In a 2005 public health advisory, the FDA warned that antipsychotics appear to increase the rate of mortality in elderly patients with dementia. The implication is that we should discontinue such agents in this population when possible. But does this actually decrease the mortality risk?