Intended for healthcare professionals


Experts disagree about usefulness of hospital mortality data

BMJ 2014; 349 doi: (Published 15 September 2014) Cite this as: BMJ 2014;349:g5658
  1. Rebecca Coombes
  1. 1The BMJ

Do hospital mortality ratios give patients meaningful data about their treatment centre, or are they over-simplistic and misleading? This question was the subject of a heated debate at the Risky Business conference on 11 September between Nick Black, professor of health services research at the London School of Hygiene and Tropical Medicine, and Roger Taylor, director of research and public affairs at the health analytics company Dr Foster.

Taylor, a cofounder of Dr Foster, defended hospital standardised mortality ratios (HSMRs) as being “incredibly useful if used as a smoke alarm—everyone from Francis to Keogh says it’s idiotic to ignore them,” referring to Robert Francis QC’s inquiry into catastrophic care failures at Mid Staffordshire hospital and Bruce Keogh’s follow-up probe into English hospitals with high death rates.

He added, “Patients want aggregate, composite measures for different procedures. Hospital standardised mortality ratios are part of that. When we started Dr Foster with Imperial College [London] there was very little information about quality of care; you mostly had to rely on waiting list data.” Taylor acknowledged that playing the system with palliative care coding was a problem, but he added that HSMRs did correlate with other measures of quality.

However, Black said that HSMRs were a good illustration of the famous warning by HL Mencken, the American journalist, about simple answers to complicated questions. Quoting Mencken, he said, “For every complex problem there is an answer that is clear, simple . . . and wrong.”

Black said that we should be cautious of oversimplifying healthcare. “A quantitative approach can contribute, but we need a wide portfolio of measures. Hospital is where 50% of people end up dying—so it is counterintuitive to use HSMRs as an indicator of poor quality, when it is actually one of hospital’s functions.”

He noted that, in Massachusetts, hospitals no longer posted overall death rates because managers did not believe them to be a valid indicator of the quality of care in hospitals. He criticised “private companies who produce mortality rates and then sell ‘solutions’ to hospitals.”


Cite this as: BMJ 2014;349:g5658


  • Letter: Hospital standardised mortality rates should not be used to make interhospital comparisons (BMJ 2013;347:f6155, doi:10.1136/bmj.f6155); Research: Including post-discharge mortality in calculation of hospital standardised mortality ratios: retrospective analysis of hospital episode statistics (BMJ 2013;347:f5913, doi:10.1136/bmj.f5913); Research methods and reporting: Strengths and weaknesses of hospital standardised mortality ratios (BMJ 2011;342:c7116, doi:10.1136/bmj.c7116); Analysis: Using hospital mortality rates to judge hospital performance: a bad idea that just won’t go away (BMJ 2010;340:c2016, doi:10.1136/bmj.c2016)

View Abstract