Electronic Health Record-leveraged, Patient-centered, Intensification of Chronic Care for HF: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Electronic Health Record-leveraged, Patient-centered, Intensification of Chronic Care for HF (NCT03334188), a randomized na study. The primary outcome is Percent of Participants Who Experienced an Initiation or Intensification of Their Guideline-Directed Medical Therapy (GDMT).

2. Clinical Question

EPIC-HF examined whether patient engagement materials were associated with more frequent initiation or intensification of guideline-directed medical therapy (GDMT) among participants with heart failure with reduced ejection fraction (HFrEF). The registry describes a randomized, two-arm study in which eligible patients at participating University of Colorado Health locations were assigned to intervention or control. The primary measure was whether a participant experienced an initiation or intensification of GDMT, assessed by reviewing medications and doses around the cardiology clinic visit at enrollment and through 30 days after enrollment.

The posted results report 29.7% in control and 49.0% in intervention. These are group-level percentages; they do not describe the experience of every participant or establish why a medication change occurred. The comparison is specific to the recorded endpoint and follow-up period, and does not by itself answer questions about longer-term medication use or clinical outcomes.

3. Trial Design

The registry identifies EPIC-HF as a randomized, parallel-group trial with no masking. It lists the study status as completed and the phase as not applicable. Enrollment was 306. The registry records a start date of 2018-01-11 and a primary completion date of 2020-11-11.

The study description says that three main regional centers in the University of Colorado Health system participated, with sites in the Metro, North, and South regions. Eligible patients at participating sites were randomly enrolled into intervention or control. The control arm received usual care. The intervention arm received a patient engagement video and HFrEF medication checklist, with the arm description specifying email delivery one week before the next clinic appointment after enrollment. The summary also describes delivery by email and/or text.

Study procedures included surveys and medical-record collection. The summary describes a baseline survey, a follow-up survey, and medical-record information at baseline, one month after the first clinic appointment post-enrollment, and one year after enrollment. The primary endpoint, however, is defined around medication and dose review from immediately before the enrollment cardiology visit through 30 days after enrollment. These time descriptions refer to different parts of the registry description; the primary endpoint timeframe is the one relevant to the main comparison.

The registry reports 145 participants in each arm for the posted primary outcome. This outcome-specific reporting should be read separately from the stated enrollment of 306. The registry record does not explain the relationship between enrollment and the outcome denominators, describe losses to follow-up, or provide a participant flow account. It also does not report the trial's original sample-size justification or detailed analysis plan.

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:

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 Percent of Participants Who Experienced an Initiation or Intensification of Their Guideline-Directed Medical Therapy (GDMT), measured over Enrollment, 1 month (30 days) after enrollment.

The primary endpoint is the percentage of participants who experienced an initiation or intensification of GDMT. It is based on medical-record review of medications and doses immediately before the cardiology clinic visit at enrollment and through 30 days after enrollment. The registry frames this outcome as a measure of medication optimization for patients with HFrEF.

This is a binary participant-level outcome: each participant is classified according to whether an initiation or intensification was recorded during the specified interval. A percentage makes the two randomized groups comparable on the proportion experiencing the defined event, while retaining the clinically relevant distinction between an event being recorded and no event being recorded. It does not measure the magnitude of a dose change, the number of medication changes per participant, adherence, or whether a change continued beyond the follow-up interval.

Medication initiation or intensification is used as a process measure of treatment optimization in this study. It is not itself a measure of symptoms, hospitalization, survival, or another patient health outcome. The primary result therefore answers a focused question about recorded medication changes during the specified period. The registry's definition does not provide further detail on which specific medicines or doses qualified, how borderline cases were adjudicated, or whether changes were attributed to the engagement materials; those details are not reported in the registry record.

7. Results Posted in the Registry

• Intervention: 49.0% (71/145) versus 29.7% (43/145) on control; risk difference 19.3 percentage points (95% CI 8.3 to 30.3), risk ratio 1.65 (1.22–2.23); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is binary: a participant either had a recorded initiation or intensification of GDMT during the defined interval or did not. For a binary outcome, each arm's event percentage is the number with the event relative to the arm's reported outcome denominator. The registry reports 29.7% for control and 49.0% for intervention, with 145 participants reported in each group. The corresponding posted counts are 43 and 71.

What does the risk difference describe?

The risk difference is the absolute difference between the two event proportions. The recalculated unadjusted estimate is 0.193, with a 95% confidence interval from 0.083 to 0.303. The estimate is on the proportion scale: it describes how far apart the observed group proportions are in absolute terms. An absolute measure is useful because it keeps the comparison tied to the event frequency in the trial population, rather than expressing the contrast only as a relative change.

The registry also reports a risk ratio of 1.651, with a 95% confidence interval from 1.222 to 2.231, and an odds ratio of 2.276, with a 95% confidence interval from 1.404 to 3.688. The risk ratio compares event probabilities as a ratio. The odds ratio compares event odds as a ratio. These relative measures answer different mathematical questions and should not be treated as interchangeable with the risk difference. For a straightforward account of the absolute group contrast, the risk difference is often easiest to interpret.

What does the confidence interval contribute?

A confidence interval is a range of values generated by a statistical procedure intended to reflect uncertainty in an estimate. In repeated use under the procedure's assumptions, intervals constructed in this way would cover the underlying parameter at the stated confidence level at the stated frequency. For an individual trial, the interval gives a range of effect sizes compatible with the data and method; it is not a guarantee that the true effect lies within those endpoints. The interval's width reflects sampling uncertainty, but not all uncertainties in study conduct, endpoint definition, or generalizability.

For the risk difference, the confidence interval has lower and upper limits of 0.083 and 0.303. Because the interval does not include zero, the data are inconsistent with a zero risk difference at the corresponding two-sided confidence level under the calculation used. That statistical observation does not establish the importance of the difference for patients, nor does it resolve possible limitations in outcome ascertainment or missing data.

What does the p-value mean?

The unadjusted p-value is <0.001. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null hypothesis and statistical model. A small p-value can be evidence against that null comparison, but it does not quantify the probability that the null is true, the probability that the result is due to chance, or the size of an effect. It also does not account automatically for multiple outcomes, subgroup analyses, or other choices made during analysis.

The reported p-value is unadjusted. No multiplicity adjustment is described in the registry record, and the registry does not report the investigators' complete statistical analysis plan. Therefore, the p-value should be understood as belonging to the posted primary comparison, not as evidence that all potential analyses were controlled for multiple testing.

Why distinguish the analysis from the design?

Random assignment provides the design basis for comparing the intervention and control groups. In general, randomization aims to balance prognostic factors across groups on average, but it does not guarantee exact balance in a particular study. The registry identifies the allocation as randomized, the model as parallel, and masking as none. It does not report details of sequence generation, allocation concealment, or the trial's prespecified statistical model. The recalculated summary comparison is unadjusted; the registry does not report an adjusted estimate or detailed modeling approach.

The endpoint is a participant-level binary event, rather than a time-to-event outcome or a continuous measurement. Its interpretation depends on consistent classification of medication initiation or intensification and complete outcome ascertainment. Detailed adjudication procedures and missing-data methods are not reported in the registry record. As general statistical context, incomplete outcome data can affect both the event proportions and the uncertainty around a group comparison, particularly if the likelihood of missingness differs by outcome or arm.

The summary comparison uses the event counts and denominators posted for each arm. It does not incorporate unreported covariates, changes after the defined follow-up period, or the other outcomes listed in the registry. This distinction matters: an estimate can be calculated accurately from available summaries while still leaving unresolved questions about analysis populations and data handling. The posted numbers support a focused comparison of the endpoint as defined, with those limitations kept in view.

10. Confidence Intervals and Precision

The primary risk-difference estimate is 0.193, and its 95% confidence interval runs from 0.083 to 0.303. The interval does not include zero, and the full range is above zero. This indicates that, under the unadjusted calculation, the data are compatible with a positive absolute difference in recorded medication initiation or intensification between the arms. The interval is not a statement that every value in its range is equally likely, nor does it include uncertainty from every possible source.

The relative estimates also have confidence intervals: 1.222 to 2.231 for the risk ratio and 1.404 to 3.688 for the odds ratio. Each interval is expressed on the scale of its corresponding measure. The risk-difference interval is on the absolute proportion scale; the other intervals are ratios. Comparing their widths directly is not meaningful because they use different scales.

Precision depends in part on the amount of outcome information and on the variability of the observed data. A confidence interval gives a useful statistical account of sampling uncertainty, but it does not answer whether the estimate is clinically important, whether the endpoint was measured without error, or whether the result applies beyond the enrolled study population. The registry does not report further information on missing outcome data or adjusted estimates that would add context to the precision assessment.

The confidence intervals are recalculated from the posted summary statistics. They should be read as uncertainty around the unadjusted comparisons, rather than as a full accounting of design, measurement, or analysis uncertainty.

11. Planning a Similar Trial: Sample Size

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

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

The 80% power column gives the number of participants per arm required under those planning assumptions for an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column gives the corresponding number per arm for a 90% chance. These are planning probabilities under a specified statistical model; they do not say how likely the observed result is to be correct.

The rows consider planning for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. For 100% of the observed effect, the figures are 100 participants per arm for 80% power and 133 participants per arm for 90% power. For 75% of the observed effect, they are 174 participants per arm and 232 participants per arm, respectively. For 50% of the observed effect, they are 380 participants per arm and 508 participants per arm, respectively. Smaller effects need more participants to detect under otherwise fixed assumptions. These are planning illustrations, not observed power.

The curve should be read as a visual version of the same planning relationship between effect size and required participants. It does not represent the trial's observed recruitment or outcome data. In general, estimates based on an observed effect can be uncertain, so a planning exercise using smaller fractions offers alternative scenarios rather than a guarantee of what a future study will observe.

12. How the Sample Size Is Derived

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

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.297; treatment proportion p₁ = 0.490; average p̄ = 0.393.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6908 = 1.3539.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6771 = 0.5699.
4. n = (1.3539 + 0.5699)² / (0.1931)² = 99.2, rounded up to 100 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 <- 0.296551724137931                          # control proportion
p1 <- 0.489655172413793                          # 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(3334188)
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 = 100 (80% power), 133 (90% power)
 75% of observed effect: n per arm = 174 (80% power), 232 (90% power)
 50% of observed effect: n per arm = 380 (80% power), 508 (90% power)
power.prop.test: 99.2 
Simulated power at n = 100 per arm: 0.807 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists four secondary outcomes: self-reported patient engagement around HFrEF medications; participants with initiation or intensification of key heart failure medications; GDMT intensifications per patient; and safety outcomes. Numeric results for these secondary outcomes are not reported in the registry record. The primary comparison is therefore the only outcome comparison that can be interpreted quantitatively from the posted results.

When a study assesses multiple outcomes or makes multiple comparisons, the chance of at least one apparently unusual result can increase. Statistical procedures can address this multiplicity, but the registry record does not report whether any adjustment was prespecified or applied. The unadjusted p-value for the primary comparison should not be interpreted as if the registry documented a complete multiplicity strategy. Details of the protocol and statistical analysis plan, including any hierarchy among outcomes, are not reported in the registry record.

The record does not describe missing-data methods, exclusions from the primary analysis, or the analysis population definition. It reports 145 participants per arm for the primary outcome, but does not explain whether these denominators include everyone randomized, how missing medical-record information was handled, or whether any participants were omitted from the outcome calculation. The registry's stated enrollment is 306, and the connection between that enrollment figure and the outcome denominators is not explained in the record.

These omissions matter because the interpretation of a group comparison depends not only on the formula used but also on which participants and observations enter the analysis. As general statistical context, complete-case analyses can be biased if missingness is related to outcome status or differs across groups. No claim about such bias can be made from the registry information alone. The posted comparison should be read as based on the group counts and denominators that are reported, with the analysis-population and missing-data details not reported in the registry record.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

The registry reports zero participants with serious adverse events in the control group and zero in the intervention group. It also reports zero participants with other adverse events in each group. Deaths are reported for one participant in control and zero in intervention. These are the safety figures recorded for the two groups; they do not establish whether an event was caused by the study intervention.

The denominators at risk are reported as 145 in each group for both serious and other adverse events. A zero count means no event was recorded in that category in the registry results; it does not prove that the underlying risk is zero or that an event could not occur in a larger or longer study. The registry does not provide serious adverse-event terms, and the list of top serious terms is empty.

Safety counts are best interpreted alongside the recorded follow-up and reporting procedures. The registry does not provide detailed definitions, event-collection methods, or a separate explanation of the death report. The figures should therefore be read narrowly as posted counts, rather than as a complete account of all possible safety outcomes.

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

The posted results show a higher percentage of participants with a recorded GDMT initiation or intensification in the intervention arm than in the control arm during the defined period. The unadjusted risk difference is 0.193, with a 95% confidence interval from 0.083 to 0.303, and the unadjusted p-value is <0.001. The comparison is based on the posted counts of 71 of 145 in intervention and 43 of 145 in control.

This finding is specific to a medication-process endpoint, the two randomized groups, and the interval through 30 days after enrollment. It does not by itself demonstrate a difference in health outcomes such as symptoms, hospitalization, or survival. Nor does it establish which components of the engagement materials, if any, contributed to the recorded medication changes. The registry does not report numeric results for the listed secondary outcomes, so those cannot broaden the quantitative interpretation.

The statistical estimate describes a group-level contrast, not a prediction for an individual participant. Its confidence interval reflects sampling uncertainty under the unadjusted calculation, but not every possible source of uncertainty. Missing-data handling, detailed analysis populations, and adjusted analyses are not reported in the registry record. These limits should accompany interpretation of the effect estimate and p-value.

The trial was randomized and parallel-group, which provides a design for comparing assignment groups. The record also identifies the trial as unmasked. The registry does not report further details on allocation procedures, outcome adjudication, or the investigators' complete analysis plan. The posted result is therefore informative about the recorded primary endpoint while leaving important methodological details unresolved.

Nothing in this statistical interpretation is medical advice. It describes the registry's reported endpoint and the precision of its group comparison, without making a recommendation about care or asserting effects beyond the measures and follow-up recorded.

16. Why This Trial Matters Statistically

This trial illustrates how a behavioral or engagement-focused intervention can be evaluated using a concrete clinical process endpoint. Rather than relying only on a survey of engagement, the primary measure uses medical-record review to classify whether GDMT was initiated or intensified during a defined period. The resulting binary outcome supports a direct comparison of the percentages in the randomized groups.

Reporting both an absolute contrast and relative measures helps keep different statistical questions distinct. The risk difference gives the absolute separation between group proportions, while the risk ratio and odds ratio describe relative contrasts on their own scales. Confidence intervals are essential for understanding the precision of these estimates; a p-value alone would not communicate the range of values compatible with the data.

The record also shows why careful attention to denominators and outcome definitions matters. Enrollment is reported as 306, while the primary outcome is reported for 145 participants in each arm. The record does not explain this relationship. Without a participant flow account or missing-data description, readers cannot determine from the registry alone how those figures relate to the full enrolled population.

Finally, the listed secondary outcomes extend beyond the primary medication-change measure, but their numeric results are not reported in the registry record. Statistical conclusions should remain tied to what can be evaluated: the posted primary comparison and its uncertainty. Distinguishing observed results from planning illustrations and recognizing unreported analysis details are central to responsible interpretation.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The registry record leaves several trial-specific details unresolved. It does not report the full eligibility criteria, the original sample-size justification, or the investigators' complete statistical analysis plan. Although allocation is identified as randomized, sequence generation and allocation concealment are not described. Masking is listed as none, and the record does not describe safeguards against bias in outcome ascertainment.

The primary endpoint is based on medical-record review, but the registry does not specify which medicines or dose changes qualified as initiation or intensification, how outcome classifications were adjudicated, or whether reviewers were masked to group assignment. The endpoint captures a medication change over the specified interval; it does not report the persistence of that change or measure a patient health outcome.

The record reports 145 participants in each arm for the primary outcome, alongside enrollment of 306, without explaining the difference. It does not provide a participant flow account, reasons for missing data, exclusions, or the primary analysis population. The effect of these unknowns on the posted comparison cannot be assessed from the registry record.

The trial took place at listed University of Colorado Health locations, so results may not automatically generalize to other settings or populations. The registry lists four secondary outcomes but does not report their numeric results. It also does not report details of multiplicity control or adjustments. These limitations do not invalidate the posted primary comparison, but they constrain how broadly and confidently it can be interpreted.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Electronic Health Record-leveraged, Patient-centered, Intensification of Chronic Care for HF — NCT03334188.
• Registry record last updated 2021-10-06; status last verified 2021-09.
• Results first posted 2021-09-17.
• 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

EPIC-HF is recorded as a completed, randomized, parallel-group trial with no masking and phase listed as not applicable. Enrollment was 306. The recorded start date is 2018-01-11, and the primary completion date is 2020-11-11. The study compared patient engagement materials with usual care among eligible patients with HFrEF at University of Colorado Health locations.

The primary endpoint was the percentage of participants with a GDMT initiation or intensification from enrollment through 30 days after enrollment. The registry reports 29.7% in control and 49.0% in intervention, with 145 participants per arm in the outcome results. The recalculated unadjusted risk difference is 0.193, with a 95% confidence interval from 0.083 to 0.303 and p < 0.001.

Secondary outcome results, detailed missing-data methods, analysis populations, and the complete statistical analysis plan are not reported in the registry record.