Engaging Patients With Mental Disorders From the ED in Outpatient Care: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Engaging Patients With Mental Disorders From the ED in Outpatient Care (NCT02989805), a randomized na study. The primary outcome is Number of Participants Attending at Least One Outpatient Visit.

2. Clinical Question

The study also examined engagement mechanisms, according to its summary, and recorded secondary outcomes relating to outpatient visits, emergency room readmission, and inpatient hospitalization. The central comparison is not care management versus no care management: both randomized groups were assigned to the CARE intervention, with the care manager role differing between groups.

3. Trial Design

The registry describes a completed, randomized, parallel-group study with no masking. The intervention was behavioral care management. Participants were allocated to either a professional care manager or a peer specialist care manager. The registry lists an enrollment of 326. The study summary describes a multisite randomized trial across 8 emergency departments in South Carolina with telepsychiatry programs.

The professional care managers were nurses or social workers. Their training was described as covering shared decision-making, action planning, motivational interviewing, and mental health as a cornerstone of recovery. Peer specialists were described as having a history of mental illness, identifying as in recovery, having at least a high school education, and having reliable transportation to the study site. The summary describes both approaches as part of the CARE Program and emphasizes shared decision-making with patients.

The summary says eligible participants would be randomized to a one-year engagement intervention and describes an intended total of 290 participants across sites. The registry enrollment is 326; the reason for the difference between these figures is not reported in the registry record. The registry lists study start as 2017-04-03, primary completion as 2019-12-31, and status as completed.

These design details establish the comparison as a randomized evaluation of two care-manager approaches within the same broader program. The registry record does not report the randomization procedure, allocation concealment, protocol deviations, or the original statistical design assumptions. No masking was used, so participants and care managers were not blinded to the assigned care-manager type.

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 completing the study are included in this analysis.

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: The baseline analysis population includes participants who completed the study.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Attending at Least One Outpatient Visit, measured over 30 days after discharge.

The primary endpoint was the number of participants attending at least one outpatient visit for a mental health problem within 30 days after discharge from the emergency department. It was operationally defined as attendance at one or more such visits during that period. The record says the data came from the South Carolina Office of Revenue and Fiscal Affairs, whose data warehouse integrates client-specific information from health and human services facilities, agencies, and organizations using an internally assigned tracking number.

This endpoint captures an early connection with outpatient mental health care after an emergency department visit. It is relevant to the study’s stated aim of improving linkage to and engagement in outpatient care. It is a threshold outcome: a participant either met the definition by attending at least one visit or did not. It does not, by itself, describe the number of visits, the timing of each visit, the quality or content of care, or engagement beyond the 30-day window.

The primary analysis population is described as participants completing the study. How completion was defined, and how participants without linked outcome data were handled, are not reported in the registry record. The registry lists the endpoint as a count of participants, so the group counts and their denominators need to be read together when interpreting the comparison.

7. Results Posted in the Registry

• Peer Specialist Care Manager: 43.2% (70/162) versus 55.2% (85/154) on control; risk difference -12.0 percentage points (95% CI -22.9 to -1.0), risk ratio 0.78 (0.62–0.98); unadjusted two-proportion z-test p = 0.033.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Percentage of Outpatient Visits Attended

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Peer Specialist Care Manager: mean 66.20 versus 66.90 on control; difference -0.70 (95% CI -8.31 to 6.91), standardized effect d = -0.03; z-test p = 0.857.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Peer Specialist Care Manager and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants Being Readmitted to the Emergency Room

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Peer Specialist Care Manager: 57.5% (77/134) versus 58.8% (77/131) on control; risk difference -1.3 percentage points (95% CI -13.2 to 10.6), risk ratio 0.98 (0.80–1.20); unadjusted two-proportion z-test p = 0.828.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Peer Specialist Care Manager and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants Being Admitted for Inpatient Hospitalization

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Peer Specialist Care Manager: 26.1% (35/134) versus 21.4% (28/131) on control; risk difference 4.7 percentage points (95% CI -5.5 to 15.0), risk ratio 1.22 (0.79–1.89); unadjusted two-proportion z-test p = 0.364.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Peer Specialist Care Manager and the control. The interval gives the range of effects that remain compatible with the data.

Across the secondary outcomes, the estimates do not form a consistent pattern of clear between-group separation. The percentage of outpatient visits attended was similar in the two groups, and the reported uncertainty around the mean difference included no difference. The emergency room readmission comparison also showed little separation in the observed group values, with uncertainty spanning both directions. The inpatient hospitalization estimate pointed in the opposite direction to the primary attendance result, but its uncertainty also included no difference. Taken together, these outcomes do not establish a coherent secondary-outcome advantage for either care-manager approach.

The secondary findings are related to, but do not directly duplicate, the primary result. The primary endpoint asks whether a participant attended at least one outpatient mental health visit within 30 days. The secondary visit-attendance measure concerns the percentage of visits attended over 6 months, while the other outcomes concern emergency room readmission and inpatient hospitalization over 6 months. They reflect different aspects and time windows of post-discharge care, so a primary difference need not be mirrored by every secondary measure.

These comparisons carry less inferential weight because the reported p-values are unadjusted and the registry record does not describe a multiplicity strategy. When multiple outcomes are examined, chance variation can make some results look more compelling than they would in a single prespecified test. The record lists no reported analyses for these secondary outcomes beyond the posted summary comparisons, and it does not identify a hierarchy that would establish which comparisons were confirmatory. The secondary results are therefore best treated as descriptive context and as questions for further evaluation, not as independent confirmation of the primary finding.

Different outcome denominators also mean that the secondary results may represent different analysis populations. The registry does not explain the differences among those denominators or report how missing outcome data were addressed. This limits direct comparisons across outcomes and leaves uncertainty about whether the observed pattern reflects outcome-specific differences in available data.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Percentage of Outpatient Visits Attended
• Number of Participants Being Readmitted to the Emergency Room
• Number of Participants Being Admitted for Inpatient Hospitalization

9. Statistical Interpretation

10. Statistical Methods Explained

For a binary outcome such as attending at least one outpatient visit, each participant is classified according to whether the event occurred. The group event count is interpreted relative to the group denominator. The primary outcome is reported as 85 participants with attendance among 154 in the professional care manager group, and 70 among 162 in the peer specialist group. These counts are the basis for comparing the observed event frequency between randomized groups.

What does a risk difference express?

A risk difference expresses the difference between the two groups’ event proportions, with direction depending on which group is subtracted from which. In the posted comparison, the peer specialist group is compared with the professional care manager group. A negative estimate therefore indicates a lower observed event proportion in the peer specialist group for this outcome. The risk difference is an absolute contrast on the event-proportion scale. It is distinct from a relative measure, which expresses one group’s event frequency in relation to the other group’s.

The registry also lists relative risk and odds ratio estimates for the primary comparison. These measures use different scales and can be useful for describing relative contrasts, but they do not replace the underlying group counts or the absolute risk difference. The registry record does not state which measure was designated as the study’s principal analysis measure in its original statistical plan; that plan is not reported in the registry record.

How should the p-value be read?

The unadjusted p-value is 0.033 for the primary comparison. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a statistical model that assumes no group difference, together with the test’s assumptions. It is not the probability that the null hypothesis is true, and it does not indicate the probability that the result will replicate. It also does not measure the magnitude of the difference. The registry labels this comparison unadjusted; the testing procedure and any covariate-adjusted analysis are not reported in the registry record.

What does a confidence interval add?

A confidence interval presents a range of estimates compatible with the data under a specified statistical procedure. It helps convey uncertainty around the point estimate: a narrow interval indicates greater precision than a wide interval, while an interval that includes no difference leaves more uncertainty about the direction of the population contrast. For the primary risk difference, the reported interval lies below zero. This supports the direction of the observed point estimate under the reported interval method, but it does not remove uncertainty about the size or generalizability of the effect. The specific interval procedure is not reported in the registry record.

What does randomization contribute?

Random allocation is intended to make the groups comparable, on average, with respect to factors measured and unmeasured at baseline. It supports a causal interpretation of a between-group contrast when the trial is conducted and analyzed appropriately. Randomization does not guarantee that every characteristic is balanced in the realized groups, and it does not correct for missing outcome data, departures from assigned care, or selective reporting. The randomization method, concealment procedures, and handling of departures from assignment are not reported in the registry record.

The study was open-label, meaning masking was not used. Knowledge of assigned care-manager type could potentially influence interactions, reporting, or other processes. The primary outcome was obtained through an administrative data source described as integrating records across facilities and organizations. The registry does not report validation details for the endpoint, the extent of missing linkage, or how records with uncertain outcome status were handled.

The secondary percentage-of-visits outcome is continuous and is reported with a mean and standard deviation for each group. A mean describes the average observed percentage; a standard deviation describes variability among observations around that mean. The posted comparison is a mean difference with a confidence interval and unadjusted p-value. The other secondary endpoints are binary participant outcomes, so their comparisons are expressed as risk differences, with relative measures also listed. The registry record does not report further details of the models or tests used for these comparisons.

All posted p-values are unadjusted. The record lists no reported analyses and does not describe whether a prespecified hierarchy, multiplicity correction, or other strategy was used to address the several outcomes. This matters because repeated testing can increase the chance of finding at least one apparently unusual result even when no genuine differences exist. Without the trial’s full analysis plan, the registry results cannot establish how outcome selection, model choice, or multiplicity was managed.

11. Confidence Intervals and Precision

Precision concerns how much uncertainty surrounds an estimated group difference. For the primary outcome, the reported risk-difference confidence interval lies entirely below zero, matching the negative point estimate and the observed lower attendance count in the peer specialist group relative to its denominator. The interval is not especially narrow in relation to its point estimate, so the exact size of the difference remains uncertain even though the interval does not cross zero. The interval method is not reported in the registry record.

The primary group counts and denominators help contextualize the estimate: 85 of 154 participants attended in the professional care manager group, compared with 70 of 162 in the peer specialist group. The point estimate should not be treated as an exact population effect. A confidence interval describes uncertainty from sampling under its assumptions; it does not account automatically for missing data, outcome misclassification, departures from the protocol, or limitations in generalizing beyond the study setting.

For the secondary outcomes, the confidence intervals are broader in relation to their point estimates and include no difference. That pattern means the posted comparisons do not pin down a clear direction or magnitude for those outcomes. In particular, a non-significant p-value does not establish that the groups are equivalent; it can also reflect uncertainty around an estimate. The registry does not report equivalence margins or a design intended to demonstrate equivalence.

Precision should also be considered alongside the analysis populations. The denominators vary across outcomes, and the registry does not explain why. Differences in available outcome information can affect precision and can also create potential bias if the unavailable observations differ systematically. The missing-data process and its handling are not reported in the registry record.

12. 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 per arm for all the listed outcomes; they are not total event counts. They illustrate participant numbers under a stated planning model, rather than documenting the original trial’s sample-size justification. The assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The target effect is set at the stated fraction of the observed effect. The original trial design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the planned participants per arm for a design with an 80% chance of detecting the specified effect under the planning assumptions, if that effect is the true one. The 90% power column makes the analogous illustration for a 90% chance. These are planning probabilities, not guarantees for a future study. Smaller target effects require more participants: planning for 75% or 50% of the observed effect entails larger per-arm numbers than planning for 100% of the observed effect.

The planning numbers use the observed effect as their starting point, so they are sensitive to the uncertainty and possible over- or underestimation in that observed result. In particular, an observed effect can be unstable, and a planning illustration based on it should not be read as a recommendation or a reconstruction of the trial’s original design. These figures are not observed power; observed power is not a useful substitute for reading the effect estimate and its confidence interval.

The assumptions listed apply to these planning illustrations only. The registry does not report whether the trial’s original sample-size plan used these assumptions, whether it allowed for loss to follow-up, or whether any other design features were incorporated.

13. 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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.552; treatment proportion p₁ = 0.432; average p̄ = 0.492.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7070 = 1.3857.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.7019 = 0.5908.
4. n = (1.3857 + 0.5908)² / (-0.1198)² = 272.0, rounded up to 272 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
p0 <- 0.551948051948052                          # control proportion
p1 <- 0.432098765432099                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(2989805)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 272 (80% power), 364 (90% power)
 75% of observed effect: n per arm = 485 (80% power), 649 (90% power)
 50% of observed effect: n per arm = 1,090 (80% power), 1,459 (90% power)
power.prop.test: 272 
Simulated power at n = 272 per arm: 0.812 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome and three secondary outcomes. It reports unadjusted comparisons and does not list reported analyses. A plan for controlling multiplicity, such as prioritizing outcomes or adjusting inference across several tests, is not reported in the registry record. As a result, the primary p-value can be read as the reported unadjusted comparison, but the record does not establish how the study’s overall risk of false-positive findings was managed. The secondary p-values should be interpreted especially cautiously because they are among several outcome comparisons.

Information about missing data is limited. The primary analysis population is described as participants completing the study, but the registry does not define completion, state how many participants lacked primary outcome data, or describe any imputation or sensitivity analyses. The denominators reported for the outcomes differ. The reasons for those differences and the treatment of unavailable observations are not reported in the registry record. If outcomes are missing for reasons related to care assignment or outcome status, a complete-case comparison may not represent the full randomized groups; whether that concern applies cannot be determined from the registry record.

The registry describes the primary endpoint data source and its integration of client-level records across facilities, agencies, and organizations. It does not report the completeness of linkage, validation procedures, or how a participant without a matching record was classified. An absence of a recorded visit could represent no visit or a visit not captured by the data source; the registry does not give enough information to assess this issue.

The primary population is described as study completers, while the original allocation population and any intention-to-treat analysis are not described in the results record. The statistical analysis plan, protocol, interim analyses, adjustment for baseline characteristics, and handling of protocol deviations are not reported in the registry record. These omissions limit assessment of whether the posted comparisons follow prespecified methods and whether the analysis preserves the benefits of randomization.

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

The primary result describes a difference in outpatient mental health visit attendance between two active care-management approaches after emergency department discharge. Attendance was observed in 85 of 154 participants assigned to professional care managers and 70 of 162 assigned to peer specialist care managers. The reported unadjusted risk-difference estimate favors the professional care manager group, and its confidence interval lies below zero. This is evidence about the groups as studied, for the recorded 30-day endpoint.

It does not establish the effect of care management compared with no intervention, because both groups were assigned to care management through the CARE Program. It also does not show that one approach is superior for every participant, every outcome, or every setting. The secondary outcomes do not establish a consistent difference, and their unadjusted nature and uncertainty mean they should not be used to make a strong claim about broader clinical consequences.

The outcome is an indicator of at least one outpatient mental health visit, not a direct measure of recovery, symptom change, continuity of care, satisfaction, or service quality. The registry summary describes the CARE Program as intended to improve linkage and engagement and to examine mechanisms, but the posted primary endpoint alone cannot establish those mechanisms. It also cannot establish what occurred after the primary 30-day window.

Randomization supports comparison of assigned groups, but interpretation still depends on appropriate follow-up and analysis. The registry does not report enough detail about missing data, intention-to-treat analysis, or the statistical plan to resolve those questions. The study was open-label and multisite within South Carolina emergency departments with telepsychiatry programs. The findings may not transfer directly to other health systems or populations; broader applicability is not established by the registry record.

The adverse-event record lists deaths in both groups but does not report causes or attribute them to the care-manager approach. It lists no serious or other adverse events in the corresponding categories. These entries do not demonstrate that either approach caused or prevented harm. Overall, the data support a cautious description of the posted comparison, not medical advice or a universal treatment conclusion.

16. Why This Trial Matters Statistically

This trial is statistically informative because it compares two ways of delivering a shared care-management program in a randomized design. The contrast isolates the care-manager type more directly than a comparison of the whole program with no program would. That distinction matters when interpreting the primary result: it concerns professional versus peer specialist care management within the described CARE approach.

The primary endpoint also illustrates the importance of defining an outcome operationally. “Linkage” is represented by at least one recorded outpatient mental health visit within 30 days, using administrative data. Such a measure is concrete and can be captured across multiple service sources, but it answers a narrower question than whether a person experienced sustained engagement or improved health. Statistical conclusions depend on the outcome definition as much as on the test used.

The contrast between a primary finding and less decisive secondary results is useful for learning how not to overgeneralize from a single p-value. The primary result has a directionally distinct unadjusted comparison, while the secondary outcomes do not create a similarly consistent pattern. Because multiplicity handling is not reported, the findings invite careful attention to prespecification, uncertainty, and the distinction between confirmatory and exploratory evidence.

Finally, differences between the enrollment figure and the total described in the study summary, as well as varying outcome denominators, underscore why transparent reporting of populations and analysis methods matters. The registry records key design and outcome information, but important details about missing data, sample-size justification, and analysis planning are not reported. Those details are necessary for a complete statistical appraisal.

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 trial affect interpretation. The study was open-label, so participants and care managers knew the assigned care-manager type. The registry does not report whether outcome assessment or data processing was blinded. Although the primary outcome came from an administrative data source, the record does not report its completeness for capturing visits or how unmatched records were treated.

The primary analysis was restricted to participants completing the study, while the registry does not define completion or describe an intention-to-treat analysis. Outcome denominators vary, and the reasons are not reported in the registry record. Missingness could affect both precision and the representativeness of the comparisons, but the available information does not show whether or how it did.

The registry enrollment is 326, whereas the study summary describes a total of 290 participants across sites. The reason for this difference is not reported in the registry record. The original sample-size justification and analysis plan are also not reported, so it is not possible to assess the intended power, prespecified analysis population, or planned approach to multiple outcomes from the registry results alone.

The study summary locates the work in 8 South Carolina emergency departments with telepsychiatry programs. The registry does not establish that results would generalize to emergency departments without those programs or to other health systems. The primary endpoint captures attendance during a short post-discharge window, and the other outcomes cover a later period; none alone describes long-term engagement. Finally, the adverse-event record provides counts but does not describe causes or attribution, limiting safety interpretation.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Engaging Patients With Mental Disorders From the ED in Outpatient Care — NCT02989805.
• Registry record last updated 2020-03-16; status last verified 2020-02.
• Results first posted 2020-03-16.
• 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

NCT02989805 is a completed randomized, parallel-group study comparing professional care manager and peer specialist care manager delivery of the CARE Program for people with mental disorders discharged from emergency departments. The registry lists enrollment of 326, no masking, and a multisite setting across 8 South Carolina emergency departments with telepsychiatry programs.

The primary outcome was attendance at at least one outpatient mental health visit within 30 days after discharge. Attendance was recorded for 85 of 154 participants in the professional group and 70 of 162 in the peer specialist group; the unadjusted comparison had a p-value of 0.033, with a reported risk-difference interval below zero. Secondary outcomes concerned outpatient visit attendance, emergency room readmission, and inpatient hospitalization during 6 months after discharge.

The registry does not report the original sample-size justification, missing-data methods, multiplicity strategy, or detailed statistical analysis plan. Its summary describes a total of 290 participants, and the reason this differs from the registry enrollment is not reported in the registry record.