Skip to main content

Table 2 Analysis methods for vaccine effectiveness against persistent HPV infections

From: Measuring vaccine effectiveness against persistent HPV infections: a comparison of different statistical approaches

  Conditional exact method for comparing two independent Poisson rates using a binomial distribution Cox proportional hazard GEE Poisson Prentice Williams Peterson-Total time
Outcome Rate ratio Hazard ratio Rate ratio Hazard ratio
Assumption(s) * Rate of events constant over time
* Groups are considered to be equally exposed [52]
* Proportional hazard assumption (hazard ratio over time should be constant)
* Independence assumption (estimate only for 1st event) [52]
* Rate of events constant over time [52]
* Measurements are independent across subjects
* Measurements may be correlated within subjects
* Event specific baseline hazard (baseline hazard for kth event allowed to be different) [51]
Check assumptions in HAVANA Assumption for constant rate over time violated among unvaccinated Proportional hazard assumption not violated Assumption for constant rate over time violated among unvaccinated Assumption for event-specific hazard not violated