Stepped Telemental Health Care Intervention for Depression: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Stepped Telemental Health Care Intervention for Depression (NCT01906476), a randomized na study. The primary outcome is Depression.

2. Clinical Question

Statistically, the comparison concerns the measured pattern of depression scores over time in the specified analysis population. It is not simply a comparison of treatment labels: the stepped approach could include either the initial Internet-based treatment with telephone coaching or a later step to therapist-delivered telephone treatment. The registry does not report how many participants actually reached that later step.

The QIDS is described as a 16-item scale with scores from 0 to 27, with higher scores representing worse outcomes. Thus, the direction of a score change has a defined interpretation, but no observed change or contrast is reported in the registry record. Without those estimates and their uncertainty, the comparative question cannot be answered from the registered results alone.

3. Trial Design

The registry describes a randomized, controlled, parallel trial comparing Stepped Care with Telephone Cognitive Behavior Therapy for depression. Allocation is recorded as randomized and masking as single. The registry does not identify who was masked, so the masking label should not be used to infer whether participants, therapists, outcome assessors, or analysts were unaware of assignment.

Enrollment is recorded as 312. The trial status is completed, with a start date of 2015-02 and primary completion recorded as 2018-04. Northwestern University is listed as the sponsor. The phase is recorded as NA; the registry does not provide a further explanation of that entry.

The two arms differ in their treatment pathway. Stepped Care starts with guided Internet cognitive behavioral treatment and telephone-coach support, with the possibility of moving to therapist-delivered telephone cognitive behavior therapy. The comparator arm is described as therapist-delivered telephone cognitive behavior therapy. The registry does not report how treatment escalation was decided or how often it occurred.

Random allocation is intended, in general, to create groups that are comparable on average at the start of a trial. It does not guarantee that every characteristic will be balanced in a particular study. A parallel design assigns participants to an arm and compares the arms over the study period, rather than having each participant receive both interventions in sequence. The registry does not report the allocation ratio, the randomization procedure, or details of any stratification.

The primary outcome assessment schedule includes baseline, midtreatment, end of treatment, and follow-up after treatment. The registry also lists two outcome measures and two analyses, but the reported information does not identify their full analytic specifications. The record describes the primary measure, but does not report the primary-outcome estimates needed to assess the comparative result.

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:

Analysis population for the primary outcome, as defined in the registry: Participants with at least one followup visit which recorded the QIDS post baseline

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 Depression, measured over Baseline, midtreatment, end of treatment, 3 month post treatment, and 6 month post-treatment follow-up.

The primary endpoint is change over time in the Quick Inventory of Depressive Symptomatology (QIDS). The instrument contains 16 items and has a possible score range of 0 to 27; higher scores represent worse outcomes. Scores were scheduled at baseline, midtreatment, end of treatment, and follow-up after treatment. The registry describes the result as a least-squares mean, with a standard error as the dispersion measure, and expresses the measure in units on a scale.

In general, a repeated measure can describe how an outcome changes across a treatment period and follow-up, rather than reducing the assessment to a single moment. A baseline measure can provide context for subsequent scores, while later assessments describe the measure during and after treatment. This is a general explanation of the role of such assessments; the registry does not state the detailed model used to combine them.

A depression symptom scale is used to quantify the measured outcome in a consistent form. Its scores are not, by themselves, a complete account of a participant’s experience or of every possible outcome. The direction is clear from the registry: higher QIDS scores indicate worse outcomes. However, the registry record does not report the observed scores, change estimates, between-arm contrast, or uncertainty for this endpoint. The scale description therefore explains what was intended to be measured, but not what the trial found.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies a least-squares mean as the primary-outcome parameter and a standard error as its dispersion measure. It does not report the numerical estimates or the detailed model that generated them. The explanations below are general statistical context, not a claim about unreported analytic choices in this trial.

What does a least-squares mean represent?

In general, a least-squares mean is a model-based mean estimate. It can represent an outcome mean after accounting for terms included in a statistical model. Its interpretation depends on the model specification, the variables included, and the way the estimate is defined. The registry does not report those model details, so the label alone cannot establish which adjustments were made or how the estimates should be interpreted beyond their role as mean estimates for a continuous measure.

For a continuous outcome such as a scale score, a mean summarizes the central value of observations. A model-based mean may be useful when an analysis incorporates more than one assessment or accounts for specified design features. It is not necessarily identical to a raw arithmetic average. The registry does not provide raw QIDS values or enough detail to compare any model-based estimate with an unadjusted average.

What does the standard error describe?

In general, a standard error describes the estimated sampling variability of an estimate. It helps indicate how precisely an estimate has been measured under the statistical model. It is different from the spread of individual participants’ scores: the standard error concerns uncertainty in the estimated mean or contrast, while a measure of individual variability describes how scores differ across observations. The registry names the standard error as the dispersion measure but does not report its value for the primary outcome.

A confidence interval is commonly formed using an estimate and a measure of its uncertainty. In general, it gives a range of values compatible with the observations and the method used, subject to the assumptions of that method. It does not show the probability that a fixed parameter lies in one particular interval. No primary-outcome confidence interval is reported in the registry record, and none can be inferred from the endpoint description alone.

How do repeated assessments affect the analysis?

The QIDS was scheduled at baseline, midtreatment, end of treatment, and follow-up after treatment. Repeated observations from the same participant are generally related; they should not automatically be treated as if each observation came from a different person. Statistical methods for repeated measurements can account for this dependence, but the specific method and covariance assumptions are not reported in the registry record.

Different analyses can target different questions. A model might estimate changes at particular assessment times, an average pattern across assessments, or a contrast between arms at a chosen time. These are not interchangeable estimands. The record’s note says the primary outcome has several categories or time points and was not reduced to one comparison. That signals that the record does not define a single numerical contrast for this summary. It does not identify the detailed time-specific estimates or the method used to compare them.

What does randomization contribute?

In general, randomization assigns treatment by chance and can reduce systematic differences between arms at baseline. The purpose is to support a fair comparison, subject to implementation and follow-up. Randomization does not prevent missing outcome data, ensure that every participant receives the assigned intervention as intended, or remove the need to define the analysis population. The registry records randomized allocation but does not describe the sequence-generation method, concealment, allocation ratio, or adherence analyses.

The parallel design means the two intervention strategies are compared as separate arms. The stepped strategy allows a possible later treatment step, while the comparator is described as therapist-delivered telephone cognitive behavior therapy. In general, the effect estimated by a trial depends on how treatment assignment, actual treatment received, follow-up, and analysis are handled. The registry does not report enough detail to determine how those features were incorporated into the primary analysis.

How should the planning illustration be read?

The registry describes a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. It says the effect to detect is a stated fraction of the observed effect. Such an exercise can illustrate the number of participants per arm needed under specified assumptions, but it is not a report of the trial’s actual design rationale. The registry does not report a sample-size justification for the trial.

In general, planning estimates depend on the assumed effect, variability, significance level, allocation, and analytic method. A calculation using an observed effect is especially dependent on how that effect was estimated and may not represent a stable expectation for another study. The planning illustration is not observed power. Observed power is not a substitute for interpreting an effect estimate and its uncertainty. Numerical effect fractions and resulting participant counts are not reported in the registry record.

The registry also identifies the sample-size assumptions as an illustration rather than the trial’s own design. This distinction is important: the illustration should not be treated as evidence that the study was planned to detect a particular effect or that it had adequate precision for every outcome and time point. The registered enrollment alone does not establish the precision of the primary comparison.

What can be concluded from the posted analysis information?

The analysis information identifies the outcome type, the summary parameter, the stated dispersion measure, and the specified follow-up population. It does not supply the numerical results or the full statistical plan. General principles can explain how estimates, uncertainty, repeated measures, and planning calculations are usually understood, but they cannot fill in the missing result. A statistical conclusion about the QIDS comparison requires the observed estimates and their uncertainty, which are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome is described across multiple assessment times: baseline, midtreatment, end of treatment, and follow-up after treatment. The registry notes that the outcome has several categories or time points and was not reduced to one comparison. It lists two outcome measures and two analyses, but does not report how the analyses correspond to the outcome measures or assessment times, or whether a specific comparison was designated as the sole confirmatory test.

In general, when several outcomes, time points, or contrasts are examined, the chance of observing a seemingly notable result can change. Methods for controlling or otherwise addressing multiplicity depend on the set of comparisons and the study’s analysis plan. The registry does not report a multiplicity procedure, so it is not possible to determine how the multiple measures or times were handled inferentially.

The specified primary-outcome population includes participants with at least one follow-up visit recording the QIDS after baseline. This definition makes clear that the primary analysis population is not described simply as everyone enrolled. The registry does not report the number meeting that definition, the amount or pattern of missing QIDS data, reasons for missingness, or how missing observations were handled. It also does not report whether sensitivity analyses assessed the effect of missing data.

In general, analyses restricted to observed follow-up information can be affected if missingness is related to the outcome or treatment experience. Different assumptions and methods can address missing data, but the appropriate interpretation depends on the analysis actually used. No particular missing-data method can be attributed to this trial from the registry record. Details of the analysis population beyond the stated criterion, including any additional exclusions, are not reported in the registry record.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry reports serious adverse events for the two arms using an at-risk count and an affected count. For Stepped Care, 157 participants are listed as at risk and 0 as affected. For Telephone Cognitive Behavior Therapy, 155 are listed as at risk and 0 as affected. The corresponding other-adverse-event entries also report 157 at risk and 0 affected for Stepped Care, and 155 at risk and 0 affected for Telephone Cognitive Behavior Therapy. Deaths are recorded as 0 in each arm.

These are reported counts, not estimates of the primary QIDS outcome. No serious adverse event terms are listed. The zero affected counts mean that the registry reports no events in those categories for these entries; they do not establish that no other symptoms, burdens, or experiences occurred. The registry does not describe event-collection procedures, severity grading beyond the serious-event category, duration of safety follow-up, or how events were attributed.

In general, a zero count in a registry entry should be read in the context of the event definition, the population monitored, and the completeness of reporting. The entries alone do not show how safety was assessed or whether all relevant events were captured. They also do not provide a basis for comparing the interventions’ safety beyond the posted counts.

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

The registry documents a randomized comparison of a stepped behavioral intervention and therapist-delivered telephone cognitive behavior therapy, with change in QIDS over time named as the primary outcome. It also gives the direction of the scale, its score range, the assessment schedule, and the analysis population criterion. These details establish what the trial set out to compare and how the primary measure is described.

They do not establish the trial’s primary finding. Numerical QIDS results, an estimated contrast between arms, and measures of uncertainty are not reported in the registry record. Therefore, the direction, magnitude, and precision of any difference cannot be characterized from these entries. The lack of a reported estimate should not be interpreted as evidence that the arms had the same outcome or that either arm had a particular outcome.

In general, statistical evidence and clinical importance are related but distinct questions. An estimate describes the measured difference; its uncertainty indicates how much precision the data provide under the chosen method. Assessing clinical importance also requires context about what magnitude of score change matters, and that threshold is not reported in the registry record. No threshold should be inferred from the scale range alone.

The safety entries report zero serious adverse events, zero other adverse events, and zero deaths in each arm. Those counts describe the registry’s safety entries, but do not establish that there were no other experiences or that all relevant events were captured. They are not a substitute for the primary outcome results.

The sample-size planning description is also not a result. It identifies assumptions for an illustrative calculation and states that the calculation is not the trial’s own design. It cannot be read as observed power, proof of adequate precision, or evidence for a treatment effect. Overall, the registry record supports a description of the design, measure, and listed safety counts; it does not report enough numerical outcome evidence to support a conclusion about comparative QIDS change.

13. Why This Trial Matters Statistically

This trial is statistically notable for the structure of its comparison: the intervention arm is a stepped pathway that begins with guided Internet treatment and telephone-coach support, with possible escalation to therapist-delivered telephone therapy, while the comparator is therapist-delivered telephone therapy. The contrast therefore concerns two treatment strategies, not simply two fixed delivery formats. The registry does not report how many participants followed each step, so the contribution of escalation cannot be assessed from the recorded description.

The primary endpoint is measured repeatedly, from baseline through follow-up after treatment. Repeated measurements can provide information about the course of a continuous outcome, but they also require clear specification of the estimand and analysis. The registry identifies least-squares means and standard errors, yet does not report the model or numerical estimates. Its note that the outcome has multiple time points and was not reduced to one comparison underscores why a single unqualified summary would be insufficient.

The trial also illustrates why a design description, an endpoint definition, and an outcome result are different kinds of evidence. Randomized allocation can support comparative inference when implementation and analysis are appropriate, but the record’s design labels cannot replace the effect estimate and its uncertainty. Similarly, enrollment and an illustrative planning method do not show how precisely the primary effect was estimated.

For statistical readers, the principal issue is the boundary between what is specified and what can be interpreted. The registry gives enough information to describe the comparison, scheduled measure, and analysis-population criterion. It does not report the results needed to judge the direction or precision of the primary contrast. That distinction prevents an endpoint definition or safety count from being mistaken for an efficacy finding.

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 interpreting this trial from the registry record is that numerical primary-outcome results are not reported. There are no QIDS estimates, between-arm contrasts, standard errors, confidence intervals, or p-values in the recorded information. Consequently, the magnitude, direction, and precision of the primary comparison cannot be assessed.

The endpoint is scheduled at multiple times, but the registry does not report time-specific estimates or explain how repeated observations were modeled. It notes that the outcome has several time points and was not reduced to one comparison. The analysis details needed to understand the target contrast, model assumptions, and any adjustment are not reported in the registry record.

The stated analysis population requires at least one post-baseline follow-up visit with a recorded QIDS. The registry does not report how many participants met that criterion, the extent or reasons for missing outcome data, or the method used to address missingness. It also does not report whether sensitivity analyses were conducted. These omissions limit assessment of how the analysis population relates to enrollment and how missing information could affect the result.

The record identifies single masking but does not state who was masked. It does not report the randomization procedure, allocation ratio, adherence to assigned interventions, or the frequency of stepping up from the initial intervention. These details would help characterize trial conduct and the treatment contrast, but should not be guessed from the design labels.

Finally, the safety entries give event counts but not the collection method, follow-up duration for safety, or detail on other event types. The planning illustration is described, but its numerical inputs and outputs are not reported. These are limitations of the information in the registry record; they do not establish that the trial lacked additional analyses or procedures.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Stepped Telemental Health Care Intervention for Depression — NCT01906476.
• Registry record last updated 2023-02-08; status last verified 2023-02.
• Results first posted 2020-04-02.
• 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

NCT01906476 is a completed randomized, controlled, parallel trial sponsored by Northwestern University. It compares Stepped Care—guided Internet cognitive behavioral treatment with telephone-coach support and possible escalation to therapist-delivered telephone cognitive behavior therapy—with therapist-delivered telephone cognitive behavior therapy.

The primary outcome is change over time in the 16-item QIDS, with higher scores indicating worse outcomes. Assessments were scheduled at baseline, midtreatment, end of treatment, and follow-up after treatment. The primary analysis population is described as participants with at least one follow-up visit recording QIDS after baseline.

The registry does not report numerical primary-outcome estimates or their uncertainty. It reports zero serious adverse events, zero other adverse events, and zero deaths in each arm. The record also describes an illustrative sample-size calculation, not the trial’s own design, without reporting its numerical results.