To the Editor: In their Special Communication,1 Dr Bach and colleagues argue that “look-back”
studies of decedents produce biased inferences regarding treatment patterns
and quality of care because patients who have their initial diagnosis of fatal
illness during their last year are not a random sample of those with a disease
and thus have systematically different durations of exposure to end-of-life
costs and utilization.
Barnato AE, Lynn J. Resurrecting Treatment Histories of Dead Patients. JAMA. 2005;293(13):1591-1592. doi:10.1001/jama.293.13.1591-b