Incorporating Patient Treatment Choice to Improve Treatment Retention in Depressed Hispanics: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Incorporating Patient Treatment Choice to Improve Treatment Retention in Depressed Hispanics (NCT00742573), a randomized phase 2 study. The primary outcome is Mean Time of Retention.

2. Clinical Question

3. Trial Design

This was a randomized, single-masked, parallel-group phase 2 study. The registry lists an enrollment of 170 and describes the study as completed. The recorded study start is 2008-08, and primary completion is recorded as 2014-01-10. The sponsor was the New York State Psychiatric Institute. These registry entries describe the study's administrative design and status; they do not establish what masking procedures were used or who was masked.

The comparison was between medication treatment following the Texas Medication Algorithm (TMA) and a patient-choice approach. In the TMA arm, participants received medication treatment according to the algorithm for depression. In the patient-choice arm, participants were offered brief interpersonal psychotherapy (IPT-B) alone or IPT-B combined with TMA medication treatment. The registry describes random allocation but does not report the allocation sequence method, whether allocation was concealed, or the procedures used to maintain masking.

The primary endpoint was mean time retained in treatment, expressed in weeks, with a timeframe of 52 weeks. The posted results list 83 participants in the TMA group and 87 in the patient-choice group for this endpoint. The registry describes the primary population as adult Hispanics with MDD. It does not report further recruitment details, the number screened or excluded, or the characteristics of participants at entry.

The trial record lists two outcome measures and no analysis records. A comparison of the primary outcome is available in the posted results, but the registry record does not report the trial's complete statistical analysis plan or the methods used for all outcomes. The original sample-size assumptions, planned interim analyses, and handling of missing outcome data are not reported in the registry record. The listed planning illustrations are based on the observed result and are not a record of the trial's original design assumptions.

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: Adult Hispanics with MDD

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.

Population: Adult Hispanics with MDD

6. Primary Endpoint

The registry lists the primary outcome as Mean Time of Retention, measured over 52 weeks.

The primary endpoint was mean time of retention in treatment for MDD, measured in weeks over a 52-week timeframe. It describes the average duration participants remained in treatment, rather than a symptom score or a direct measure of depression improvement. Retention can be useful as an outcome because it captures whether participants continue receiving care during the study period. In a trial that changes the treatment options available to participants, the endpoint can describe how treatment engagement differed between the approaches.

The registry reports a mean of 27.08 weeks in the TMA group and 25.84 weeks in the patient-choice group. The corresponding standard deviations were 22.4 and 20.54 weeks. A standard deviation describes how spread out individual observations were around a group mean; it is not the uncertainty around the mean itself. The sizable within-group spread relative to the reported means indicates that participants' retention times varied, while the endpoint summaries alone do not show the reasons for that variation.

The record does not define the operational rules used to determine when retention ended, how temporary interruptions were treated, or how follow-up ending before 52 weeks was handled. It also does not report whether the endpoint used all randomized participants or a different analysis population. Those details matter when interpreting an average retention time, because outcome definitions and missing-data handling can affect which observations contribute to a mean. A posted result for the trial's stated improvement aim is not reported in the registry record.

7. Results Posted in the Registry

• 2 Patient Choice: mean 25.84 versus 27.08 on control; difference -1.24 (95% CI -7.71 to 5.23), standardized effect d = -0.06; z-test p = 0.707.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a difference between means represent?

The primary outcome is continuous: each participant's retention is expressed as a duration in weeks, and each group is summarized by a mean. A mean difference subtracts one group mean from the other. For this comparison, the stated direction is patient choice minus TMA. The resulting estimate of -1.24 weeks describes the difference between the two posted averages. It does not imply that every participant's experience differed by that amount, and it does not describe the distribution of individual outcomes.

In general, a group mean is a compact description of the center of observed values. It can be informative, but it does not reveal the full shape of the data. A small number of very long or very short retention times can influence an average. The registry provides standard deviations, which describe within-group variation, but does not report medians, ranges, distributional plots, or other summaries that would help assess the shape of retention times.

What does the standard deviation tell the reader?

The standard deviations were 22.4 weeks for TMA and 20.54 weeks for patient choice. In general, a standard deviation measures how dispersed individual observations are around the group mean. It is not a standard error, and it is not a confidence interval for the mean. A standard deviation describes participant-level variability; a standard error describes uncertainty in an estimate, given the data and a statistical model. The registry also lists standard errors for the group means, but those values do not replace the information in the standard deviations.

When variability is large relative to the difference between group means, a study may estimate the average difference imprecisely unless the data are sufficiently informative. The observed variation is relevant context for the interval around the difference. This is a general statistical point, not a conclusion about any unreported features of the trial or its design.

How should the interval be read?

The interval for the recalculated mean difference runs from -7.709 to 5.229 weeks. In general, a confidence interval gives a range of values compatible with the observed data under the method and assumptions used to calculate it. It conveys uncertainty about the estimated difference more directly than a point estimate alone. Because this interval includes zero, the data do not clearly distinguish the estimated group difference from no difference under that calculation. The interval also spans values in both directions, so the posted result does not settle which group would have a higher mean in a larger or different sample.

The confidence level, method used to construct the interval, and any adjustments are not reported in the registry record. The interval should therefore be interpreted as the reported uncertainty range, without assuming an undocumented analysis method. Its width matters: a broad interval indicates less precision than a narrow interval, even when the point estimate is easy to state.

What does the p-value contribute?

The reported unadjusted p-value is 0.707. In general, a p-value is calculated under a null hypothesis and a statistical model. It describes how often results at least as extreme as the observed result would occur if the null hypothesis and model were true. It does not give the probability that the treatment approaches are equal, and it does not establish that a clinically meaningful effect is absent. A large p-value can occur when the estimated difference is small, when uncertainty is substantial, or both.

The p-value and interval are complementary descriptions of the comparison. The p-value gives one measure of compatibility with a null hypothesis, while the interval conveys a range of effect estimates compatible with the data and method. Neither statistic alone answers whether a difference matters clinically. No threshold for a meaningful retention difference is reported in the registry record, and no improvement outcome result is posted.

What does unadjusted mean in this comparison?

The mean difference and its associated interval and p-value are simple unadjusted re-calculations from the posted group summaries. An unadjusted comparison does not account for other participant characteristics through a covariate model. The registry record does not report whether the sponsor's analyses used adjustment, which covariates might have been included, or how the reported summaries were analyzed in the original study. The recalculation is therefore a transparent comparison of the posted means, not a substitute for a complete trial analysis plan.

The trial is randomized, which is intended in general to distribute participant characteristics across treatment groups, but randomization does not guarantee that all characteristics are balanced in a particular study. The record does not include enough baseline information or analysis detail to assess balance or the consequences of any imbalance. No claim about adjusted results can be made from the posted summaries.

Why can variation beyond a mean matter?

General statistical methods for continuous outcomes often rely on assumptions about how observations behave and how uncertainty is estimated. If observations are highly skewed, have influential extreme values, or do not fit the assumed model, a conventional mean-based comparison may not capture all relevant features of the data. The registry does not report the outcome distribution or diagnostics, so it is not possible to evaluate those issues from the record. Retention durations may also have features related to follow-up or censoring, but the registry does not state how such features were handled. These are questions for the analysis documentation, not details that can be inferred from the group means alone.

10. Confidence Intervals and Precision

The estimated mean difference was -1.24 weeks for patient choice minus TMA. The interval extends from -7.709 to 5.229 weeks. This range conveys uncertainty around the estimate and is more informative than the point estimate on its own. It includes zero, so the posted results are compatible with no average difference, while also allowing for differences favoring either group's mean. The registry record does not report the interval's confidence level or the full method used to calculate it.

The interval is broad relative to the point estimate. That limits precision: the result does not tightly identify the size or direction of the underlying average difference. The interval should not be interpreted as a range containing every participant's retention time, nor as a statement about where most individual observations fell. It pertains to uncertainty about the group-level mean difference under the calculation used.

The standard deviations, 22.4 weeks for TMA and 20.54 weeks for patient choice, describe variation among participant outcomes. The posted group sizes for the endpoint were 83 in TMA and 87 in patient choice. These summaries provide context for understanding the estimate, but they do not disclose the full outcome distributions or missing-data patterns. The registry does not report how many participants contributed complete retention data under a defined analysis population or whether the endpoint was adjusted for other factors.

A wide interval is not proof that the treatment approaches have the same effect. Nor does an interval that includes zero establish equivalence. Equivalence or noninferiority would require a prespecified margin and an appropriate design and analysis; such details are not reported in the registry record. The appropriate reading is that the observed difference is uncertain and the available interval does not resolve a precise treatment contrast.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. They are illustrations based on fractions of the observed effect, not the trial's original design assumptions. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the arms, and a normal approximation. The effect to detect is specified as a fraction of the observed effect. The trial's own original sample-size assumptions are not reported in the registry record.

For the row based on 100% of the observed effect, the planning figures are 4,706 participants per arm for 80% power and 6,299 participants per arm for 90% power. For 75% of the observed effect, they are 8,365 participants per arm for 80% power and 11,199 participants per arm for 90% power. For 50% of the observed effect, they are 18,821 participants per arm for 80% power and 25,196 participants per arm for 90% power.

In plain language, the 80% power column gives the illustrated number of participants per arm for a design with an 80% chance of detecting the specified effect under the listed assumptions, if that effect and the statistical model were correct. The 90% column gives the corresponding planning target for a higher chance of detection under those assumptions. These are prospective planning illustrations, not observed power for the completed trial.

Smaller effects are harder to distinguish from random variation, so detecting a smaller fraction of the observed effect requires more participants per arm in these illustrations. The figures are especially sensitive to the observed effect used as their starting point. They should not be read as evidence that the trial was designed with these targets or that a study of this size would necessarily reproduce the observed result.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -1.240; pooled standard deviation σ = 21.468.
2. n = 2 × (1.96 + 0.842)² × 21.468² / -1.240² = 4,705.2, rounded up to 4,706 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -1.24                      # observed difference in means
sigma <- 21.4679991083074                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(742573)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 4,706 (80% power), 6,299 (90% power)
 75% of observed effect: n per arm = 8,365 (80% power), 11,199 (90% power)
 50% of observed effect: n per arm = 18,821 (80% power), 25,196 (90% power)
power.t.test: 4706.2 
Simulated power at n = 4,706 per arm: 0.797 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures, but a posted result for the primary retention endpoint is the only result discussed in the record. A secondary result for improvement is not reported in the registry record. The record lists no analysis records. It therefore does not establish which outcomes were included in the original analysis plan, whether additional analyses were performed, or whether all planned outcomes were reported.

Multiplicity arises when several outcomes, treatment comparisons, time points, or subgroup analyses are examined. In general, conducting many statistical tests can increase the chance of at least one apparently notable result by chance. Statistical approaches can address that issue, but the appropriate method depends on the planned analyses and scientific questions. The registry does not report a multiplicity strategy, a hierarchy among outcomes, or any adjustment to the posted p-value. The p-value is explicitly unadjusted and should be read as such.

The handling of missing data is not reported in the registry record. The record does not say how participants who left treatment, missed assessments, or had incomplete retention information were represented in the primary analysis. It also does not report an imputation method, sensitivity analysis, or assumptions about missingness. These details can influence continuous-outcome estimates, particularly when missingness is related to treatment experience or retention itself.

The analysis population is also not reported. The endpoint summaries include 83 participants in TMA and 87 in patient choice, but the record does not clarify whether these figures represent all randomized participants, participants with outcome data, or another population definition. The registry does not report whether the analysis followed randomized assignment regardless of treatment received, or whether any participants were excluded after randomization. No inference about the intended analysis population or adherence to it should be made from the counts alone.

Further statistical details that are not reported include the original analysis plan, any interim analyses, the adjustment set for possible covariates, and any subgroup analyses. This absence limits independent assessment of whether the posted primary result was prespecified and how it relates to other analyses. It does not, by itself, establish that any particular analysis was or was not performed.

14. 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 affecting 0 of 83 participants at risk in the TMA group and 0 of 87 at risk in the patient-choice group. It also lists 0 other adverse events affecting participants at risk in each group, and 0 deaths in each group. These are the posted counts for the categories recorded in the registry.

Zero reported events in these groups means that the registry lists no affected participants for the specified categories. It does not establish that no symptoms or other unwanted experiences occurred, nor does it show that the approaches have identical safety profiles. The record's serious adverse event terms list is empty. Definitions, event-collection procedures, attribution rules, and the completeness of reporting are not described in the registry record, so the counts should be interpreted within those limits.

The at-risk counts are group-specific and are reported separately from the primary endpoint summaries. The safety figures do not provide information about treatment adherence, retention, or depression improvement. They should not be combined with the retention result to draw conclusions about clinical benefit or risk.

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

The result describes mean time retained in treatment during a 52-week timeframe for the two randomized approaches. The patient-choice group had a posted mean of 25.84 weeks, and the TMA group had a posted mean of 27.08 weeks. The recalculated unadjusted mean difference was -1.24 weeks for patient choice minus TMA. The interval, from -7.709 to 5.229 weeks, includes zero and leaves uncertainty about the size and direction of the average difference.

This finding does not establish that the approaches are equivalent. A result that does not clearly distinguish a difference from zero can reflect uncertainty, and equivalence requires a prespecified margin and suitable design and analysis details that are not reported in the registry record. The unadjusted p-value of 0.707 is not the probability that there is no difference, and it does not quantify the likelihood that a treatment approach will work for an individual.

The retention outcome is not a direct measure of symptom improvement. Although the study summary describes an aim to retain and improve treatment for Hispanic patients with MDD, the registry record does not report a result for improvement. It also does not identify why participants discontinued treatment or describe which patient-choice options participants selected. The retention summaries therefore cannot explain the mechanisms behind participants' treatment duration.

The results apply to the population and setting represented in this trial record, described as adult Hispanics with MDD. The registry does not provide further participant characteristics or enough design detail to establish how well the findings apply to other groups or care settings. The available comparison is unadjusted and based on posted summary statistics; analysis details, missing-data procedures, and the original statistical plan are not reported.

Overall, the posted result is an uncertain estimate of average retention, not a conclusive statement about treatment benefit, symptom response, or safety. It is useful as a description of the recorded group means and their uncertainty, while leaving important questions about clinical outcomes and analysis methods unanswered.

16. Why This Trial Matters Statistically

This trial raises a statistical question that is easy to oversimplify: whether offering participants a treatment choice changes retention compared with a standardized medication approach. The primary outcome is a mean duration, so the comparison focuses on average time in treatment. The group means are close in the posted summaries, but the interval is broad enough to include differences in both directions. Considering the estimate and its precision together is more informative than treating a single p-value as a verdict.

The patient-choice approach is not a single fixed treatment exposure: participants could receive IPT-B alone or IPT-B combined with TMA medication treatment. The trial therefore compares treatment strategies as described in the registry, not a single medication against a single psychotherapy regimen. The record does not report how participants selected or used the offered options. That makes the distinction between assignment to a strategy and actual treatment received important when interpreting retention.

The endpoint also illustrates the difference between continued engagement and clinical improvement. Retention can be relevant to the delivery of care, but it does not by itself show symptom change. The trial summary mentions both aims, while the registry record reports a primary retention result and no improvement result. A careful statistical reading keeps those outcomes separate.

Finally, the planning illustrations show how sample-size requirements depend on the effect chosen for detection. Smaller effects require more participants per arm under the stated assumptions. Because those illustrations use the observed effect, they are not evidence of the completed trial's design or an assessment of its power. The study is statistically informative as a randomized comparison with a defined continuous primary outcome, but the unavailable analysis and missing-data details constrain how fully the result can be evaluated.

17. 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.

Several limitations specific to this record affect interpretation. The primary endpoint is described as mean time retained in treatment over 52 weeks, but the registry does not provide the operational definition of retention. It does not explain how treatment interruption, withdrawal, transfer of care, or incomplete follow-up affected the endpoint. Without those rules, it is difficult to assess exactly what the reported duration represents.

The patient-choice arm offered IPT-B alone or with TMA medication treatment, but the registry does not report how participants chose between those options, how many received each, or whether choices could change. Consequently, the group-level comparison does not reveal the outcome associated with each option separately. The record also does not describe medication implementation, psychotherapy delivery, adherence, or fidelity to either approach.

The posted summaries do not include baseline characteristics, detailed eligibility criteria, outcome distributions, or reasons for leaving treatment. Although group-specific endpoint counts are reported, the analysis population and treatment of missing data are not specified. These omissions make it hard to assess comparability at entry, possible attrition effects, or whether the mean is an adequate summary of the observed durations.

The record does not report the original statistical plan, confidence level for the interval, any adjustment for other variables, or multiplicity handling. It lists no analysis records and does not report a secondary result for improvement, despite the study summary's interest in improvement as well as retention. Safety event definitions and ascertainment procedures are also not detailed. These are limits of the available registry information; they do not establish that the corresponding procedures or analyses were absent from the study.

Finally, the population is characterized as adult Hispanics with MDD, with no further demographic or clinical detail reported. The record therefore does not allow readers to judge the relevance of the result across different participant profiles or settings. The statistical comparison should be interpreted in light of these missing design and reporting details.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Incorporating Patient Treatment Choice to Improve Treatment Retention in Depressed Hispanics — NCT00742573.
• Registry record last updated 2020-03-03; status last verified 2020-02.
• Results first posted 2020-03-03.
• 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.

21. Record Summary

The registry describes NCT00742573, “Incorporating Patient Treatment Choice to Improve Treatment Retention in Depressed Hispanics,” as a completed, randomized, single-masked, parallel-group phase 2 trial sponsored by the New York State Psychiatric Institute. Its recorded enrollment was 170. The study compared medication treatment according to the Texas Medication Algorithm with a patient-choice approach offering brief interpersonal psychotherapy alone or combined with TMA medication treatment.

The primary endpoint was mean time retained in MDD treatment, in weeks, over 52 weeks, among adult Hispanics with MDD. Posted means were 27.08 weeks for TMA and 25.84 weeks for patient choice. The recalculated unadjusted mean difference was -1.24 weeks, with an interval from -7.709 to 5.229 weeks and an unadjusted p-value of 0.707. The registry reports zero serious adverse events and zero deaths in each group.

The registry record does not report the original statistical plan, missing-data methods, analysis population, or a result for improvement. The retention comparison is informative as a summary of the posted outcome, but its precision and the missing analysis details limit broader conclusions.