Early Detection of GEnetic Risk (EDGE): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Early Detection of GEnetic Risk (EDGE) (NCT04746794), a randomized na study. The primary outcome is Rates of Screening.

2. Clinical Question

The screening pathway was intended to identify patients at high familial cancer risk. Those found to be at high risk would be offered genetic testing for a panel of hereditary cancers, and a “previvor” plan would support appropriate follow-up for patients with an identified mutation. The primary outcome, however, is screening completion—not genetic testing, mutation identification, or follow-up care. Those steps should not be conflated with the principal comparison.

3. Trial Design

EDGE, titled “Early Detection of GEnetic Risk,” is recorded as a randomized, parallel-design study with no masking. Its intervention is classified as behavioral, and the registry lists population-level screening. The University of Washington is the sponsor. The study status is completed; the recorded start date is 2020-09-25 and primary completion date is 2023-05-31.

The comparison is between two ways of reaching patients for familial cancer risk screening. In the Point of Care arm, patients were approached when attending a routine primary care clinic visit, with screening offered using a tablet in the waiting room or an online method for telehealth. In the Direct Patient Engagement arm, active patients were identified from clinic records and contacted by postal mail and email with a screening link. The registry also describes stakeholder interviews and surveys involving samples of patients, providers, and clinic leaders, using quantitative surveys and qualitative interviews at baseline and follow-up.

The primary outcome is titled Rates of Screening. It is defined as the fraction of the active clinic patient population completing screening, with a one-year timeframe. Eligibility is described as patients 25 years of age or older, comfortable speaking English, and having an appointment at a participating clinic during the 12 month recruitment window.

The registry lists enrollment as 20184. The posted arm denominators for the primary outcome are different from that enrollment figure; the registry record does not explain how the enrollment figure relates to those outcome denominators. It also does not report the original sample-size justification, the allocation unit, or further details about implementation across clinics. These details should not be inferred from the arm descriptions alone.

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: As randomization was done by clinic, 115,484 individuals started the study (115,319 patients; 165 primary care providers). The 95,623 patients seen during the 12 month window set by the study protocol serve as the denominator for the main outcomes (proportion of patients who completed the cancer risk assessment screening; proportion of patients who completed genetic testing). 20,184 individuals were officially enrolled through completing surveys or the cancer risk assessment screening.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Patients 25 years of age or older, comfortable speaking English, who had an appointment at one of the participating clinics during the 12 month recruitment window.

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: Baseline characteristics are provided for all patients who completed the cancer risk assessment screening. And for patients and primary care providers who completed a baseline survey prior to study launch. Patients, Providers, and Clinic Leaders who enrolled in the study by completing surveys post-intervention did not have Baseline data collected (2,286).

6. Primary Endpoint

The registry lists the primary outcome as Rates of Screening, measured over 1 year.

The primary endpoint is the number of participants who completed familial cancer risk screening, expressed in relation to the active clinic patient population. The registry’s endpoint description calls this a fraction and sets the timeframe at one year. The posted results give the screened counts and corresponding population denominators for both engagement strategies.

This endpoint addresses whether patients were reached and completed the initial risk assessment. It is relevant to the trial’s implementation question because the two arms use different routes to invite patients: an approach around a clinical visit versus outreach based on clinic records. A screening-completion measure can therefore describe the uptake associated with each route.

The endpoint does not measure whether screening accurately identifies inherited risk, whether eligible patients proceed to genetic testing, whether a mutation is found, or whether follow-up occurs. The study description includes those later steps as part of the intended pathway, while the primary outcome is limited to screening. Secondary outcomes include screening and testing rates at Healthcare System A, patients screened and tested, and costs or cost-effectiveness for the engagement strategies. Those outcomes address additional questions, but they do not change what the primary endpoint measures.

The registry record does not report further operational detail about how completion was verified or how incomplete screening attempts were handled. Those details matter when interpreting a process measure, and should not be assumed.

7. Results Posted in the Registry

• Direct Patient Engagement: 8.7% (3813/43930) versus 19.1% (9892/51693) on control; risk difference -10.5 percentage points (95% CI -10.9 to -10.0), risk ratio 0.45 (0.44–0.47); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What kind of outcome is being compared?

The primary outcome is binary at the participant level: a patient either completed screening or did not. The posted arm results give event counts—the number completing screening—and denominators for the active clinic patient populations. The outcome is titled Rates of Screening, and its stated timeframe is one year. Although the endpoint is described as a fraction, the registry also reports counts and a comparison of rates.

For a binary outcome, a rate is interpreted as the share of the relevant group with the event. The comparison depends on both the number who completed screening and the denominator used for that arm. Looking at counts alone can be misleading when the groups have different denominators, so the arm rates and a rate-based contrast are useful complementary summaries.

What does the risk difference express?

The unadjusted recalculation expresses the contrast as a risk difference, with Direct Patient Engagement compared with Point of Care. A risk difference subtracts one arm’s event rate from the other arm’s rate, using the stated comparison direction. A value below zero in this orientation means the event rate is lower in DPE. Unlike a relative measure, a risk difference describes an absolute gap between rates. It is generally easier to relate to a difference in event frequency, although its practical importance still depends on the setting and decision context.

The registry also includes relative measures, but interpretation of those measures is not necessary to understand the primary contrast. Any such measure compares the arms on a relative scale rather than an absolute scale. Different scales answer related but not identical descriptive questions.

What does a confidence interval add?

A confidence interval gives a range of estimates compatible with the observed data under the assumptions of the method used to construct it. It communicates uncertainty around the point estimate, rather than only classifying the result by a significance threshold. A narrower interval indicates more precision on that scale; a wider interval indicates less precision. For the posted unadjusted risk difference, the interval is on the negative side of no difference. This supports the direction of the observed contrast while leaving uncertainty about its exact size.

A confidence interval does not mean that there is a stated probability that the fixed true effect lies within this particular interval. Nor does it account automatically for selection, measurement problems, clustering, or other design features not represented by its calculation. The registry record does not report enough detail to assess every such assumption.

How should the p-values be read?

A p-value describes how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the analysis assumptions were correct. It is not the probability that the null hypothesis is true, and it does not measure the size or practical value of the observed difference. The registry reports a chi-squared analysis with a p-value, while the unadjusted recalculation reports a separate p-value. The values differ. The record does not explain the difference in detail, so each should be attributed to its own analysis rather than combined.

The sponsor-reported chi-squared result is a test of association between arm and outcome under the test’s assumptions. The recalculated risk difference provides an effect estimate and interval based on the posted summary statistics. One gives a test result; the other emphasizes the size and uncertainty of an absolute contrast. These analyses can be read together as complementary descriptions, but one should not be substituted for the other.

Why can over-dispersion or clustering matter in general?

In general, outcomes from patients treated in the same clinic may be more alike than outcomes from patients in different clinics. If so, treating every patient as statistically independent can understate uncertainty. This is one reason cluster-aware methods may be considered in studies involving multiple clinics. This is a general statistical point, not a claim about the method used in EDGE. The registry record does not report the allocation unit, whether clustering was accounted for, or whether an over-dispersion adjustment was used.

Similarly, adjustment for patient or clinic characteristics can change an estimate when those characteristics relate to outcome and differ across groups. The registry record does not report adjusted estimates or covariates for the primary analysis. Therefore, the recalculation should be understood as unadjusted, not as a replacement for any unreported adjusted analysis.

10. Confidence Intervals and Precision

Precision is reflected in how tightly the estimated contrast is bounded by its confidence interval. The unadjusted risk-difference interval is entirely below zero, so the posted interval supports a lower screening rate in Direct Patient Engagement than in Point of Care. Its width describes the remaining uncertainty about the magnitude of that rate difference. The interval should be considered alongside the point estimate rather than reduced to a significance label.

The arm denominators are large, and the posted interval is relatively narrow on the risk-difference scale. That does not remove uncertainty arising from the design or reporting. In particular, the record does not state whether the analysis accounted for clinic-level clustering, whether it adjusted for baseline characteristics, or how missing outcome information was handled. The interval from an unadjusted calculation does not automatically address these issues.

The sponsor’s chi-squared analysis and the unadjusted recalculation report different p-values. A difference between p-values does not by itself show that the estimated direction is different; p-values can depend on the test and its assumptions. The registry record does not provide enough analytic detail to reconcile the values. The confidence interval belongs to the recalculated risk difference and should not be presented as though it were an interval from the sponsor’s chi-squared analysis.

Precision is also distinct from generalizability and practical importance. The record describes patients meeting specific age, language, appointment, and recruitment-window criteria at participating clinics. A precise estimate for that setting need not apply equally to other patient populations or systems. No threshold for an important difference is reported, so statistical precision alone does not establish whether the observed gap is operationally consequential.

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, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is set at the stated fraction of the observed effect. These are planning assumptions for the illustration; the trial’s own original design assumptions are not reported in the registry record.

The 80% power column gives the number of participants per arm that the illustration estimates would be needed for an 80% chance of detecting the specified effect under those assumptions, if that effect and assumptions were true. The 90% power column gives the corresponding planning figure for a 90% chance. For the full observed effect, the listed figures are 171 participants per arm for 80% power and 229 participants per arm for 90% power. For planning for 75% of the observed effect, the figures are 329 and 439 participants per arm. For planning for 50% of the observed effect, they are 791 and 1,059 participants per arm.

Smaller effects are harder to distinguish from random variation, so the illustration requires more participants per arm as the assumed effect becomes smaller. The curve should be read as a visual expression of the same planning relationship, not as another observed trial result. These figures are not observed power; they are sample-size planning illustrations based on the observed effect and the stated assumptions. The registry record does not report whether the original trial design used these assumptions or these sample-size targets.

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 = 0.191; treatment proportion p₁ = 0.087; average p̄ = 0.139.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4894 = 0.9591.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4837 = 0.4071.
4. n = (0.9591 + 0.4071)² / (-0.1046)² = 170.7, rounded up to 171 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 <- 0.191360532373822                          # control proportion
p1 <- 0.0867971773275666                          # 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(4746794)
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 = 171 (80% power), 229 (90% power)
 75% of observed effect: n per arm = 329 (80% power), 439 (90% power)
 50% of observed effect: n per arm = 791 (80% power), 1,059 (90% power)
power.prop.test: 170.7 
Simulated power at n = 171 per arm: 0.816 (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. Analyses Reported by the Sponsor

The registry reports a sponsor analysis using a chi-squared test for the comparison of Point of Care and Direct Patient Engagement. It labels the analysis as a superiority test. Separately, the unadjusted recalculation uses the posted arm summary statistics to estimate a risk difference and its confidence interval, with a corresponding p-value. The two p-values differ, and the registry record does not explain why.

These results should be attributed to their respective analyses. The chi-squared result is the sponsor-reported test; the risk difference and interval describe the unadjusted recalculation from the posted counts and denominators. The recalculation does not replace the sponsor’s analysis or establish that the sponsor used the same method. Because the discrepancy indicator is not flagged, there is no registry designation that the recalculation conflicts with the sponsor analysis. Still, the record does not provide enough detail to establish whether the analyses used identical populations, assumptions, or handling of clinic-level structure.

Neither p-value conveys the practical importance of the screening-rate contrast. The effect estimate, interval, endpoint definition, and limits of reporting are needed for interpretation.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists seven outcome measures, five posted secondary outcomes, and two analyses. The secondary topics include costs for each engagement strategy, incremental costs, screening and testing rates at Healthcare System A, numbers of additional patients screened or tested, and cost-effectiveness ratios per patient screened or tested. These are different questions from the primary comparison of screening completion.

The registry record does not report whether the primary outcome was designated as the sole confirmatory endpoint, how secondary outcomes were ordered, or whether any adjustment for multiple comparisons was planned. When many outcomes or analyses are examined, the chance of at least one apparently notable result can increase. Adjustment procedures or a prespecified hierarchy can address that issue, but their use in this trial is not reported in the registry record. The secondary findings should therefore be considered in the context of their status and any unreported multiplicity plan, rather than treated as equivalent to the primary outcome.

Missing-data handling is also not reported in the registry record. The record does not specify whether screening status was available for every eligible patient, how incomplete surveys were classified, or whether missing records were excluded or otherwise handled. No analysis-population definition is given beyond the endpoint description’s reference to the active clinic patient population and the eligibility criteria. It is not reported whether the primary analysis followed an intention-to-treat principle or another approach.

The record also does not report whether the analysis accounted for the clinic structure or any differences in patient mix. This is relevant because the intervention strategies are described at the clinic level, but the allocation unit and analytic treatment of clinic-level correlation are not specified. These omissions limit assessment of how the reported uncertainty should be interpreted. They do not establish that the trial failed to address these issues; the details are simply not reported in the registry record.

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

The result describes screening completion under two engagement strategies in the study’s eligible patient population. The posted data show a higher screening rate for Point of Care than for Direct Patient Engagement, and the unadjusted contrast is in the same direction. The confidence interval for that recalculated contrast lies below no difference. This is evidence about the measured screening process, not a direct measure of the later genetic testing pathway.

The finding does not establish how well either strategy identifies people with inherited cancer risk. It does not show how many patients were offered genetic testing, completed testing, received a mutation result, or followed a previvor plan. Those topics appear in the study description or secondary outcomes, but the primary endpoint is screening completion. The registry record does not provide enough posted results to draw conclusions about the full downstream care pathway.

The result also does not establish that the observed difference will be the same in other clinics, health systems, or populations. Eligibility was limited to patients meeting the stated age and language criteria and having an appointment during the recruitment window at participating clinics. The record does not report detailed implementation fidelity, patient characteristics, or the reasons patients did not complete screening.

Statistical evidence of a difference is not the same as evidence of practical importance. The registry record does not state what magnitude of difference would matter for clinic operations, costs, or patient outcomes. Secondary outcomes address costs and cost-effectiveness, but their results and analysis details are not included among the registry entries for interpretation. Accordingly, the primary result should be kept narrow: it compares completion of risk screening by the two approaches, under the recorded study conditions.

16. Why This Trial Matters Statistically

This trial raises a statistical question that is common in implementation research: how should uptake be compared when different strategies use different routes to reach a clinic population? Point of Care approached patients around visits, while Direct Patient Engagement used clinic records and postal and email contact. The primary endpoint connects those implementation choices to a concrete process outcome—completion of risk screening.

The choice of denominator is important. The endpoint concerns the active clinic patient population, and the posted results include both screened counts and arm denominators. Reporting a rate-based contrast alongside counts helps readers avoid treating raw event counts as though they alone describe uptake. The absolute risk difference also provides a direct scale for describing the gap between strategies, while the confidence interval shows the uncertainty around that contrast.

The design and reporting raise further questions relevant to clinical researchers. The registry describes randomization and a parallel design but does not report the allocation unit, missing-data approach, or whether clinic-level correlation was addressed. In a multisite implementation context, those details can affect both uncertainty and interpretation. The record also lists cost and cost-effectiveness outcomes, illustrating that uptake is only one dimension of comparing engagement approaches.

Finally, separating screening from genetic testing and follow-up keeps the analysis aligned with its endpoint. A screening-rate comparison can inform an implementation question without answering questions about diagnostic yield, downstream care, or health outcomes. That distinction is essential when communicating what a trial’s primary result does and does not establish.

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.

The registry record leaves several trial-specific limitations unresolved. It records randomized allocation but does not state whether clinics or individual patients were randomized. The arm descriptions refer to clinics in each strategy, making the missing allocation-unit detail important for evaluating both design and statistical independence. Whether the analysis accounted for clustering by clinic is not reported in the registry record.

The record lists enrollment as 20184, while the primary outcome has separate arm denominators. Their relationship is not explained. The record also does not report whether all patients in those denominators had a known screening status, how missing or incomplete screening data were treated, or which analysis population was used. The eligibility description and active-population endpoint definition therefore do not fully specify the flow from eligible patients to the analyzed outcome.

The intervention descriptions are not equally detailed in the registry record. Point of Care screening is described through a clinic visit and electronic tools; Direct Patient Engagement is described through record-based identification and mail and email outreach. The registry does not report adherence, contact success, exposure to both approaches, or differences in implementation between clinics. Such factors could shape observed uptake, but their role cannot be assessed from the posted information.

The primary outcome is a process measure over one year. It does not capture the accuracy of risk assessment, genetic testing completion, mutation findings, or follow-up. Although the study lists cost and cost-effectiveness outcomes, the registry entries do not include their results. Generalizability is also limited by the stated age and English-language criteria, appointment requirement, recruitment window, and participating-clinic setting.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Early Detection of GEnetic Risk (EDGE) — NCT04746794.
• Registry record last updated 2024-10-09; status last verified 2024-10.
• Results first posted 2024-10-09.
• 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

EDGE is a completed randomized, parallel study sponsored by the University of Washington, with no masking. It compares Point of Care and Direct Patient Engagement approaches for familial cancer risk screening. The registry lists enrollment as 20184 and records the study dates as 2020-09-25 to 2023-05-31.

The primary outcome is the rate of screening over one year among the active clinic patient population, with eligibility described by age, English-language comfort, clinic appointment, and recruitment-window criteria. The posted results show a higher screening rate in Point of Care than in Direct Patient Engagement. The registry reports a chi-squared analysis, while an unadjusted recalculation gives a risk-difference estimate and confidence interval; their p-values differ, and the record does not explain the difference.

The result concerns screening completion, not genetic testing or downstream follow-up. Details about missing data, analysis populations, allocation unit, and adjustment for clinic structure are not reported in the registry record.