Text Messaging and Cognitive Behavioral Therapy for Depression: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Text Messaging and Cognitive Behavioral Therapy for Depression (NCT01083628), a non-randomized na study. The primary outcome is PHQ-9 Depression Symptoms.

2. Clinical Question

The question concerns an adjunct to a behavioral therapy program, rather than a comparison between different forms of psychotherapy. The two groups both received group CBT; the described difference was whether automated text messaging accompanied it. The record identifies the intended purpose of the messages as monitoring mood and reminding participants about session content. It does not report message frequency, participant engagement with the messages, or how engagement was defined or measured.

The comparison was nonrandomized. That matters when interpreting any apparent difference: group membership may be associated with participant or treatment circumstances other than the messaging adjunct. The record does not report how those circumstances were measured or addressed. More importantly, it does not give group-level PHQ-9 results or an estimated difference, so the registered question cannot be answered from the posted outcome information.

3. Trial Design

This was a completed, parallel-group behavioral study sponsored by the University of California, San Francisco. The registry lists an enrollment of 85, a start date of 2014-01, and primary completion in 2016-09. The allocation is classified as nonrandomized, and masking is listed as none. The record does not report a clinical trial phase beyond listing it as not applicable.

One group received group CBT for depression with MoodText. The description says the therapy used the BRIGHT manual and included automated messaging for mood monitoring and reminders of session content. The other group received standard group CBT for depression using the BRIGHT manual. Thus, both groups shared the described CBT approach, while the messaging adjunct distinguished the intervention.

The registry lists PHQ-9 depression symptoms at 16 weeks as the primary outcome, reported as a mean with standard deviation. Attendance and duration of therapy attended are listed as secondary outcomes. There are 3 outcome measures in the record, but no analysis is listed. Outcome results by group, participant flow, and the number included in each outcome analysis are not reported in the registry record.

Because allocation was nonrandomized, the arms should not be read as if they were necessarily exchangeable at the start of the comparison. A difference in later outcomes, if reported, could reflect the adjunct, differences between people entering each group, or other aspects of care. The record does not describe the allocation process, adjustment methods, or a statistical analysis plan, so it is not possible to determine how the study intended to address these concerns. The record also does not report whether follow-up or therapy delivery differed between groups.

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

Analysis population for the primary outcome, as defined in the registry: The number of participants analyzed in this intention-to-treat analyses represents the number of participants that provided PHQ-9 scores at the end of the study period, at 16 weeks. Because the PHQ-9 was filled in during psychotherapy sessions, this number is smaller than the original sample (n=44, n=38) due to patient drop-out of psychotherapy.

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 PHQ-9 Depression Symptoms, measured over 16 weeks.

The primary endpoint was PHQ-9 depression symptoms at 16 weeks, expressed in units on a scale and summarized as a mean with standard deviation. The PHQ-9 is a questionnaire commonly used to screen for depression and monitor changes in depressive symptoms. The record describes score ranges as indicating minimal, mild, moderate, moderately severe, or severe depression, and notes that a score of 10 is often used as a cutoff for identifying an episode that may require treatment.

A mean describes the group’s average score, while a standard deviation describes how much scores vary among participants. Those two summaries answer different questions: the mean indicates the average symptom score, and the standard deviation conveys spread around that average. Neither is reported by group in the posted results, and no between-group estimate or uncertainty interval is reported. Consequently, the endpoint definition is available, but the measured outcome comparison is not.

The registry says the analyzed participants were those who provided PHQ-9 scores at the end of the study period, at 16 weeks, and that this number was smaller than the original sample. The exact analyzed number is not reported in the registry record. Although the record describes the PHQ-9 as an intention-to-treat analysis, its description ties the analyzed set to having an endpoint score; the precise handling of participants without such scores is not explained.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a continuous primary outcome summarized by a mean and standard deviation. It does not report the analysis used to compare the groups. The statistical ideas below are general explanations of how such results are commonly interpreted, not a description of methods confirmed for this study.

What does a mean and standard deviation convey?

For a continuous measure such as a questionnaire score, the mean is the arithmetic average among the participants included in the analysis. The standard deviation describes the spread of individual scores around that average. A group mean alone does not show whether most participants had scores close to the average or whether scores varied substantially. The standard deviation adds information about that variability, but it is not itself a measure of uncertainty about the group mean.

When two groups are compared, the difference between their means is often used to express the contrast on the original outcome scale. Its interpretation depends on which group is subtracted from which, the scale’s meaning, and the analysis population. The registry does not report either group mean or a difference between them, so no such contrast can be interpreted for this study.

How should uncertainty be understood?

General statistical explanation: a confidence interval is a range of values generated by a statistical procedure to express uncertainty around an estimate. A narrow interval indicates greater precision than a wide interval, given the assumptions and data used. It is not a guarantee that the true effect lies within the particular interval. A p-value, when reported, describes how incompatible the observed data are with a specified null model under the analysis assumptions; it does not measure the size or clinical importance of an effect.

Neither confidence intervals nor p-values for the primary comparison are reported in the registry record. Without an estimate and an uncertainty measure, readers cannot judge the precision of a possible group difference. A significance label alone, even if one were available, would not substitute for the estimated effect and its uncertainty.

Why does the allocation method matter?

General statistical explanation: randomization helps make comparison groups similar in expectation at baseline, while nonrandom allocation can leave them different. Statistical adjustment can sometimes address measured baseline differences, but adjustment cannot automatically remove bias from factors that were not measured or were modeled inadequately. The allocation is listed as nonrandomized, and the record does not report assignment procedures, baseline comparisons, covariate adjustment, or other methods for addressing group differences. No specific adjusted or unadjusted estimate can therefore be attributed to the study.

The parallel design means the groups were compared as separate arms rather than as repeated treatment periods in the same participants. That description alone does not specify whether every participant was assessed on the same schedule, how repeated measurements were handled, or whether the primary analysis used endpoint scores, change from baseline, or another approach. Those details are not reported in the registry record.

What does an analysis population mean?

General statistical explanation: an analysis population identifies which participants contribute to a particular estimate. An intention-to-treat approach is generally intended to retain participants in the groups to which they were assigned, regardless of treatment adherence, but a missing endpoint still requires a defined approach. The registry calls the PHQ-9 analysis intention-to-treat while describing the analyzed participants as those who provided 16-week scores. It also says that this number was smaller than the original sample. The exact number analyzed and the missing-score procedure are not reported, so the label does not by itself clarify how unavailable scores were handled.

Missing outcomes can affect both the size and the representativeness of an analysis. For example, if participants with missing scores differ systematically from those with scores, an analysis limited to observed scores may not reflect the full enrolled groups. This is a general concern, not evidence that such a pattern occurred in this study. The registry does not report reasons for missing scores or whether missingness differed by group.

What can planning calculations say?

General statistical explanation: a sample-size calculation is a planning exercise based on assumptions about an effect, variability, allocation, and a chosen statistical criterion. A calculation based on an observed effect is an illustration of the sample size that might be needed under those assumptions; it is not observed power and does not establish that the study had a particular probability of detecting an effect. The registry notes that planning illustrations use an observed effect and specified assumptions, but the outcome effect itself is not reported in the registry record. The trial’s own sample-size justification is not reported.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Group CBT With Text Messaging Adjunct", "Group CBT for Depression")
n   <- c(18, 9)
m   <- c(6, 12.67)
sp  <- c(4.2, 6.4)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Group CBT With Text Messaging Adjunct  18  6.000  4.200  3.911  8.089
CARM  Group CBT for Depression  9  12.670  6.400  7.751  17.589

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 3 outcome measures, including the primary PHQ-9 measure and the secondary measures of attendance and duration of therapy attended. It does not report an analysis for any outcome, and no multiplicity procedure or prespecified hierarchy among outcomes is reported in the registry record. Without those details, it is not possible to tell whether the study planned to control the chance of false-positive findings across multiple outcome comparisons or how it would prioritize them.

Multiplicity is a general statistical concern when many outcomes, time points, or comparisons are examined. The more analyses considered, the greater the chance that at least one result may appear notable by chance if no adjustment or other strategy is used. This general principle does not imply that the study conducted unreported analyses. The record lists 0 analyses, so it supplies no outcome-level inferential results to assess.

The registry’s PHQ-9 description says the analyzed number consisted of participants who posted a score at 16 weeks and was smaller than the original sample. The exact count and reasons for missing scores are not reported. It is also not reported whether the number with scores differed between groups, whether missing observations were imputed, or whether sensitivity analyses examined alternative missing-data assumptions. The record’s use of the phrase intention-to-treat does not resolve these details because the analysis description is tied to availability of an endpoint score.

General statistical explanation: missing-data methods rely on assumptions about why observations are unavailable. Complete-case analysis uses participants with observed outcomes, while imputation or model-based methods use additional assumptions to account for missing values. Different assumptions can lead to different estimates, particularly when missingness is related to outcome or treatment group. No particular method can be assigned to this study from the record.

12. Safety Results

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

Other (non-serious) adverse events

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

The serious adverse event figures list 0 affected participants among 45 at risk in the group receiving CBT with text messaging, and 0 affected participants among 40 at risk in the group receiving CBT without text messaging. The other adverse-event figures also list 0 affected participants among 45 at risk and 0 among 40 at risk, respectively. These are the figures recorded for the two groups.

These counts should be read as registry entries, not as proof that no adverse events occurred under every possible definition or observation process. The record does not report event definitions, ascertainment procedures, follow-up details for safety, or whether events were judged related to treatment. Deaths are not reported in the registry record. The absence of listed serious events does not establish equivalent safety between groups, especially since no comparative safety analysis is reported.

In general, an event count depends on how events are defined, sought, and recorded, as well as on the participants observed. A count of zero in a registry entry is not an estimate of risk with reported uncertainty. The record does not provide a safety analysis or detail how the at-risk counts were determined beyond the numbers shown.

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

The record identifies a study designed to examine whether automated text messaging added to group CBT could increase engagement and potentially improve outcomes. It specifies a primary outcome—PHQ-9 depression symptoms at 16 weeks—and lists attendance and duration of therapy attended as secondary outcomes. It also describes the two treatment groups and records an enrollment of 85.

It does not, however, report numerical results for those outcomes. There is no reported PHQ-9 mean by group, no estimated contrast, and no uncertainty interval or statistical test. Attendance and therapy duration results are likewise not reported. Consequently, the posted results do not establish the direction or size of any symptom or engagement difference. The lack of a reported result should not be interpreted as evidence that groups were equal or that the adjunct had no effect.

The nonrandomized design is important to the scope of any possible interpretation. Even with a reported difference, group composition or other factors could contribute to it. The record does not describe how participants entered or were assigned to each arm, whether important baseline characteristics were balanced, or whether analyses adjusted for differences. These omissions limit what can be concluded about cause and effect from the registry information.

The endpoint population is also incompletely described. The registry says that the PHQ-9 analysis included participants with scores at the 16-week endpoint and that the analyzed number was smaller than the original sample, but the exact count and handling of missing scores are not reported. As a result, the extent to which the primary analysis represents all enrolled participants cannot be assessed.

The most defensible reading is therefore descriptive: the registry documents the intended question, design, outcome definitions, and recorded safety counts, but it does not contain the outcome estimates needed to characterize treatment-group differences. It cannot establish efficacy, lack of efficacy, or comparative safety.

14. Why This Trial Matters Statistically

This study raises a useful statistical distinction between measuring an outcome and reporting an analyzable comparison. A registry can identify a primary endpoint, its time point, and its summary measure, yet those details alone do not show what participants experienced. Here, the PHQ-9 outcome is defined as a mean with standard deviation at 16 weeks, but the numerical outcome comparison is not reported. Statistical interpretation requires the estimates as well as the endpoint definition.

The study also illustrates why design labels matter. A parallel-group comparison may appear straightforward, but nonrandom allocation means that an observed contrast could reflect differences between groups in addition to the intervention. Without assignment details, baseline information, and analysis methods, readers cannot judge how those risks were addressed. This is a general lesson in interpreting comparative clinical research, not evidence of a particular bias in this study.

Engagement is central to the stated rationale: the text messages were intended to monitor mood and remind participants about session content, while attendance and therapy duration were listed as secondary outcomes. Those outcomes could help distinguish a proposed pathway—greater engagement—from symptom outcomes. However, the registry does not report their results or specify how engagement was operationalized beyond those listed measures. The proposed relationship between engagement and improved outcomes therefore remains an aim, not a demonstrated finding in the posted record.

Finally, the record identifies a smaller endpoint analysis population than the original sample but does not report the exact count or missing-data approach. That makes completeness of follow-up a key interpretive issue. Together, the absent outcome estimates and incomplete analysis details show why transparent reporting of effect estimates, uncertainty, participant flow, and methods is necessary for a useful statistical reading.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record leaves several trial-specific details unresolved. Although the study is marked completed and has an enrollment of 85, it does not report participant flow by group or the number contributing to each outcome. The PHQ-9 description says that the analyzed number was smaller than the original sample because scores were collected during psychotherapy sessions, but the exact analyzed count is not reported. Reasons for missing scores and whether missingness differed between arms are also not reported.

Allocation is classified as nonrandomized, but the mechanism for assigning participants to the two groups is not reported. Baseline characteristics are not reported, so the degree of similarity between groups cannot be assessed. The record does not report statistical adjustment, an analysis plan, or whether the primary comparison was based on endpoint scores or change over time.

The primary endpoint is described, but its group-level values and any comparative estimate are absent. Results for the secondary outcomes of attendance and duration of therapy attended are also not reported. The record lists 3 outcome measures and 0 analyses. Consequently, outcome direction, magnitude, and precision cannot be assessed from the posted results.

The registry does not report the messaging schedule, participant exposure to messages, adherence to CBT, or a detailed definition of engagement. These omissions limit assessment of whether the intervention was delivered as described and how the study’s stated engagement objective was evaluated. Safety event definitions and ascertainment procedures are not reported, and deaths are not reported in the registry record. These limitations concern information available in the record; they do not establish that the study itself did not collect or analyze the information.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Text Messaging and Cognitive Behavioral Therapy for Depression — NCT01083628.
• Registry record last updated 2019-08-12; status last verified 2019-07.
• Results first posted 2019-08-12.
• 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.

19. Record Summary

The completed study compared group CBT for depression using the BRIGHT manual with the same described CBT approach accompanied by automated MoodText messages for mood monitoring and reminders. It was a nonrandomized, parallel-group behavioral study with no masking, sponsored by the University of California, San Francisco. The registry lists enrollment of 85, a start date of 2014-01, and primary completion in 2016-09.

The primary outcome was mean PHQ-9 depression symptoms at 16 weeks, with standard deviation. Attendance and duration of therapy attended were listed as secondary outcomes. The record lists 3 outcome measures and 0 analyses, and does not report outcome estimates or the exact number analyzed for PHQ-9. It records 0 serious adverse events among 45 at risk in this discussion-messaging group and 0 among 40 at risk in the CBT group; deaths are not reported in the registry record.