100 Apples Divided by 15 Red Herrings: A Cautionary Tale from the Mid-19th Century on Comparing Hospital Mortality Rates

doi:10.1059/0003-4819-124-12-199606150-00009

100 Apples Divided by 15 Red Herrings: A Cautionary Tale from the Mid-19th Century on Comparing Hospital Mortality Rates

Lisa I. Iezzoni, MD, MSc

From Beth Israel Hospital, Boston, Massachusetts. For the current author address, see end of text. Acknowledgments: In a brief overview of Nightingale's contribution to examining outcomes of care, I fell into the trap of seizing her “bottom line” on hospital mortality rates without investigating her methods. Duncan Neuhauser, PhD, alerted me to my failing and kindly pointed me to sources that revealed the full story. I thank Dr. Neuhauser for his gentle rebuke and for giving me a rich opportunity to investigate and learn the lessons of history. I also thank Kristin Miller, who scoured the Countway Medical Library Rare Books Room and the Harvard University libraries, for her assistance in this research. Requests for Reprints: Lisa I. Iezzoni, MD, MSc, Division of General Medicine and Primary Care, Department of Medicine, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215.

Abstract

In 1863, Florence Nightingale argued that London hospitals were dangerous, especially compared with provincial facilities.She bolstered this contention with statistics published in William Farr's Registrar-General report, which claimed that 24 London hospitals had mortality rates exceeding 90%, whereas rural hospitals had an average mortality rate of 13%. Farr had calculated mortality rates by dividing the total number of patients who died throughout the year by the number of inpatients on a single day. When calculated as the annual number of deaths divided by the total number of inpatients during the year, the mortality rate of London hospitals was 10%. A raucous debate erupted in the London medical press over how best to calculate hospital mortality rates. Critics claimed that Farr had not adjusted for differences in severity of illness between urban and rural hospitals and that his figures would mislead the public. Farr and Nightingale, in turn, criticized the poor quality of hospital data. This story reinforces the need to understand the methodologic derivation of statistics intended to compare provider quality.