Impact of Emergency Physician-Performed Ultrasound for the Evaluation of Patients With Acute Abdominal Pain: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Impact of Emergency Physician-Performed Ultrasound for the Evaluation of Patients With Acute Abdominal Pain (NCT04912206), a randomized na study. The primary outcome is Exact Diagnostic Rate.

2. Clinical Question

3. Trial Design

The registry describes a prospective, randomized, parallel-group study conducted at two emergency departments. Its official title describes it as a dual-centre study. The record lists the trial as completed, with a start date of 2021-06-11 and a primary completion date of 2022-06-23. The listed enrollment is 256.

Participants were assigned to an intervention group or a control group. The intervention was point-of-care ultrasound added to the diagnostic work-up; the control was usual diagnostic work-up without point-of-care ultrasound. Masking is listed as none. The registry record does not provide further details about how randomization was generated or concealed, or whether the clinicians, adjudicators, or participants knew the assigned group.

The primary outcome was exact diagnostic rate. The registry describes comparing diagnosis after clinical examination and biological results, with POCUS in the intervention group and without it in the control group, against a reference diagnosis from an adjudication committee. The committee’s reference diagnosis was to be determined from an exhaustive review of the patient’s file at Day 28.

The posted primary results list 70 participants with an exact diagnosis among 123 in the intervention group and 78 among 126 in the control group. The record also lists a mixed models analysis with p = 0.42. It does not report secondary outcome results, although the study description names time in the emergency department, disposition, and additional examinations—particularly radiologic examinations—as secondary objectives. The registry record does not report the original sample-size justification or detailed analysis protocol.

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 Exact Diagnostic Rate, measured over The reference diagnosis will be determined by an adjudication committee on the basis of an exhaustive analysis of the patient's file at Day 28.

The primary endpoint is the exact diagnostic rate, reported as a count of participants with an exact diagnosis. It asks whether the diagnosis reached during the initial clinical work-up matches the reference diagnosis determined by the adjudication committee. The record specifies that the comparison concerns diagnosis after clinical examination and biological results, with POCUS added in the intervention group and omitted in the control group. The reference diagnosis is based on review of the patient’s file at Day 28.

An exact-match endpoint turns diagnostic agreement into a binary outcome: each participant either has an exact diagnosis under the study definition or does not. This makes the group-level rate directly interpretable as the frequency of exact diagnoses in each group. As a general statistical point, a binary endpoint can be summarized with counts, group risks, and a difference between those risks. These summaries describe the observed classification outcome; they do not by themselves explain which diagnostic decisions led to agreement or disagreement.

The registry record does not report more detail about the adjudication criteria, how disagreements among reviewers were resolved, or whether adjudicators were masked to group assignment. Those details matter when interpreting an endpoint based on a reference diagnosis. The record also does not provide participant-level diagnostic categories or the reasons a diagnosis was judged inexact. The endpoint therefore describes exact diagnostic agreement at the group level, rather than the clinical course or the accuracy of any specific diagnosis.

7. Results Posted in the Registry

• Intervention: 56.9% (70/123) versus 61.9% (78/126) on control; risk difference -5.0 percentage points (95% CI -17.2 to 7.2), risk ratio 0.92 (0.75–1.13); unadjusted two-proportion z-test p = 0.422.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the risk difference describe?

For a binary outcome, each participant is classified as having or not having the event—in this study, an exact diagnosis under the endpoint definition. A risk is the proportion of a group with that event. The risk difference compares the group risks by subtraction. The recalculated estimate is -0.050 for intervention minus control, so the observed direction favors a higher exact-diagnosis rate in the control group. The estimate is an unadjusted comparison from the posted counts and does not account for any other variables.

As a general statistical explanation, a risk difference is expressed on the probability scale and can communicate how far apart two event rates are. A risk ratio instead expresses one group’s risk relative to the other group’s risk, while an odds ratio compares odds rather than risks. These measures can have different numerical values and interpretations, especially when outcomes are not rare. They are alternative descriptions of an association, not independent findings.

How should the confidence interval be read?

A confidence interval gives a range of estimates compatible with the data under the assumptions of the method used. The recalculated risk-difference interval is -0.172 to 0.072. Because it contains zero, the data do not rule out a zero difference under this interval calculation. It also contains values on both sides of zero, which indicates uncertainty about the direction as well as the size of the group difference. The interval should not be interpreted as a range containing a fixed proportion of individual participant outcomes.

Confidence intervals depend on the analysis method and its assumptions. The registry record does not describe the method used for the recalculated interval, nor does it provide a confidence interval for the sponsor’s mixed models analysis. The recalculated interval is therefore an unadjusted summary based on the posted outcome counts, not a replacement for an adjusted model-based estimate.

What does a p-value contribute?

A p-value describes how unusual data at least as discrepant as the observed data would be under a specified null hypothesis and statistical model. As a general explanation, it is not the probability that the null hypothesis is true, and it does not measure the size or importance of an effect. The unadjusted recalculated p-value is 0.422; the registry reports p = 0.42 for its mixed models analysis. The record does not specify the test or model details behind the reported analysis, so the result should be read with that limitation in mind.

Why distinguish the sponsor analysis from the recalculation?

The sponsor’s analysis is described as a mixed models analysis, whereas the recalculated effect measures are simple unadjusted comparisons from the posted counts. A mixed model can account for a structure in the data or covariates when specified to do so, but the registry record does not say what structure or variables were included. It is not possible to infer the model specification from the method label alone. The two p-values are similar, but they are not interchangeable evidence from an identical procedure. The recalculated comparison helps readers interpret the posted counts; the sponsor’s reported analysis remains the trial’s own analysis.

What can be concluded about the binary outcome?

The registry classifies the primary outcome as a count of participants with an event. The values 70 and 78 are event counts; 123 and 126 are the corresponding group sizes. As a general principle, the same counts can be presented as risks or used to derive risk differences, risk ratios, and odds ratios. Each summary answers the comparison in a different way. None removes uncertainty, and none reveals the patient-level diagnostic pathway or the adjudication details that produced each classification.

These methods describe the primary endpoint only. The registry record does not report secondary outcome results, an analysis population definition, a missing-data method, or the details needed to reproduce the mixed models analysis. Statistical interpretation must therefore stay within the posted aggregate results and the stated endpoint definition.

10. Confidence Intervals and Precision

The unadjusted risk-difference estimate is -0.050, with a 95% confidence interval from -0.172 to 0.072. This range crosses zero, the value corresponding to no difference on the risk-difference scale. It includes estimates in both directions, so the observed counts do not pin down whether the underlying group difference is negative or positive.

The interval also indicates that the point estimate should not be treated as precise. Its span allows for effects meaningfully different from the estimate itself. The interval is a measure of uncertainty about the group-level effect under the recalculation’s assumptions; it does not describe variation among individual participants or predict the result for a future participant.

The alternative unadjusted measures are also uncertain: the risk ratio is 0.919, with a 95% confidence interval from 0.748 to 1.13, and the odds ratio is 0.813, with a 95% confidence interval from 0.490 to 1.349. Both intervals include the no-association value for their respective scales. These measures present the same binary outcome on different scales and should not be mistaken for separate confirmations.

The registry reports p = 0.42 for the mixed models analysis but gives no corresponding effect estimate or confidence interval. The unadjusted recalculation has p = 0.422. A p-value alone does not communicate the range of plausible effect sizes, which is why the effect estimate and interval are important. The record does not provide enough model detail to assess whether the sponsor analysis has different precision from the unadjusted calculation.

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 for the binary outcome, not total events. They illustrate the sample sizes associated with different fractions of the observed effect under the stated planning assumptions. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The figures use the stated fraction of the observed effect as the effect to detect.

The 80% power column means the illustrated number of participants per arm for a design with an 80% chance of detecting the specified effect under those assumptions, if that effect were the true effect. The 90% power column gives the analogous planning target for 90% power. These are conditional design calculations, not guarantees that a future study will produce a particular result.

The planning rows refer to 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The corresponding participant requirements increase as the assumed effect becomes smaller. This is a general property of power calculations: smaller differences are harder to distinguish from random variation, so they generally require more participants for the same target power and significance level.

These calculations are illustrations based on the observed effect, not the original trial’s design justification. The original sample-size assumptions are not reported in the registry record. This is not observed power; it does not tell whether the completed trial had adequate power or explain the observed p-value.

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 = 0.619; treatment proportion p₁ = 0.569; average p̄ = 0.594.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6945 = 1.3612.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6936 = 0.5837.
4. n = (1.3612 + 0.5837)² / (-0.0499)² = 1,516.5, rounded up to 1,517 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 <- 0.619047619047619                          # control proportion
p1 <- 0.569105691056911                          # 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(4912206)
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 = 1,517 (80% power), 2,030 (90% power)
 75% of observed effect: n per arm = 2,684 (80% power), 3,592 (90% power)
 50% of observed effect: n per arm = 6,007 (80% power), 8,041 (90% power)
power.prop.test: 1516.5 
Simulated power at n = 1,517 per arm: 0.799 (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 the sponsor’s analysis as a mixed models analysis with p = 0.42. The unadjusted recalculation from the posted counts gives a risk difference of -0.050 and p = 0.422. The reported p-value is rounded to two decimal places, while the recalculated value is displayed to three decimal places. The values are close, but the registry does not give enough detail to determine the exact relationship between the analyses.

The recalculation is a simple unadjusted comparison of the posted counts. The sponsor analysis is identified as mixed models, but the record does not report the model specification, variables included, handling of centre, or an adjusted effect estimate and confidence interval. It would therefore be inappropriate to treat the calculations as identical or to infer what the mixed model adjusted for.

The record marks no discrepancy between the sponsor analysis and the recalculation. The sponsor’s analysis is the trial’s reported analysis, while the unadjusted calculation provides a transparent comparison based on the group counts. Neither p-value by itself establishes the size or clinical importance of a difference. The confidence interval from the unadjusted calculation supplies additional information about uncertainty, but it is not the sponsor model’s interval.

14. Multiplicity, Missing Data and Analysis Populations

The registry record reports one outcome measure and one analysis. It does not report a multiplicity adjustment or a hierarchy for testing multiple outcomes. The study description names secondary objectives involving time in the emergency department before diagnosis and disposition, as well as complementary examinations, particularly radiologic examinations. However, no secondary results are posted in the record. Without reported estimates and methods for those outcomes, their findings cannot be assessed or placed alongside the primary endpoint.

The record does not report how missing outcome data were handled, whether any participants lacked an adjudicated diagnosis, or whether an imputation or sensitivity analysis was used. The listed group counts are available, but the analysis population—such as whether the primary analysis followed all randomized participants—is not described. The registry record also does not report a participant flow account or reasons for exclusions from the primary analysis. These omissions limit assessment of how missingness or post-randomization exclusions might affect the result.

The record does not provide details of the mixed models analysis, including its model form, included variables, assumptions, or treatment of centre. It does not report whether prespecified subgroup analyses were planned or conducted. These details should not be inferred from the method label or from the two-centre design.

Adverse-event fields list zero counts in both groups for serious events, other events, and deaths. The registry record does not provide event descriptions or additional detail about ascertainment and reporting. Those entries should be understood as the counts recorded in the registry, not as proof that no harms occurred.

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

The posted data show that exact diagnoses were recorded for 70 of 123 participants in the POCUS group and 78 of 126 participants in the usual-work-up group. The unadjusted risk-difference estimate is negative, but its confidence interval extends across zero. The sponsor’s reported mixed models analysis has p = 0.42, and the unadjusted recalculation has p = 0.422. Together, these results indicate substantial uncertainty about the size and direction of the group difference in the primary endpoint.

This is not evidence that the groups are identical. An interval that includes zero can arise when the data are compatible with no difference and with effects in either direction. Conversely, the observed point estimate is not proof of a causal effect. The trial’s random allocation is relevant to comparing groups, but the registry record does not contain all the information needed to evaluate implementation, missing data, or the analysis in detail.

The endpoint concerns exact diagnostic agreement with a reference diagnosis based on file review at Day 28. It does not directly measure time to diagnosis, emergency-department disposition, additional imaging, patient experience, or longer-term clinical outcomes. Those outcomes appear as objectives in the study description, but results are not posted. A group-level exact-diagnosis count also cannot show which diagnostic categories were more or less often correctly identified.

The result should not be interpreted as a recommendation about whether POCUS should be used. It describes the primary outcome and statistical uncertainty reported in this registry record. The lack of detailed methods and secondary results places limits on what can be concluded beyond this comparison.

16. Why This Trial Matters Statistically

This trial asks a diagnostic-process question in emergency care: whether adding clinician-performed bedside ultrasound to the usual work-up changes the rate of exact diagnosis. That question is statistically distinct from asking whether a treatment improves a health outcome. The endpoint is diagnostic agreement against a reference adjudication, so the interpretation depends both on the observed group counts and on how the reference diagnosis was established.

The two-centre randomized design offers a structured comparison between POCUS added to usual work-up and usual work-up without POCUS. The posted result is useful for seeing how a binary endpoint can be expressed as counts, an unadjusted risk difference, and alternative relative measures. It also illustrates why uncertainty matters: the point estimate alone suggests a direction, while the confidence interval spans values on either side of no difference.

The record also demonstrates the importance of keeping the sponsor analysis distinct from a recalculation. A mixed models analysis is reported, while the displayed effect measures are unadjusted. Similar p-values do not make the methods identical, and missing model details constrain independent interpretation.

Finally, the registry describes secondary questions about time, disposition, and further testing, but provides no results for them. A primary diagnostic endpoint cannot stand in for those outcomes. Statistical reading is strongest when the endpoint, its uncertainty, the analysis method, and the limits of the posted record are considered together.

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 gives limited information about who was eligible and enrolled. It identifies acute abdominal pain as the condition and describes the study as concerning non-traumatic abdominal pain, but detailed inclusion and exclusion criteria and participant characteristics are not reported. Generalizability beyond the two emergency departments therefore cannot be assessed from this record alone.

Masking is listed as none. The record does not say whether the adjudication committee was masked to group assignment, how the reference diagnosis was determined in difficult cases, or what criteria defined an exact diagnosis. Because the endpoint depends on adjudication, these missing details are important when considering possible assessment bias.

The registry lists enrollment of 256, while the posted primary outcome counts are presented with group denominators of 123 and 126. The record does not explain the relationship between the enrollment figure and the outcome denominators, or report participant flow and exclusions. It also does not describe missing-data handling or define the primary analysis population.

The sponsor’s mixed models analysis is not specified beyond its method label and p-value. The record does not provide the model, adjusted effect estimate, confidence interval, or analysis assumptions. The reported secondary objectives have no posted results, and the adverse-event fields provide counts without further descriptions. These omissions limit assessment of the full benefit-risk context and the reproducibility of the analysis.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Impact of Emergency Physician-Performed Ultrasound for the Evaluation of Patients With Acute Abdominal Pain — NCT04912206.
• Registry record last updated 2026-11-06; status last verified 2026-11.
• Results first posted 2026-11-06.
• 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

NCT04912206 is a completed randomized, parallel-group, unmasked study described as a prospective dual-centre trial in emergency departments. It compares point-of-care ultrasound added to usual diagnostic work-up with usual work-up without ultrasound for patients with acute, non-traumatic abdominal pain.

The primary endpoint is exact diagnostic rate, assessed against a reference diagnosis determined by an adjudication committee from review of the patient’s file at Day 28. The posted counts are 70 of 123 in the intervention group and 78 of 126 in the control group. The unadjusted risk difference is -0.050, with a 95% confidence interval from -0.172 to 0.072; the sponsor reports a mixed models p-value of 0.42.

The registry record does not report detailed analysis methods, the original sample-size justification, or results for the described secondary objectives. Interpretation is consequently limited to the posted primary outcome and its reported uncertainty.