Age as Time-Scale: An Application to the Survival Analysis of Chronic Diseases

Abstract

Background: Cohort study survival analysis does not take into account the age at which an individual dies, but rather the age throughout which the individual has been included in the follow-up. This study is aimed at assessing the resulting differences in the calculation of the relative risk (RR) when the age is used instead of follow-up as a timescale in the chronic diseases survival analysis. Methods: The five-year cohort follow-up data for the Catalan Health Interview Survey (COHESCA) is analyzed. Cox models are adjusted to the follow-up time («standard» approach) and the age («alternative» approach) as the timescale. The main covariable used is the comorbility index, which is adjusted by demographics, lifestyles and self-perceived health. Results: Among men in the 40-64 age range, using the «standard » approach, no significant increase in the RR of death was found. Among women within the 40-64 age range with a 3-4 score on the index, there was a significant increase in the RR. Among those over age 64, a significant increase in the risk was found for the 3-4 score (RRmen= 2.1(1.3-2.5); RRwomen=2.1(1.2-3.7)) and >=5 on the index (RRmen= 3.4(1.7-6.9); RRwomen=4.0(2.1-7.7)). Using the «alternative» approach, the RR was 1.0 (0.7-1.4) among men with a score of 1-2, 1.5 (1.0-2.3) for those with a score of 3-4 and 2.6 (1.4-4.9) for those with a score of >=5. Among women, the RR were respectively 0.8 (0.5-1.2), 1.7 (1.1-2.7) and 2.6 (1.5-4.8). Conclusions: Both approaches showed an increase in the RR according to the comorbility index. However, in the «alternative» approach, it is adjusted by the confounding effect of age.