Zambia Chlorhexidine Application Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Zambia Chlorhexidine Application Trial (NCT01241318), a randomized phase 2, phase 3 study. The primary outcome is All-cause Neonatal Mortality.

2. Clinical Question

3. Trial Design

The Zambia Chlorhexidine Application Trial was a randomized, parallel-group study with no masking. Its registry phase is listed as phase 2 and phase 3. The registry describes the study as cluster-randomized: health facility catchment areas, rather than individual mothers, were assigned to chlorhexidine cord care or dry cord care. This distinction is important because participants within the same catchment area may share features that make their outcomes more alike than outcomes from different areas.

The chlorhexidine group was assigned daily application of chlorhexidine gluconate (4%) until three days after the cord completely separated. The comparator followed dry cord care, keeping the stump clean and dry as routine standard care under Zambia Ministry of Health policy. The registry lists 77535 participants as the enrollment. The study sponsor is Boston University. The recorded start was 2011-02, primary completion was 2013-09, and status is completed.

The primary outcome was all-cause neonatal mortality, assessed using vital status at 28 days post-partum. Listed secondary outcomes include incidence of omphalitis, place of delivery, factors influencing delivery location, and health facility characteristics. Although the registry lists six outcome measures and three analyses, the recorded outcome information does not supply a complete account of the analysis methods or of how each outcome was handled. The analytic method and further details of the cluster-level allocation procedure are 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: Mothers were enrolled in the study while pregnant and infants were enrolled after delivery; thus the total protocol enrollment is women + infants. Study reporting is completed at the woman level and thus the total enrolled is listed at the woman level in the outcomes tables.

Overall Study

Reasons for not completing this period:

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: Pregnant women attending health facilities for antenatal care

6. Primary Endpoint

The registry lists the primary outcome as All-cause Neonatal Mortality, measured over 28 days post-partum.

The primary endpoint is all-cause neonatal mortality, defined in the registry as vital status at 28 days post-partum. It counts whether a participant experienced the event during that stated period; the registry describes the endpoint parameter as a number of participants. “All-cause” means the endpoint is not restricted to a particular recorded cause of death. The record does not provide a cause-specific breakdown for this primary outcome.

A mortality endpoint gives a direct count-based summary of the event under study, while a relative measure such as a risk ratio compares event risk between groups. The endpoint's fixed assessment time also gives the comparison a defined follow-up horizon. These features make the outcome interpretable as a group-level measure of mortality over the specified period. The registry does not state additional details about how vital status was ascertained, how deaths with uncertain timing were treated, or how participants without known status were handled.

The reported estimate is a risk ratio, but the registry entry does not identify the ordering of the two analytic groups in that estimate. Therefore, the numerical ratio alone cannot be attributed to chlorhexidine versus dry cord care. The outcome definition and comparison are clear at a broad level; further operational and analytic details are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

A risk ratio compares the risk of an event in one group with the risk in another. A value of 1 corresponds to equal risks; values on either side describe a relative difference, with the direction depending on which group is in the numerator. The registry reports a risk ratio of 1.12 for all-cause neonatal mortality, but does not identify the arm ordering in the analysis entry. The ratio should therefore be read as a reported relative estimate without assigning its direction to chlorhexidine or dry cord care.

What does the confidence interval convey?

The reported 95% confidence interval runs from 0.88 to 1.44. In general, a confidence interval gives a range of estimates compatible with the observed data under the model and its assumptions, and communicates precision more fully than a point estimate alone. Here the interval crosses 1, so it includes the equal-risk reference and values on either side. That does not establish equality between the groups. It indicates that the registry estimate does not pin down the relative comparison narrowly enough to exclude both lower and higher values under the stated interval. The registry provides no p-value for this analysis.

A confidence interval is not a probability statement that the true value has a 95% chance of falling inside these particular limits. Its usual interpretation concerns the long-run performance of the interval procedure under repeated sampling and the model assumptions. Those assumptions matter, including how the cluster-randomized design was represented in the analysis. The model, variance estimator, and other technical details are not reported in the registry record.

Why does clustering matter?

In a cluster-randomized trial, assignment takes place at the level of groups such as health facility catchment areas. Outcomes among participants in the same group may be correlated because they share care settings, local practices, or other conditions. General statistical reasoning implies that an analysis should account for this dependence. If it instead treats every participant as an unrelated observation, it may overstate the amount of independent information and produce misleading uncertainty estimates. The registry explicitly cautions that a simple individual-level re-analysis would be misleading and states that only the sponsor's own analysis is shown. The exact adjustment used by the sponsor is not reported in the registry record.

Cluster assignment also affects how readers should think about the effective information in a trial. The enrollment count alone does not reveal how many independent randomized units there were or how much within-cluster correlation was present. The registry record does not report the number of clusters or a measure of within-cluster correlation. No such quantities can be inferred from the participant counts.

What information is available about the analysis?

The registry identifies the primary measure as a risk ratio and reports its point estimate and confidence interval. It does not report the analytic model, covariate adjustment, method for handling cluster dependence, or a detailed definition of the analysis population. It also does not identify the analytic group ordering in the reported estimate. These omissions limit the ability to reproduce the analysis or assess whether alternative modeling choices would materially change its precision.

The record lists six outcome measures and three analyses, but the information available for the primary result does not describe a complete analysis plan. The existence of multiple outcomes and analyses makes it relevant to ask how the primary outcome was distinguished from secondary or additional analyses, and whether any adjustment or hierarchy was prespecified. Such details are not reported in the registry record. Without them, the reported confidence interval should be interpreted as the interval attached to the sponsor analysis, not as a complete account of all inferential decisions.

How should planning calculations be understood?

Any sample-size values displayed with the results are planning illustrations based on the observed effect, rather than the trial's original design justification. The stated illustration assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. For time-to-event outcomes, the stated convention is a total event count; for other outcomes, it is participants per arm. These calculations are not observed power and do not show the probability that the completed trial detected an effect. A planning illustration answers a hypothetical question about the sample size under its assumptions, not a question about the study's actual design rationale. The trial's own sample-size justification is not reported in the registry record.

10. Analyses Reported by the Sponsor

The registry presents the sponsor's own analysis as the relevant result because the trial was cluster-randomized and a simple individual-level re-analysis would be misleading. The reported estimate for all-cause neonatal mortality is a risk ratio of 1.12, with a 95% confidence interval from 0.88 to 1.44. No p-value is recorded for this analysis. The group ordering behind the estimate is not identified in the entry, so the relative direction cannot be assigned to a named arm from this information.

No recalculated result is presented for this primary outcome. In general, recalculation from posted summary statistics would not reproduce a cluster-adjusted analysis: a calculation that ignores the randomized clusters would use a different statistical basis and could misrepresent uncertainty. The sponsor's result should therefore remain the focus rather than an unadjusted individual-level reconstruction. The precise method used in the sponsor's analysis, including its handling of clustering, is not reported in the registry record, which limits independent assessment of the model and its assumptions.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists six outcome measures and three analyses. The primary outcome is all-cause neonatal mortality at 28 days post-partum. Listed secondary outcomes include incidence of omphalitis, place of delivery, factors influencing delivery location, and health facility characteristics. The record does not explain how the primary outcome was prioritized among the analyses, whether a hierarchy of outcomes was used, or whether statistical adjustments were made for multiple comparisons. Those details are not reported in the registry record.

Multiplicity matters because examining several outcomes or analyses can increase the chance that at least one result appears notable by chance, depending on the procedures used. This general concern does not establish that a reported result is false; it makes prespecified outcome selection and transparent reporting important for interpreting a collection of findings. The registry information does not describe interim analyses, stopping rules, or a multiplicity strategy.

The record also does not describe missing outcome data, loss to follow-up, or the method used to handle participants whose vital status was unknown. Details about exclusions, adherence to assigned care, and the precise analysis population are not reported in the registry record. Although the enrollment is listed as 77535, the primary outcome entry does not give the number analyzed in each arm or explain the relationship between enrollment and the group-specific at-risk counts reported for adverse events.

These omissions affect assessment of both bias and precision. Without a description of missingness and analysis populations, readers cannot determine from the registry entry whether all enrolled participants contributed to the primary estimate or how incomplete outcome information was handled. The available result remains interpretable as the registry's reported sponsor analysis, but the supporting analysis details are incomplete.

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 non-serious adverse event terms

The adverse-event entries report zero serious adverse events among 18510 participants at risk in the chlorhexidine cord care group and zero among 19346 at risk in the dry cord care group. The entries also report 9 other adverse events among 18510 at risk in the chlorhexidine group and 0 among 19346 at risk in the dry cord care group. These are recorded counts, not evidence on their own about causation or comparative risk.

The same entries list 282 deaths for chlorhexidine cord care and 263 for dry cord care. The registry's primary outcome is all-cause neonatal mortality at 28 days post-partum, but the adverse-event entries do not provide enough detail to establish that these death counts use the same outcome definition, follow-up window, or analysis population. They should not be substituted for the primary outcome estimate. The registry does not report further detail about the classification, severity criteria, timing, attribution, or collection process for the adverse events. Zero recorded serious events also does not establish that no serious events occurred outside the captured reporting or observation process.

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

The result describes a sponsor-reported comparison for all-cause neonatal mortality at 28 days post-partum in a cluster-randomized trial of chlorhexidine cord care and dry cord care. The registry gives a risk ratio of 1.12 and a 95% confidence interval from 0.88 to 1.44. Because the interval spans the equal-risk reference, it does not precisely resolve the relative comparison. The group ordering is not reported in the analysis entry, so the estimate cannot be assigned a direction for either named arm from this entry alone.

This result does not establish that the two assigned care approaches have identical outcomes. A confidence interval that includes 1 can reflect uncertainty, and its range contains values on either side of that reference. Nor does the point estimate alone establish a benefit or harm for a named approach. The analysis method and group ordering would be needed to interpret direction, while the uncertainty interval and model details are needed to assess precision.

The registry result also does not answer every question about the trial. It does not provide detailed information on the analysis population, missing outcome data, cluster-level modeling, adjustment for multiple analyses, or interim analyses. The registry lists secondary outcomes, but the primary result should not be taken as a summary of all those outcomes. The adverse-event counts are reported separately and do not establish that an event was caused by an assigned care approach.

These statements concern the posted statistical record, not an individual care decision. The numerical estimate is a summary of the comparison as recorded, with important analytic details unavailable. Any stronger conclusion about the relative effect, the reasons for the observed estimate, or its application beyond the trial population would go beyond what the registry record reports.

14. Why This Trial Matters Statistically

This trial illustrates why the unit of randomization matters when interpreting clinical research. Although the outcome is measured for individual infants, assignment occurred by health facility catchment area. Outcomes among people in one cluster may be related, so a statistical analysis needs to reflect the design. The registry explicitly warns that a simple individual-level re-analysis would be misleading. This is a useful distinction between the number of enrolled participants and the independent units contributing to a randomized comparison.

The reported risk ratio is accompanied by a confidence interval rather than a point estimate alone. The interval from 0.88 to 1.44 conveys substantial uncertainty around the estimate and crosses the equal-risk reference. This makes clear why statistical interpretation should consider both magnitude and precision, not just the central value. The missing group ordering also shows that even a familiar effect measure cannot be fully interpreted without knowing how the groups were coded.

The record further demonstrates the importance of transparent analysis reporting. It identifies the primary outcome, trial design, and sponsor estimate, but does not describe the model, cluster adjustment, missing-data handling, analysis populations, or multiplicity strategy. Those details help readers understand how the estimate was obtained and what uncertainty it represents. Their absence does not erase the reported result, but it limits independent verification and the scope of conclusions that can be drawn from the registry entry.

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 gives limited detail about the statistical analysis of this cluster-randomized trial. It does not report the number of randomized catchment areas, the cluster-level analysis method, the within-cluster correlation, or the variance estimation approach. These details are important for evaluating how the sponsor accounted for dependence among participants assigned within the same area. The record itself cautions against a simple individual-level re-analysis, and no such re-analysis is used to replace the sponsor result.

The reported risk ratio is not linked in the analysis entry to the named intervention and comparator through an explicit numerator-denominator order. This prevents a reliable statement about which arm had the higher estimated risk. The registry also does not report a p-value, adjusted estimates, or model covariates for the primary comparison.

There is no reported account of the primary analysis population, missing vital-status data, exclusions, or follow-up completeness. The listed enrollment is 77535, while separate adverse-event entries give smaller at-risk counts; the record does not explain how these counts relate to the primary analysis. The adverse-event death counts likewise lack detail connecting them to the primary endpoint's definition and follow-up period.

The registry lists multiple outcome measures and analyses but does not describe the prespecified analysis hierarchy, multiplicity handling, or interim monitoring. Further details about intervention delivery, adherence, and fidelity are also not reported in the registry record. These limitations constrain interpretation of the estimate and its precision; they do not, by themselves, establish the presence or direction of bias.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Zambia Chlorhexidine Application Trial — NCT01241318.
• Registry record last updated 2020-08-31; status last verified 2020-07.
• Results first posted 2019-05-03.
• 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 Zambia Chlorhexidine Application Trial was a completed, randomized, parallel-group, unmasked cluster trial sponsored by Boston University. It enrolled 77535 participants and compared daily chlorhexidine gluconate (4%) cord care with routine dry cord care in health facility catchment areas in Zambia. The recorded start was 2011-02, and primary completion was 2013-09.

The primary outcome was all-cause neonatal mortality based on vital status at 28 days post-partum. The registry reports a risk ratio of 1.12 with a 95% confidence interval from 0.88 to 1.44. It does not identify the analytic group ordering or report a p-value. Because assignment was by cluster, the registry cautions that a simple individual-level re-analysis would be misleading and presents the sponsor's analysis. Details of the analysis model, missing data, analysis populations, and multiplicity procedures are not reported in the registry record.