Competing risks analysis using Markov chains: impact of cerebrovascular and ischaemic heart disease in cancer mortality

Abstract

A decrease in cerebrovascular disease (CVD) and ischaemic heart disease (IHD) mortality can produce an increase in mortality from other causes, even cancer. This problem is called the competing risks problem. A Markov chain is used to analyse the interrelation between CVD, IHD and cancer mortalities in Spanish women in 1981 and 1994. We compare the results using two models: discarding CVD and IHD mortality (the elimination model) and substituting CVD and IHD 1981 mortality rates in 1994 figures (the constant model). Removing mortality from CVD and IHD increases cancer mortality rates in women aged > or = 70, and the probability of death from cancer rises from 10.7% to 13.3%. In the second model, the use of CVD and IHD 1981 mortality rates in 1994 data yields slightly lower mortality rates and so the impact of CVD and IHD mortality changes in the period 1981 to 1994 is negligible except in elderly women. Although IHD and CVD mortality have decreased in all age groups of Spanish women from 1981 to 1994, this has not had a great impact on cancer mortality.