Implementing Effective, Collaborative Care for Schizophrenia (EQUIP-2): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Implementing Effective, Collaborative Care for Schizophrenia (EQUIP-2) (NCT00137280), a non-randomized na study. The primary outcome is The Effect of Care Model Implementation on Treatment Appropriateness: Weight Service Utilization.

2. Clinical Question

The project concerns both implementation and care delivery: it evaluates a model intended to reorganize services within diverse VA practices and priorities. Its stated focus includes wellness and work services for Veterans with schizophrenia or schizoaffective disorder. The primary measure addresses one specific aspect of that broader effort—use of weight services among those eligible at baseline—rather than the full range of wellness, employment, or clinical outcomes.

3. Trial Design

The registry record identifies this as a completed, nonrandomized study with a factorial model and single masking. Enrollment is listed as 1067. The study is sponsored by the VA Office of Research and Development. The recorded start date is 2007-06-13, and the primary completion date is 2010-03-05. The phase is listed as NA.

The comparison contrasts sites implementing the Collaborative Chronic Illness Care Model with control sites receiving Usual Care. Since assignment was not randomized, the arms should not be treated as if they were created by a randomized allocation process. Site differences, participant characteristics, and other contextual factors may be associated with both implementation and outcomes. The registry record does not report the methods used to account for those differences, so it is not possible to determine from this information how fully they were addressed.

The model is identified as factorial, but the registry record does not describe the factors, their levels, or how they were incorporated into the analysis. It also does not explain which party was masked. The “single” masking label alone does not identify whether participants, care providers, outcome assessors, or another group was masked. Further details about site assignment, recruitment, follow-up, or the organizational structure of the comparison are not reported in the registry record.

The intervention is behavioral. Its description refers to increased availability of clinical information, changed practice systems and provider roles, care coordination, and protocols focused on supported employment and wellness services. Usual Care is the comparator label; a more detailed description of usual-care services is not reported in the registry record. The posted information thus identifies the broad service model and comparison, but leaves important operational and analytic details unspecified.

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: Patients were excluded if they did not meet inclusion criteria, refused to participate, or were not approached about participation. Staff were excluded if they refused to participate.

Overall Study

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: People with schizophrenia

6. Primary Endpoint

The registry lists the primary outcome as The Effect of Care Model Implementation on Treatment Appropriateness: Weight Service Utilization, measured over 1 year.

The primary endpoint is the number of participants with one or more weight-service appointments during 1 year of implementation. It applies only to participants who were overweight at the baseline interview, described as eligible for weight services. The comparison is between implementation sites and control sites. The unit is participants, and the outcome is recorded as a count of people meeting the appointment criterion.

This measure addresses whether eligible participants used a relevant service during the implementation period. It is an implementation and service-utilization measure: it records appointment use, not the amount of weight change, the quality of an appointment, or a health outcome. A participant with one or more appointments meets the stated criterion; the registry record does not specify further distinctions by appointment number, service type, or timing within the period.

Such an endpoint can help assess whether a service model is associated with delivery or uptake of targeted care. General interpretation still depends on the comparison design and on the completeness and consistency of appointment ascertainment. Because this study is nonrandomized, a difference in use between sites would not by itself establish that the care model caused the difference. The registry record identifies the endpoint but does not report its observed group counts, effect estimate, or uncertainty interval.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The endpoint is defined as a participant-level event: one or more weight-service appointments during 1 year. The registry record expresses the result as a number of participants, so the underlying classification is whether each eligible participant did or did not meet the appointment criterion. This kind of binary outcome can be summarized with counts and, when denominators and suitable analyses are reported, estimates of the probability of meeting the criterion. Here, the record does not report the observed primary counts or a model-based estimate.

What does the nonrandomized comparison mean?

In a randomized study, allocation by chance is intended to balance measured and unmeasured participant characteristics across groups on average. That protection does not apply in the same way when allocation is nonrandomized. In this study, implementation sites are compared with control sites, so factors that distinguish sites could also be related to the opportunity to receive weight services. General statistical practice is to describe relevant baseline characteristics and use a prespecified analytic approach to address measured differences where appropriate. The registry record does not report whether such adjustment was done or what variables it might have included.

Adjustment cannot automatically remove all concerns. Statistical models can address only factors that are measured appropriately and represented adequately. Differences that are unmeasured, poorly measured, or modeled incorrectly may remain. This is a general limitation of observational and nonrandomized comparisons, not a finding that any particular bias occurred in this study. Consequently, a reported association would need to be interpreted with more caution about causality than a result from a randomized allocation.

Why does the site comparison matter?

The primary description explicitly compares implementation sites with control sites. Outcomes among people treated within the same site may be related because they share service systems, staff, and local practices. General statistical methods can account for clustering when the design and data support it, but the registry record does not say whether clustering was modeled, how sites were assigned, or how many sites contributed. No particular site-level analysis can therefore be inferred.

The study is also labeled factorial. In general, a factorial design evaluates multiple factors or components within a study framework, and analysis may need to distinguish the effects of those factors and possible interactions. The registry record does not identify the factors, their assignment, or any interaction analysis. The label alone is not enough to reconstruct the design or statistical model.

What is the role of outcome definition?

A clear outcome definition helps determine which observations count as events. Here, the stated threshold is one or more weight-service appointments, and the eligible population is limited to participants overweight at baseline. This differs from measuring the number of appointments per participant, time to first appointment, or change in weight. Those alternative outcomes are not the primary endpoint described in the record. The stated measure can capture whether any service use occurred during the period, but it does not convey how much service was received or whether the service changed health outcomes.

How should uncertainty be understood?

General explanation: a confidence interval describes a range of values compatible with the observed data under the model and assumptions used. Its width reflects statistical precision; it does not account automatically for biases caused by nonrandomized allocation, incomplete measurement, or an unsuitable model. No confidence interval or primary effect estimate is reported in the registry record, so the precision of the primary comparison cannot be assessed.

General explanation: for a binary endpoint, the analysis might compare event counts or proportions, or fit a model that accounts for relevant design features. The appropriate choice depends on the allocation structure, the outcome data, and the analysis plan. The registry record does not identify the test, model, covariates, or inferential method used. It would therefore be inappropriate to attribute a particular statistical procedure to the study.

What does over-dispersion mean in general?

General explanation: over-dispersion occurs when observed variability is greater than a statistical model assumes. It can make standard errors too small if ignored, producing confidence intervals that are too narrow and tests that appear more conclusive than warranted. This concern is especially relevant to some count-data models, but the primary endpoint is described as the number of participants who had one or more appointments, rather than the number of appointments. The registry record does not report an over-dispersion assessment or identify a count model, so over-dispersion should not be presented as a demonstrated issue in this trial.

The sample-size note describes planning illustrations based on an observed effect, with two-sided significance level 0.05, equal allocation, and a normal approximation. It distinguishes participant counts per arm from total event counts for time-to-event tables. Such illustrations are planning aids, not observed power and not evidence that the original study was designed using those assumptions. The registry record does not report the trial’s own sample-size justification, so the illustration cannot establish the rationale for enrollment.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Collaborative Chronic Illness Care Model", "Usual Care")
n      <- c(308, 315)
events <- c(73, 38)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Collaborative Chronic Illness Care Model  308  73  23.7  19.1  28.9  19.3  28.8  19.3  28.8
ARM  Usual Care  315  38  12.1  8.7  16.2  8.9  16.1  8.9  16.1

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists 4 outcome measures and 4 analyses. It does not report how many of these were designated primary in the analysis plan, how the analyses were related, or whether a multiplicity procedure was used. When multiple outcomes or comparisons are examined, general statistical practice considers the chance of false-positive findings across the set of tests. Approaches may include prespecifying a primary comparison, adjusting significance thresholds, or interpreting findings as exploratory. The registry record does not state whether any of these approaches was used.

The primary measure is explicitly described, but the analysis details are not. The record does not report the statistical model, covariate adjustment, treatment of site structure, analysis population, or whether analyses followed an intention-to-treat principle or another approach. Because allocation was nonrandomized, knowing which participants and sites entered each analysis would be important for evaluating comparability and interpretation.

Missing-data methods are not reported in the registry record. The record does not indicate whether appointment information was complete, how unavailable or incomplete records were handled, or whether participants without recorded appointments were considered to have no appointment. These distinctions matter for service-use outcomes: lack of documentation is not necessarily equivalent to verified non-use. No assumptions about missing observations or their impact can be made from the information reported.

The record also does not describe interim analyses or stopping rules. Their presence or absence should not be inferred. Overall, the listed counts of outcomes and analyses establish that multiple measures and analyses are recorded, but do not reveal the prespecified hierarchy, multiplicity control, missing-data strategy, or analysis population.

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 adverse-event listing records 0 serious events among 389 at risk in the Collaborative Chronic Illness Care Model group and 0 serious events among 412 at risk in the Usual Care group. The listing also records 0 other events in each group, with at-risk counts of 389 and 412, respectively. No serious-event terms are listed.

These are event counts as recorded, not estimates of comparative safety. The registry record does not provide a statistical comparison, confidence interval, event definitions, or details about how events were collected and adjudicated. A recorded count of 0 does not establish that no event occurred outside the collection or reporting process, and it does not establish that the groups had identical risks. The figures should be read narrowly as the counts posted for the listed groups.

The at-risk counts in the safety listing should not be assumed to be the denominators for the primary weight-service endpoint, which is restricted to participants overweight at baseline. The registry record does not report the primary endpoint denominators or counts. Nor should the adverse-event listing be treated as a complete account of all possible harms: broader safety definitions and ascertainment details are not reported in the registry record.

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

The study addresses a practical question about implementation: whether a collaborative care model is associated with use of weight services among participants who were overweight at baseline. The intervention description connects service organization, coordination, and evidence-based protocols with wellness and supported employment services. The endpoint focuses on one specific part of that broader approach—having one or more weight-service appointments during 1 year.

The registry record does not report the primary outcome counts or a treatment-effect estimate. It therefore cannot establish whether appointment use was more common at implementation sites, more common at control sites, or similar between them. No conclusion about the direction or magnitude of the primary result is supported by the information recorded. In particular, an absent result is not evidence of no difference.

Even with a reported difference, this nonrandomized site comparison would call for caution about causal interpretation. Implementation and control sites may differ in ways that affect both service delivery and observed use. Without reported baseline information and analysis methods, it is not possible to assess the extent to which those differences were accounted for. The factorial label and single-masking designation provide limited design information because the factors and the party masked are not identified.

The measure also has a narrow meaning. It records whether a participant had at least one appointment, not the number or quality of services, changes in weight, broader health outcomes, or return to competitive work. The project’s stated goals extend beyond this primary measure, but the endpoint alone cannot establish whether those broader goals were achieved.

The serious-event listing records 0 in each group, but that does not answer the primary service-use question or establish overall safety. The most appropriate conclusion is that the registry record documents the study’s setting, intervention, comparison, and target endpoint, while leaving the primary result and key analysis details unreported. Interpretation should remain within those limits.

14. Why This Trial Matters Statistically

This trial is statistically instructive because it brings an implementation question into a nonrandomized site comparison. Many real-world care models are introduced within existing organizations rather than assigned to individuals by random allocation. Such evaluations can address questions about service delivery in practice, but their comparisons require careful attention to how sites and participants differ before implementation.

The endpoint is also a useful example of matching an outcome to a service question. Counting participants with one or more appointments captures whether eligible people reached a targeted service. It is not interchangeable with counting appointments or measuring clinical change. A precise endpoint definition helps readers understand what a result would represent and what it would leave unanswered.

At the same time, the registry record illustrates why design labels are not a substitute for analysis details. “Factorial” and “single” masking identify aspects of the recorded design, but do not reveal the factors, masked party, or statistical model. Similarly, a count of outcomes and analyses does not show whether multiplicity was addressed, and an enrollment figure does not explain the study’s sample-size rationale.

The record’s missing primary results place a clear boundary on interpretation. Without outcome counts and estimates of uncertainty, neither the direction nor precision of the comparison can be evaluated. The listed serious-event counts are not a substitute for primary results. Statistically, this case underscores the distinction between a well-defined research question and an interpretable estimate: both are needed to assess what a study’s evidence supports.

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 primary limitation for interpreting the endpoint is that the registry record does not report the observed group counts, effect estimate, confidence interval, or test result for weight-service use. As a result, the primary comparison cannot be quantified or evaluated for precision. The record identifies the endpoint and its target population but not the outcome values needed to interpret it.

Nonrandomized allocation is an additional design limitation. Implementation sites and control sites may differ in relevant ways, and the registry record does not report baseline comparisons or adjustment methods. The possibility of site-level clustering also matters for a site-based comparison, but the record does not describe the number of sites or how site structure was handled in analysis. These are interpretive uncertainties, not evidence that a particular bias occurred.

The factorial model and single-masking designation are not elaborated. The factors and their analysis are not reported, and the masked party is not identified. The registry record also does not describe recruitment, detailed eligibility criteria, usual-care content, or the process for measuring appointments. These omissions restrict assessment of implementation fidelity and outcome ascertainment.

Further details about missing data, analysis populations, covariate adjustment, multiplicity, interim analyses, and the trial’s own sample-size justification are not reported in the registry record. The sample-size note concerns planning illustrations and should not be interpreted as observed power or as a description of the original design rationale. Finally, the primary endpoint captures any qualifying appointment, not service intensity or a clinical outcome, so it provides only a limited view of the project’s broader aims.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Implementing Effective, Collaborative Care for Schizophrenia (EQUIP-2) — NCT00137280.
• Registry record last updated 2018-02-22; status last verified 2018-01.
• Results first posted 2014-10-15.
• 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 registry record describes Implementing Effective, Collaborative Care for Schizophrenia (EQUIP-2), a completed, nonrandomized study sponsored by the VA Office of Research and Development. Enrollment is listed as 1067. The study compares implementation sites using the Collaborative Chronic Illness Care Model with control sites receiving Usual Care, and is labeled factorial with single masking.

The primary endpoint is the number of participants with one or more weight-service appointments during 1 year, among those overweight at baseline. The record lists 4 outcome measures and 4 analyses, but does not report the primary outcome counts, effect estimate, uncertainty interval, or analysis methods. Serious adverse event counts are listed as 0 in both groups, with at-risk counts of 389 and 412. Those safety figures do not provide the missing primary endpoint results or establish overall safety.