Baseline Oral Health Study: UnCoVer the Connections to General Health: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Baseline Oral Health Study: UnCoVer the Connections to General Health (NCT04954313), a randomized phase 4 study. The primary outcome is Impact of Treatment on Change in Oral Health Measures.

2. Clinical Question

The trial asks whether adding the treatment-group intervention to periodontal treatment is associated with a different course of oral-health measurements over follow-up. The registry frames the study more broadly as examining oral health and whether changes in oral health are associated with changes in biomarkers of systemic health. The reported primary comparison concerns probing depth, not a direct test of all systemic-health outcomes.

The posted primary estimate is a difference in change between groups. Its confidence interval extends to both negative and positive values, and the reported p-value is 0.957. This is evidence that the posted analysis did not distinguish the groups statistically on this endpoint; it is not, by itself, proof that the interventions are equivalent or that no meaningful difference is possible. The precision and scope of the result should be considered alongside the endpoint definition, analysis population, missing-data approach, and the registry’s incomplete explanation of enrollment and intervention details.

3. Trial Design

The registry describes a Phase 4, randomized, double-masked, parallel-group trial. Participants were allocated to a control group or a treatment group. The record status is completed. The registry lists a start date of 2021-09-28 and a primary completion date of 2025-02-01.

The stated enrollment is 786. Separately, the brief summary says the study planned to enroll approximately 200 participants, with approximately 100 in each group. It also says recruitment might be expanded to University of North Carolina participants if screening from the North Carolina Project Baseline Health Study pool did not fulfill study enrollment powering. The registry record does not explain the difference between the stated enrollment and the approximate enrollment described in the summary, or how the enrollment figure relates to the analysis populations.

The trial is described as using an intent-to-treat group assignment for the primary longitudinal comparison. The primary analysis description says it includes all subjects who were randomized and their data from visits T1 through T4. The primary outcome figures list 73 participants in the control group and 80 in the treatment group. Those outcome-specific figures should not be treated as a reconciliation of the enrollment figures: the registry does not explain the difference.

The treatment group is described as receiving the same periodontal treatment as the control group in addition to subgingival chlorhexidine irrigation. The control group is described as receiving an initial periodontal treatment involving a pre-procedural rinse and scaling and root planing. The intervention and arm descriptions are incomplete in the registry record, so a more detailed account of all products and procedures cannot be established from the record.

Masking is listed as double, but the identities of the masked parties and how masking was maintained are not reported in the registry record. The method used to generate the random allocation sequence, allocation concealment, and any stratification factors are also not reported in the registry record. A sample-size justification for the trial itself is not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: Of the 786 participants who signed informed consent, 194 completed the required assessments, and remained eligible. Of the 194 eligible participants, 39 declined to be randomized. Of the 155 randomized participants, two did not continue following randomization and did not receive any exposure to a study intervention.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Data include imputation. Predictions based on linear mixed models were used for imputation.

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: two participants withdrew between randomization and first intervention

6. Primary Endpoint

The registry lists the primary outcome as Impact of Treatment on Change in Oral Health Measures, measured over Comparing T1 (day 90) and T4 (day 360).

The primary outcome is titled “Impact of Treatment on Change in Oral Health Measures.” Its stated unit is millimeters, described as a probing-depth difference. Periodontal probing depths are measured per sextant, and the registry says whole-mouth mean probing depths are used as a measure of oral health. The reported time frame compares T1 (day 90) with T4 (day 360), while the analysis description uses longitudinal data from visits T1 through T4.

The endpoint is used to quantify change in a clinical measure of periodontal status and to compare that change between randomized groups. Using a repeated-measures approach can incorporate measurements across follow-up rather than treating each visit as an unrelated observation. The registry reports least-squares means and a 95% confidence interval for the primary outcome, along with a difference-in-differences estimate from generalized estimating equations.

For the control group, the posted estimate is 0.186, with a 95% confidence interval from 0.086 to 0.285. For the treatment group, it is 0.19, with a 95% confidence interval from 0.074 to 0.306. The reported between-group difference in change is -0.004, with a 95% confidence interval from -0.157 to 0.149. The registry does not fully specify the precise construction of the visit-to-visit change underlying the reported group estimates, or how individual sextant measurements are combined beyond stating that whole-mouth mean probing depth is used.

Probing depth is the primary oral-health measure in the posted result. It should not be conflated with the additional listed outcomes, which include bleeding on probing, gingival index, clinical attachment level, dental plaque, flow-mediated dilation, glycated hemoglobin, emotional wellbeing, and beliefs about periodontal disease.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary analysis estimate?

The registry reports a difference in difference of least-squares means for the primary comparison. In general, a difference-in-differences estimate compares the amount of change in one group with the amount of change in another. It focuses on the contrast in change rather than simply comparing one group’s follow-up value with the other group’s follow-up value. The posted estimate is -0.004 millimeters. Its 95% confidence interval is -0.157 to 0.149, and the reported p-value is 0.957.

Least-squares means are model-based group estimates. They can account for features included in the statistical model, unlike a simple arithmetic mean calculated without adjustment. The registry does not report all model terms or covariates, so the precise adjustments behind these estimates are not reported in the registry record. The group estimates and the between-group contrast should be read according to the model and outcome definitions stated in the registry.

Why use generalized estimating equations?

GEE is a general approach for estimating average effects when observations are correlated, such as repeated measurements from the same participant. Measurements from one person over time are commonly related rather than independent. GEE accounts for that within-participant dependence when estimating population-average relationships and their uncertainty. This makes it suitable in principle for longitudinal outcomes, as described for this trial.

Over-dispersion is a general statistical concern in some outcome models: observed variability can exceed what a model’s assumed distribution allows, which can make standard errors too small if not handled appropriately. The primary outcome is continuous, and the registry does not describe an over-dispersion adjustment. It also does not report the working correlation structure, variance estimator, or other implementation details for GEE. Those elements are not reported in the registry record and should not be inferred from the method label alone.

How should the confidence interval and p-value be read?

A confidence interval expresses uncertainty around an estimated effect under the assumptions of the model. Here, the reported 95% interval spans negative and positive differences. The result therefore does not isolate a direction of between-group difference. A p-value measures how incompatible the observed result is with a specified null hypothesis under the analysis assumptions; it is not the probability that the null hypothesis is true, nor is it a measure of clinical importance. The reported p-value of 0.957 does not establish equivalence. To support an equivalence conclusion, the record would need a prespecified margin and an analysis designed to evaluate that margin; these are not reported in the registry record.

The reported confidence interval and p-value belong to the registry’s GEE analysis. Recalculation from posted summary statistics is not used for this primary endpoint. The recalculated secondary comparisons are simple, unadjusted comparisons and should not be confused with the primary longitudinal model.

What does imputation contribute?

The primary outcome population is described as including imputed data. The registry says predictions based on linear mixed models were used for imputation. In general, imputation fills in missing values using a model rather than discarding every participant with incomplete outcome data. A linear mixed model can account for repeated observations and participant-level patterns, depending on its specification. Imputation can help retain information, but its validity depends on assumptions about why values are missing and whether the imputation model represents the data adequately.

The registry does not report the number of missing measurements, the timing or pattern of missingness, the variables included in the imputation model, the number of imputed data sets, or sensitivity analyses under alternative missing-data assumptions. It is therefore not possible to assess those aspects from the registry record. The primary analysis is described as intent-to-treat by randomized group assignment, but the record does not provide a fuller accounting of the participants contributing observed and imputed measurements at each visit.

What do the secondary recalculations show methodologically?

The secondary continuous-outcome comparisons are presented as mean differences between treatment and control, with confidence intervals and unadjusted p-values recalculated from posted summary statistics. In general, a mean difference keeps the result in the outcome’s original units. A standardized effect such as Cohen’s d expresses a difference relative to a pooled standard deviation and can help describe scale-independent magnitude, but it does not replace the original-scale estimate or establish clinical importance.

The recalculated results are simple comparisons, not adjustments for the trial’s longitudinal structure or for other covariates. For emotional wellbeing, the recalculated mean difference is -1.85, with a 95% confidence interval from -6.79 to 3.09 and an unadjusted p-value of 0.463. For beliefs about periodontal disease, the recalculated mean difference is -3.16, with a 95% confidence interval from -8.008 to 1.688 and an unadjusted p-value of 0.201. For bleeding on probing, the recalculated mean difference is -0.103, with a 95% confidence interval from -0.183 to -0.023 and an unadjusted p-value of 0.011. These are unadjusted recalculations; the sponsor’s reported analyses use two-sided t-tests and report p-values of 0.469, 0.220, and 0.014, respectively.

Planning illustrations based on observed secondary effects are not observed power calculations and do not describe the trial’s original design. In general, planning calculations estimate the participants per arm that would be needed under stated assumptions to target a chosen power for a chosen effect. The registry’s illustrations use a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects defined as stated fractions of observed effects. These are prospective planning illustrations based on the observed estimates; the trial’s own sample-size justification is not reported in the registry record.

10. Analyses Reported by the Sponsor

For the primary endpoint, the registry reports the sponsor analysis using generalized estimating equations: a difference in difference of least-squares means of -0.004, a 95% confidence interval from -0.157 to 0.149, and p = 0.957. No separate recalculation is presented for this primary result. The reported analysis is therefore the basis for interpreting the primary comparison.

For the three secondary outcomes with detailed summary statistics, simple unadjusted recalculations produce p-values of 0.463 for emotional wellbeing, 0.201 for beliefs about periodontal disease, and 0.011 for bleeding on probing. The sponsor’s reported two-sided t-test p-values are 0.469, 0.220, and 0.014, respectively. The small differences are consistent with the recalculations and sponsor analyses not being identical procedures: the recalculations use posted summary statistics, while the record identifies the sponsor’s method as a two-sided t-test. The registry does not provide enough detail to determine every reason for the differences.

The discrepancy indicator for these recalculations is false. Even so, the sponsor’s reported analyses remain the registry’s formal analyses; the recalculated results are supplementary, simple unadjusted checks rather than replacements for the sponsor’s analysis. They also do not address multiplicity across secondary outcomes or the primary analysis’s longitudinal modeling.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 16 outcome measures and 16 analyses, and it identifies 13 secondary outcomes as posted. Multiple outcome tests raise a general multiplicity concern: as the number of statistical tests increases, the chance of at least one apparently positive result from random variation also increases. Statistical approaches can account for that issue through adjustments, testing hierarchies, or other prespecified strategies. A multiplicity plan, including whether any adjustment was used, is not reported in the registry record.

The recalculated secondary p-values are explicitly unadjusted. The sponsor’s secondary analyses are described as two-sided t-tests, but the record does not report a correction for multiple comparisons or a prespecified ordering of secondary hypotheses. Consequently, a nominally small secondary p-value should be interpreted with caution and in the context of the full set of outcomes, rather than as independent confirmation of the primary result.

For the primary endpoint, the registry states that data include imputation and that predictions based on linear mixed models were used for imputation. This indicates that missing values were addressed, but it does not establish the assumptions or robustness of that approach. The amount and pattern of missing data, the variables used in the imputation model, and sensitivity analyses are not reported in the registry record. The effect of missingness on the reported estimates therefore cannot be assessed fully.

The primary analysis is described as using all randomized subjects and their data from T1 through T4 according to intent-to-treat group assignment. This is a useful statement of the intended analysis population, but the registry does not report a participant flow account or explain how many randomized participants had observed, imputed, or unavailable measurements at each visit. The primary outcome groups list 73 control participants and 80 treatment participants; the registry does not reconcile those figures with the stated enrollment of 786 or with the brief summary’s approximate enrollment of 200.

For the secondary outcomes, analysis sizes differ by measure. The reasons for the differences, handling of missing secondary measurements, and analysis populations for each outcome are not reported in the registry record. No assumptions about exclusions, protocol deviations, interim analyses, or sample-size justification should be inferred from the available descriptions.

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: pancreatitis (1.4%).

The registry reports serious adverse events for the control and treatment groups separately. In the control group, one serious adverse event is reported among 73 participants at risk, and one death is listed. In the treatment group, zero serious adverse events are reported among 80 participants at risk, and zero deaths are listed. The registry reports no other adverse events in either group.

Pancreatitis is listed as the top serious adverse event term, at 1.4%. The record does not clarify in the posted serious-event summary whether this term accounts for the single control-group event, nor does it provide further event descriptions or assessments of relatedness. The listed death is a separate safety outcome in the record; its cause and relationship to study treatment are not reported in the registry record.

These figures describe the events recorded in the registry, not a complete estimate of comparative safety. The groups have different numbers at risk, and the event counts are sparse. A zero count in one group does not establish that the risk is absent. The registry record does not report a statistical comparison of serious adverse events, event timing, severity details beyond the serious-event classification, or investigator judgments about causality. Safety interpretation should therefore remain descriptive and limited to the posted counts and term.

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

The primary result means that the registry’s analysis did not detect a statistically discernible difference between the treatment and control groups in change in probing depth over the reported follow-up. The estimated difference in change is close to zero, and the confidence interval includes both negative and positive values. The p-value of 0.957 is consistent with a result that offers little evidence against the null hypothesis of no between-group difference under the stated analysis.

It does not mean that the groups have been shown to be identical. A nonsignificant result can arise when the true difference is small, when uncertainty is substantial, or when the available analysis has limited information. The confidence interval communicates the range of effects compatible with the analysis, but the registry does not report a clinical-importance threshold that would determine whether any part of that range matters clinically. An equivalence conclusion would require a specific margin and an appropriate analysis, neither of which is reported in the registry record.

The result also does not establish whether either group improved or worsened in a clinically meaningful sense. The primary comparison is a between-group difference in change, and the registry does not provide a threshold for interpreting the direction or magnitude as clinically important. The posted group estimates should be understood as model-based values on the reported probing-depth scale, not as a broad judgment about oral or systemic health.

The registry’s broader study purpose includes examining whether oral-health effects are associated with changes in systemic-health biomarkers. The primary result presented concerns probing depth; it does not by itself answer that broader question. Secondary outcomes include several systemic and behavioral measures, but a full interpretation of those outcomes requires attention to their analysis sizes, uncertainty, missing data, and multiplicity.

Finally, the primary result is tied to the registry’s stated analysis: GEE, intent-to-treat group assignment, longitudinal data from T1 through T4, and imputation based on linear mixed-model predictions. Details needed to fully evaluate those modeling and missing-data assumptions are not reported in the registry record. The result is informative as the posted analysis, but it should not be extended to outcomes, populations, or time points that the record does not establish.

14. Why This Trial Matters Statistically

This trial illustrates why the estimated contrast, its uncertainty, and the analysis plan all matter more than a binary reading of statistical significance. The primary point estimate is close to zero, but the confidence interval extends in both directions. A reader needs both pieces: the point estimate describes the center of the result, while the interval indicates that the analysis does not pin down the between-group difference precisely enough to support a definitive direction or an equivalence claim.

It also illustrates the distinction between a primary endpoint and a collection of secondary outcomes. The primary endpoint concerns longitudinal change in probing depth, while secondary measures cover different oral, behavioral, and systemic domains. Those outcomes can add context, but multiple testing and varying analysis sizes make it important to know which results were prespecified as confirmatory and how multiplicity was handled. That information is not reported in the registry record.

The trial record also highlights the importance of a clear participant flow and consistent enrollment accounting. It lists enrollment of 786, while the study summary describes an anticipated enrollment of approximately 200. The primary outcome groups contain 73 and 80 participants. The registry does not explain how these figures relate to one another. Readers cannot determine from the record alone why the analysis populations are smaller than the stated enrollment or how missingness and eligibility affected the posted outcomes.

Finally, methods labels are not a substitute for implementation details. GEE and model-based imputation are relevant approaches for longitudinal data, but their interpretation depends on model specification, missing-data assumptions, and sensitivity analyses. Clear reporting of those details, alongside clinically meaningful thresholds and multiplicity plans, would make it easier to judge how strongly the results answer the study’s intended questions.

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 registry record leaves several trial-specific questions unresolved. The enrollment figure is 786, but the brief summary describes an anticipated enrollment of approximately 200, with approximately 100 in each group. The primary outcome groups list 73 and 80 participants. The record does not reconcile these counts or provide a participant flow account explaining the numbers who were randomized, followed, or included in each analysis.

The descriptions of the interventions are incomplete. The record identifies periodontal treatment and subgingival chlorhexidine irrigation, and describes an initial control-group procedure, but portions of the arm descriptions are truncated. The details of treatment delivery, adherence, and any differences in the care received beyond the stated components are not fully reported in the registry record. Although the trial is listed as double-masked, the masked parties and procedures used to maintain masking are not reported.

The primary analysis description does not report the GEE correlation structure, all model terms, or the variance estimation details. The record says the data include imputation using predictions from linear mixed models, but does not report the amount or pattern of missing measurements, the imputation model variables, or sensitivity analyses. These omissions limit assessment of the robustness of the primary estimate to modeling and missing-data assumptions.

The registry does not report a clinical-importance threshold for probing depth, an equivalence margin, or a trial-specific sample-size justification. It also does not report a multiplicity plan for the 13 posted secondary outcomes. The recalculated secondary analyses are unadjusted comparisons from summary statistics and do not reproduce the primary longitudinal analysis. These are important boundaries on the conclusions; they do not establish that the analysis is invalid, but they limit what can be concluded from the registry record alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Baseline Oral Health Study: UnCoVer the Connections to General Health — NCT04954313.
• Registry record last updated 2026-02-12; status last verified 2025-05.
• Results first posted 2026-02-12.
• 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.

19. Record Summary

The registry describes NCT04954313, a completed Phase 4 randomized, double-masked, parallel-group trial sponsored by the University of North Carolina, Chapel Hill. It compares a control periodontal-treatment group with a treatment group that also received subgingival chlorhexidine irrigation. The registry lists enrollment of 786, while the brief summary describes an anticipated enrollment of approximately 200; the difference is not explained.

The primary outcome is change in oral-health measures based on whole-mouth mean probing depth, comparing T1 (day 90) and T4 (day 360), with longitudinal data from T1 through T4 analyzed using GEE. The reported difference in change is -0.004 millimeters, with a 95% confidence interval from -0.157 to 0.149 and p = 0.957. The analysis description specifies intent-to-treat group assignment and imputation using predictions based on linear mixed models.

The registry lists 13 secondary outcomes, with detailed summaries for three in the available results. Serious adverse event and death counts are reported separately by group. Important details about enrollment accounting, model implementation, missingness, multiplicity, masking, and safety-event attribution are not reported in the registry record.