Fig. 2
Temporal representation of the non-concomitant measurement times of the exposure and the subsequent health outcome considered in our study framework. For both exposure and health outcome, measurements are collected at discrete and individual-specific occasions with error. The exposure history is constructed according to the time tU prior to the first outcome assessment at time 0 (−S≤tU≤0). The longitudinal health outcome Y is modeled prospectively according to time t (t≥0)