Dissemination of CVD Risk Factor Treatment Among Diabetic Patients in Safety Net Clinics: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Dissemination of CVD Risk Factor Treatment Among Diabetic Patients in Safety Net Clinics (NCT02299791), a randomized na study. The primary outcome is Patients Indicated for ACE/ARB and Statin Who Had an Active Prescription for Both.

2. Clinical Question

The question concerns dissemination of the ALL intervention from an integrated care setting into Community Health Centers. Its focus is a population-level prescribing measure among patients with diabetes who were eligible for both medication types, rather than a direct comparison of individual patients assigned to receive a medication.

The registry describes a pre-post comparison within clinics alongside randomized, staggered implementation across clinics. The primary outcome is framed as the proportion of eligible patients who had an active prescription for both medications. This measure reflects prescribing status among the defined clinic population; it does not itself establish whether a prescription was filled, whether a medication was taken, or what happened to patients’ health.

The record identifies 6 study clinics in the early-intervention group and 5 in the late-implementation group. It does not report numerical changes in the outcome, an estimated difference between implementation groups, or an uncertainty interval. The clinical question is therefore identifiable from the registry, but the magnitude and precision of any observed change are not reported in the registry record.

3. Trial Design

The registry describes a randomized, parallel study with no masking. Its brief summary characterizes the implementation as staggered and randomized across clinics, with a pre-post comparison within clinics. Because implementation was assigned at the clinic level, the design is cluster-randomized: the clinics, rather than individual patients, are the units whose implementation timing was randomized.

Six clinics received the ALL intervention starting 6/1/11, and five clinics received it starting 6/1/12. The registry calls these groups Early Intervention and Late implementation. This timing created an early and a later implementation group, with the stated aim of examining changes in population prescription rates as the intervention was disseminated into Community Health Centers. The intervention is classified as “OTHER” in the record.

The registry lists enrollment as 4856 and the study status as completed. The listed study start is 2010-09, and the listed primary completion is 2014-05. The phase is recorded as NA. These administrative details do not establish how many participants contributed to each monthly outcome assessment, whether the enrolled population changed over time, or how missing observations were handled.

Randomization at clinic level matters for interpretation: patients within the same clinic may share care processes, so their outcomes are not necessarily statistically independent. A simple individual-level re-analysis could therefore give misleading uncertainty estimates. The registry record does not report the allocation procedure in further detail, the number of clinics considered before assignment, or the sponsor’s model specification and numerical treatment-effect estimate.

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: clinic patients with diabetes who had a clinic encounter (in person or by telephone) within the previous year and were indicated for ACE/ARB and statin per current national care guidelines

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Patients Indicated for ACE/ARB and Statin Who Had an Active Prescription for Both, measured over Percent of clinic patients prescribed guideline-concordant cardioprotective medications, as of the 1st day of each month, from up to 36 months.

The primary endpoint is the number of patients indicated for ACE/ARB and statin who had an active prescription for both, described as a proportion of patients indicated for both. The population includes clinic patients with diabetes who had an in-person or telephone encounter within the previous year and met the stated guideline-based indication for both medication types.

The timeframe describes the percentage of clinic patients prescribed guideline-concordant cardioprotective medications, assessed as of the first day of each month, for up to 36 months. Repeated monthly measurement can describe how a clinic-level prescribing proportion changes over the intervention period, rather than reducing the outcome to a single assessment point.

This endpoint is used to assess prescribing practice because it captures whether the eligible clinic population had an active prescription for both medication types. It is a process measure, not a measure of medication use or clinical events. An active prescription does not show whether a patient filled or took the medication, and the endpoint does not directly measure cardiovascular outcomes.

The outcome title identifies patients with active prescriptions, while the description specifies the proportion of eligible patients. The registry lists the measure unit as Participants and the parameter as a count of participants, but it does not provide the observed counts or percentages. The numerical endpoint results are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Why does the clinic-level assignment matter?

In a cluster-randomized study, groups such as clinics are assigned to implementation conditions, while outcomes are measured on people or on the group population. Patients within one clinic can share staff, workflows, and local practice patterns. As a result, their measurements may be more alike than measurements from different clinics. This dependence is called within-clinic correlation.

In general, a statistical analysis that treats every patient as independent can understate uncertainty when outcomes are correlated within clinics. Appropriate methods account for the assignment unit and the correlation structure, for example by modeling clinics or using a cluster-aware variance estimate. The registry record identifies the cluster-randomized design and cautions that a simple individual-level re-analysis would be misleading. It does not report which statistical method the sponsor used, the correlation estimate, or the model assumptions.

What does staggered implementation add?

The early-intervention clinics began the ALL intervention starting 6/1/11, and the late-implementation clinics began starting 6/1/12. The brief summary describes both a pre-post comparison within clinics and randomized, staggered implementation across clinics. In general, repeated measurements before and after implementation can help describe changes over time, while the later-implementing clinics offer a comparison during the staggered rollout.

To interpret that design statistically, a reader would need to know how calendar time and implementation timing were represented in the analysis, how monthly observations were handled, and whether differences between clinics or changes over time were modeled. Those details are not reported in the registry record. The design description does not by itself establish that any particular model or causal interpretation was used.

How should the endpoint be understood?

The endpoint concerns a monthly clinic-population prescribing proportion: eligible patients with active prescriptions for both medication types relative to the eligible population. In general, a proportion has a numerator and a denominator, and both can vary over repeated assessments. Changes can therefore reflect changes in prescribing status, in the eligible population, or in both. The registry defines the population and the intended measure but does not provide the monthly counts or estimates.

The measure is a prescribing process outcome. In general, process measures and health outcomes answer different questions: an active prescription records prescribing status, while medication use and clinical outcomes require other measurements. This distinction limits what can be inferred from the endpoint even if its numerical results were available.

What do uncertainty estimates communicate?

In general, a confidence interval presents uncertainty around an estimate under a specified analysis and its assumptions. Its width depends on the information in the data and the analysis, including the number of clusters and the extent of within-clinic correlation in a cluster-randomized study. A p-value evaluates how compatible data are with a stated null hypothesis under a statistical model; it does not quantify the clinical importance of an effect.

The registry record reports no confidence interval, p-value, or effect estimate for the primary outcome. Thus, there is no trial-specific uncertainty estimate to interpret. The enrollment figure alone cannot provide that information, especially when assignment occurred by clinic and the outcome was assessed repeatedly.

What do planning calculations mean?

In general, a sample-size calculation describes the amount of information a hypothetical design would require for a specified effect, significance level, allocation, and statistical approximation. The record’s planning assumptions characterize any such illustration as based on a fraction of an observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. These are planning assumptions, not a description of the trial’s original design justification.

For non-time-to-event outcomes, such illustrations are expressed as participants per arm. They are not observed sample sizes, observed power, or evidence about whether the trial detected an effect. The registry does not report an original sample-size justification, and numerical effect estimates for the primary outcome are not reported in the registry record. Therefore, no trial-specific planning quantity or detectable effect can be interpreted from the registry record alone.

Overall, the design indicates that a suitable analysis should respect clinic-level assignment and the repeated outcome structure. However, the registry does not provide enough statistical analysis detail or outcome results to assess the sponsor’s model, its uncertainty estimates, or the robustness of its findings.

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("Early Intervention", "Late Implementation")
n      <- c(2053, 2803)
events <- c(1279, 1318)
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  Early Intervention  2053  1279  62.3  60.2  64.4  60.2  64.4  60.2  64.4
ARM  Late Implementation  2803  1318  47.0  45.2  48.9  45.2  48.9  45.2  48.9

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure, but it lists zero analyses. It does not report a multiplicity plan, such as whether multiple time points, contrasts, or secondary outcomes were tested, or whether any adjustment was made for multiple testing. Without reported analyses, it is not possible to determine which comparisons were conducted or how multiplicity was handled.

The record describes monthly measurement for up to 36 months. It does not state which monthly observations were included in an analysis, whether one time point was selected as primary for inference, or whether the trajectory over time was modeled as a whole. Repeated observations raise questions about correlation and the choice of summary comparison; the registry record does not report the corresponding analysis details.

Missing-data methods, the amount of missing outcome data, and reasons for missingness are not reported in the registry record. The record also does not describe exclusions or provide an analysis population, such as whether analyses included all eligible patients, only patients with complete observations, or another defined group. Enrollment is listed as 4856, but this does not establish how many observations were available for each monthly assessment.

These omissions prevent assessment of whether missing data could have influenced estimates or whether the reported analysis, if any, followed a prespecified plan. The registry record does not report a statistical analysis plan, interim analyses, or a procedure for handling multiple comparisons.

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

The registry record establishes that the study examined dissemination of the ALL intervention into Community Health Centers and defined a clinic-population prescribing outcome among patients with diabetes who were indicated for both ACE/ARB and statin medications. It also describes an early and a later implementation group and a repeated, monthly outcome timeframe. These registry entries identify the study question and the intended measurement.

The record does not report numerical outcome results or analyses. It therefore does not show how the prescribing proportion changed, whether changes differed by implementation timing, or how precisely any estimated difference was measured. The absence of numerical results in the registry record should not be interpreted as evidence of no change, and it does not establish that a change occurred.

Even a reported difference in this endpoint would concern active prescriptions, not necessarily medication filling, taking, or clinical benefit. The endpoint is a prescribing measure among an eligible clinic population. It does not directly report cardiovascular events or other patient health outcomes.

In general, study design and results answer different questions. Randomized clinic-level implementation describes how clinics were allocated to timing groups; only reported estimates and suitable uncertainty measures can describe the observed comparison. The registry’s design description does not replace those results. Nor can the enrollment figure alone establish the precision of a cluster-level comparison.

Any planning illustration should be understood as hypothetical rather than as observed power or as a trial result. The registry record does not provide an original sample-size justification or the primary outcome estimates needed to assess the observed effect. Accordingly, the record supports a description of the trial’s question and design, but not a conclusion about the size, direction, or precision of the outcome finding.

13. Why This Trial Matters Statistically

This study raises a central statistical issue in implementation research: the unit receiving an intervention may differ from the unit whose outcome is measured. Here, implementation timing was assigned across clinics, while the primary measure concerns patients in clinic populations. An analysis that ignores the clinic-level assignment can misrepresent uncertainty because patients in the same clinic may share care processes.

The staggered rollout also makes time an important part of the design. The early and late implementation groups began the intervention at different times, and the outcome was intended to be assessed monthly. Interpreting such a design requires clarity about how clinic differences, calendar time, repeated measurements, and implementation timing are handled. The registry describes the structure but does not report the statistical model or numerical results.

The endpoint illustrates why outcome definitions matter. A proportion of eligible patients with active prescriptions is a population-level prescribing measure. It can assess a care process, but it is not equivalent to medication use or a clinical outcome. Keeping those distinctions clear prevents a process measure from being interpreted as evidence about an outcome it does not measure.

Finally, completeness of reporting is essential for statistical interpretation. Without estimates and measures of uncertainty, readers cannot assess the observed magnitude or precision of the primary outcome. The registry identifies a meaningful design and outcome, but its record does not report an outcome analysis. That difference between a clearly described question and an interpretable numerical result is itself important when evaluating trial records.

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 main limitation of the registry record for statistical interpretation is that it lists one outcome measure but zero analyses. It gives no numerical primary outcome results, effect estimate, confidence interval, or p-value. As a result, the magnitude, direction, and precision of any change in the prescribing outcome cannot be assessed from the record.

The study randomized implementation timing at the clinic level, with 6 clinics in the early-intervention group and 5 in the late-implementation group. The record does not report the sponsor’s statistical model, how within-clinic correlation was handled, or how repeated monthly measurements were analyzed. Those omissions prevent evaluation of whether the analysis appropriately reflected the design.

The registry does not report the number of clinics or participants contributing outcome data at each assessment, the amount or handling of missing data, or the analysis population. The listed enrollment of 4856 does not establish the number of observations available for the primary endpoint or the information supporting a clinic-level comparison.

Multiplicity procedures, interim analyses, and a statistical analysis plan are not reported in the registry record. Nor is an original sample-size justification reported. The record’s description of a planning illustration does not supply an observed outcome result or establish observed power. These reporting gaps limit assessment of the analytical choices and the robustness of any trial finding.

Finally, the endpoint captures active prescription status, not medication use or clinical events. This is a limitation of what the measure can establish, separate from the absence of numerical results.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Dissemination of CVD Risk Factor Treatment Among Diabetic Patients in Safety Net Clinics — NCT02299791.
• Registry record last updated 2017-10-04; status last verified 2017-10.
• Results first posted 2017-10-04.
• 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

NCT02299791 is a completed, randomized, parallel study of dissemination of the ALL intervention among diabetic patients in safety net clinics. The registry lists Kaiser Permanente as sponsor, enrollment of 4856, a start of 2010-09, and primary completion in 2014-05. The phase is listed as NA, and masking as none.

The early-intervention group comprised 6 clinics starting the intervention on 6/1/11; the late-implementation group comprised 5 clinics starting on 6/1/12. The primary outcome concerns the proportion of eligible clinic patients with active prescriptions for both ACE/ARB and statin medications, assessed monthly for up to 36 months.

The record lists one outcome measure and zero analyses. Numerical primary outcome results, uncertainty estimates, analysis details, missing-data methods, and an original sample-size justification are not reported in the registry record. The available record therefore describes the study question, design, and endpoint, but does not provide results for a statistical assessment of the observed outcome.