Table 2 Statistical analysis plan
Variable/outcome | Hypothesis | Outcome measure | Methods of analysis |
|---|---|---|---|
1. Primary | |||
Resolution without relapse | Intervention not inferior to control | Percent cured on day 14 without relapse by day 28 [binary] | Absolute risk difference |
2. Secondary | |||
Relapse by day 90 | No difference | Percent with relapse [binary] | Chi2 ➔ Relative risk |
Time to relapse | No difference | Relapse after cure [time to event] | Kaplan-Meier survival analysis |
Objective speed of recovery | No difference | Cellulitis severity score [continuous] | Student’s T-test or Mann–Whitney U test |
Subjective speed of recovery | No difference | Visual analogue scales of pain and swelling [continuous] | Student’s T-test or Mann–Whitney U test |
Quality of life | No difference | SF-36 questionnaire score at day 1, 28 and 90 [continuous] | Student’s T-test or Mann–Whitney U test |
Additional antibiotic usage | No difference | Total amount of additional antibiotic use, in DDD [continuous] | Student’s T-test or Mann–Whitney U test |
Health care utilization | No difference | Total treatment associated costs, Health Care Utilization Questionnaire [continuous] | Student’s T-test or Mann–Whitney U test |
3. Subgroup analyses | |||
Cellulitis severity score on a continuous scale | Severity score affects cure rate | Regression analysis, with interaction term for severity score | |
Diabetes mellitus vs no diabetes mellitus | Diabetes affects cure rate | Regression analysis, with interaction term for (no) diabetes | |
4. Sensitivity analyses | Intervention not inferior to control | All outcomes | |
Per protocol analysis | Chi2 ➔ Relative risk | ||
Primary outcome adjusted for covariates | Logistic regression analysis |