Efficacy, Tolerability, and Safety Study of DFN-15: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy, Tolerability, and Safety Study of DFN-15 (NCT03006276), a randomized phase 3 study. The primary outcome is Percentage of Subjects Who Are Pain-free at 2 Hours Postdose (DB1).

2. Clinical Question

3. Trial Design

This completed phase 3 study, NCT03006276, compared DFN-15 with placebo in a randomized, triple-masked, parallel-group trial. The registry record reports enrollment of 622 participants and describes the study as taking place at multiple centers in the United States. The record lists BioDelivery Sciences International as sponsor. It gives a start date of 2016-12 and a primary completion date of 2017-11.

The intervention groups were labeled DFN-15 Active and DFN-15 Placebo. Although the record identifies the active intervention as a drug, the dose, route, and detailed dosing schedule are not reported in the registry record. Further details about how randomization was carried out, including allocation procedures, are not reported in the registry record. The record identifies the trial as triple-masked but does not specify which groups of people were masked.

The main endpoint concerned pain freedom 2 hours after dosing during DB1. The endpoint description defines pain freedom as a change from moderate or severe pain before dosing to no pain. The main analysis population was the Full Analysis Set 1. Results are also labeled LOCF and OC, terms that distinguish the recorded analysis sets; the registry record does not give a fuller explanation of how these labels were applied.

The record lists 17 outcome measures and reports several secondary endpoints concerning pain relief, pain freedom, migraine-associated symptoms, functional disability, and recurrence. Secondary outcomes include assessments in DB1 and DB2. The primary completion date and completed status describe the trial record, but detailed operational information such as the treatment period and follow-up schedule is not reported in the registry record.

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.

Double-blind Period 1 (DB1)

Double-blind Period 2 (DB2)

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set 1

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 list the safety population (i.e., all subjects who were dosed during DB1). Note that there are 2 subjects in the placebo group and 2 subjects in the DFN-15 group, for whom no postdose efficacy data were collected. These individuals are, therefore, included in the baseline subject counts but are not included subsequent results reporting.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Subjects Who Are Pain-free at 2 Hours Postdose (DB1), measured over 2 hours post dose.

The primary endpoint was the percentage of subjects who were pain-free 2 hours after dosing in DB1. In the registry description, pain freedom means that a participant had moderate or severe pain before dosing and no pain at the assessment. The endpoint therefore represents a binary outcome: a participant either met the stated pain-free definition at the assessment or did not.

Reporting a percentage makes the frequency of this outcome interpretable across the two randomized groups. The accompanying confidence intervals communicate uncertainty around each group estimate, while a comparison between groups addresses whether the observed proportions differ beyond what might be expected from sampling variability. These are general statistical roles of the measures; they do not establish a treatment mechanism or the experience of every participant.

The assessment was specific to DB1 and the 2-hour postdose time point. It should not be read as a measure of pain throughout the entire study, nor as a measure of all migraine symptoms. Other symptom and functional outcomes were listed separately as secondary outcomes. The registry record identifies the primary analysis population as the Full Analysis Set 1, but does not report further details defining that population in the information available for this analysis.

7. Results Posted in the Registry

• DFN-15 LOCF: 35.7% (101/283) versus 21.8% (61/280) on control; risk difference 13.9 percentage points (95% CI 6.5 to 21.3), risk ratio 1.64 (1.25–2.15); unadjusted two-proportion z-test p = <0.001.
• Placebo OC: 22.1% (62/280) versus 21.8% (61/280) on control; risk difference 0.4 percentage points (95% CI -6.5 to 7.2), risk ratio 1.02 (0.74–1.39); unadjusted two-proportion z-test p = 0.919.
• DFN-15 OC: 35.7% (101/283) versus 21.8% (61/280) on control; risk difference 13.9 percentage points (95% CI 6.5 to 21.3), risk ratio 1.64 (1.25–2.15); unadjusted two-proportion z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Time to Headache Pain Relief Postdose (DB1 and DB2)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • DFN-15 DB1: mean 68.90 versus 59.50 on control; difference 9.40 (95% CI 3.23 to 15.57), standardized effect d = 0.25; z-test p = 0.003.
  • Placebo DB2: mean 55.30 versus 59.50 on control; difference -4.20 (95% CI -9.24 to 0.84), standardized effect d = -0.14; z-test p = 0.103.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for DFN-15 DB1 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Placebo DB2 and the control. The interval gives the range of effects that remain compatible with the data.

Time to Headache Pain Freedom Postdose (DB1 and DB2)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • DFN-15 DB1: mean 68.50 versus 66.60 on control; difference 1.90 (95% CI -17.41 to 21.21), standardized effect d = 0.06; z-test p = 0.847.
  • Placebo DB2: mean 68.80 versus 66.60 on control; difference 2.20 (95% CI -16.39 to 20.79), standardized effect d = 0.08; z-test p = 0.817.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between DFN-15 DB1 and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Placebo DB2 and the control. The interval gives the range of effects that remain compatible with the data.

Headache Pain Recurrence Postdose (DB1 and DB2)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • DFN-15 DB1: 9.2% (26/283) versus 13.9% (39/280) on control; risk difference -4.7 percentage points (95% CI -10.0 to 0.5), risk ratio 0.66 (0.41–1.05); unadjusted two-proportion z-test p = 0.078.
  • Placebo DB2: 6.5% (16/248) versus 13.9% (39/280) on control; risk difference -7.5 percentage points (95% CI -12.6 to -2.4), risk ratio 0.46 (0.27–0.81); unadjusted two-proportion z-test p = 0.005.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between DFN-15 DB1 and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Placebo DB2 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcome set broadens the assessment beyond the primary pain-free endpoint. It includes time to pain relief and pain freedom, postdose pain relief and freedom, migraine-associated symptoms, functional disability, recurrence, and a pain-freedom assessment among subjects with cutaneous allodynia. These outcomes address related but distinct aspects of an episode, at different time points and in different analysis populations.

The numerical secondary summaries available in the registry show a varied pattern rather than one uniform result across every measure. The time-to-pain-relief comparison in DB1 has an unadjusted recalculated p-value of 0.003, while the reported time-to-pain-freedom comparisons have unadjusted recalculated p-values of 0.847 and 0.817. Recurrence comparisons also differ by period: the DB1 unadjusted recalculation has p = 0.078, and the DB2 comparison has p = 0.005. These are recalculations, not sponsor analyses, and they should not be treated as a single body of confirmatory evidence.

These outcomes can provide context for the primary result because they measure other dimensions of pain and symptoms. However, differences in endpoint definitions, periods, and analysis populations mean that they do not directly confirm or contradict the primary endpoint. Outcome-specific numerical results for other listed secondary measures are not reported in the registry record information available for this analysis.

Unadjusted secondary p-values carry less weight when many outcomes are tested. Each additional test creates another opportunity to observe a low p-value by chance, while unadjusted values do not account for that multiplicity. The registry record does not report whether a multiplicity procedure was planned or used. The secondary results are therefore best read as supporting context and hypothesis-generating information, rather than as a set of equally definitive tests.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Freedom From Nausea, Photophobia, and Phonophobia Postdose (DB1 and DB2)
• Time to Headache Pain Relief Postdose (DB1 and DB2)
• Time to Headache Pain Freedom Postdose (DB1 and DB2)
• Headache Pain Relief Postdose (DB1 and DB2)
• Headache Pain Freedom Postdose (DB1 and DB2)
• Absence of Screening MBS at Time Points Postdose (DB1 and DB2)
• Change in Functional Disability Score Postdose (DB1 and DB2)
• Headache Pain Freedom Among Subjects With Cutaneous Allodynia (DB1 and DB2)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is binary: each participant is classified according to whether the pain-free definition was met at the specified assessment. The registry reports a percentage for each group and a 95% confidence interval. A percentage expresses the observed frequency in the analyzed group. A confidence interval is a range of values compatible with the data under the method used; it conveys precision rather than guaranteeing that the true value lies within the range.

What does the odds ratio describe?

The sponsor's reported primary comparison uses an odds ratio and the Fisher Exact method. An odds ratio compares the odds of the outcome between groups. It is not the same measure as a difference in percentages, and its numerical size should not be translated directly into a percentage-point change. The sponsor reports an odds ratio of 2.00 with a 95% confidence interval of 1.36 to 2.94 and p < 0.001 for the DB1 comparison. The confidence interval gives a range of estimates under the reported method; its position relative to the null value for an odds ratio is relevant to interpreting the comparison.

Fisher Exact is a method for comparing categorical outcomes. It can be used when the data are counts of participants classified by outcome status. The registry's own analysis, rather than a separate calculation from rounded summaries, is the appropriate source for the trial's reported inferential result. The detailed statistical analysis plan and any additional adjustment procedures are not reported in the registry record.

What is the recalculated risk difference?

A risk difference expresses the difference between the proportions experiencing an outcome in two groups. The posted recalculation for DFN-15 LOCF versus Placebo LOCF gives a risk difference of 0.139 and an unadjusted p < 0.001. This recalculation is based on posted summary counts and is not an adjusted analysis. It is a different scale from the sponsor's odds ratio, so the estimates answer related but not identical descriptive questions. Because the registry analysis and recalculation differ, the sponsor's own analysis takes precedence.

The registry also reports a separate sponsor analysis for the OC-labeled groups: an odds ratio of 1.97, a 95% confidence interval of 1.34 to 2.89, and p < 0.001. The OC label is reported as part of the group name. The registry record does not explain the label in more detail, so its exact handling should not be inferred from the label alone.

Why are outcome types important?

Secondary measures use different statistical forms. Time to pain relief and time to pain freedom are reported as means with standard deviations. A mean describes the group average, while a standard deviation describes variability among observations. A mean difference compares averages, but its interpretation depends on the underlying measurement and analysis population. The time-to-pain-freedom groups contain relatively few observations compared with the primary groups, so their estimates are less precise. The registry does not report sponsor inferential analyses for these secondary comparisons in the information available for this analysis.

Binary secondary outcomes, such as recurrence, are reported as percentages and can be compared using a risk difference or other relative measure. A confidence interval around a risk difference communicates uncertainty in the group contrast. When it includes the null value of no difference, the data are compatible with a range of contrasts that includes no difference under that method. The posted recalculations for secondary outcomes are unadjusted; they should not be mistaken for sponsor analyses.

Statistical significance and effect magnitude are different concepts. A p-value reflects how unusual the observed data would be under a specified null model and assumptions; it does not measure the probability that the null model is true, nor does it measure clinical importance. Confidence intervals and the underlying outcome definitions are needed to understand the scale and precision of estimates.

Several details that could affect the interpretation are not reported in the registry record. These include the complete statistical analysis plan, any interim analyses, the trial's original sample-size justification, and the full handling of missing observations. The primary endpoint labels include LOCF and OC, but those labels alone do not establish every rule used to construct the analysis populations or handle incomplete data.

11. Confidence Intervals and Precision

The primary group estimates have reported 95% confidence intervals. For the LOCF-labeled results, the interval is 30.0 to 41.6 for DFN-15 and 16.8 to 27.1 for placebo. For the OC-labeled results, the intervals are 30.1 to 41.8 and 17.2 to 27.7, respectively. These intervals communicate uncertainty around the percentages; they are not ranges containing every participant's outcome.

The sponsor's primary comparison reports an odds ratio of 2.00 with a 95% confidence interval of 1.36 to 2.94. The interval gives a sense of the precision of the relative comparison on the odds-ratio scale. It is important to keep that scale distinct from a difference in percentages. The posted recalculation also reports a risk difference, but its interval is not repeated because the registry's own analysis takes precedence when the analyses differ.

In general, a narrower confidence interval indicates greater statistical precision than a wider interval, for a given measure and method. Precision is affected by the amount of information in the analysis and by the variability of the outcome. It is not the same as freedom from bias: a precise estimate can still be affected by limitations in design, measurement, analysis, or missing data.

Confidence intervals are estimates under a statistical procedure, not guarantees about the unknown value. Their interpretation depends on the method and assumptions used. The registry record identifies confidence intervals for the primary outcome, but a complete account of the analysis plan and missing-data procedures is not reported in the registry record. The confidence intervals should therefore be read alongside the sponsor's reported analysis and the stated endpoint and population.

12. 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.

Sample-size figures are planning illustrations based on the observed effect, not the original design justification for this trial. For the binary and continuous comparisons, the figures are participants per arm. A time-to-event planning table, if used, gives the total number of events instead. The calculation assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to 100%, 75%, or 50% of the observed effect.

The 80% power column gives the illustrated number of participants per arm for a design with an 80% chance of detecting the stated effect if that effect were the true one and the assumptions held. The 90% power column gives the corresponding illustration for a 90% chance. These are planning concepts, not guarantees; actual performance depends on the true effect, outcome variability, analysis choices, and other design features.

The rows for 100%, 75%, and 50% of the observed effect show how the illustration changes when the effect to detect is smaller than the observed estimate. Smaller effects generally require more participants to detect with the same target power and significance level. Estimates based on an observed trial effect are uncertain, so these calculations should not be read as a reliable forecast of a future result.

These values are not observed power. Observed power is not a useful substitute for interpreting the estimate and its confidence interval after a study is completed. The trial's own original sample-size justification and design assumptions are not reported in the registry record. The planning calculations should therefore be understood as separate illustrations, not as evidence about what enrollment the investigators originally intended or why they selected it.

13. 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 = 0.218; treatment proportion p₁ = 0.357; average p̄ = 0.287.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6400 = 1.2543.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6324 = 0.5322.
4. n = (1.2543 + 0.5322)² / (0.1390)² = 165.1, rounded up to 166 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 <- 0.217857142857143                          # control proportion
p1 <- 0.356890459363958                          # 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(3006276)
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 = 166 (80% power), 221 (90% power)
 75% of observed effect: n per arm = 284 (80% power), 379 (90% power)
 50% of observed effect: n per arm = 613 (80% power), 819 (90% power)
power.prop.test: 165.1 
Simulated power at n = 166 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.

14. Analyses Reported by the Sponsor

The sponsor's reported primary DB1 comparison uses Fisher Exact and reports an odds ratio of 2.00, a 95% confidence interval of 1.36 to 2.94, and p < 0.001. For the OC-labeled comparison, the reported odds ratio is 1.97, with a 95% confidence interval of 1.34 to 2.89 and p < 0.001. These are the registry's own inferential analyses.

The posted recalculation uses unadjusted comparisons from summary statistics and expresses the primary contrast as a risk difference. It reports a risk difference of 0.139 and p < 0.001 for the LOCF-labeled groups, with the same recalculated estimate and p-value for the OC-labeled groups. The risk difference and odds ratio use different scales, and the recalculation does not reproduce every feature that may have been used in the sponsor's analysis.

The record flags a discrepancy between the sponsor analysis and recalculation. The sponsor's own analysis takes precedence for the trial's reported inferential conclusion. Recalculated statistics can provide a transparent, simple comparison of posted summaries, but they are not a correction to or replacement for the sponsor's analysis. The complete analysis plan and details needed to investigate the discrepancy are not reported in the registry record.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary endpoint and multiple secondary outcomes, but does not report a complete statistical analysis plan in the information available for this analysis. In particular, the record does not report whether the analysis controlled the overall chance of a false-positive finding across the primary and secondary tests, or whether a prespecified hierarchy governed interpretation. Without such information, secondary unadjusted p-values cannot be assumed to have been adjusted for multiplicity.

Multiplicity matters because testing multiple outcomes can increase the chance that at least one result appears statistically notable even when the underlying effects are absent. This is a general statistical concern, not a claim that any particular secondary result is false. The primary endpoint should be distinguished from the broader secondary outcome set, and the results from the latter should be interpreted with appropriate caution.

The primary endpoint is identified as belonging to the Full Analysis Set 1. The registry record does not give further details about the exact inclusion rules for that population. The primary outcome groups are also labeled LOCF and OC. Those labels are recorded, but the detailed definitions and implementation are not reported in the registry record. Consequently, the labels alone do not establish how every missing observation was handled.

The registry record does not report the extent or pattern of missing primary endpoint data, the assumptions used for missingness, or sensitivity analyses addressing missing data. It also does not report interim analyses or the trial's original sample-size justification. Those omissions limit the ability to assess whether the posted results reflect a prespecified analysis strategy and how robust the results might be to alternative assumptions. They do not, by themselves, show that a particular analysis choice was or was not made.

16. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: road traffic accident (0.4%), non-cardiac chest pain (0.4%).

The serious adverse event figures are counts of participants affected, alongside the number at risk in each group. In DB1, the registry reports 2 serious adverse events among 282 placebo participants at risk and 0 among 285 DFN-15 participants at risk. In DB2, it reports 0 among 249 placebo participants and 0 among 244 DFN-15 participants. Overall, 2 serious adverse events were reported among 567 participants at risk. The record reports 0 deaths.

The registry lists Road traffic accident and Non-cardiac chest pain among the top serious event terms, each at 0.4%. The number of events and the number of participants affected are different ways of describing safety information; the record's figures should be read according to their labels. The posted safety counts do not establish whether an event was caused by the study intervention.

These counts describe what was recorded in the stated groups and periods. They are not sufficient on their own to characterize all aspects of safety, including event severity beyond the serious-event classification, timing, duration, or attribution. Further details about those matters are not reported in the registry record information available for this analysis.

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

The posted primary results describe the proportion of analyzed participants who met a specific pain-free definition at a single postdose assessment in DB1. The sponsor's analysis reports an odds ratio of 2.00, a 95% confidence interval of 1.36 to 2.94, and p < 0.001 for the DFN-15 versus placebo comparison. The group percentages and confidence intervals provide complementary descriptions: one shows observed proportions, and the other reports the sponsor's relative comparison with its uncertainty.

This result does not establish that every participant had the same outcome, that pain remained absent beyond the assessment, or that every symptom of migraine changed. The primary endpoint concerns pain status at 2 hours after dosing and is not a comprehensive measure of an entire episode. The listed secondary outcomes address other questions, but their different definitions and the absence of reported multiplicity information mean they should not be treated as interchangeable confirmations.

A statistically notable comparison is not equivalent to a judgment about clinical importance. The registry record does not report a threshold for a clinically meaningful difference, and no such threshold should be inferred from the p-value alone. The confidence interval describes precision under the analysis method; it does not resolve limitations that may arise from missing data, population definitions, or other design and analytic choices.

The results also do not answer questions beyond the studied setting and recorded outcomes. Details such as the dose, complete eligibility criteria, and full statistical analysis plan are not reported in the registry record information available for this analysis. Safety counts are descriptive and do not by themselves establish causal attribution. The appropriate reading is therefore bounded: the record reports a primary difference on the specified endpoint, alongside uncertainty and limits on what that endpoint and the available analyses can establish.

18. Why This Trial Matters Statistically

This trial is statistically informative because it reports a defined binary primary endpoint, group-level percentages, confidence intervals, and a sponsor-reported comparison. Keeping those quantities distinct helps readers understand both the observed outcome frequency and the uncertainty in the relative analysis. The record also includes results labeled LOCF and OC, highlighting that analysis labels can matter when interpreting a reported endpoint.

The comparison illustrates why measures on different scales should not be conflated. A risk difference describes an absolute contrast in proportions, while an odds ratio describes a relative contrast in odds. The recalculated risk difference is a simple unadjusted summary; the sponsor's odds-ratio analysis is the registry's own inferential result and takes precedence because a discrepancy is flagged. Readers should evaluate the sponsor estimate, its confidence interval, and the outcome definition together rather than selecting a single statistic in isolation.

The secondary outcomes demonstrate the challenge of drawing a unified conclusion from several measures that assess different symptoms, times, and periods. Unadjusted secondary analyses may be useful as context, but without a reported multiplicity strategy their evidential weight is different from that of the designated primary endpoint. Sample-size illustrations based on observed effects also have a distinct role: they show how planning estimates respond to assumed effects, but they do not reveal the trial's original rationale.

Finally, the registry record makes clear that interpretation depends on information beyond outcome estimates. The original sample-size justification, detailed missing-data procedures, complete analysis plan, and multiplicity handling are not reported in the available record. Recognizing these gaps is part of sound statistical interpretation, not a reason to substitute speculation for the posted analyses.

19. 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.

Several trial-specific details needed for a fuller assessment are not reported in the registry record information available for this analysis. These include the DFN-15 dose and dosing schedule, full eligibility criteria, the randomization procedure, and which roles were masked under the triple-masked designation. The record identifies a multicenter study in the United States but does not give further site-level information.

The primary endpoint applies to DB1 and the Full Analysis Set 1, while secondary outcomes include both DB1 and DB2 and may use different analysis populations. The record does not provide enough detail to establish how the sets relate to one another or how participants entered each analysis. The group labels LOCF and OC are present, but their full definitions and the missing-data procedures are not reported. This limits assessment of how incomplete observations influenced the endpoint estimates.

The sponsor's primary analysis and the simple recalculation do not match exactly, and the record flags a discrepancy. The sponsor's analysis takes precedence, but the complete statistical analysis plan and detailed data needed to investigate the difference are not reported in the registry record. The recalculations are unadjusted and rely on posted summary statistics; they cannot reproduce adjustments or other analytic choices that may be part of the sponsor's method.

Several secondary summaries are based on smaller analysis groups than the primary endpoint. The registry does not report sponsor inferential analyses for the listed time-to-event summaries in the information available for this analysis. The multiplicity strategy, interim analyses, and original sample-size justification are also not reported. These limitations affect how broadly the numerical results can be interpreted, particularly for secondary outcomes, without changing the descriptive details recorded for the primary endpoint.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Efficacy, Tolerability, and Safety Study of DFN-15 — NCT03006276.
• Registry record last updated 2023-01-10; status last verified 2022-12.
• Results first posted 2023-01-10.
• 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.

23. Record Summary

NCT03006276 is a completed randomized, triple-masked, parallel-group phase 3 trial of DFN-15 and placebo in episodic migraine with or without aura. The registry reports enrollment of 622 and lists BioDelivery Sciences International as sponsor. The study start date is 2016-12, and the primary completion date is 2017-11.

The primary endpoint was pain freedom 2 hours after dosing in DB1, reported for the Full Analysis Set 1. The sponsor's reported DB1 comparison gives an odds ratio of 2.00, a 95% confidence interval of 1.36 to 2.94, and p < 0.001. Because the record flags a discrepancy with recalculated statistics, the sponsor's analysis takes precedence.

The record lists secondary symptom, pain, functional, and recurrence outcomes, along with adverse event counts. The original sample-size justification, full analysis plan, multiplicity handling, and detailed missing-data procedures are not reported in the registry record information available for this analysis.