This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
TOBRA was a randomized, parallel-group phase 3 trial evaluating whether adding local tobramycin powder to local vancomycin powder affected the occurrence of deep surgical site infection after definitive fracture fixation in patients at high risk of infection.
| Feature | TOBRA |
|---|---|
| Trial name | TOBRA |
| ClinicalTrials.gov identifier | NCT04597008 |
| Phase | Phase 3 |
| Condition | Post Operative Surgical Site Infection |
| Design | Randomized, parallel-group |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 1660 |
| Primary endpoint type | Binary |
| Primary effect measure | Risk ratio / relative risk |
| Hypothesis type | Superiority |
| Lead sponsor | Major Extremity Trauma Research Consortium |
| Status | Completed |
2. Clinical Question
The central statistical question was whether the proportion of patients experiencing a deep surgical site infection within 182 days of definitive fracture fixation surgery differed between patients treated with local vancomycin powder and those treated with a combination of local vancomycin and tobramycin powders.
Population
Patients enrolled in a phase 3 randomized trial for post-operative surgical site infection following definitive fracture fixation, with the primary analysis restricted to participants adjudicated as eligible.
Intervention
Local treatment with a combination of vancomycin and tobramycin powders.
Comparator
Local treatment with vancomycin powder.
Primary question
Does the combination of local vancomycin and tobramycin powders change the proportion of deep SSIs compared with local vancomycin powder alone?
3. Trial Design
Treatment groups
Local Vancomycin
- Local vancomycin powder
- Comparator group
Local Vancomycin + Tobramycin
- Local vancomycin powder
- Local tobramycin powder
- Combination treatment group
The registry describes the intervention categories as a treatment group and a control group, each involving a drug intervention. The primary statistical comparison is explicitly reported as Control vs Treatment.
4. Randomization, Stratification, and Analysis Population
The primary analysis states that randomization was stratified by clinical site and fracture location, with fracture location defined as tibial plateau or tibial pilon. The analysis therefore recognized the randomized structure across these strata rather than treating the study population as a single unstratified sample.
| Population / stage | Participants | Role |
|---|---|---|
| Met eligibility criteria | 2310 | Patients meeting the trial's eligibility criteria between June 18, 2021, and January 31, 2025 |
| Provided informed consent and were randomized | 1660 | Randomized trial population |
| Adjudicated as eligible | 1528 | Included in the primary analysis |
This distinction is statistically important. The enrolled population and the primary analysis population are not identical. The posted analysis explicitly identifies 1528 adjudicated eligible participants as the population used for the primary analysis.
5. Primary Endpoint
| Endpoint | Registry definition | Type | Time frame |
|---|---|---|---|
| Number of Participants With Deep Surgical Site Infection (SSI) | Compare the proportion of deep SSIs of the study injury in patients treated with local Vancomycin powder compared to those treated with a combination of local Vancomycin and Tobramycin powders. A deep SSI is an SSI treated with operative debridement. In current CDC terminology this includes all deep organ space, deep incisional, and superficial infections treated with surgery. | Binary | Within 182 days of definitive fracture fixation surgery |
The endpoint is therefore patient-level and binary: each participant contributes an outcome classification based on whether the qualifying deep SSI occurred within the specified 182-day window.
6. Statistical Methodology
Binary outcome analysis
The primary endpoint is a binary outcome rather than a time-to-event endpoint. The principal reported effect measure is the risk ratio (RR), comparing the probability of the specified deep SSI between the two randomized treatment groups.
For this analysis, the reported comparison is explicitly labeled Control vs Treatment. An RR of 1 represents equal observed risk between the groups. Values above 1 indicate a higher estimated risk in the control group relative to the treatment group under this comparison direction.
Stratified analysis
The primary analysis states that randomization was stratified by clinical site and fracture location. The analysis text identifies stratified analysis as an analysis concept. The specific statistical method used to generate the reported risk ratio is not reported in the registry entry.
Superiority framework
The registered hypothesis type is superiority. In a superiority trial, the statistical question is whether the outcome differs between the randomized groups rather than whether one treatment is merely no worse than another within a prespecified non-inferiority margin.
A confidence interval that contains 1 remains compatible with equal risk as well as with effects in either direction, depending on the width and location of the interval.
Confidence interval
The posted primary analysis reports a two-sided 95% confidence interval for the risk ratio. The interval provides a measure of statistical precision around the point estimate. A wider interval indicates greater uncertainty about the underlying relative risk than a narrower interval would.
What is not specified in the posted analysis
The registry does not report the formal statistical method used to calculate the risk ratio and its confidence interval. It does, however, identify the analysis as stratified and specifies the clinical site and fracture location strata used in randomization.
7. Primary Result
Deep Surgical Site Infection
Risk ratio: Control vs Treatment
95% CI: 0.75–1.66 · Two-sided 95% confidence interval
Primary analysis population: 1528 adjudicated eligible participants
| Endpoint | Analysis population | Comparison | Effect measure | Estimate | 95% CI |
|---|---|---|---|---|---|
| Number of Participants With Deep Surgical Site Infection (SSI) | 1528 adjudicated eligible participants | Control vs Treatment | Risk Ratio | 1.11 | 0.75–1.66 |
The reported RR of 1.11 means that, under the stated Control-versus-Treatment comparison, the estimated risk of deep SSI in the control group was 11% higher than in the treatment group. This is a relative comparison of the two group-level risks, not an 11-percentage-point difference in risk.
The estimate does not mean that 11% of participants developed infection, that the treatment prevented 11% of infections, or that every individual participant experienced an 11% change in risk. A risk ratio summarizes the relative difference between groups.
The 95% CI of 0.75–1.66 indicates substantial uncertainty around the point estimate. Because the interval includes the null value of 1, the reported interval is compatible with equal risk as well as with a lower or higher risk in the control group relative to treatment.
No p-value is reported in the posted statistical analysis. A p-value, when available, would address compatibility with a specified null hypothesis under the statistical model; it would not measure the magnitude or clinical importance of the treatment effect. The risk ratio and its confidence interval provide the effect-size information here.
The analysis is based on the adjudicated eligible population of 1528 participants, not simply the total enrollment of 1660. The stratified randomization by clinical site and fracture location is also relevant when interpreting the reported comparison.
8. Understanding the Primary Effect Estimate
The primary endpoint is binary, so the risk ratio is a natural way to summarize the relative occurrence of deep SSI between the two groups. The direction of the reported ratio depends on which group is placed in the numerator. Here, the registry explicitly reports Control vs Treatment.
RR = 1
The estimated risk would be the same in the control and treatment groups.
RR > 1
The estimated risk is higher in the control group under the reported comparison direction.
RR < 1
The estimated risk is lower in the control group under the reported comparison direction.
RR = 1.11
The point estimate places the control-group risk 11% above the treatment-group risk.
The confidence interval is especially important because the point estimate alone does not communicate how precisely the treatment groups were compared. An RR of 1.11 with a very narrow interval would convey a different degree of statistical precision than the same point estimate accompanied by a wide interval such as 0.75–1.66.
9. Statistical Methods Explained
Why is the primary endpoint binary?
The registry defines the primary endpoint as the number of participants with a deep surgical site infection within 182 days of definitive fracture fixation surgery. Each participant can therefore be classified according to whether the specified event occurred during the prespecified time window. This produces a binary outcome suitable for comparison of proportions or risks between randomized groups.
Why was a risk ratio used?
A risk ratio directly compares the probability of the outcome between two groups. This makes it straightforward to describe the relative occurrence of deep SSI in the control group versus the treatment group. Unlike an odds ratio, the risk ratio is expressed directly in terms of the event probability.
What does an RR of 1.11 mean?
An RR of 1.11 means that the estimated risk in the control group was 1.11 times the estimated risk in the treatment group. Expressed as a relative difference, the point estimate corresponds to an 11% higher estimated risk in the control group. It does not establish the absolute difference in risk because the underlying group-specific event proportions are not reported in the posted analysis.
Why does the confidence interval include 1?
For a risk ratio, 1 is the null value because a ratio of 1 corresponds to equal risk between groups. The reported 95% CI extends from 0.75 to 1.66, so it includes the possibility of no relative difference as well as values representing either direction of association.
Why does stratification matter?
The trial randomized participants with stratification by clinical site and fracture location. Stratified analysis can preserve that structure when comparing outcomes and can account for differences across the prespecified strata. In TOBRA, those factors were clinical site and fracture location, specifically tibial plateau or tibial pilon.
Why is the analysis population important?
The registry states that 1660 participants were randomized but that 1528 participants were adjudicated as eligible and included in the primary analysis. An effect estimate therefore needs to be interpreted in the context of the population from which it was calculated. Using the enrollment number as though it were the primary analysis population would describe a different statistical quantity.
10. Safety Results
The registry reports serious adverse events by randomized arm using affected participants over participants at risk. One serious adverse event was reported among 775 participants in the control group, and one was reported among 753 participants in the treatment group.
| Safety measure | Control | Treatment |
|---|---|---|
| Serious adverse events | 1 / 775 | 1 / 753 |
The reported serious-adverse-event counts are descriptive safety information. They should not be treated as a statistical comparison of the primary infection endpoint. The denominators also differ from the 1528-participant primary analysis population because the safety measure is reported separately by arm using its own at-risk denominators.
11. Trial Timeline
Trial start
The registry identifies June 18, 2021 as the study start date.
Eligibility and randomization period represented in the analysis
A total of 2310 patients met the trial's eligibility criteria during this period. Of these, 1660 provided informed consent and were randomized.
Adjudicated eligible population
A total of 1528 participants were adjudicated as eligible and included in the primary analysis.
Primary completion
The registry identifies July 30, 2025 as the primary completion date.
Registry status
The study is listed as completed.
12. Interpreting the Analysis Population
The relationship among the enrollment, eligibility, and primary analysis counts is an important part of understanding the statistical record.
| Stage | Count | Statistical meaning |
|---|---|---|
| Patients meeting eligibility criteria | 2310 | Patients described as meeting the trial's eligibility criteria during the stated period |
| Patients who consented and were randomized | 1660 | Participants entering the randomized comparison |
| Participants adjudicated eligible for primary analysis | 1528 | Population used for the posted primary endpoint analysis |
These counts should not be conflated. The 2310 figure describes patients meeting eligibility criteria during the specified period; the 1660 figure is the randomized enrollment; and the 1528 figure is the adjudicated eligible population used for the primary analysis.
13. Stratified Analysis: Clinical Site and Fracture Location
TOBRA used stratified randomization by two factors: clinical site and fracture location. Fracture location was categorized as tibial plateau or tibial pilon.
Clinical site
Randomization was stratified by the clinical site at which participants were enrolled.
Fracture location
Randomization was stratified by tibial plateau versus tibial pilon fracture location.
Stratification is a design feature established before outcome comparison. Its purpose is to ensure that important structural factors are represented appropriately across randomized treatment groups and to support an analysis that reflects the way randomization was performed.
The posted analysis identifies stratified analysis as an analysis concept, but does not report the specific mathematical procedure used to obtain the risk ratio and confidence interval. That distinction matters because several statistical procedures can be used for stratified binary outcomes, and the precise method determines details such as weighting and variance estimation.
14. Superiority Testing and the Null Hypothesis
The trial's registered hypothesis type is superiority. For the binary primary endpoint, the natural null comparison for a risk ratio is equality of risk, represented by RR = 1.
The two-sided 95% confidence interval reported for the primary analysis extends from 0.75 to 1.66. Because the interval includes 1, the reported interval includes the null value.
The registry does not report a p-value for the primary endpoint. Consequently, the confidence interval is the principal inferential information available in the posted statistical analysis. The interval should be interpreted together with the point estimate rather than using the point estimate alone.
15. What the Risk Ratio Does — and Does Not — Mean
The reported risk ratio of 1.11 indicates an estimated control-group risk 11% higher than the treatment-group risk under the stated comparison direction. This is a relative measure, not an absolute risk difference.
The risk ratio cannot by itself tell a reader how many additional infections occurred per 100 treated participants. That requires the underlying event proportions in each group, which are not included in the posted statistical analysis.
The 95% CI of 0.75–1.66 spans a substantial range of possible relative effects. The lower portion of the interval is compatible with lower control risk relative to treatment, while the upper portion is compatible with higher control risk relative to treatment.
No p-value is reported in the posted primary analysis. When a p-value is available, it addresses evidence against a null hypothesis under a specified statistical model; it does not quantify the size, clinical relevance, or precision of an effect.
16. Limitations
- Analysis population: 1660 participants were randomized, while 1528 adjudicated eligible participants were included in the primary analysis. The primary result therefore applies specifically to the reported analysis population.
- Unreported event counts: the posted primary statistical analysis provides the risk ratio and confidence interval but does not provide the underlying deep-SSI event counts or group-specific risks needed to calculate an absolute risk difference from the registry entry.
- Unreported formal method: the analysis identifies a stratified analysis but does not specify the statistical procedure used to calculate the risk ratio and its confidence interval.
- Confidence-interval width: the 95% CI of 0.75–1.66 indicates appreciable uncertainty around the point estimate.
- Stratification: the randomized design incorporated clinical site and fracture location, so interpretation should recognize those prespecified strata.
- Safety denominators: the serious-adverse-event denominators of 775 and 753 differ from the 1528 participants in the primary efficacy analysis and should not be substituted for one another.
- Endpoint timing: the primary endpoint is defined within 182 days of definitive fracture fixation surgery. Outcomes outside that window are not represented by this endpoint.
17. Why This Trial Matters Statistically
TOBRA is a useful teaching example because it illustrates several core principles of randomized-trial analysis without relying on a time-to-event endpoint. The primary outcome is binary, the effect measure is a risk ratio, the randomization incorporates stratification, and the posted result provides a confidence interval that can be interpreted directly against the null value of 1.
| Concept | How it appears in TOBRA |
|---|---|
| Randomization | 1660 participants were randomized to two parallel treatment groups. |
| Binary endpoint | Deep SSI within 182 days of definitive fracture fixation surgery. |
| Risk ratio | Primary effect measure reported as Control vs Treatment. |
| Confidence interval | Two-sided 95% CI of 0.75–1.66 around the RR estimate. |
| Superiority | The registered hypothesis type is superiority. |
| Stratification | Randomization was stratified by clinical site and fracture location. |
| Analysis population | 1528 adjudicated eligible participants were included in the primary analysis. |
| Safety analysis | Serious adverse events were reported by arm using affected/at-risk counts. |
The trial also illustrates why a point estimate should rarely be interpreted in isolation. The RR of 1.11 is the center of an interval extending from 0.75 to 1.66. The interval supplies information about precision and about the range of relative effects compatible with the statistical analysis.
18. Related Tutorials
Learn more about the methods used in this trial:
19. Related Statistical Calculators
20. Sources
- ClinicalTrials.gov: TOBRA, NCT04597008.
- PubMed: PubMed-indexed publication, PMID 40667718.
Continue through Clinical Biostats
Explore additional clinical-trial results, statistical tutorials, and analysis tools for understanding randomized evidence.
21. Record Summary
TOBRA is a randomized phase 3 parallel-group trial addressing a binary clinical endpoint: deep surgical site infection within 182 days of definitive fracture fixation surgery. The trial randomized 1660 participants, with 1528 adjudicated eligible participants included in the primary analysis. Randomization was stratified by clinical site and fracture location, and the primary effect measure was a risk ratio comparing control with treatment.
The posted primary analysis reported an RR of 1.11 with a two-sided 95% CI of 0.75–1.66. The point estimate corresponds to an estimated 11% higher risk in the control group relative to the treatment group, while the confidence interval indicates substantial uncertainty and includes the null value of 1. The registry does not report a p-value or the formal statistical procedure used to calculate the risk ratio and confidence interval.
Statistically, the key lessons are the distinction between randomized enrollment and the analysis population, the interpretation of a binary endpoint through a risk ratio, the role of stratification in randomized trials, and the importance of considering confidence-interval precision rather than relying on a point estimate alone. The reported serious-adverse-event counts provide a separate descriptive view of safety, with 1/775 affected in the control group and 1/753 in the treatment group.