Fit Proportional Hazards/ Cox proportional Hazard/ risk ratio/ Likelihood ratio
Jul 4, 2019 12:14 PM(210 views)
I am using Fit Proportional Hazard platform for checking if a test (continuous variable which is divided into terttiles-T1,T2,T3) is associated with the risk of any event (E) in a follow up period of 1year.
1. During the analysis, Whole model P= 0.08, Likelihood ratio is 0.08, since these P values are insignificant so in that case can I still compare the risk ratio between the tertiles(T1,T2,T3)?
2. If the Risk ratio between any of the two tertiles for example T1 vs T3 is signficant, can that be reported?