This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Active Bathing to Eliminate Infection (ABATE Infection) Trial (NCT02063867), a randomized na study. The primary outcome is MRSA and VRE Clinical Cultures.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02063867 |
| Phase | NA |
| Condition | Healthcare Associated Infections, Methicillin Resistant Staphylococcus Aureus, Multi Drug Resistant Organisms |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 53 |
| Arms in the primary outcome | n/a |
| Lead sponsor | University of California, Irvine |
| Results posted | Yes |
2. Clinical Question
Population
The trial concerned patients in non-critical care units at participating hospitals. The randomized subjects were hospitals, not individual patients: 53 hospitals, representing approximately 190 non-critical care units, were enrolled. The primary outcome concerned clinical cultures attributable to participating units during the specified unit-stay window.
Intervention
The decolonization strategy used daily chlorhexidine showering or a chlorhexidine cloth bath for all non-critical care patients. Patients positive for methicillin-resistant Staphylococcus aureus by history, culture, or screening were also assigned topical intranasal mupirocin in both nares twice daily for five days.
Comparator
The comparator was usual care: the hospital’s routine policy for showering or bathing non-critical care patients. The record does not describe a single standardized bathing product or schedule for usual care; it characterizes the arm as routine policy.
Primary question
The primary question was whether the hospital-level decolonization strategy was associated with a change in MRSA and vancomycin-resistant enterococci clinical cultures, compared with usual care, over the 21-month outcome timeframe. The registered outcome was expressed as a hazard ratio for intervention versus baseline. The posted analysis reports p = 0.17, but no hazard ratio or confidence interval is reported in the registry record, so the size and precision of the estimated effect cannot be described from that record.
3. Trial Design
ABATE Infection was a randomized, parallel, cluster trial of hospitals. Randomization at the hospital level means that the assigned strategy applied within hospitals rather than being allocated independently to each patient. The registry record lists no masking. It describes 53 hospitals as the enrolled subjects, representing approximately 190 non-critical care units. The distinction between hospital-level allocation and patient-level outcomes is important: patients treated in the same hospital share a care setting, so their outcomes may be correlated.
The trial compared hospitals following usual bathing policy with hospitals implementing a decolonization strategy. That strategy included daily chlorhexidine bathing for all non-critical care patients and intranasal mupirocin twice daily for five days for patients identified as MRSA-positive by history, culture, or screening. The record describes the intervention as a quality-improvement strategy rather than a comparison of isolated medications.
The primary outcome was MRSA and VRE clinical cultures attributable to participating units, with a 21-month timeframe. The outcome definition included cultures occurring >2 days into a participating unit stay through 2 days following unit discharge. The primary population description identifies an unadjusted intention-to-treat analysis and says the analysis compared baseline and intervention patients using difference in differences. The precise implementation details of that approach are not reported in the registry record.
The registry record lists a mixed models analysis and p = 0.17 for the primary comparison. It does not report the model specification, the estimated hazard ratio, a confidence interval, or the number of primary events in each group. The trial is marked completed, with a start in 2014-04 and primary completion in 2019-02. Those dates describe the registered trial timeline; they do not fill in the missing statistical details.
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.
Overall Study
| Stage | Arm 1: Usual Care | Arm 2: Decolonization |
|---|---|---|
| STARTED | 156889 | 183013 |
| COMPLETED | 152598 | 180048 |
| NOT COMPLETED | 4291 | 2965 |
Analysis population for the primary outcome, as defined in the registry: Unadjusted intention-to-treat analysis. Number of patients in the intervention period provided. Analysis involves comparison of baseline and intervention patients (difference in differences)
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: These numbers reflect the number of participants in the baseline period across all participating facilities enrolled.
| Characteristic | Unit | Arm 1: Usual Care | Arm 2: Decolonization | Total |
|---|---|---|---|---|
| Participants analysed | 156889 | 183013 | 339902 | |
| Age, Continuous | years | 62.3 (SD 18.2) | 62.6 (SD 18.6) | 62.4 (SD 18.4) |
| Sex: Female, Male: Female | Participants | 84585 | 100249 | 184834 |
| Sex: Female, Male: Male | Participants | 72288 | 82748 | 155036 |
| Race/Ethnicity, Customized: White | Participants | 107261 | 112067 | 219328 |
| Race/Ethnicity, Customized: Hispanic | Participants | 22581 | 38704 | 61285 |
| Race/Ethnicity, Customized: Black | Participants | 21501 | 22462 | 43963 |
| Race/Ethnicity, Customized: Asian | Participants | 2021 | 4175 | 6196 |
| Race/Ethnicity, Customized: Other/Unknown | Participants | 3525 | 5605 | 9130 |
| Region of Enrollment: United States | participants | 156889 | 183013 | 339902 |
6. Primary Endpoint
The registry lists the primary outcome as MRSA and VRE Clinical Cultures, measured over 21 months.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | 95% Confidence Interval |
| Unit | Hazard Ratio (Intervention vs Baseline) |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint combines clinical cultures for methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococci that were attributed to participating units. The registry defines the qualifying time window as cultures occurring >2 days into a unit stay through 2 days after discharge from that unit. The registered timeframe is 21 months.
A clinical-culture endpoint focuses on organisms identified through clinical testing rather than treating every positive screening result as an event. In general, a defined time window helps distinguish outcome events associated with a unit stay from events detected outside that period. The record does not give further detail on culture collection practices, adjudication, or how multiple cultures from the same patient were handled.
The endpoint is expressed in the registry as a hazard ratio comparing intervention with baseline. In general, a hazard ratio summarizes the relative rate at which events accrue over time under the modeled comparison; its interpretation depends on the model and the comparison being made. Here, the posted result does not include the hazard-ratio estimate or its confidence interval. Therefore the direction and magnitude of the estimated endpoint effect, and the precision around that estimate, are not reported in the registry record.
The primary endpoint should be read as one defined measure of culture outcomes, not as a complete measure of all infections, all organism carriage, or every consequence of the bathing strategy. Other registered outcomes include Gram-negative multidrug-resistant organism clinical cultures and all-cause bloodstream infections.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The posted primary analysis is a mixed models analysis with p = 0.17. This p-value describes the compatibility of the observed data with the model’s null hypothesis, under the assumptions of that analysis. It is not an estimate of the size of any effect and does not show how precisely the effect was measured. The registry record does not report the corresponding hazard ratio or its 95% confidence interval.
Without the point estimate, the direction and magnitude of the primary comparison cannot be stated from the posted result. Without the confidence interval, readers cannot assess the range of effect sizes that would be consistent with the analysis. The p-value alone cannot establish whether any difference was small or large, or whether the data rule out effects that might matter. It also does not establish that the two strategies were equivalent.
The design randomized hospitals, while the measured outcomes arose among patients and units. A mixed model can account for dependence in clustered data when specified appropriately. The registry record identifies the analysis as mixed models, but it does not describe the model structure, adjustment variables, covariance assumptions, or handling of the baseline and intervention periods. Those details are not reported in the registry record.
The registered primary-population description refers to an unadjusted intention-to-treat analysis and a comparison of baseline and intervention patients using difference in differences. In general, difference in differences compares the change over time in one group with the change over time in another group. Its interpretation depends on the design and assumptions, including how time trends are represented. The registry record does not state the detailed assumptions or diagnostics used for this analysis.
The reported p-value is the available inferential result, not a complete account of the outcome. The record gives no event counts for the primary endpoint, no hazard-ratio estimate, and no confidence interval. It would therefore be inappropriate to infer an absolute change in cultures or quantify the uncertainty in that change from p = 0.17 alone.
Nor can this p-value be interpreted as the probability that the intervention works, the probability that the null hypothesis is true, or the chance that the result occurred randomly. A p-value is conditional on a statistical model and a null hypothesis. It measures how unusual the data or more extreme data would be under those conditions; it does not measure clinical importance or prove a causal explanation.
The sponsor’s own analysis is the only primary analysis reported in the registry record. Because the trial randomized clusters, a simple individual-level re-analysis would not appropriately reflect the design. No independent recalculated effect estimate is available to compare with the sponsor’s analysis. The posted result should therefore be described narrowly: the mixed-model analysis reported p = 0.17, while the estimated effect and its precision are not reported in the registry record.
9. Statistical Methods Explained
Why does the cluster design matter?
In an individually randomized trial, each participant is assigned independently to a study group. In a cluster-randomized trial, groups such as hospitals are randomized, and multiple patients receive care within each assigned cluster. Patients within one hospital may share staff, routines, policies, and local conditions. Their outcomes can therefore be more alike than outcomes from patients in different hospitals.
This dependence is called within-cluster correlation. General statistical consequence: if it is ignored, the analysis may treat correlated observations as though they contributed more independent information than they actually do. Standard errors and confidence intervals may then be misleading. A mixed model can represent clustering by including model components for groups or other sources of shared variation. The registry record names mixed models but does not report the model specification or the estimated degree of clustering.
Because hospitals were randomized, an analysis that treats every patient as independently randomized would not match the allocation design. The record specifically notes that a simple individual-level re-analysis would be misleading and lists only the sponsor’s own analysis. That design consideration limits the scope for a straightforward recalculation from posted summary statistics.
What does the mixed-model result say?
The primary result is a mixed models analysis with p = 0.17. A p-value is calculated under a null hypothesis and a statistical model. In general, it measures how compatible the observed data are with that null model, relative to a specified test. It does not state the probability that the null hypothesis is true, and it does not report the size of the estimated effect.
Effect estimates and confidence intervals answer different questions. In general, an effect estimate gives the direction and scale of the modeled association, while a confidence interval conveys the uncertainty around that estimate under the model. For the primary endpoint, the registry record does not report the hazard ratio or its 95% confidence interval. Consequently, p = 0.17 cannot be used to reconstruct those missing quantities or to judge the range of effects supported by the data.
The registry names the primary measure as a hazard ratio for intervention versus baseline. In general, time-to-event methods account for when events occur as well as whether they occur during follow-up. Hazard-ratio interpretation depends on the time-to-event model and how the groups and periods are defined. The record also describes a baseline and intervention comparison using difference in differences. It does not provide enough model detail to explain exactly how these labels were combined in the reported analysis.
What is difference in differences?
In general, difference in differences compares a change over time in one group with the change over time in another group. It can help separate a change associated with an intervention period from a common time trend, provided the study design and analysis support that interpretation. The registry record says the primary population was analyzed using a baseline and intervention comparison described as difference in differences. It does not report the underlying period-specific estimates, the model equation, or the assumptions and diagnostics used. No further protocol detail should be inferred.
The analysis is described as unadjusted intention-to-treat. In general, intention-to-treat analysis retains units or participants according to their randomized assignment, regardless of later adherence. In a cluster trial, the meaning of the analysis population and the unit used for each component of the model matter. The record does not report details about eligibility, exclusions, missing observations, or how deviations from assigned strategy were handled.
Why does over-dispersion matter?
The endpoint is a count-related clinical-culture outcome, and count outcomes can vary more than a simple statistical model expects. General explanation: over-dispersion occurs when observed variability exceeds the model’s assumed variability. If a model does not account for that extra variation, uncertainty can be understated. Cluster effects, uneven hospital sizes, and differences in event occurrence may also affect the variability of count data. The registry record does not report whether over-dispersion was assessed or what distributional assumptions were used.
These are general considerations, not claims about the exact model used for ABATE Infection. The record identifies mixed models but does not say whether it used a particular count distribution, link function, random-effects structure, or correction for over-dispersion. Those statistical details are not reported in the registry record.
How should planning illustrations be read?
General sample-size illustrations based on an observed effect are planning exercises, not observed power. The stated assumptions for such illustrations are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation, with the effect to detect expressed as a fraction of the observed effect. For time-to-event outcomes, the illustrations refer to total events; for other outcomes, they refer to participants per arm.
Those assumptions do not establish the trial’s original design rationale. The record does not report the protocol’s sample-size justification, target effect, or any interim analysis plan. Any planning illustration based on the observed effect should be kept distinct from the trial’s actual enrollment of 53 hospitals and should never be interpreted as observed power. The registry record also does not report enough information to evaluate whether the assumptions fit the trial’s cluster structure.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| Mixed Models Analysis | not reported | 0.170 |
The sponsor’s own primary analysis is reported as a mixed models analysis with p = 0.17. The registry record does not report a hazard-ratio estimate, its confidence interval, or primary event counts, so there is no posted numerical effect estimate against which to compare a recalculation. A p-value alone is insufficient to recalculate an effect or its precision.
More importantly, the randomized unit was the hospital. Patient outcomes within a hospital are not necessarily independent, and the registry record cautions that a simple individual-level re-analysis would be misleading. Recalculation from aggregate summaries would also require suitable cluster-level information and enough detail about the analysis model. Those details are not reported in the registry record.
Accordingly, the sponsor’s mixed-model result is the only primary inferential result available in the record. There is no independent unadjusted recalculated result presented alongside it. The distinction is methodological: a cluster-aware analysis must respect the hospital-level assignment rather than treating each patient as independently randomized. The available result can be reported as p = 0.17, but it cannot be expanded into an effect estimate or precision statement without further reported information.
11. Multiplicity, Missing Data and Analysis Populations
The registry record lists 13 outcome measures and 7 analyses. It identifies MRSA and VRE clinical cultures as the primary outcome, while Gram-negative multidrug-resistant organism clinical cultures and all-cause bloodstream infections are listed as secondary outcomes. The record does not state whether the analyses were organized into a prespecified testing hierarchy or whether any correction for multiple comparisons was applied.
When several outcomes or analyses are tested, the chance of obtaining at least one small p-value can increase if each is interpreted independently. General statistical approaches can address this issue through a prespecified primary comparison, adjusted testing procedures, or cautious interpretation of secondary and exploratory results. The registry record does not report which, if any, approach was used. The presence of multiple registered measures alone does not show that the primary analysis was improperly conducted; it does mean that the overall analysis plan cannot be assessed from the available record.
Missing-data methods are also not reported in the registry record. There is no description of missing outcome observations, loss to follow-up, incomplete hospital-period data, or sensitivity analyses for missingness. Without those details, readers cannot determine how missing observations were handled or whether the reported analysis depended on assumptions about them.
The record describes the primary population as an unadjusted intention-to-treat analysis and mentions a baseline and intervention comparison using difference in differences. It does not report the exact analysis population, exclusions, protocol deviations, or the number of observations entering each component of the model. The phrase intention-to-treat identifies an analysis principle, but it does not by itself resolve these operational details in a cluster-randomized setting.
Likewise, interim analyses and stopping rules are not reported in the registry record. No sample-size justification or detailed statistical analysis plan is reported. The absence of these details in the record should not be treated as evidence that the procedures were or were not used; it limits what can be evaluated from the posted information.
12. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Arm 1: Usual Care | 0 | 156889 | 0.0% |
| Arm 2: Decolonization | 0 | 183013 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Arm 1: Usual Care | 0 of 156889 |
| Arm 2: Decolonization | 0 of 183013 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Arm 1: Usual Care | 0 of 156889 |
| Arm 2: Decolonization | 25 of 183013 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Arm 1: Usual Care | Arm 2: Decolonization |
|---|---|---|---|
| Mild rash on limbs, trunk | Skin and subcutaneous tissue disorders | 0 of 156889 | 6 of 183013 |
| Diffuse rash | Skin and subcutaneous tissue disorders | 0 of 156889 | 5 of 183013 |
| Mild rash on trunk | Skin and subcutaneous tissue disorders | 0 of 156889 | 4 of 183013 |
| Mild rash on limbs | Skin and subcutaneous tissue disorders | 0 of 156889 | 2 of 183013 |
| Diffuse itching | Skin and subcutaneous tissue disorders | 0 of 156889 | 1 of 183013 |
| Mild rash, location unknown | Skin and subcutaneous tissue disorders | 0 of 156889 | 1 of 183013 |
| Mild rash on groin | Skin and subcutaneous tissue disorders | 0 of 156889 | 1 of 183013 |
| Mild rash on limbs, face | Skin and subcutaneous tissue disorders | 0 of 156889 | 1 of 183013 |
| Mild rash on limbs, groin | Skin and subcutaneous tissue disorders | 0 of 156889 | 1 of 183013 |
| Mild rash on limbs, trunk, face | Skin and subcutaneous tissue disorders | 0 of 156889 | 1 of 183013 |
The registry lists zero serious adverse events in the usual-care group among 156889 at risk, and zero in the decolonization group among 183013 at risk. It also lists zero deaths in both groups. For other adverse events, the listed counts are zero in usual care and 25 in decolonization, with the same respective at-risk counts.
These entries should be read as registry-reported counts, not as a complete safety comparison. The record does not provide event definitions, event-level descriptions, attribution, severity breakdowns beyond the serious-event category, or a statistical comparison. It also does not report whether the listed counts represent unique people or events. The at-risk counts are reported as given and should not be interpreted as randomized hospital counts.
A zero count in the serious-event field means no serious events are listed for that group in this record; it does not prove that no event could occur or that there is no uncertainty about safety. Similarly, the other adverse-event counts do not by themselves establish a causal relationship to the assigned strategy. These figures do not describe treatment effectiveness, and they are not a substitute for the primary clinical-culture analysis.
13. What the Result Does, and Does Not, Mean
The posted primary result establishes that the registry reports a mixed-model p-value of 0.17 for the primary comparison. It does not establish the direction or size of the estimated difference because the hazard ratio is not reported. It also does not establish the precision of that difference because the 95% confidence interval is not reported. No primary event counts are reported in the registry record.
A p-value is not a direct measure of clinical importance. It does not state the probability that the intervention is effective, the probability that the null hypothesis is true, or the probability that the result arose by chance. It is interpreted in relation to a defined null hypothesis and model. Without the effect estimate and its interval, the posted result cannot show which effect sizes are most consistent with the data.
The result does not demonstrate equivalence between usual care and decolonization. Equivalence or non-inferiority requires a suitable design and prespecified margin, and the relevant design details are not reported in the registry record. A p-value that does not meet a conventional threshold, by itself, cannot establish that the strategies have the same outcome.
The result also cannot be generalized beyond the outcome and population described without additional evidence. The primary outcome concerns MRSA and VRE clinical cultures attributed to participating non-critical care units during a defined unit-stay window. It is not a report of every infection, all organism carriage, or every clinical outcome. Other registered endpoints include Gram-negative multidrug-resistant organism clinical cultures and all-cause bloodstream infections, but their results are not detailed in the registry record for this analysis.
The trial’s cluster-randomized design matters to interpretation. Hospitals, rather than individual patients, were randomized, and the sponsor used a mixed-model analysis. The registry does not provide enough detail to independently evaluate the model or reproduce the estimate. The most defensible summary is therefore narrow: p = 0.17 is reported for the primary mixed-model analysis, while the estimated effect, its uncertainty, and key analytical details are not reported in the registry record.
14. Why This Trial Matters Statistically
ABATE Infection is statistically notable because it randomized hospitals while measuring outcomes among patients and units. That structure makes the level of randomization central to analysis. Treating every patient as an independent randomized observation would ignore shared hospital context and could misstate uncertainty. The registry record recognizes this issue by identifying a mixed-model analysis and cautioning against simple individual-level re-analysis.
The trial also illustrates why a p-value is not a sufficient statistical summary. The primary result includes p = 0.17, but no hazard ratio or confidence interval is reported. Without the effect estimate, readers cannot describe its direction or magnitude; without the interval, they cannot assess precision. The result consequently demonstrates a reporting limitation as well as an inferential result.
The registered outcome combines MRSA and VRE clinical cultures and uses a time window tied to the participating unit stay. Clearly defining the outcome and its attribution window is important when interpreting events in hospital settings. It also limits what the endpoint represents: it is a specific culture-based outcome, not a comprehensive measure of every infection-related consequence.
Finally, the record describes a baseline and intervention comparison using difference in differences, alongside a mixed-model result. These methods depend on details of implementation, assumptions, and model specification. Because those details are not reported, the registry result should be interpreted with care rather than extended into claims about the exact causal estimate. The central statistical lesson is to keep design, effect size, uncertainty, and reporting completeness distinct.
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.
Several trial-specific limitations to interpretation arise from incomplete statistical reporting. The primary outcome is named as a hazard ratio for intervention versus baseline, but the estimate and its 95% confidence interval are not reported in the registry record. The primary event counts are also absent. This prevents a direct assessment of effect direction, magnitude, and precision from the posted result.
The analysis is described as mixed models and as an unadjusted intention-to-treat comparison of baseline and intervention patients using difference in differences. The record does not report the model specification, the random-effects structure, covariates, link function, event distribution, or handling of over-dispersion. It also does not give the period-specific data or diagnostics needed to evaluate the stated comparison.
The registry lists multiple outcomes and analyses but does not report a multiplicity plan. Missing-data handling, exclusions, protocol deviations, interim analyses, and stopping rules are not reported in the registry record. These omissions do not prove that such procedures were absent; they constrain independent evaluation of the analysis.
Cluster allocation is an additional practical constraint. The 53 randomized subjects were hospitals, representing approximately 190 non-critical care units, and the record cautions against a simple patient-level re-analysis. Without suitable cluster-level data and model details, independent recalculation is not justified by the reported result alone.
The adverse-event entries are counts without detailed definitions or a statistical comparison. The record does not explain whether counts represent events or people, or report attribution. The registry also does not provide additional detail about usual-care bathing practices, intervention adherence, or the implementation of the strategy at each hospital. These points limit interpretation of both the primary result and the safety entries.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: Active Bathing to Eliminate Infection (ABATE Infection) Trial — NCT02063867.
- Registry record last updated 2019-07-12; status last verified 2019-06.
- Results first posted 2019-07-12.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
19. Record Summary
NCT02063867, the Active Bathing to Eliminate Infection (ABATE Infection) Trial, was a completed, randomized, parallel cluster trial with no masking listed. The University of California, Irvine is identified as sponsor. The registry records 53 hospitals as enrolled subjects, representing approximately 190 non-critical care units.
The comparison was usual bathing policy versus daily chlorhexidine bathing for all non-critical care patients, with intranasal mupirocin twice daily for five days for patients identified as MRSA-positive by history, culture, or screening. The primary outcome was MRSA and VRE clinical cultures attributable to participating units, over a 21-month timeframe.
The posted primary analysis is a mixed models analysis with p = 0.17. The registry record does not report the hazard-ratio estimate, confidence interval, or primary event counts. It also does not report detailed model specifications, multiplicity procedures, or missing-data methods. The registry cautions that simple individual-level re-analysis would be misleading because hospitals were randomized.