Lexis Expansion- Age-at-Risk Adjustment for Survival Analysis

Cox proportional hazards model is a widely used method to analyse time-to-event data for survival analysis. In the medical field, age is the most common confounder of any association between risk factor and disease since most risk factors and the risk of disease change with age. Usually, age recorded at the start of the study (age-at-baseline) will be adjusted for in the model to account for the effect of age on risk, but the effect of 'acquired' age during the study (age-atrisk) may be more appropriate because it is age-at-risk that determines current risk rather than the age when the risk factor was measured or determined. Lexis expansion is a method that can facilitate the adjustment of age-at-risk by expanding the observation of each individual into several observations of different age-at-risk bands; age-at-risk can then be adjusted using Cox proportional hazards model. In this paper, Lexis expansion will be explained with examples as well as the implementation of the Cox model using the expanded data.