I have to compare relative survival (have 5-year estimates with SEs and patients number per group) across 3 time periods. Cannot use log-rank etc as (probably) it is impossible to operate with raw data for this job. Found this (https://www.ncbi.nlm.nih.gov/pubmed/6671129) but the mathematical prerequisites were too much). I feel that I should use Z-scores or a Poisson model; however I did not find a simple algorithm (suitable for a medical professional) to do this. What are my options?