Reducing Cardiovascular Risk in Adults With Serious Mental Illness: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Reducing Cardiovascular Risk in Adults With Serious Mental Illness (NCT02451670), a randomized na study. The primary outcome is Predicted Annual Rate of Change in 10 Year Risk of Fatal or Nonfatal Heart Attack or Stroke.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group study with no masking. Clinics were randomized to the prioritized clinical decision support intervention or usual care. The registry describes patients and primary care providers in intervention clinics as receiving patient-specific written advice, while those in usual-care clinics were not presented with that support. The trial therefore evaluated an electronic health record-based decision-support strategy in clinical practice, rather than assigning individual patients to a drug or procedure.

The registry lists enrollment as 10347 and the status as completed. The study began on 2016-01-20, with primary completion on 2018-09-19. The listed phase is NA. The sponsor was HealthPartners Institute. The registry record identifies chronic disease, mental disorder, health behavior, bipolar disorder, and schizophrenia among the conditions associated with the study.

The reported primary outcome values are based on patients who met eligibility criteria, had an eligible index visit at a randomized clinic, and had at least one follow-up visit in a randomized primary care clinic. The registry reports 4550 in the intervention arm and 4387 in usual care for this outcome. These are outcome-specific arm counts; they should not be treated as interchangeable with overall enrollment. The registry does not report further detail about how clinic-level assignment was incorporated into the statistical analysis, or whether any adjustment was made for similarities among patients treated at the same clinic.

Although the study is described as randomized, randomization alone does not explain how the primary comparison was estimated. The registry record does not report the analysis model, adjustment variables, or the handling of the clinic assignment structure. It also reports no posted analysis results beyond the arm-level primary outcome values.

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: The patients whose data were included in the primary outcome analysis met eligibility criteria, had an index visit at a randomized clinic at which they were eligible for the Cardiovascular Wizard intervention, and had at least one follow-up visit in a randomized primary care clinic.

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: Patients whose data was included in the baseline analysis population met eligibility criteria, had an index visit at a randomized clinic at which they were eligible for the CV Wizard intervention, and had at least one follow-up visit in a randomized primary care clinic.

6. Primary Endpoint

The registry lists the primary outcome as Predicted Annual Rate of Change in 10 Year Risk of Fatal or Nonfatal Heart Attack or Stroke, measured over Index to 12 months post index visit.

The primary endpoint is titled predicted annual rate of change in 10-year risk of fatal or nonfatal heart attack or stroke. It concerns modifiable cardiovascular risk, rather than the occurrence of heart attacks or strokes during follow-up. The registry describes calculating, for each cardiovascular risk factor not at its optimal goal, the difference between total 10-year atherosclerotic cardiovascular disease risk using the patient’s actual values and risk using the goal value. These risk components were summed to form total modifiable cardiovascular risk at the index visit and at subsequent encounters.

Annual rate of change was estimated from patient encounters during the intervention period. The primary comparison was the difference in the model-estimated rate of change at 12 months after the index visit. In general, a risk-factor-based endpoint can reflect changes in measured or recorded risk factors without requiring enough clinical events to estimate an event rate. It is therefore a modeled measure of risk, not a count of observed cardiovascular events.

The registry reports values of 14.2 for prioritized clinical decision support and 20.8 for usual care, with the unit described as percentage of annual rate of change. It does not report the uncertainty around these values or a between-group estimate. The endpoint description does not state a sign convention that would establish, from the two values alone, how the direction should be interpreted. The registry record also does not report how missing encounter measurements were handled.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is a model-based measure of change in cardiovascular risk, not a directly counted clinical event. The registry describes deriving modifiable risk from the difference between estimated 10-year risk using a patient’s actual risk-factor values and estimated risk using goal values. Components for risk factors not at goal were summed, and the resulting measure was calculated at the index visit and subsequent encounters. A rate of change was then estimated from encounters during the intervention period.

What does a rate of change represent?

In general, a rate describes how an outcome changes over time. A rate estimated from repeated encounters can use information from more than a single baseline and follow-up measurement. That can be useful when measurements are collected at several clinical visits. The estimate nevertheless depends on how time and repeated measurements are represented in the model. The registry record does not report the specific model, assumptions about the pattern of change, or how irregular visit timing was handled.

The stated primary comparison was the difference in model-estimated rate of change at 12 months after the index visit. This is a comparison of modeled trajectories, not a comparison of the number of heart attacks or strokes observed during that period. The endpoint refers to the predicted 10-year risk of fatal or nonfatal heart attack or stroke, while the timeframe for the analysis is index to 12 months post index. These are different time concepts: the risk score projects risk over 10 years, while the study evaluates change over the stated follow-up period.

Why is the analysis unit important?

Clinics, rather than individual patients, were randomized. In general, when assignment occurs by clinic, outcomes for people treated at the same clinic may be correlated. A statistical analysis can account for that dependence through an appropriate model or variance calculation. If dependence is ignored, estimated uncertainty may be too small; the size and direction of any impact depend on the data and analysis. The registry record does not report whether or how clinic-level dependence was accounted for, so the precision of the comparison cannot be assessed from the posted values alone.

The intervention was delivered through clinical decision support and written advice to patients and providers. This means the comparison concerns a care-process strategy implemented in clinics. The registry record does not report whether exposure to the advice was measured, how consistently it was displayed, or whether patients and providers acted on it. Such details can help distinguish assignment to an intervention strategy from actual receipt or use of its components.

What can the arm values establish?

The registry reports 14.2 for the intervention arm and 20.8 for usual care. These values describe the primary endpoint as posted, but they are not accompanied by a standard error, confidence interval, or p-value. In general, a confidence interval describes the range of values compatible with an estimate under a specified statistical procedure and its assumptions. It communicates precision as well as direction and magnitude. Without an interval or another measure of uncertainty, the precision of these arm-level estimates and of their contrast is not available.

The record lists 4550 and 4387 patients in the primary outcome arms. It does not report additional details about the distribution of measurements, missingness, or the model used to obtain the arm-level values. The endpoint description says that risk was calculated at the index and subsequent encounters, but it does not explain how patients with incomplete follow-up measurements were treated. The record also does not identify any covariates used in estimation.

How should planning illustrations be understood?

General statistical planning calculations can illustrate the number of participants per arm associated with a specified power under stated assumptions. The assumptions described for the planning illustrations include a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects expressed as a stated fraction of an observed effect. Such calculations are planning illustrations, not observed power and not evidence that the trial had a particular probability of detecting an effect. The registry record does not report that these calculations formed part of the trial’s original design or sample-size justification.

For the primary endpoint, the registry indicates that a two-arm comparison was not estimable from the posted summary statistics. Accordingly, no inferential comparison can be reconstructed from the arm-level values alone. Simple recalculation from summary values, where possible, would not substitute for the sponsor’s own analysis. The registry record reports no discrepancy flag, but it also does not post the sponsor’s statistical analysis or its uncertainty estimates.

These points illustrate a general distinction: descriptive summaries show what values were recorded or estimated in each group, whereas inferential analyses quantify uncertainty in a comparison. A randomized design informs the intended comparison, but interpretation still depends on the statistical model, handling of the clinic-level allocation, outcome data completeness, and estimation procedure. Those details are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and no posted secondary outcomes. It reports zero analyses. The record therefore does not describe a family of posted hypothesis tests, a multiplicity adjustment, or a hierarchy of secondary endpoints. It would be inappropriate to infer that no other analyses were conducted; the registry record simply does not report them. Any multiplicity strategy or interim-analysis plan is not reported in the registry record.

Multiplicity is a general concern when many statistical hypotheses are tested. As the number of tests grows, the chance of at least one apparently unusual result can increase unless the analysis plan accounts for that testing. The available record does not identify additional posted analyses, so the extent of any multiplicity issue cannot be assessed. The absence of reported secondary results is not evidence that secondary analyses were negative or were not performed.

The registry defines the primary outcome population as patients who met eligibility criteria, had an eligible index visit at a randomized clinic, and had at least one follow-up visit in a randomized primary care clinic. This definition indicates that inclusion in the primary outcome analysis required follow-up. The record does not report how many eligible patients lacked a follow-up visit, how much outcome information was missing, why it was missing, or how missing measurements were handled. It also does not report whether the primary analysis followed an intention-to-treat principle or used another analysis population definition.

In general, missing data can affect both the representativeness of the analyzed group and the precision of an estimate. The extent of concern depends on how much information is missing and on the reasons it is missing. Without those details, the potential influence of missing follow-up on the primary outcome cannot be evaluated. The record also does not report imputation procedures, sensitivity analyses, or assumptions about missingness.

11. 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: in patient hospitalization (16.7%), suicide attempt (2.7%).

The registry’s adverse-event listing gives, for prioritized clinical decision support, 4550 at risk, 722 with a serious adverse event, 1593 with another adverse event, and 526 deaths. For usual care, it gives 4387 at risk, 763 with a serious adverse event, 1882 with another adverse event, and 581 deaths. These are reported counts, not estimates of a causal effect of the intervention.

The record lists “In Patient Hospitalization” at 16.7% and “Suicide Attempt” at 2.7% among the top serious adverse event terms. It does not specify in this listing the denominator or further details needed to interpret those percentages. The registry also does not explain whether the death counts were judged related to the intervention, or how serious and other adverse events were defined, collected, or attributed.

General safety interpretation requires attention to event definitions, follow-up time, ascertainment, and whether events were considered treatment-related. Counts alone do not establish that an intervention caused or prevented an event. The listed groups also have different at-risk counts, so the counts should not be compared as if they were rates. No rate comparison, uncertainty estimate, or causal safety analysis is reported in the registry record.

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

The posted results establish that the registry reports primary outcome values of 14.2 in the prioritized clinical decision support arm and 20.8 in usual care, for an endpoint describing the predicted annual rate of change in modifiable 10-year cardiovascular risk. The primary outcome analysis counts are 4550 and 4387, respectively. The outcome was derived from risk-factor information across encounters, according to the registry description.

Those values do not, by themselves, establish the statistical evidence for a difference between the randomized groups. The registry does not report a contrast, confidence interval, standard error, or p-value. It also does not provide a sign convention sufficient to determine from the values alone which direction represents a more favorable change. As a result, the posted record does not support a conclusion that the decision-support strategy did or did not change cardiovascular risk.

The endpoint is a calculated risk measure rather than a count of observed heart attacks or strokes. A modeled risk outcome may be useful for describing changes in risk factors, but it should not be described as evidence that clinical events were prevented or caused. The registry record does not report an event-based primary endpoint or an analysis connecting the modeled endpoint to later cardiovascular events.

The randomized clinic design identifies the intended comparison, but assessment of the result also requires information about the analysis model and how clinic-level assignment was handled. Those methods are not reported. The primary outcome population required at least one follow-up visit, while the record does not explain the handling or extent of missing outcome data. These are limitations in what can be interpreted from the registry, not proof that the underlying trial analysis was invalid.

Overall, the registry provides descriptive arm-level primary outcome values and adverse-event counts. It does not report the inferential analysis needed to judge the magnitude, direction, or precision of the primary between-group comparison. Any stronger interpretation would require results and methods that are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically notable because the intervention was assigned at the clinic level, while the primary outcome was calculated from patient encounters. In general, the level at which randomization occurs is important to analysis: patients treated in the same clinic can share care processes or other influences, making their outcomes statistically dependent. An analysis that reflects the assignment structure is needed for appropriate uncertainty estimates. The registry record does not report how this feature was handled.

The outcome also illustrates the distinction between a clinical event and a modeled risk measure. The primary endpoint used cardiovascular risk-factor values to estimate modifiable risk and its rate of change. That approach can produce an outcome from routine clinical information without relying on observed heart attack or stroke events as the primary measure. At the same time, a modeled risk change is not equivalent to a demonstrated change in clinical events. The interpretation should remain aligned with what was measured.

There is also a useful distinction between arm-level summaries and a statistical comparison. The registry reports values for each arm, but not a contrast or uncertainty interval. In general, the two group summaries do not convey how precisely the difference is estimated. Randomization provides a design framework for comparison, but the result still depends on the analysis method, measurement completeness, and treatment of the clinic structure.

Finally, the registry record illustrates why transparent reporting matters. The stated primary hypothesis and endpoint description identify what the trial aimed to compare, while the absence of posted inferential results leaves the primary question unresolved in this record. Reporting the model, comparison estimate, uncertainty, missing-data approach, and analysis population would allow readers to evaluate the result more fully.

14. 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 principal limitation for interpretation is that the registry reports arm-level primary outcome values without a between-group estimate or measure of uncertainty. There is no posted confidence interval, standard error, or p-value, and the registry lists zero analyses. Consequently, the statistical evidence for the primary comparison cannot be evaluated from the record.

The registry describes randomization of clinics but does not report how the analysis accounted for clinic-level assignment. This is important because the primary outcome was measured for patients receiving care in those clinics. The record also does not report the statistical model, any covariate adjustment, or the assumptions used to estimate the annual rate of change.

The primary outcome depends on risk-factor measurements across encounters. The record does not state how incomplete or irregular measurements were handled, how much outcome information was missing, or whether sensitivity analyses were conducted. The primary analysis population required at least one follow-up visit, so the results may not represent every enrolled patient; the record does not provide enough information to assess the possible effect of that restriction.

The endpoint is a model-based estimate of 10-year risk, not an observed count of cardiovascular events. The registry does not report a sign convention that clarifies how the values map to favorable or unfavorable change. The adverse-event listing supplies counts and selected terms but not the details needed for a comparative safety analysis, including event attribution and the denominator for the listed percentages.

Finally, the record does not report an original sample-size justification, interim-analysis plan, or multiplicity approach. The planning illustrations described for the site are not a substitute for the trial’s own design justification or results. These limitations concern the information reported in the registry record; they do not establish what analyses may have appeared elsewhere.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Reducing Cardiovascular Risk in Adults With Serious Mental Illness — NCT02451670.
• Registry record last updated 2019-11-05; status last verified 2019-03.
• Results first posted 2019-11-05.
• 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.

18. Record Summary

NCT02451670 evaluated prioritized electronic clinical decision support for cardiovascular risk management in adults with serious mental illness. Clinics were randomized in a parallel, unmasked design to the intervention or usual care. Enrollment was 10347, and the study status is completed.

The primary endpoint was the predicted annual rate of change in modifiable 10-year risk of fatal or nonfatal heart attack or stroke, assessed from the index visit through 12 months. The registry reports values of 14.2 for the intervention arm and 20.8 for usual care, with primary outcome counts of 4550 and 4387. It does not report an inferential comparison or uncertainty estimates.

The record lists adverse-event counts and selected serious-event terms but does not provide the detail needed for causal interpretation. It reports no secondary outcomes or analyses. Important statistical details—including the analysis model, clinic-level adjustment, and missing-data handling—are not reported in the registry record.