A study published recently in Pediatrics cited the "statistical uncertainty" of using mortality rates in children's hospitals to rank their quality. The study of approximately 473,000 U.S. patients discharged from children's hospitals in 2008 found that adjusted mortality rates at the 42 hospitals studied did not vary much by the hospital's ranking ("Statistical Uncertainty of Mortality Rates and Rankings for Children's Hospitals").
In a cogent Reuters article on the topic, Frederik Joelving points out that death rates, " which carry heavy weight in commercial rankings like the U.S. News & World Report Best Hospitals, are mostly indistinguishable from a statistical point of view" ("Hospital ratings for kids a roll of the dice: study").
Dr. Chris Feudtner, lead author on the Pediatrics article, told Reuters that when patients look for a hospital, along with rankings they should also consider other issues, such as its proximity to their home. Statistics seem to promise clear answers to complicated questions, but sometimes they can't deliver that.