Veterans Response to Dosage in Chiropractic Therapy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Veterans Response to Dosage in Chiropractic Therapy (NCT04087291), a randomized na study. The primary outcome is Number of Participants With Clinically Meaningful Improvement in Back-Related Disability From Baseline to Week 10.

2. Clinical Question

3. Trial Design

VERDICT is a pragmatic randomized trial with a parallel-group model. The registry describes the masking as double. It records 766 participants enrolled and a completed study status. Participants were randomized in Phase 1 to a lower dose of chiropractic care, 1-5 visits, or a higher dose, 8-12 visits, over 10 weeks. Both groups received multimodal, evidence-based chiropractic care.

The trial also included a second randomization. After Phase 1, participants who completed that phase and consented to continue were randomized to chronic chiropractic pain management (CCPM) or no CCPM for 10 months. CCPM is described as one scheduled chiropractic visit per month. The no-CCPM group had no scheduled chiropractic visits during that follow-up period, in addition to usual VA care; participants could continue to access other VA health services. This later randomization creates four Phase 2 groups defined by the initial dose and subsequent CCPM assignment.

The primary outcome concerns the Phase 1 randomization only: clinically meaningful improvement in back-related disability from baseline to week 10. The record separately lists outcomes concerning pain intensity, PEG pain severity, PROMIS physical, social, and mental health, and days per week with low back pain, including outcomes at week 52. Those measures should not be conflated with the primary endpoint. The registry record does not report the original sample-size justification or a detailed statistical analysis plan.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Potential participants were screened using a standardized five-step process: identification and notification, phone screening, informed consent and baseline visit (virtual or in person), eligibility questionnaires, and in-person chiropractic consultation to confirm clinical eligibility. Veterans who were uninterested, ineligible, or unable to complete screening before consultation were excluded. A total of 3,667 veterans were screened; 766 met criteria and were enrolled and randomized.

Phase 1 (0-10 Weeks)

Reasons for not completing this period:

Phase 2 (10-52 Weeks)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All participants randomized to Phase 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: All participants randomized to Phase 1 with baseline data collected prior to initiation of study interventions.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Clinically Meaningful Improvement in Back-Related Disability From Baseline to Week 10, measured over Baseline to Week 10.

The primary endpoint counts participants who achieved a clinically meaningful improvement in low back-related disability between baseline and week 10. Improvement is defined in the registry as at least a 30% relative improvement on the Roland Morris Disability Questionnaire (RMDQ). The questionnaire ranges from 0-24, with higher scores indicating greater disability.

This is a binary endpoint at the participant level: each person either meets the registry definition of improvement or does not. Reporting the count alongside the number of participants in each group lets readers see how many participants met the threshold in relation to each group’s size. The registry reports 140 of 382 participants in the lower-dose group and 157 of 384 in the higher-dose group.

A threshold-based outcome can be easier to interpret than a mean score change because it expresses whether participants crossed a prespecified level of relative improvement. It also simplifies a range of possible score changes into two categories, so it does not describe the size of improvement among participants who met the definition, or the exact score changes among those who did not. The registry record does not report the clinical rationale for choosing the 30% threshold or further details about how baseline scores were handled.

7. Results Posted in the Registry

• Phase 1: Higher Dose Chiropractic Care (8-12 Visits): 40.9% (157/384) versus 36.6% (140/382) on control; risk difference 4.2 percentage points (95% CI -2.7 to 11.1), risk ratio 1.12 (0.93–1.33); unadjusted two-proportion z-test p = 0.229.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the risk difference represent?

In general, a risk difference is the difference between the proportions of participants with an event in two groups. For this endpoint, the event is meeting the definition of clinically meaningful improvement by week 10. A positive estimate in the posted comparison indicates a higher observed response proportion in the higher-dose group than in the lower-dose group. The measure is expressed on an absolute scale, so it describes the separation in the proportions rather than a relative increase. The registry reports the risk difference and its confidence interval; its point estimate is positive, while the interval extends below and above zero.

The risk difference is useful for a binary outcome because it keeps the comparison tied to the observed frequency of participants meeting the threshold. It is distinct from the risk ratio, which compares proportions as a ratio, and the odds ratio, which compares event odds. The record includes relative measures as well, but the primary interpretation focuses on the reported risk difference. Different measures can express the same group comparison on different scales and need not look numerically alike.

What does the p-value tell readers?

In general, a p-value measures how unusual results at least as discrepant as those observed would be if a specified null model were true and the analysis assumptions held. The posted unadjusted p-value is 0.229. It does not express the probability that a treatment effect exists, and it does not tell readers the chance of obtaining the same result in a repeat trial. A p-value also does not show the magnitude or precision of the estimated difference; those are better conveyed by the effect estimate and its confidence interval.

The registry record does not report the complete analysis specification behind this p-value. In particular, the detailed model, any covariate adjustment, and the approach to missing outcomes are not reported in the registry record. The term unadjusted indicates that the comparison is not presented as an adjusted estimate in the posted results. It should not be assumed that the result accounts for baseline differences or other variables.

Why is a confidence interval important?

In general, a confidence interval gives a range of effect estimates compatible with the data and the procedure used to calculate the interval. It helps convey uncertainty around a point estimate. A narrow interval indicates more precision than a wide interval, while an interval that includes no difference allows for both directions of effect as well as no difference under the stated scale. For this result, the risk-difference interval reaches below and above zero. That pattern signals uncertainty about the direction and size of the underlying difference; it does not establish that the groups are identical.

A confidence interval is not a guarantee that the true effect lies inside its limits, and it does not summarize every source of uncertainty. Its interpretation depends on the analysis method and assumptions. Those details are not reported in the registry record, so the interval can be described as posted but not fully assessed for methodological choices.

How should the endpoint’s binary form be understood?

The RMDQ score itself is a scale, but the primary endpoint reduces change on that scale to whether a participant meets the specified relative-improvement threshold. This makes the endpoint a count of participants with an event. It also discards information about the degree of change on either side of the threshold. Two participants with very different score changes could be classified alike if both meet the definition, and participants with changes close to the threshold could fall into different categories. The registry’s secondary outcome list includes change in low back pain intensity and other measures, but those do not replace the primary endpoint.

For a binary endpoint, group counts and group denominators are central to interpretation. Counts alone can mislead when denominators differ. The posted comparison therefore reports the arm sizes and a risk-difference estimate with an interval, in addition to the event counts. The registry does not report further endpoint-level information about missing observations or whether every randomized participant contributed an outcome.

What do the two randomizations mean for interpretation?

The initial randomized comparison addresses the two Phase 1 dose ranges. The second randomization occurred after Phase 1 and was limited to participants who completed that phase and consented to continue. It assigned CCPM or no CCPM for the 10-month follow-up. These are different randomized questions and populations. The primary endpoint is explicitly defined for all participants randomized to Phase 1 and measured through week 10; it is not a Phase 2 outcome. The registry record does not report Phase 2 efficacy comparisons in the posted primary result.

Randomization is intended, in general, to make groups comparable at assignment with respect to both measured and unmeasured factors. It does not guarantee that the groups will be identical, nor does it address every issue arising after randomization, such as missing measurements or departures from assigned care. The record identifies randomized allocation but does not report the details needed to evaluate adherence, missing data, or any analysis that accounts for these issues.

10. Confidence Intervals and Precision

Precision concerns how much uncertainty surrounds an estimated effect. The primary comparison reports a risk difference with a confidence interval. The interval extends from a negative value to a positive value, so it includes no difference and values in either direction. This means the posted results do not tightly establish which dose group had the higher underlying proportion meeting the improvement threshold, or how large the difference might be.

The lower-dose group had 140 participants meeting the endpoint among 382, and the higher-dose group had 157 among 384. The counts are close in scale to their group sizes, but the observed counts should not be treated as exact estimates of what would occur in another study. Sampling variation is one reason the confidence interval has a range rather than a single value. The posted unadjusted p-value of 0.229 is another summary of the data under a null model, but it does not replace the interval as a description of uncertainty.

Confidence intervals are most informative when considered with the effect measure, its direction, and the outcome definition. Here, the endpoint is whether participants crossed a threshold, so the interval concerns the difference in proportions meeting that definition. It does not provide an interval for mean RMDQ change or for any secondary outcome. Nor does it capture uncertainty about the endpoint definition, missing data, or analysis choices that are not detailed in the registry record.

The interval’s inclusion of no difference should not be read as proof of equivalence. Equivalence requires a prespecified margin and an analysis designed to test whether effects fall within that margin; such details are not reported in the registry record. Conversely, the positive point estimate alone does not establish a dependable advantage for the higher-dose assignment. The interval is central because it communicates that the observed difference is uncertain across a range that includes no difference and both possible directions.

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 sample-size table and curve are planning illustrations based on fractions of the observed effect, not results from the trial’s original design calculations. The stated 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 the stated fraction of the observed effect. For this non-time-to-event comparison, the listed numbers are participants per arm, not total event counts. The total-events convention applies to time-to-event tables, which is not the endpoint presented in this comparison.

The 80% power column gives the planned participants per arm for a design with an 80% probability of detecting the specified effect under the planning assumptions, if that effect is the true one and the model assumptions hold. The 90% power column gives the corresponding planning number for a 90% probability. These are design probabilities, not the probability that the current result is true. This is not observed power.

Rows that plan for 100%, 75%, or 50% of the observed effect ask how many participants per arm would be needed to target an effect at that fraction of the observed estimate. Smaller effects are harder to distinguish from random variation, so they require more participants to achieve the same planned power. The listed requirements rise as the assumed fraction gets smaller. These values are illustrations for future planning and do not show that the study was originally designed around the observed effect. The original sample-size assumptions and justification are not reported in the registry record.

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 = 0.366; treatment proportion p₁ = 0.409; average p̄ = 0.388.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6890 = 1.3505.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6884 = 0.5794.
4. n = (1.3505 + 0.5794)² / (0.0424)² = 2,075.3, rounded up to 2,076 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 <- 0.366492146596859                          # control proportion
p1 <- 0.408854166666667                          # 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(4087291)
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 = 2,076 (80% power), 2,778 (90% power)
 75% of observed effect: n per arm = 3,672 (80% power), 4,915 (90% power)
 50% of observed effect: n per arm = 8,218 (80% power), 11,001 (90% power)
power.prop.test: 2075.3 
Simulated power at n = 2,076 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. Multiplicity, Missing Data and Analysis Populations

The registry lists several secondary outcomes alongside the primary endpoint, including measures of pain, physical and social health, mental health, and days with low back pain. When many outcomes or comparisons are examined, the chance of at least one apparently unusual result can increase. In general, multiplicity procedures may adjust the interpretation or statistical thresholds to account for multiple tests. The registry record does not report whether any such procedure was planned or used, or how the primary endpoint was prioritized relative to the secondary outcomes beyond its designation as primary.

Missing outcome data can affect a randomized comparison if the availability of measurements differs between groups or is related to outcomes. The record states that the primary population is all participants randomized to Phase 1, but does not report how missing week 10 RMDQ measurements were handled, how many were missing, or whether the posted arm counts include all randomized participants with complete endpoint data. It also does not report sensitivity analyses addressing missing outcomes. These omissions limit assessment of whether the reported counts represent the full randomized population as intended.

The primary outcome description identifies all Phase 1 randomized participants as the population, but the detailed analysis population and method are not reported in the registry record. The registry record also does not report an intention-to-treat procedure, per-protocol analysis, covariate adjustment, or a model specification. Those labels should not be inferred solely from the randomized design or the stated population. The posted result includes an unadjusted p-value and a risk-difference interval; the record does not explain how either was calculated.

Phase 2 involved a further randomization among participants who completed Phase 1 and consented to continue. That eligibility condition means Phase 2 groups represent a different selection of participants from the full Phase 1 randomized population. Interpretations of Phase 2 comparisons would need to respect that distinction. The primary result under discussion concerns Phase 1 through week 10, and no Phase 2 primary efficacy result is reported in the record. Overall, the registry gives enough information to identify the primary outcome and the posted group counts, but not enough to reconstruct the full analysis plan or evaluate all missing-data and multiplicity decisions.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: major orthopedic surgery (2.9%), other major cardiac events (non-arrhythmia) (2.4%), serious infection (including respiratory and systemic) (2.4%), other major surgery (non-orthopedic) (2.3%), gastrointestinal emergency (non-cancer) (1.3%).

The registry reports serious adverse events separately for the Phase 1 and Phase 2 groups. In Phase 1, 11 participants in the lower-dose group and 13 in the higher-dose group were recorded as affected by serious adverse events, with 382 and 384 participants at risk, respectively. Deaths were 0 in the lower-dose group and 1 in the higher-dose group. The record reports no other adverse events for either Phase 1 group in the listed other-adverse-event category.

For Phase 2, serious adverse events were reported for 19 of 172 participants assigned to lower dose followed by CCPM, 15 of 172 assigned to lower dose followed by no CCPM, 19 of 169 assigned to higher dose followed by CCPM, and 13 of 164 assigned to higher dose followed by no CCPM. Deaths were 0, 0, 0, and 2, respectively. In the other-adverse-event category, the counts were 0, 0, 2, and 0, respectively. The Phase 2 groups followed a second randomization among participants who completed Phase 1 and consented to continue, so their denominators differ from the Phase 1 randomized groups.

These counts describe events recorded in the registry; they do not establish that an event was caused by the assigned care. The registry record does not report event-level timing, exposure duration, or a causal assessment for these counts. The listed serious-event terms should also be read as descriptive categories, not as comparisons between randomized groups.

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

The result addresses a focused question: whether the proportion of Veterans with chronic low back pain meeting a defined threshold for improvement in back-related disability by week 10 differed between two assigned doses of chiropractic care. The posted counts are 140 of 382 in the lower-dose group and 157 of 384 in the higher-dose group. The risk-difference interval includes no difference and extends in both directions, while the unadjusted p-value is 0.229. Taken together, the posted estimates leave substantial uncertainty about the size and direction of the underlying difference.

The result does not compare chiropractic care with no care during Phase 1. Both groups were assigned chiropractic care, with different visit ranges. It also does not answer whether Phase 2 CCPM changes disability, pain, or health-service use, because the primary endpoint is limited to baseline through week 10 and the registry record does not report a Phase 2 efficacy result in this primary comparison.

The endpoint records whether a participant achieved at least a 30% relative improvement on the RMDQ. It does not show average score change, the degree of improvement for each participant, or whether changes short of the threshold were important. The secondary measures named in the record address other outcomes and time points; they should not be substituted for the primary endpoint or treated as evidence for it.

The posted result also does not establish equivalence between the dose groups. A confidence interval spanning no difference indicates uncertainty, not proof that effects are the same. Nor does the positive point estimate by itself demonstrate a reliable difference. Further interpretation depends on details not reported in the registry record, including how missing week 10 outcomes and the analysis population were handled. These results describe the posted data and their statistical uncertainty; they are not medical advice.

16. Why This Trial Matters Statistically

This trial is statistically informative because its primary question is a dose comparison rather than a simple treatment-versus-no-treatment contrast. Both Phase 1 groups were assigned chiropractic care, but to different visit ranges. The distinction matters when interpreting the estimand: the result concerns the relative frequency of meeting a disability-improvement threshold under the two assigned dose strategies, not the effect of receiving care compared with receiving none.

The endpoint also illustrates how a continuous scale can be converted into a binary outcome. The RMDQ has a range of scores, but the primary result counts participants who cross the specified relative-improvement threshold. This makes the outcome easier to express as an event count, while losing information about the amount of change. Reading the counts together with the interval helps avoid overemphasizing the raw difference in event numbers.

The two-stage design creates an additional statistical distinction. Phase 1 asks about the initial dose assignment, while Phase 2 re-randomizes eligible participants to ongoing CCPM or no scheduled CCPM. Those stages involve different assignment questions and populations. Keeping them separate prevents the primary week 10 result from being incorrectly interpreted as an estimate of the later management strategy.

Finally, the record demonstrates why effect estimates, uncertainty, and reporting detail all matter. The point estimate is positive, but the interval reaches across no difference, and the unadjusted p-value does not independently establish a clear separation. Missing-data handling and the complete analysis plan are not reported in the registry record, so readers should distinguish the information that is posted from analysis details that remain unavailable.

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 does not report the original sample-size justification, so the planning illustrations based on the observed effect cannot show what assumptions guided trial design. The detailed statistical analysis plan is also not reported in the registry record. In particular, the record does not specify the analysis model, whether baseline covariates were used, or the methods used to calculate the posted interval and p-value.

Missing outcome information is another limitation. The primary population is identified as all participants randomized to Phase 1, but the record does not report the number with unavailable week 10 RMDQ data or how such data were treated. Without that information, it is not possible to determine from the registry record whether the reported counts include the intended population or how sensitive the comparison may be to missing outcomes.

The threshold-based endpoint simplifies interpretation but does not capture the full distribution of RMDQ change. The registry record does not provide the continuous score changes within the primary result, so the result cannot describe the magnitude of change among participants who did or did not cross the threshold. The record also lists multiple secondary outcomes, but it does not report whether multiplicity was addressed in the analysis.

The trial’s second randomization took place among Phase 1 completers who consented to continue, so Phase 2 groups are not the same population as all participants initially randomized. The primary result should not be generalized to that later comparison. Finally, registry adverse-event counts are descriptive and do not by themselves establish causation. Further details about event timing, exposure, and attribution are not reported in the registry record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Veterans Response to Dosage in Chiropractic Therapy — NCT04087291.
• Registry record last updated 2026-08-11; status last verified 2026-07.
• Results first posted 2026-08-11.
• 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

The registry record describes VERDICT as a completed, pragmatic, randomized, double-masked, parallel-group trial with 766 participants enrolled. Phase 1 compared lower-dose chiropractic care, 1-5 visits, with higher-dose care, 8-12 visits, over 10 weeks. Participants who completed Phase 1 and consented to continue were randomized again to CCPM or no scheduled CCPM for 10 months.

The primary endpoint was the number of Phase 1 randomized participants with at least a 30% relative improvement in RMDQ disability from baseline to week 10. The posted counts were 140 of 382 in the lower-dose group and 157 of 384 in the higher-dose group. The comparison’s unadjusted p-value was 0.229, and the risk-difference confidence interval included no difference.

The record includes secondary outcomes and adverse-event counts, but detailed analysis methods, missing-data handling, multiplicity procedures, and the trial’s original sample-size justification are not reported in the registry record.