Figure 1From: Design of cohort studies in chronic diseases using routinely collected databases when a prescription is used as surrogate outcome Definition of outcome event and censoring in the sub-cohort I (non-users of OS at RA index date).Back to article page