COVID-19 Vaccine Hesitancy Counseling Intervention for Pharmacists: A Stepped-Wedge Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for COVID-19 Vaccine Hesitancy Counseling Intervention for Pharmacists: A Stepped-Wedge Trial (NCT06547814), a randomized na study. The primary outcome is Mean Fidelity Score.

2. Clinical Question

3. Trial Design

This is a randomized, sequential, stepped-wedge trial of implementation approaches in rural community pharmacies. The registry record describes participants as beginning with standard implementation and then moving to virtual facilitation. The random allocation is described through block assignment: standard implementation and virtual facilitation each last 8-16 weeks, depending on the assigned block. The model is listed as sequential, and the record identifies the design as cluster-randomized. The unit of randomization and the number of clusters are not reported in the registry record.

Standard implementation consists of a webinar on COVID-19 vaccinations and a 30-minute online training on vaccine hesitancy communication. In the later stage, participants receive a virtual coach to help implement the counseling intervention. Fidelity is observed six times per each 8-week intervention period under each approach. When a pharmacy has more than one pharmacist, ratings are averaged to produce a pharmacy-level measure. This aggregation matters because a pharmacy, rather than each individual rating, may be the relevant unit for comparing approaches.

The trial is listed as single-masked, but who was masked is not reported in the registry record. The enrollment is 20, and the status is completed. The listed start date is 2024-08-05 and the primary completion date is 2025-03-07. These record details describe the study’s conduct and timing; they do not establish the size or precision of an effect. A sequential design can use observations collected over time, but the approaches also occur at different stages. The registry record does not report how the analysis addressed time trends, clustering, repeated observations, or other features of the 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: Pharmacies begin with the standard implementation interventions for 8-16 weeks and after completing the standard implementation, then complete the virtual facilitation intervention for 8-16 weeks.

Early Fall 2024

Late Fall 2024

Winter 2024-2025

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: The unit of analysis is pharmacy and no baseline characteristics were collected.

6. Primary Endpoint

The registry lists the primary outcome as Mean Fidelity Score, measured over From baseline through the end of the virtual facilitation periods, up to 24 weeks maximum..

The primary endpoint is the mean fidelity score for pharmacists’ delivery of the vaccine hesitancy counseling intervention. Its seven items assess competence and skill, including empathy, a non-confrontational manner, confidence without jargon, support for patient autonomy, accurate reflection of patient statements, and respectful demeanor. Each item is scored from 1 to 3; the composite is the sum, with scores from 7 to 21. Higher scores reflect greater competency in delivering the intervention.

Fidelity is useful in an implementation study because it describes whether the intended intervention is being delivered as designed. A counseling approach can be introduced without being implemented consistently, so measuring delivery helps distinguish implementation from the intervention’s possible downstream consequences. In this trial, the endpoint is intended to capture whether adding virtual coaching is associated with a difference in how pharmacists carry out counseling, rather than directly measuring vaccine uptake as the primary outcome.

The endpoint is measured repeatedly during standard implementation and virtual facilitation, with six fidelity observations per each 8-week intervention period. For pharmacies with more than one pharmacist, individual ratings are averaged to a pharmacy-level measure. The listed timeframe runs from baseline through the end of virtual facilitation, up to 24 weeks maximum. The registry record does not report the observed mean scores, standard deviations, or a between-condition estimate. The secondary outcome concerns the proportion of individuals receiving counseling who chose to get the COVID-19 vaccine; its results are also not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Understanding the design helps explain why a direct comparison of individual observations may not represent the trial’s intended analysis. The registry identifies the study as cluster-randomized and sequential. Fidelity ratings may be related when they come from the same pharmacy, and ratings from the same pharmacist may also be related across repeated observations. General statistical methods for such data account for this dependence rather than treating every rating as independent. The specific model used in this trial is not reported in the registry record.

Why does clustering matter?

General: Clustered designs assign or organize treatment at a group level, such as a pharmacy, rather than independently assigning every observation. People in one group can share working conditions, training, local practices, or other influences. As a result, their measurements may be more alike than measurements from different groups. An analysis that ignores this similarity can misrepresent the amount of information in the data and the precision of an estimate. The registry record identifies the design as cluster-randomized and explicitly cautions that a simple individual-level re-analysis would be misleading.

For pharmacies with more than one pharmacist, the record says ratings are averaged to make a pharmacy-level measure. This is one way the endpoint reflects the pharmacy as a unit. It does not establish how the final analysis handled clustering across pharmacies, the repeated observations, or the randomized blocks. Those details are not reported in the registry record. The sponsor’s own analysis takes precedence over a recalculation that ignores the cluster structure.

What does the sequence imply?

General: In a stepped-wedge or sequential implementation design, groups move between conditions over time rather than remaining in separate parallel conditions throughout. Such designs can provide information from more than one stage for a group, but time and intervention may be associated. For example, practice may change as staff gain experience, regardless of additional coaching. An analysis may need to distinguish changes associated with the approach from changes associated with time or order. The registry record describes standard implementation first and virtual facilitation afterward. It does not report the model or assumptions used to separate these influences.

The trial uses randomized block assignment to determine intervention durations of 8-16 weeks. This describes how the sequence is organized; it does not tell readers how many pharmacies were assigned to each block or whether the analysis adjusted for block. Those details are not reported in the registry record. Likewise, single masking is listed, but the identity of the masked party and the procedures used to maintain masking are not reported in the registry record.

How should a mean score be read?

General: A mean is an average that describes the central level of a continuous measurement. The standard deviation describes how spread out individual measurements are around that mean. In this trial, the composite score sums seven item ratings and ranges from 7 to 21, with higher scores representing greater competency. An average score can summarize fidelity, but the mean alone does not show how much scores vary across pharmacies or observations. The registry record does not report the condition-specific means or standard deviations, so the distribution and average level cannot be assessed.

A contrast between means would describe the average difference between the conditions on the score scale. A confidence interval, in general, describes a range of estimates compatible with the observed data under a specified statistical method. Its width communicates precision: a wider interval indicates greater uncertainty, while a narrower interval indicates greater precision. A confidence interval is not a guarantee that the true effect lies within its bounds. No confidence interval or mean contrast is reported in the registry record, so neither magnitude nor precision can be discussed for this trial.

What can planning illustrations tell readers?

General: Sample-size calculations explore how much information might be needed to detect an effect under stated assumptions. They are planning aids, not results from the trial and not evidence that the study had a particular probability of finding an effect. The planning assumptions associated with this record specify a two-sided significance level of 0.05, equal allocation, and a normal approximation; the target effect is framed as a fraction of the observed effect. Any resulting count is an illustration under those assumptions. It is not observed power and cannot substitute for an estimate and uncertainty interval from the trial.

For cluster-randomized data, planning and analysis may also depend on how similar observations are within clusters and on the number and size of clusters. The relevant inputs and the trial’s own sample-size justification are not reported in the registry record. The enrollment of 20 alone cannot establish whether the study was adequately powered, because the information available for a clustered comparison depends on more than the total number enrolled. No claim about the adequacy of the planned sample can be made from enrollment alone.

Why do repeated observations and over-dispersion matter?

General: Repeated measurements can help describe whether an outcome changes over time, but they are not equivalent to collecting the same number of independent participants. Measurements from the same person or pharmacy may share unmeasured influences. Appropriate methods account for that dependence when estimating uncertainty. The registry describes six fidelity observations per each 8-week intervention period and pharmacy-level averaging when more than one pharmacist participates. It does not report how repeated observations were incorporated into the sponsor’s analysis.

Over-dispersion is a general concern when observed variability exceeds what a chosen statistical model expects. In a count or proportion analysis, ignoring over-dispersion can understate uncertainty. The primary outcome is a continuous mean score, and the registry record does not report that over-dispersion was assessed or that a particular adjustment was used. The secondary outcome is a proportion, but its analysis method is not reported in the registry record. No specific model, standard-error method, or distributional assumption for either outcome can therefore be attributed to this trial.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Standard Implementation (Stage 1)", "Virtual Facilitation (Stage 2)")
n   <- c(17, 17)
m   <- c(18.1, 18.4)
sp  <- c(1.7, 1.5)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Standard Implementation (Stage 1)  17  18.100  1.700  17.226  18.974
CARM  Virtual Facilitation (Stage 2)  17  18.400  1.500  17.629  19.171

11. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures: the primary mean fidelity score and the secondary proportion of individuals receiving vaccine hesitancy counseling who chose to get the COVID-19 vaccine. It also lists one analysis. The statistical plan, including whether the analysis addresses multiple outcomes or multiple comparisons, is not reported in the registry record. Without that plan, the role of any significance threshold across outcomes cannot be determined. General: When several outcomes or comparisons are tested, the chance of at least one apparently notable result can increase; a prespecified multiplicity approach can clarify how such tests are interpreted. No such approach should be assumed for this trial.

How missing observations were handled is not reported in the registry record. The record describes repeated fidelity observations and a pharmacy-level average when more than one pharmacist is present, but it does not say how incomplete ratings, missed observations, or incomplete participation were treated. It also does not report whether missingness was assessed for patterns that could affect the comparison. This matters because an average based on available observations may represent a different set of observations across conditions.

The analysis population is not reported in the registry record. It is not possible to determine from the described details whether all enrolled participants or pharmacies were included in the analysis, or whether inclusion depended on completing a particular stage or providing enough fidelity observations. The numerical results and participant flow needed to assess these issues are not reported in the registry record. These gaps prevent a full appraisal of the analysis, even before considering the cluster-randomized sequence and its implications for uncertainty.

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

The trial is designed to evaluate an implementation strategy, not directly to test whether COVID-19 vaccination prevents illness or whether counseling changes health outcomes. Its primary measure concerns pharmacists’ fidelity and competency in delivering vaccine hesitancy counseling. That focus can help distinguish whether implementation support is associated with how the intervention is delivered. It does not, on its own, establish whether patients choose vaccination or experience a health effect.

The secondary outcome concerns the proportion of people receiving counseling who chose to get the COVID-19 vaccine. The registry record names this outcome but does not report its result. Even if a result were available, interpreting it would require attention to the comparison, uncertainty, and analysis methods. The record details do not establish that any change in vaccine choice was caused by the implementation approach.

The registry record also does not report the comparative primary outcome estimate, so it is not possible to say whether fidelity was higher, lower, or similar under virtual facilitation, or to judge the precision of such a comparison. A lack of a reported result should not be read as evidence of no effect. It means the available record details do not support a conclusion about the estimated effect.

Likewise, enrollment of 20 should not be treated as proof that the study was either adequate or inadequate for its question. Precision depends on the design, clustering, outcome variability, and analysis, among other factors. The sample-size illustrations are planning exercises based on stated assumptions, not observed power. Finally, zero entries in the listed adverse-event count fields do not establish that no events occurred or that an approach is safe; the event data and their completeness cannot be assessed from those entries alone. The appropriate interpretation is limited to the trial’s stated question, methods, and the results actually reported in the registry record.

13. Why This Trial Matters Statistically

This trial illustrates a statistical challenge common to implementation research: an intervention is introduced in real-world groups, while outcomes are collected repeatedly and may be related within groups. Treating every rating as an independent piece of information can distort the estimated precision. Recognizing the cluster-randomized design is therefore central to interpreting any comparison, and the registry record explicitly cautions against a simple individual-level re-analysis.

The sequence also makes time an important consideration. Participants move from standard implementation to virtual facilitation, so the stage of the trial and the implementation approach are connected. Experience, local conditions, or other changes over time could influence fidelity alongside coaching. A sound interpretation depends on an analysis that accounts for the design, but the specific model and its treatment of time are not reported in the registry record.

The outcome focuses on fidelity rather than directly on patient decisions. That distinction is statistically and substantively useful: implementation support may affect whether counseling is delivered as intended, while downstream outcomes require separate measurement and analysis. The registry includes a secondary vaccine-choice outcome, but its results are not reported in the registry record. Keeping these outcomes distinct prevents an implementation measure from being treated as evidence of a downstream effect.

Finally, a completed trial record and a defined endpoint do not, by themselves, communicate the strength of evidence. Readers need an estimate, a measure of uncertainty, and enough information about analysis and missing data to judge what the data support. Since those comparative results and methods are not reported in the registry record, the design offers a valuable statistical case for careful interpretation, but no conclusion about the effect of virtual facilitation can be drawn from the available results.

14. 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 information available for interpreting this trial is limited because the comparative results for the primary fidelity outcome are not reported in the registry record. Condition-specific means, standard deviations, an estimated contrast, uncertainty intervals, and a test result are not reported in the registry record. Without them, the size, direction, and precision of the comparison cannot be evaluated.

Important design details are also not reported in the registry record. These include the number of pharmacies, the unit and allocation details for the randomized blocks, the identity of the party masked, and the statistical model used for the cluster-randomized sequence. The record does not explain how the analysis handled time, clustering, pharmacy-level averaging, or repeated observations. These details are particularly relevant because the study compares successive stages and collects multiple fidelity observations.

The record does not report the sample-size justification, analysis populations, missing-data procedures, or a multiplicity plan. Although the enrollment is 20, the number of pharmacies and distribution of participants across them are not reported in the registry record. Enrollment alone cannot describe the effective information available for a cluster-randomized analysis. Likewise, the record does not provide the result for the secondary vaccine-choice outcome.

The primary endpoint is a competency score based on seven items, but the registry record does not report validation details, assessor procedures, or information about consistency of ratings. The stated single masking does not clarify whether outcome assessors were masked. Finally, the listed adverse-event counts do not include enough context to assess completeness or interpret event risk. These limitations constrain appraisal of both the estimated effect and the reliability of the measurement, without implying a particular direction of bias.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: COVID-19 Vaccine Hesitancy Counseling Intervention for Pharmacists: A Stepped-Wedge Trial — NCT06547814.
• Registry record last updated 2026-01-21; status last verified 2025-11.
• Results first posted 2026-01-12.
• 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.

18. Record Summary

The registry record describes a completed, randomized, single-masked, sequential, cluster-randomized trial in rural community pharmacies, with enrollment of 20. Participants began with standard implementation, including a webinar and a 30-minute online training, then moved to virtual facilitation with a virtual coach. Randomized block assignment determined intervention periods of 8-16 weeks.

The primary outcome is mean fidelity score, based on seven items and scored from 7 to 21, with higher scores reflecting greater competency. Fidelity observations are collected during both approaches, with pharmacy-level averaging when more than one pharmacist participates. A secondary outcome concerns vaccine choice among people receiving counseling.

The comparative outcome results, statistical model, missing-data methods, and analysis populations are not reported in the registry record. The cluster structure makes simple individual-level re-analysis misleading, so the sponsor’s own analysis takes precedence over an unadjusted recalculation.