Transforaminal Versus Lateralized Interlaminar Cervical Epidurals: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Transforaminal Versus Lateralized Interlaminar Cervical Epidurals (NCT03389620), a randomized phase 2 study. The primary outcome is Change in Pain Numerical Rating Scale (Averaged Over Past 24 Hours).

2. Clinical Question

The clinical comparison concerns two ways of delivering an epidural corticosteroid injection for the same condition. The registry description frames the question in terms of effectiveness and safety. Statistically, the primary outcome is a continuous pain score, and the reported comparison focuses on average change rather than on the proportion of participants meeting a threshold.

The posted arm means are changes of -3.38 for the transforaminal group and -2.3 for the lateralized interlaminar group. The unadjusted mean difference is reported as 1.08, with the comparison oriented to lateralized interlaminar relative to transforaminal. These summaries describe the participants with posted primary outcome data; the registry notes that primary pain data were not collected for 3 transforaminal participants.

The question is not limited to whether the group means differ. It also concerns the uncertainty around that estimated difference, the interpretation of the sponsor's noninferiority analysis, and how missing outcome data and multiple posted analyses affect what can be concluded from the record.

3. Trial Design

The registry describes NCT03389620 as a randomized, double-masked, parallel-group phase 2 study. It is listed as completed. The study began on 2018-01-05, and its primary completion date is 2023-07-16. Duke University is named as the sponsor. Enrollment is recorded as 74.

Participants were assigned to one of two groups: transforaminal cervical epidural corticosteroid injections or lateralized interlaminar epidural corticosteroid injections. Dexamethasone appears in the intervention list. The arm descriptions themselves are blank, so procedural details beyond the route and the intervention name are not reported in the registry record. The registry does not report the allocation ratio or how randomization was implemented.

Double masking indicates that the registry classifies the study as masked for two groups of people, but the identities of those groups and the practical masking procedures are not reported in the registry record. The record also does not report how treatment assignment was concealed before allocation.

The primary outcome was change in the pain numerical rating scale averaged over the past 24 hours, assessed from baseline to 2 months. The registry lists additional outcomes involving pain, neck disability, quality of life, and work ability or impairment. The record lists 13 outcome measures and 26 analyses; it also lists 10 posted secondary outcomes. These counts do not by themselves indicate which analyses were prespecified as primary or how results across outcomes were handled.

The original sample-size justification, the target enrollment rationale, and the trial's full statistical analysis plan are not reported in the registry record. Consequently, the planning illustrations calculated from the observed effect should not be read as a description of the trial's original design.

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: This is not a crossover study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Data not collected on 3 Transforaminal Cervical Epidural Corticosteroid Injection participants.

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 Change in Pain Numerical Rating Scale (Averaged Over Past 24 Hours), measured over Baseline to 2 months.

The primary endpoint is change in the Pain Numerical Rating Scale (NRS), averaged over the past 24 hours, from baseline to 2 months. The NRS asks participants to rate current pain intensity on an 11-point scale from 0, “no pain,” to 10, “worst possible pain.” The registry describes the NRS as a well-validated tool for quantitatively assessing pain.

A change score captures how a participant's recorded pain score differs between the baseline assessment and the follow-up assessment. Reporting the mean change for each group gives a summary of the average change among participants with data. The posted mean changes are -3.38 in the transforaminal group and -2.3 in the lateralized interlaminar group. The standard deviations are 2.5 and 2.2, respectively, describing variation in individual change scores around each group's mean.

A continuous score preserves more information than a binary classification such as improved or not improved, because it retains the scale's graded values. The mean difference then expresses the contrast between group averages in the same score units. Its interpretation depends on the direction used to calculate change and on which group is subtracted from the other; the registry gives a mean difference of 1.08 for lateralized interlaminar relative to transforaminal.

The registry notes that primary pain data were not collected for 3 participants in the transforaminal group. The reason, timing, and handling of those missing data are not reported in the registry record.

7. Results Posted in the Registry

• Lateralized Interlaminar Epidural Corticosteroid Injections: mean -2.30 versus -3.38 on control; difference 1.08 (95% CI -0.02 to 2.18), standardized effect d = 0.46; z-test p = 0.054.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous, so its arm-level summaries are means and standard deviations. A mean describes the arithmetic average of participants' change scores. A standard deviation describes how spread out those individual scores are around the mean; it is not the uncertainty of the mean itself. In the registry, the transforaminal group has a mean change of -3.38 and standard deviation of 2.5, while the lateralized interlaminar group has a mean change of -2.3 and standard deviation of 2.2.

What does a mean difference represent?

A mean difference subtracts one group's average from the other's, using a stated direction. The posted comparison is expressed as lateralized interlaminar relative to transforaminal and has an estimate of 1.08. That estimate is in NRS change-score units. It does not describe the experience of every participant: individual changes can vary around each group mean, as the standard deviations indicate.

A comparison of means is useful when the outcome is recorded on a numeric scale and the question concerns average differences. General statistical considerations include whether observations are independent, whether the score distributions and sample sizes support the chosen inferential approximation, and whether the analysis accounts for features of the study design. Which diagnostics or adjustments the sponsor used is not reported in the registry record.

How should the confidence interval be read?

A confidence interval expresses uncertainty around an estimate under the assumptions of the statistical procedure. The posted interval for the mean difference runs from a value just below zero to 2.179. It therefore contains zero, the value corresponding to no difference in group means. The interval also contains positive values, so the posted data are compatible with a range of possible differences. The interval does not assign a probability to the true effect being within its endpoints after the data have been observed.

The reported unadjusted p-value for the mean comparison is 0.054. A p-value describes how unusual data at least as incompatible with a specified null hypothesis would be if that null and the test assumptions held. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The registry does not state a multiplicity adjustment for the posted analyses.

What is different about a noninferiority test?

The sponsor reports a one-sided t-test of noninferiority with p = 0.057. A superiority question typically tests whether the groups differ; a noninferiority question instead evaluates whether the intervention under consideration is not worse than its comparator by more than a prespecified margin. The margin defines how much difference is acceptable for that test. The margin, direction of the sponsor's formal hypothesis, and decision rule are not reported in the registry record.

A one-sided test evaluates evidence in a specified direction rather than splitting the testing threshold across both directions. Its p-value should not be interpreted as interchangeable with the unadjusted p-value for the mean-difference comparison. The registry reports both results, but the sponsor's noninferiority conclusion cannot be reconstructed from its p-value without the margin and analysis details.

Why can dispersion matter?

General statistical explanation: when participant outcomes vary more widely, the group averages are less precisely estimated for a given sample size. In count outcomes, over-dispersion means that observed variability exceeds what a simple count model expects; it can make conventional standard errors too small if not addressed. Over-dispersion is not the issue described for this primary continuous endpoint, and the registry does not report a count-model analysis for it.

For this outcome, the arm-level standard deviations describe observed variability in change scores. The posted comparison also provides standard errors of 0.429 and 0.362 for the two arm means, but the registry does not describe how they were used in the sponsor's analysis. The recalculated comparison is an unadjusted calculation from posted summary statistics; it is not a replacement for any adjusted or prespecified sponsor analysis.

The record does not report whether the primary analysis adjusted for baseline score or other covariates, how missing outcomes were handled, or which analysis population was used. Those choices can affect an estimate and its uncertainty. They cannot be inferred from the means, standard deviations, and group sizes alone.

10. Confidence Intervals and Precision

The unadjusted mean difference is 1.08, and its 95% confidence interval extends from just below zero to 2.179. The lower endpoint is negative, so the interval includes zero; its upper endpoint is positive. The interval therefore includes both no difference and positive differences in the stated contrast direction.

Precision concerns how narrowly the data locate an estimate. A confidence interval is more informative than the point estimate alone because it displays a range of values compatible with the data under the analysis assumptions. Here, the interval reaches across zero and extends to 2.179, so the point estimate should not be read as an exact or certain difference. The lower endpoint is described without additional decimal places because it is a small negative value.

The reported unadjusted p-value is 0.054. It complements the interval but does not explain the practical importance of the estimate. The interval and p-value relate to the unadjusted mean comparison; they should not be substituted for the sponsor's one-sided noninferiority analysis, which has a reported p-value of 0.057 and a margin not reported in the registry record.

These precision statements also depend on the analyzed data. The posted primary summaries use 34 transforaminal participants and 37 lateralized interlaminar participants, and the registry notes missing primary pain data for 3 transforaminal participants. The consequences of those missing outcomes for the estimate and interval cannot be determined from the posted summary alone.

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 this continuous pain outcome, not total events. The planning illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers effects equal to 100%, 75%, and 50% of the observed effect. These calculations are planning illustrations based on the observed effect, not the trial's original design assumptions; those original assumptions are not reported in the registry record.

The 80% power column gives the planned participants per arm under the stated assumptions for an 80% chance of detecting the selected effect if that effect and the model assumptions are correct. The 90% power column gives the corresponding participants per arm for a 90% chance. These are planning probabilities, not probabilities that a particular study result is correct.

For planning at 100% of the observed effect, the figures are 75 participants per arm for 80% power and 100 participants per arm for 90% power. For planning at 75% of the observed effect, they are 132 and 177 participants per arm. For planning at 50% of the observed effect, they are 297 and 398 participants per arm.

The pattern reflects a general feature of planning: smaller effects are harder to detect, so they require more participants under otherwise unchanged assumptions. The curve represents how the planning requirement changes as the assumed effect becomes smaller. These illustrations are not observed power.

The trial enrolled 74 participants, but the registry does not report that the planning illustration's assumptions were used to design the trial. It would be inappropriate to treat the illustrated requirements as evidence about the study's original target or rationale.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 1.080; pooled standard deviation σ = 2.348.
2. n = 2 × (1.96 + 0.842)² × 2.348² / 1.080² = 74.2, rounded up to 75 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
delta <- 1.08                      # observed difference in means
sigma <- 2.34826485433116                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(3389620)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 75 (80% power), 100 (90% power)
 75% of observed effect: n per arm = 132 (80% power), 177 (90% power)
 50% of observed effect: n per arm = 297 (80% power), 398 (90% power)
power.t.test: 75.2 
Simulated power at n = 75 per arm: 0.807 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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 sponsor's reported analysis is a one-sided t-test for noninferiority, with p = 0.057. The registry does not report its test statistic, confidence interval, noninferiority margin, analysis population, or complete decision rule. Without the margin and decision rule, the p-value alone does not establish whether the sponsor's noninferiority criterion was met.

The unadjusted recalculation from the posted arm summaries estimates a mean difference of 1.08, with a 95% confidence interval extending from just below zero to 2.179 and an unadjusted p-value of 0.054. This calculation addresses a mean-difference comparison and is not the same as the sponsor's one-sided noninferiority test. Their p-values answer different statistical questions and should not be treated as competing versions of one test.

The registry does not flag a discrepancy between the sponsor's analysis and the recalculation. The sponsor's result remains the reported trial analysis; the recalculation is a simple, unadjusted interpretation of the posted summary statistics. The record does not provide enough detail to reproduce the sponsor's analysis fully or to determine how missing primary outcomes were handled in it.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists secondary outcomes concerning pain, neck disability, EQ-5D visual analog scale and index scores, and Work Ability Index measures of absenteeism, presenteeism, productivity loss, and activity impairment. It records 13 outcome measures, 26 analyses, and 10 posted secondary outcomes. This indicates that the record contains multiple outcome and analysis entries, but those counts do not show which analyses were prespecified or how results were prioritized.

When many hypotheses are tested, the chance of at least one apparently unusual result can increase. General statistical practice may address this through prespecified primary outcomes, adjustment procedures, or an explicitly exploratory interpretation. Whether any multiplicity adjustment or hierarchical testing plan was used is not reported in the registry record. The posted primary comparison is therefore best interpreted as one result within a record that includes multiple analyses, rather than as isolated evidence across all outcomes.

The primary outcome population has a specific missing-data note: pain data were not collected for 3 participants assigned to the transforaminal group. The posted primary summaries include 34 participants in that group and 37 in the lateralized interlaminar group. The reason the data were not collected, whether missingness was related to outcomes, and whether any imputation or sensitivity analysis was performed are not reported in the registry record.

The record does not identify the primary analysis population as intention-to-treat, modified intention-to-treat, or per-protocol. It also does not say whether participants were analyzed according to assigned group or treatment received. These details matter particularly for noninferiority analyses, where the analysis population and missing-data handling can influence interpretation.

The registry does not report a detailed statistical analysis plan, baseline-adjusted estimates, or the method used to handle missing data. Those omissions limit independent assessment of whether the posted estimates correspond to the intended analysis and how robust they are to alternative assumptions. No further conclusions about the missing observations or multiplicity strategy can be drawn from the registry record.

15. 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 0% reporting threshold.

Most frequent non-serious adverse event terms

The safety summaries list 37 participants at risk in each group. Serious adverse events affected 0 participants in the transforaminal group and 0 in the lateralized interlaminar group. The record lists 6 participants with other adverse events in the transforaminal group and 8 in the lateralized interlaminar group. Deaths are listed as 0 in both groups.

These are counts of affected participants, not percentages or event rates. The safety denominators are 37 in each group, while the primary outcome summaries use 34 transforaminal participants and 37 lateralized interlaminar participants. The record does not explain that difference in denominators, so the safety counts should be kept distinct from the primary outcome analysis counts.

No serious adverse-event terms are listed. The registry record does not provide event-level descriptions, timing, severity classifications for other adverse events, or a statistical comparison of safety outcomes. Counts of zero serious events in these groups do not establish that the routes have identical safety profiles, and the posted counts alone do not characterize risks beyond this study record.

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

The posted primary results describe average changes in pain scores for participants with available data: -3.38 in the transforaminal group and -2.3 in the lateralized interlaminar group. The estimated mean difference, expressed as lateralized interlaminar relative to transforaminal, is 1.08. The confidence interval reaches from just below zero to 2.179, so the estimate is uncertain and the interval includes no difference.

The sponsor also reports a one-sided noninferiority test with p = 0.057. Noninferiority is a question defined by a margin, and that margin and the decision rule are not reported in the registry record. The sponsor's result therefore cannot be translated into a clear noninferiority conclusion from the posted p-value alone. It is also not interchangeable with the unadjusted mean-difference result.

These findings do not establish that the two injection routes are equivalent. A failure to demonstrate a difference in a particular comparison is not the same as evidence that differences are absent. Conversely, the estimated mean difference is not proof of a clinically important difference. A threshold for clinical importance is not reported in the registry record.

The primary data were not collected for 3 transforaminal participants, and the missing-data method and analysis population are not reported. The study record also lists multiple outcomes and analyses without reporting a multiplicity strategy. Those uncertainties limit how broadly the primary result can be generalized within the trial's own analytic framework.

The safety record lists no serious adverse events or deaths among the participants at risk in either group, and lists other adverse events in both groups. These counts do not provide a detailed comparison of event types or establish that safety is the same. Overall, the record supports a description of the posted estimates and their limitations, rather than a treatment recommendation.

17. Why This Trial Matters Statistically

This trial presents a useful statistical distinction between estimating an average difference and testing noninferiority. The posted summary-statistic comparison gives a mean difference and confidence interval, while the sponsor reports a one-sided noninferiority test. Those analyses address related but different questions. A noninferiority conclusion requires a prespecified margin, which is not reported in the registry record.

The uncertainty around the mean difference is important. The confidence interval includes zero and extends to 2.179, so the point estimate alone does not capture the full range of values compatible with the comparison. Reading the interval alongside the means and the unadjusted p-value helps prevent a single number from being interpreted as a definitive result.

The record also illustrates why denominators and missingness need close attention. The primary summaries include 34 transforaminal participants and 37 lateralized interlaminar participants, with a registry note that primary data were not collected for 3 transforaminal participants. The safety summaries use 37 participants at risk in each group. Keeping these analysis populations distinct avoids treating all reported counts as if they came from the same set of observations.

Finally, a registry record can list multiple outcomes and analyses without giving enough information to assess multiplicity, analysis populations, or missing-data handling. Those details are not minor technicalities: they affect the scope and precision of the conclusions that can be drawn from posted results. Careful interpretation means separating what the record quantifies from what it does not report.

18. 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 does not report the original sample-size justification or the assumptions used to choose the enrollment target. The planning figures associated with the observed effect are illustrations and do not establish how the trial was designed.

Primary pain data were not collected for 3 transforaminal participants. The reason for missingness, its timing, and the method used to handle missing outcomes are not reported in the registry record. The analysis population for the sponsor's result is also not specified.

The sponsor's one-sided noninferiority analysis is reported with p = 0.057, but the noninferiority margin, test decision rule, confidence interval, and other details needed to reproduce the analysis are not reported. The unadjusted recalculation is not a substitute for those missing details.

The record lists multiple outcomes and analyses but does not report a multiplicity adjustment or a complete statistical analysis plan. It also does not report baseline outcome values or whether the primary comparison was adjusted for baseline or other participant characteristics. The arm descriptions do not provide procedural detail beyond the injection routes, and the masking procedures are not described.

The safety summaries provide counts of affected participants but not event-level descriptions or a statistical comparison. These limitations do not negate the posted results; they define what can be assessed from the registry record and what remains uncertain.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Transforaminal Versus Lateralized Interlaminar Cervical Epidurals — NCT03389620.
• Registry record last updated 2025-12-19; status last verified 2025-12.
• Results first posted 2025-12-19.
• 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.

22. Record Summary

NCT03389620 is a completed randomized, double-masked, parallel-group phase 2 trial sponsored by Duke University. The registry records enrollment of 74 and compares transforaminal with lateralized interlaminar cervical epidural corticosteroid injections for cervicogenic upper extremity radiculopathy.

The primary outcome is change in pain numerical rating scale score from baseline to 2 months. The posted means are -3.38 and -2.3, with an unadjusted mean difference of 1.08 and a confidence interval extending from just below zero to 2.179. The sponsor reports a one-sided noninferiority test with p = 0.057; its margin is not reported in the registry record.

Primary pain data were not collected for 3 transforaminal participants. Safety summaries list 0 serious adverse events and 0 deaths in each group, alongside other adverse events in both groups. Missing-data handling, analysis populations, and the multiplicity strategy are not reported.