To the Editor.
—The thoughful article by Brennan et al1 reported data on the distribution of negligent AEs in 52 hospitals in New York. The range of this distribution was 1% to 60% (mean, 25%). With these data collected over 1 year, the authors asserted that this distribution was significantly nonrandom (P<.001). We might easily conceive of a random process that generates samples, with a mean of 25 and a range of 1% to 60%. Only by understanding the processes that generate negligent AEs and by studying rates over time might we determine whether the rates are part of a common cause variation (ie, due to chance alone) or are due to special causes (ie, nonrandom).2 Park et al3 illustrated this issue with hospital death rates. Contrary to the assertion by Brennan et al, high outlier hospitals should not "be the focus of greater internal and
Schlosser J. Hospital Events Associated With Adverse Events and Substandard Care. JAMA. 1991;266(21):2983. doi:10.1001/jama.1991.03470210051018