Improving COVID-19 Vaccine Uptake in Nursing Homes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Improving COVID-19 Vaccine Uptake in Nursing Homes (NCT04732819), a randomized na study. The primary outcome is Number of Residents Who Received SARS-CoV-2 Vaccine.

2. Clinical Question

3. Trial Design

The registry describes a randomized, parallel-group, single-masked trial, with facilities as the units of assignment. It identifies the design as cluster randomized: facilities, rather than individual residents, were assigned to usual care or the high-touch intervention. This matters because residents within one facility share an environment and may be exposed to the same local vaccination efforts. Their outcomes therefore cannot automatically be treated as statistically independent observations.

The comparison was between usual-care electronic messaging and education and the same material supplemented by a multi-pronged behavioral intervention. The registry describes recruitment of a random sample of facilities across four skilled nursing facility chains. It does not state the number of facilities randomized, the number assigned to each arm, or how allocation was generated or concealed. The meaning of “single” masking—specifically, which people were masked—is not reported in the registry record.

The registry lists an enrollment of 23768. It records a start date of 2020-12-21, primary completion on 2021-03-31, and a completed status. The primary outcome timeframe is 15 weeks. The record describes the primary population as long-stay residents in the nursing homes at the time of the first of three vaccine clinics. It does not give the count of residents in that population or clarify how residents who entered or left facilities during follow-up were handled.

The intervention was delivered at facility level, while the primary measure concerns residents. This alignment makes the cluster structure central to interpretation: an analysis would need to account for similarity among residents within a facility. The registry record does not report the statistical model or other details of the sponsor’s analysis. The posted results contain no sponsor analysis of the primary outcome.

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: We first identified 154 facilities within 4 nursing home chains that were eligible to participate in the trial. After talking with corporate leadership in the four chains, we excluded 21 facilities with institutional instability. This left us with 133 facilities for randomization and a total of 23,768 staff and residents across all facilities (16,272 staff and 7,496 residents).

Overall Study

Analysis population for the primary outcome, as defined in the registry: Total number of long-stay residents in the nursing homes at the time of the first of three vaccine clinics.

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 overall number of baseline participants includes the number of residents and staff in each arm. The unit analyzed was nursing home facility. Baseline characteristics were obtained for residents and nursing home facilities, but not for staff.

6. Primary Endpoint

The registry lists the primary outcome as Number of Residents Who Received SARS-CoV-2 Vaccine, measured over 15 weeks.

The primary endpoint is titled “Number of Residents Who Received SARS-CoV-2 Vaccine.” It is defined as a binary yes-or-no measure of whether a long-stay nursing home resident received any dose during the study period. Receipt was identified through electronic medical records. The stated population is the total number of long-stay residents in the nursing homes at the time of the first of three vaccine clinics, and the outcome timeframe is 15 weeks.

Although the endpoint is described as a number of residents, the underlying resident-level status is binary: each resident either had a recorded vaccine dose or did not. A group-level count can describe how many residents met that definition, but a comparative interpretation also depends on the relevant population size and the analysis used. The registry record does not report the primary outcome counts, group-specific denominators, or an estimated comparison.

Vaccination receipt is a direct measure of uptake, which is the behavioral target described for this intervention. It does not, by itself, measure vaccine hesitancy, the quality of consent, or reasons for accepting or declining vaccination. The registry record does not describe validation of electronic-record ascertainment or how missing vaccination records were classified. Those details would help readers understand how the recorded endpoint corresponds to actual vaccination.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Randomization assigns study units to comparison groups by chance, helping make groups comparable on average. In a cluster-randomized trial, the assigned unit is a group, such as a facility, rather than each person. Here, facilities were assigned to usual care or high touch, while the main outcome was measured for residents. The facility-level assignment is essential when interpreting the analysis: the number of residents is not equivalent to the number of independently randomized units.

People within a facility may have related outcomes because they share staff, local policies, access to clinics, and communication. This within-facility similarity is called intracluster correlation. When it is present, the effective information in a set of residents is less than it would be if every resident were fully independent. General statistical practice therefore accounts for clustering, for example with an analysis that models facilities or uses standard errors adjusted for clustering. Which approach the sponsor used is not reported in the registry record.

The record explicitly cautions that a simple individual-level re-analysis would be misleading and states that only the sponsor’s own analysis is shown. At the same time, the posted results list no analyses. Thus, no sponsor estimate, confidence interval, or test result for the primary comparison is available in the registry record. A confidence interval, in general, describes a range of values compatible with the data and analysis under stated assumptions. Without a reported estimate and interval, the magnitude and precision of the primary comparison cannot be evaluated.

The endpoint is a binary resident-level status—receipt of any vaccine dose or no recorded receipt—summarized as a number of residents. General approaches to binary outcomes can compare proportions, counts with appropriate denominators, or model-based effects. In a cluster trial, the chosen method must also account for the facility assignment and within-facility dependence. The registry does not report the primary outcome counts, denominators, model, effect measure, or uncertainty estimate, so it is not possible to identify or assess the sponsor’s specific approach.

Over-dispersion is a general concern when count variability exceeds what a simple model assumes. More broadly, a model that underestimates variability can yield confidence intervals that are too narrow and tests that appear more decisive than the data warrant. Cluster structure is one reason a simple individual-level model may misrepresent uncertainty. The registry’s warning about individual-level re-analysis is therefore pertinent, but the record does not specify whether over-dispersion was assessed or how any extra variability was handled.

Masking can reduce some sources of bias when people involved in care, outcome assessment, or analysis are unaware of assignments. The design is labeled single-masked, but the registry does not identify who was masked or how masking was maintained. Because the intervention included communication, opinion leaders, champions, and support, masking of facility participants may have been difficult; whether any particular group was masked cannot be inferred from the label alone.

Multiplicity refers to the risk of misleading findings when many outcomes, comparisons, or analyses are examined. A prespecified primary outcome can help focus interpretation, while secondary outcomes can provide additional context. The record identifies a primary resident vaccination outcome and a secondary staff vaccination outcome, and reports two outcome measures. It does not describe a multiplicity adjustment, a hierarchy for testing outcomes, or a prespecified statistical analysis plan. The absence of that information does not show that no plan existed; it means its details are not reported in the registry record.

Planning calculations are distinct from observed trial results. A sample-size illustration based on an observed effect is not evidence that the trial had a particular probability of detecting an effect before it was conducted. It also cannot replace the actual estimated effect and its uncertainty. General power calculations depend on assumptions, including the target effect and the analysis structure; for a cluster trial, the clustering must be reflected in design assumptions. The sponsor’s sample-size justification and the trial’s original planning assumptions are not reported in the registry record.

Finally, statistical interpretation depends on who is included in the analysis and how incomplete outcome information is handled. An intention-to-treat approach generally retains participants according to their assigned group, while other analysis populations may use different inclusion rules. The registry does not report the analysis population, missing-data methods, or the number of residents with unavailable outcome information. These omissions limit assessment of both the reported endpoint’s completeness and the applicability of any comparative result.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Usual Care", "High Touch")
n      <- c(3865, 3631)
events <- c(3083, 2994)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Usual Care  3865  3083  79.8  78.5  81.0  78.5  81.0  78.5  81.0
ARM  High Touch  3631  2994  82.5  81.2  83.7  81.2  83.7  81.2  83.7

11. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary outcome for resident vaccination and a secondary outcome concerning staff vaccination, and lists two outcome measures. It does not report whether the outcomes were tested in a prespecified order, whether any adjustment for multiple comparisons was planned, or how multiple analyses would be handled. These details are not reported in the registry record. Without the sponsor’s analysis results, there are no reported primary or secondary estimates to interpret in relation to multiplicity.

Missing-data handling is also not described. The primary outcome was identified from electronic medical records, but the registry does not say how unavailable, incomplete, or conflicting records were treated. It does not report the number of residents with missing outcome status or whether missingness differed between arms or facilities. In general, missing outcomes can affect a comparison if their occurrence or handling differs between groups; the direction and size of any such effect depend on information not reported.

The record does not state whether the analysis followed intention-to-treat principles, define an alternative analysis population, or describe exclusions after randomization. It also does not report a statistical analysis plan or explain how residents who moved, were newly admitted, or were otherwise not observed throughout follow-up were handled. These details matter because assignment occurred by facility, while outcome status was measured for residents. Their absence limits evaluation of the analysis population and the robustness of any result.

12. 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

Serious events with the highest arm-level rate: all-cause hospitalizations (4.9%).

The serious adverse event information lists Usual Care with 3865 at risk and 158 affected, and High Touch with 3631 at risk and 179 affected. Deaths are listed as 259 for Usual Care and 224 for High Touch. The leading serious-event term is All-Cause Hospitalizations, reported as 4.9%. The record lists zero other, non-serious events in both arms.

These entries are descriptive summaries, not an adjusted comparison. The registry record does not provide a statistical test, confidence interval, event timing, or an explanation of whether events were judged related to vaccination or to the intervention. “All-cause” hospitalizations should be read as hospitalizations from any cause in the reported category, not as events attributed to a particular cause. The listed counts alone cannot establish whether the groups differed beyond chance or whether an event was caused by the intervention.

The at-risk figures are not the primary-outcome denominators; the registry does not explain their relationship to the resident and staff populations in detail. The serious-event information should therefore be interpreted on its stated descriptive terms, without deriving additional rates or causal conclusions from it.

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

The trial addresses a practical statistical question: whether adding a facility-level, multi-component behavioral strategy to vaccine messaging and education changes recorded vaccine uptake among long-stay nursing home residents. Its cluster-randomized structure is suited to assigning an intervention delivered across facilities, rather than treating each resident as independently assigned.

The registry record does not report the primary outcome counts or the sponsor’s comparative analysis. It therefore does not establish the direction, magnitude, or precision of a difference in resident vaccination between the arms. The overall enrollment figure cannot fill that gap. Nor can descriptive serious-event counts answer the primary effectiveness question.

Even when a trial reports a statistically estimated difference, such a result describes the outcome defined and measured in that study; it does not automatically explain why the difference occurred or show that every intervention component contributed equally. Here, the intervention combined several forms of communication, support, and facility-level activity, so a comparative result alone would not separate their individual contributions. The registry does not report component-specific evaluations.

The endpoint records whether a resident received any dose during the study period. It does not measure completion of a vaccine series, changes in attitudes, quality of informed consent, or longer-term health outcomes. The listed secondary outcome concerns staff vaccination, but its comparative results are also not reported. The registry record therefore supports description of the study question and design, but not a conclusion about the intervention’s observed effect.

Nothing in the record warrants a recommendation for an individual, facility, or clinical program. The appropriate statistical conclusion is narrower: the trial was designed to compare the two facility-level strategies, but the comparative primary result and its uncertainty are not available in the posted registry results.

14. Why This Trial Matters Statistically

This trial is statistically notable because the treatment strategy was delivered to facilities while the main outcome was measured among residents. That cluster structure changes how evidence should be analyzed: residents from the same facility may share exposures and outcomes, so counting each resident as an independent randomized observation can overstate the information available. The registry itself warns against a simple individual-level re-analysis.

The design also illustrates how intervention delivery and outcome measurement can operate at different levels. Facility-wide training, local opinion leaders, champions, communication, and support are intended to influence decisions made by many individuals. A resident-level vaccination count is a relevant uptake measure, but interpretation depends on facility-level assignment, the resident population at risk, and appropriate handling of clustering.

At the same time, a well-framed statistical question does not replace a reported result. The registry identifies the comparison, population, and endpoint, but does not include the sponsor’s primary analysis. Readers therefore cannot assess the observed contrast or its uncertainty. This is an important distinction between a completed trial record and an interpretable comparative result.

The record also shows why safety summaries require restraint. Counts of hospitalizations and deaths provide descriptive context, but without attribution, comparative analyses, and uncertainty, they cannot establish cause or a reliable difference between groups. Statistical literacy requires separating design features, recorded outcomes, and evidence about effects rather than treating each as interchangeable.

15. 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 principal limitation for interpretation is that the registry record does not report the primary outcome results or sponsor analysis, despite noting that a simple individual-level re-analysis would be misleading. No primary effect estimate, confidence interval, or test result is listed. The record also gives no arm-specific primary-outcome denominators or resident vaccination counts, preventing assessment of the observed contrast and its precision.

Important design details are absent. The number of randomized facilities and their distribution between arms are not reported. The registry labels the trial single-masked but does not state who was masked. It does not give the allocation procedure, identify the analysis model, or explain how clustering was incorporated. The statistical analysis plan and original sample-size justification are not reported in the registry record.

Outcome ascertainment relied on electronic medical records, but validation procedures and treatment of missing or uncertain records are not described. The record does not define the analysis population or state how residents who were not observable for the full period were handled. These gaps make it difficult to judge completeness and potential differences in outcome ascertainment across facilities.

The intervention comprised several components, so a comparison of the overall strategy would not isolate the contribution of each element. The registry also does not provide secondary staff vaccination results. Safety information is descriptive and lacks causal attribution and a reported comparative analysis. These limitations reflect what the registry record reports; they do not establish that the underlying trial lacked procedures or analyses that were not entered in the record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Improving COVID-19 Vaccine Uptake in Nursing Homes — NCT04732819.
• Registry record last updated 2021-09-05; status last verified 2021-09.
• Results first posted 2021-09-05.
• 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.

19. Record Summary

NCT04732819, titled “Improving COVID-19 Vaccine Uptake in Nursing Homes,” is a completed randomized, parallel, single-masked, cluster trial sponsored by Brown University. The registry lists an enrollment of 23768, a start date of 2020-12-21, and primary completion on 2021-03-31.

Facilities were assigned to usual-care electronic messaging and education or to the same material plus a high-touch multi-pronged behavioral intervention. The primary outcome was resident receipt of any SARS-CoV-2 vaccine dose, identified from electronic medical records over 15 weeks. Staff vaccination was a secondary outcome.

The registry record does not report the primary outcome results or the sponsor’s statistical analysis. It cautions that simple individual-level re-analysis would be misleading because of the cluster-randomized design. Serious adverse event counts are listed descriptively, without a reported comparative analysis or attribution.