry entries.