Geriatrician assessment and immortal time bias in the FiTR 2 study

Detecting Selection Bias in Observational Studies—When Interventions Work Too Fast

“…Instead, the results suggest residual confounding. Examples include the association of bisphosphonates with mortality 4 and the effect of geriatrician involvement following serious traumatic injury, 5 among others. 6 The Figure shows a visual representation of this construct, in which the curves separate from the beginning of the follow-up period.…”