Abstract
Colossus is designed to meet a growing need for survival analysis software capable of analyzing tens of millions of rows of radiation epidemiological data. Colossus is an R package devised to offer scalable survival analysis for the Million Person Study. The total and relative rate equations available in Colossus are outlined in this article, which are used in conjunction with Cox proportional hazards, Poisson, and Fine-Grey regression models. Following a comparison with existing software, validation with epidemiological cohort data is described. Exposure data and specific causes of death among workers at Los Alamos National Laboratory and U.S. nuclear power plants were analyzed by Colossus and 32-bit Epicure and compared with published results. Colossus results agreed with the results of existing software and previous publications.