Skip to main content

Table 1 Summary of the statistical analysis plan

From: Prophylactic Antibiotic Regimens In Tumor Surgery (PARITY): a multi-center randomized controlled study comparing alternative antibiotic regimens in patients undergoing tumor resections with endoprosthetic replacements—a statistical analysis plan

Objective

Outcome

Hypothesis

Method of analysis

Name

Variable type

Primary objective

 To compare the surgical site infection rates at 1 year

Surgical site infection

Binary

A longer duration (5 days) of post-operative antibiotics will reduce the risk of surgical site infection compared to a shorter duration (1 day) of post-operative antibiotics.

Cox proportional hazards

Secondary objectives

 To compare the risk of antibiotic-related complications at 1 year

Antibiotic-related complications

Binary

A longer duration (5 days) of post-operative antibiotics will result in more antibiotic-related complications compared to a shorter duration (1 day) of post-operative antibiotics.

Cox proportional hazards

 To compare the risk of unplanned re-operations at 1 year

Unplanned re-operations

Binary

A longer duration (5 days) of post-operative antibiotics will result in fewer unplanned re-operations compared to a shorter duration (1 day) of post-operative antibiotics.

Cox proportional hazards

 To compare the risk of oncologic events at 1 year

Oncologic events

Binary

There will be no difference in the number of oncologic events irrespective of post-operative antibiotic duration.

Cox proportional hazards

 To compare the mortality at 1 year

Mortality

Binary

There will be no difference in the risk of death irrespective of post-operative antibiotic duration.

Cox proportional hazards

 To compare the patient functional outcomes at 1 year

MSTS-87, MSTS-93

Continuous

A longer duration (5 days) of post-operative antibiotics will result in better patient functional outcomes compared to a shorter duration (1 day) of post-operative antibiotics.

Multiple linear regression

 To compare the patient quality-of-life outcomes at 1 year

TESS

Continuous

A longer duration (5 days) of post-operative antibiotics will result in better patient quality-of-life outcomes compared to a shorter duration (1 day) of post-operative antibiotics.

Multiple linear regression

Sub-group analyses

 Tumor type (bone sarcoma, soft tissue sarcoma, and oligometastatic bone disease)

Surgical site infection

Binary

There will be no difference between the tumor types in the association between surgical site infection and post-operative antibiotic duration.

Cox proportional hazards

 Tumor location (femur or tibia)

Surgical site infection

Binary

A longer duration (5 days) of post-operative antibiotics will be more effective at reducing surgical site infections relative to a shorter duration (1 day) of post-operative antibiotics in tibial reconstructions than in femoral reconstructions.

Cox proportional hazards

 Sex (male or female)

Surgical site infection

Binary

There will be no difference between the sexes in the association between surgical site infection and post-operative antibiotic duration.

Cox proportional hazards

 Age (pediatric and young adults [12–30 years of age] or older adults [≥ 31 years of age])

Surgical site infection

Binary

A longer duration (5 days) of post-operative antibiotics will be more effective at reducing surgical site infections relative to a shorter duration (1 day) of post-operative antibiotics in the older adult population than in the pediatric and young adult population.

Cox proportional hazards

 Peri-operative chemotherapy

Surgical site infection

Binary

A longer duration (5 days) of post-operative antibiotics will be more effective at reducing surgical site infections relative to a shorter duration (1 day) of post-operative antibiotics in patients who received chemotherapy than in those who did not receive chemotherapy.

Cox proportional hazards

Sensitivity analyses

 Competing risks (death and amputation)

Surgical site infection

Binary

We do not expect the association between post-operative antibiotic duration and surgical site infections to change substantially once we take into account the competing risk of death.

Competing risks analysis

 Trial site (center-effects)

Surgical site infection

Binary

We do not expect the results to change substantially when center-effects are removed from the primary analysis.

Cox proportional hazards (with center-effects removed)

 Potential prognostic imbalances at baseline

Surgical site infection

Binary

Results will remain robust after adjusting for the following prognostic baseline imbalances: total operative time, tumor location, diabetes status, chemotherapy regimen, and radiation treatment.

Cox proportional hazards