Survivor Treatment Selection Bias in Observational Studies: Examples from the AIDS Literature | Synapse