Intended for healthcare professionals

Research News

Weekend hospital death rates are not just an English problem

BMJ 2015; 351 doi: https://doi.org/10.1136/bmj.h3659 (Published 07 July 2015) Cite this as: BMJ 2015;351:h3659

Re: Weekend hospital death rates are not just an English problem

26/7/15

First do no harm

David Cameron stated that there is a 16 percent increased chance of dying in hospital if you are admitted on a Sunday rather than a weekday with the implication that this justified an overhaul of the NHS weekend hospital working practices. Either the government are using percentages to portray a situation worse than it is or they do not know the difference between percentages and percentage points.

If one consults the source paper (1) there were 14,217,640 admissions with 187,377 (1.3 percent) inpatient deaths within 30 days of admission. The paper does not state the number of admissions on each day of the week and the number of deaths subsequent to these admissions. What we are told are the 30 day hazards of deaths was Monday vs Wednesday (1.02) Tuesday vs Wednesday (1.0), Thursday vs Wednesday (1.0), Friday vs Wednesday (1.0), Saturday vs Wednesday (1.11) and Sunday vs Wednesday (1.16), the last providing the 16 percent increased chance of dying following a Sunday admission. We can only guess the numbers but, for the sake of illustration, if we hypothesize that there were an equal number of admissions each day of the week, we can estimate the same number of deaths, 26,768, should occur following 2,031,091 admissions each day, giving an inpatient death rate of about 1.3 percent of admissions for each day. If there were 16 percent more deaths, 31,051 (26,768 increased by 16 percent), following Sunday admissions, that is a death rate of about 1.52 percent following Sunday admissions. Thus whilst there was a 16 percent increase in deaths Sundays vs Wednesday, given the large number of admissions this equates to a difference of about 0.22 percentage points. Sufficient to radically change clinical practice?

The source paper does not offer explanations for differences between Wednesday, Sundays, Saturdays and, to a lesser degree, Mondays. The discussion and conclusions includes seven “mays,” one “possibly,” one “suggests,” and one “could speculate.” The 16 percent difference in Sunday and Wednesday admission deaths may be caused by factors not analysed. Indeed the extra 16 percent of deaths contains a range of 0-63 percent of extra risk for various clinical conditions. Importantly, neither the numbers nor the hazard factor for Sunday vs Saturday are given. Nevertheless we can infer the hazard for Saturday is less than that for Sunday. This contradicts the suggestion that absence of weekday type services on weekends causes increased risk. If this absence did cause increased risk then the risk on Saturdays (an extra day before normal weekday services resumed) should be higher than on Sundays.

Whilst every extra Sunday death within 30 days is a tragedy, attempting to remove a 0.22 percent point inequality between deaths following Sundays and Wednesdays may do more harm than good. This is a distinct possibility because the source paper also reports the likelihood of deaths occurring on Sunday were lower than those on other days of the week! Curiously the government did not mention this. Making more hospital staff work at the weekend would mean less working on weekdays and this probably would worsen the risk of dying after a weekday admission.

The authors’ end statement is that “the economics for such a change (in weekend services) needs further evaluation to ensure that such reorganization represents efficient use of scarce resources.” The evaluation has not occurred and there is no evidence that radical policy changes would be beneficial.

Jeremy Hunt should be congratulated. He wants to reduce the death rates somewhere in the order of 0.22 percentage points without realizing the financial costs, the opportunity costs, and the now irreversible costs of alienating hospital staff. Hunt may not know that most acute service medical consultants are actively involved at the weekend when on call. I and my colleagues used to do a full ward round on Saturdays and Sundays when on call and would go in immediately when required.

For the government to think that their initiatives will achieve much is deluded and there is the possibility more harm might result.

Philip D Welsby
Honorary Teaching Fellow,
Edinburgh University
Previously Consultant in Infectious Diseases, Edinburgh
Philipwelsby@aol.com

Reference
Weekend Hospitalization and Additional Risk of Death: an Analysis of inpatient data. Freemantle N, Richardson M, Wood J et al, J R Soc Med 2012;106:74-84.

Competing interests: No competing interests

28 July 2015
Philip D Welsby
Honorary teaching fellow Edinburgh University
Edinburgh