Contact: Developing New Clinical Management Strategies: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Contact: Developing New Clinical Management Strategies (NCT03812588), a randomized phase 4 study. The primary outcome is Change From Baseline Hamilton Rating Scale for Depression 24-Item Scale to Study Completion (8 Weeks).

2. Clinical Question

3. Trial Design

The registry classifies the study as phase 4, randomized, quadruple-masked, and parallel. It lists New York State Psychiatric Institute as the sponsor and Major Depressive Disorder as the condition. Enrollment was 29, and the study status is completed. The recorded start date is 2019-01-30 and the primary completion date is 2020-05-27.

The four arm descriptions pair two visit schedules with two treatments: Clinical Frequency Management with placebo or escitalopram, and Research Frequency Management with placebo or escitalopram. Clinical Frequency Management used monthly site visits at Week 0, 4, and 8, plus phone visits every other week at Week 2 and 6. Research Frequency Management used weekly study visits. In both schedule descriptions, the treatment began at 10mg/day, with a possible increase to 20mg/day at Week 4 for participants described as non-responders. The descriptions also mention a 3-month continuation phase with monthly site visits for participants who respond.

The primary outcome was assessed through 8 weeks, and the listed endpoint is change from baseline on the first 24 items of the Hamilton Rating Scale for Depression. Group-level endpoint summaries are reported for the four arms. The sponsor’s posted inferential analysis uses a linear regression and includes all four groups, but the registry does not describe how visit frequency and treatment were encoded in that regression or which coefficients or contrasts were prespecified. The protocol’s allocation ratio, detailed randomization procedure, sample-size justification, and planned statistical analysis are not reported in the registry record. The record also does not specify how the stated continuation phase relates to the primary 8-week analysis population.

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: In total, 29 subjects were enrolled. Of the 29 enrolled, 4 subjects did not continue in the study after enrolling. These 4 subjects did not appear for their randomization visit at which they would have been assigned to a study arm. Thus, 29 patients enrolled and 25 patients started.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Two CFM, Placebo, two RFM, Placebo, and two CFM, Escitalopram participants did not complete the Hamilton Rating Scale for Depression at Week 8, and so their change from baseline Hamilton Rating scale for Depression score could not be calculated.

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 Change From Baseline Hamilton Rating Scale for Depression 24-Item Scale to Study Completion (8 Weeks), measured over Up to 8 Weeks.

The primary endpoint is change from baseline through study completion, up to 8 weeks, on the 24-item Hamilton Rating Scale for Depression. This is a clinician-administered measure of depressive symptoms, with scoring based on the first 24 Hamilton items. The registry describes the total score as ranging from 0 to 74 and gives score categories from normal through very severe. The endpoint uses the sum of item scores, while the reported arm summaries describe the mean change and its standard deviation.

The registry says this scale was chosen because it is a standard measure of depression severity in clinical trials of antidepressants. It also states that the measure was selected to make the study results compatible with meta-analyses and ongoing studies of expectancy. That rationale describes why the measure was selected; it does not establish that the study’s treatment comparisons are precise or that the scale captures every clinically relevant aspect of depression.

Change scores compare each participant’s assessment with their own baseline score, then summarize those changes by group. The record does not state the sign convention used to identify improvement or worsening. Accordingly, the reported positive group means should be read as positive changes on the recorded scale, without assigning a clinical direction to the sign. The endpoint summaries are based on the participants with calculable Week 8 change scores, and the registry notes missing Week 8 assessments in three groups.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant has a numerical change score, and the registry reports group means and standard deviations. In general, a mean describes the center of observed values, while a standard deviation describes how dispersed individual values are around that mean. Neither quantity alone describes uncertainty in the estimated mean or treatment contrast. A confidence interval is generally useful because it expresses a range of values compatible with the data and model. No confidence interval is reported for the sponsor’s regression slope.

The four reported arm means are 1.33, 11.4, 5.33, and 9.25, with standard deviations of 3.51, 4.04, 13.65, and 9.21. The endpoint counts are 3, 5, 3, and 8. These summaries provide a descriptive view of each group, but they do not by themselves answer whether the visit schedule or medication caused a difference. The groups are small, and the variability within groups is substantial relative to the precision information available. In general, a group mean based on few observations can be sensitive to individual values, while a standard deviation describes observed spread rather than the precision of the mean.

The registry reports a linear regression with a slope of 7.900 and p = 0.292. In general, linear regression estimates a relationship between an outcome and one or more predictors under a specified model. The slope describes the modeled change in outcome associated with a change in a predictor, but its meaning depends on how that predictor is defined and coded. Here, the registry does not explain the coding, the regression formula, or the contrast represented by the reported slope. It is therefore not possible to identify the slope as a medication effect, visit-frequency effect, or interaction from the record alone.

The p-value of 0.292 is a test result under the regression model. In general, a p-value measures how compatible the observed data are with a specified null hypothesis, assuming the model and its assumptions. It does not measure the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. A result that does not meet a conventional significance threshold does not establish equivalence or absence of an effect. The registry does not report a confidence interval to show the uncertainty around the slope, nor does it report the regression’s residual diagnostics or other model checks.

Two-factor designs can be analyzed in ways that distinguish the effects of treatment, visit schedule, and whether the effect of one factor varies by the other. In general, an interaction term addresses whether the association with one factor differs across levels of another. The study’s four arms have a structure that could motivate such questions, but the registry’s analysis description only names linear regression. It does not specify whether treatment and visit schedule were entered as separate predictors, whether an interaction was included, or what the analysis coefficient represents. No more specific interpretation should be assigned to the slope without that information.

There is no reported two-arm recalculated estimate for a continuous endpoint. The registry record indicates that values are missing or not estimable for a two-arm comparison. A recalculation based on group summary statistics would be an unadjusted comparison and would not replace the sponsor’s model. When such calculations are possible, they can help illustrate the scale and uncertainty of a simple contrast, but they do not reproduce adjustments, modeling choices, or handling of missing observations that may have been used in a sponsor analysis. No recalculated comparison should be inferred from the four group means alone.

In general, over-dispersion means that observed variability is greater than a model’s assumptions would predict. It can make standard errors too small if ignored, leading to uncertainty that appears narrower than warranted. The primary outcome is continuous, and the registry reports group standard deviations, but it does not report a formal dispersion assessment or model diagnostics. Thus, there is no basis in the record for concluding that over-dispersion was evaluated or that a particular variance adjustment was used. This general concern should not be mistaken for evidence that the regression was misspecified.

Planning calculations answer a different question from trial results. In general, a sample-size illustration estimates the number of participants needed under stated assumptions to detect a chosen effect with a selected significance level and power. Such an illustration is not observed power and is not evidence about the result of the completed trial. The registry describes assumptions for planning illustrations, including a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects defined as fractions of an observed effect. However, no usable two-arm continuous endpoint comparison is estimable in this record, and no numerical planning sample sizes are reported for such a comparison. The trial’s own sample-size justification is not reported in the registry record.

Finally, the meaning of any regression test depends on the analysis population, missing-data handling, and prespecified hypothesis. In general, missing outcomes can affect a comparison if the likelihood of missingness is related to treatment, visit schedule, or outcome. The registry notes uncompleted Week 8 assessments in three arms, but does not report a missing-data method, imputation, or sensitivity analysis. It also does not identify an analysis population or state whether an interim analysis was planned or conducted. These are not details that can be reconstructed from the reported means and regression result.

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("Clinical Frequency Management: Placebo", "Research Frequency Management: Placebo", "Clinical Frequency Management: Escitalopram", "Research Frequency Management: Escitalopram")
n   <- c(3, 5, 3, 8)
m   <- c(1.33, 11.4, 5.33, 9.25)
sp  <- c(3.51, 4.04, 13.65, 9.21)
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  Clinical Frequency Management: Placebo  3  1.330  3.510  -7.389  10.049
CARM  Research Frequency Management: Placebo  5  11.400  4.040  6.384  16.416
CARM  Clinical Frequency Management: Escitalopram  3  5.330  13.650  -28.578  39.238
CARM  Research Frequency Management: Escitalopram  8  9.250  9.210  1.550  16.950

11. Analyses Reported by the Sponsor

The registry reports the sponsor’s own analysis as a linear regression across the four groups, with a slope of 7.900 and p = 0.292. A confidence interval for that slope is not reported in the registry record. The registry’s analysis description does not define the coding of the regression predictor or the particular contrast represented by the slope, which limits its substantive interpretation.

A recalculated two-arm estimate is not available for this continuous endpoint because values are missing or not estimable for a two-arm comparison. Consequently, there is no recalculated contrast to compare numerically with the sponsor’s regression result. The group means and standard deviations are descriptive summaries; comparing them informally would not reproduce a specified regression contrast or account for any model features beyond those summaries.

The sponsor’s analysis is therefore the available inferential result in the registry record, but its scope should not be overstated. The reported p-value does not show the size or precision of a clinically defined treatment effect, and the absent model coding prevents a clear interpretation of the slope as a medication or scheduling effect. The registry does not report a separate adjusted comparison or a confidence interval that would resolve those limitations.

12. Multiplicity, Missing Data and Analysis Populations

The record lists one primary outcome and one secondary outcome: change from baseline on the 14-item Hamilton Anxiety Rating Scale. It also states that there are 3 outcome measures and 1 analysis. The relationship among those counts, the listed outcomes, and the reported regression is not explained. No multiplicity adjustment or hierarchy of outcome tests is reported in the registry record.

Multiplicity matters in general when several outcomes, time points, treatment contrasts, or analysis choices are tested. The more questions examined, the greater the chance that at least one result may appear notable by chance if no adjustment or prespecified hierarchy is used. The registry does not say how many contrasts were planned, whether the regression was the only inferential test, or how any multiple comparisons were handled. The p-value should therefore be read as the reported result for the stated regression, not as evidence that all possible comparisons were tested or controlled.

The primary outcome population note says that two participants in Clinical Frequency Management: Placebo, two in Research Frequency Management: Placebo, and two in Clinical Frequency Management: Escitalopram did not complete the Week 8 Hamilton assessment, so their change scores could not be calculated. The registry does not describe an imputation method, a model-based approach to incomplete outcomes, or sensitivity analyses. It also does not specify whether these participants were excluded from all analyses or whether other observations were used in a broader analysis population.

The analysis population definition, treatment of protocol deviations, and handling of withdrawals are not reported in the registry record. Nor is an interim analysis plan reported. Without these details, the posted summaries cannot establish whether the analysis followed a prespecified intention-to-treat approach or another population definition. These missing details limit assessment of bias and of how robust the reported regression is to incomplete outcome data and analytic choices.

13. 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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: dehydration (12.5%).

The safety record lists serious adverse events separately for the four groups. The numbers at risk were 5 for Clinical Frequency Management: Placebo, 7 for Research Frequency Management: Placebo, 5 for Clinical Frequency Management: Escitalopram, and 8 for Research Frequency Management: Escitalopram. The serious adverse event counts were 0, 0, 0, and 1, respectively. Deaths were 0 in each group. The listed top serious event term is Dehydration, reported as 12.5%.

These counts should be read as registry-reported event counts, not as a comparative safety estimate. The group sizes are small, and no confidence intervals or formal between-group safety analysis are reported. A count of zero does not establish that an event cannot occur, while a count in one group does not by itself demonstrate a difference between treatments or schedules. The record does not provide enough detail to infer causality or to determine whether the serious event was related to treatment.

The registry also reports other adverse events in the four groups: 2, 6, 1, and 6, respectively, among the listed numbers at risk. Those figures are counts, not rates calculated for a shared follow-up period. The record does not describe event severity beyond the serious/non-serious distinction, duration, recurrence, or attribution. The listed safety data should therefore be considered descriptive and interpreted alongside the limited group sizes and the absence of detailed comparative analysis.

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

The posted results describe a small completed randomized study with four combinations of visit schedule and treatment, and a primary endpoint based on change in a depression rating scale through 8 weeks. They show arm-level mean changes and standard deviations, as well as a linear regression result with slope 7.900 and p = 0.292. These are the results reported in the registry; they do not, on their own, establish which factor accounts for any observed pattern in the means.

The study’s stated rationale was to investigate whether differing visit schedules might affect antidepressant and placebo response and their separation. The reported analysis does not specify a clear contrast for that question. In particular, the record does not identify whether the slope corresponds to medication assignment, visit frequency, an interaction, or another coding of the four groups. The registry also gives no confidence interval for the slope. Accordingly, the result cannot be treated as a precise estimate of the effect of a particular treatment or visit strategy.

The p-value should not be read as proof that there is no effect, nor should the group mean pattern be taken as proof of benefit or harm. A non-significant test can be compatible with a range of effects, especially when the endpoint group counts are small and uncertainty is not quantified by a confidence interval. The direction of the change-score scale is not described, so positive means cannot be labeled as improvement or worsening from this record alone.

The findings also do not establish that the two visit schedules are equivalent, that medication and placebo perform similarly, or that one schedule improves treatment response. Those conclusions would require defined contrasts, appropriate estimates of uncertainty, and enough information about the analysis and missing data. The registry does not report a numerical two-arm recalculation for the continuous outcome, a detailed statistical plan, or the sample-size justification.

For readers, the clearest interpretation is descriptive: the registry records the group summaries and a four-group regression test, but key information needed to translate the regression into the study’s motivating clinical question is absent. The data and analysis reported are not a basis for medical advice or for choosing a treatment schedule.

15. Why This Trial Matters Statistically

The design addresses a methodological question about how the conduct of a clinical trial may influence measured response. The registry’s study description argues that weekly research visits differ from standard clinical practice and that visit frequency could affect responses in both medication and placebo groups. This makes the trial statistically interesting because the treatment being studied is not the only feature that could shape the measured outcome: contact frequency is also part of the randomized group structure.

That structure creates a need to distinguish medication effects, visit-schedule effects, and possible interaction between them. In general, a factorial arrangement can help separate these questions when the analysis is specified to estimate the relevant contrasts. The registry gives four groups that combine two schedules and two treatments, but its analysis record does not explain whether the regression separated those factors or how the slope maps to a scientific question. This illustrates why a named statistical method is not enough for interpretation: the model’s coding and target contrast matter.

The trial also highlights the difference between descriptive evidence and inferential evidence. Group means can reveal the observed pattern, but small groups and substantial within-group variability make precision important. The sponsor reports a p-value but no confidence interval, leaving the uncertainty around the slope difficult to assess. The missing Week 8 assessments in three groups further emphasize that endpoint summaries depend on who contributes data.

For researchers, the record underscores the value of transparent reporting of estimands, model terms, missing-data procedures, analysis populations, and multiplicity handling. Those details are especially relevant when multiple dimensions of a design are being studied and the central question concerns their separate or joint effects. The study’s conceptual motivation is clear from the record, while the reported analysis details are too limited to fully connect the result to that motivation.

16. 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 endpoint group counts are small: 3, 5, 3, and 8 participants. The reported standard deviations vary across groups, and the standard deviation for Clinical Frequency Management: Escitalopram is 13.65. These summaries suggest that individual outcomes varied, but the registry does not provide confidence intervals for the group means or the regression slope. Precision therefore cannot be assessed fully from the inferential result.

The record does not define the regression slope’s coding or target contrast. Although the four groups combine visit schedule and treatment, the analysis description does not say whether it included separate schedule and treatment effects or an interaction. This prevents a direct interpretation of the reported slope in terms of the study’s main scientific motivation. The record also does not give the direction convention for change scores, so the sign of a mean change cannot be translated into improvement or worsening.

Missing Week 8 assessments are noted for the placebo groups under both schedules and for Clinical Frequency Management: Escitalopram. The registry does not report how missing values were handled, whether sensitivity analyses were performed, or which analysis population was used. It does not report a sample-size justification, allocation ratio, detailed randomization procedure, protocol-level statistical plan, or interim analysis information. Multiplicity procedures are also not described.

Safety counts are descriptive, with no reported comparative analysis or uncertainty estimates. The continuation phase mentioned in the arm descriptions is not connected to the primary outcome analysis in the record. These limitations concern what can be interpreted from the registry record; they do not establish that any unreported procedure was or was not conducted.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Contact: Developing New Clinical Management Strategies — NCT03812588.
• Registry record last updated 2022-04-21; status last verified 2022-03.
• Results first posted 2022-04-21.
• 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.

20. Record Summary

NCT03812588 is a completed, randomized, quadruple-masked, parallel phase 4 study sponsored by New York State Psychiatric Institute. Enrollment was 29, and the condition was Major Depressive Disorder. The study compared two visit schedules across placebo and escitalopram groups.

The primary endpoint was change from baseline through 8 weeks on the 24-item Hamilton Rating Scale for Depression. The registry reports four arm-level means and standard deviations and a linear regression slope of 7.900 with p = 0.292. A confidence interval for the slope is not reported in the registry record, and the regression contrast is not defined clearly enough to attribute the result to a particular treatment or visit-schedule effect.

The record notes missing Week 8 assessments in three groups and does not describe the missing-data method, analysis population, multiplicity approach, or sample-size justification. Serious adverse event counts were 0, 0, 0, and 1 across the groups, with no deaths reported. Overall, the posted information supports a descriptive account of the outcomes and the reported regression, but leaves substantial uncertainty about precision and the intended interpretation of the model.