Different statistics tell you different things. The event rate tells us about events, while the cumulative incidence tells us about people. A person going in for an operation might like to know the chances they they will experience one or more complications. However, a hospital administrator would want to know how many adverse events occurred per 1,000 admissions. 

In terms of identifying risk factors, relative and absolute risk are useful, but what is more interesting sometimes is attributable risk in the exposed (how does the factor affect risk in a person who has it) and the population (what proportion of the cases in the population are attributable to the factor)

You probably need to read up on the literature. The best answer depends on the question! So you need to formulate your question(s)

Dear Dr Ronán Michael Conroy,

Thanks for your answer.

I need to calculate event of an outcome for postoperative complications in one group  at  various follow up times? 

Dear Dr. Al-Moraissee,

It would help you to define your time period.  Like Dr. Conroy mentioned, if you are interested in the number of patients who have a complication in the first 30 days after surgery, you would calculate the proportion of patients with a complication in that time period.  This is the cumulative incidence, also known as the 30-day risk of complication.  

Let me recommend to you an excellent text I use with my medical students:  Clinical Epidemiology--the Essentials by Fletcher, Fletcher, and Fletcher.  

Kim

Dear Keturah Faurot ,

Thank you for your answer. I thought that I have to calculate the event rate per 100 by dividing an occurrence of complications on the total sample size . Then to aggregate the whole included studies  using event rate with sample size to estimate the overall effect size .

Thanks for the all answers