Aviation safety revisited

An article from today's New York Times describes a 65% drop in plane-crash fatalities over 10 years, accomplished in large part by identifying seemingly small problems that often precede crashes. Now that there are so few crashes, and data are sparse, the industry has reversed its learning patterns and analyzes flights that go well.

Apparently aviation is learning from its mistakes. Can hospitals - where, in an era of tens of thousands of deaths due to errors per year, data are not sparse - do the same?