Re: Margaret McCartney: The zombie statistic behind the push for seven day working
Margaret McCartney has cogently debunked the robustness of the evidence quoted in support of the Government’s unfunded initiative to impose 7-day working on the NHS , yet there has been little mention of the most obvious flaw in the studies purporting to show an increased mortality risk for patients admitted to hospital at weekends [2,3]. Patients with milder illnesses are less likely to call the doctor or come into hospital at weekends, so there are fewer admissions and the average severity of illness is likely to be greater than in those admitted on weekdays.
The Freemantle study  does not give the actual numbers of admissions for each day of the week, but another large study of 4.3M emergency admissions  showed a mortality rate of 5.2% for those admitted on Saturdays or Sundays vs. 4.9% for weekday admissions (adjusted OR 1.1, 95% CI 1.08 – 1.11). The most striking finding from this study, however, was that there were 1/3 more emergency admissions on average weekdays than on weekend days. If this extra 33% of weekday cases (“optional emergency admissions”) had a slightly lower mortality risk at 4%, say, this would be enough to account for the whole of the difference. This confounding by illness severity is likely to be even stronger for non-emergency admissions.
According to Freemantle et al , acute cerebrovascular disease is the biggest single contributor to the excess of weekend deaths, so I checked the in-hospital death rates for patients admitted to hospital in Gateshead with suspected acute stroke, using our hospital stroke register . From 1997 to early 2011, 1169.4 admissions were registered for each weekday, compared with 859.5 for each weekend day, giving a 36% higher weekday admission rate. Crude death rates were 15.5% for weekday admissions and 17.5% at weekends (OR 1.16 95% CI 1.001 – 1.34, p=0.049). The table shows that the weekday admissions included higher proportions of TIAs and stroke mimics, among whom death rates were much lower . Adjusting for this difference alone gave a Mantel-Haenszel common odds ratio of 1.125 (95%CI 0.97 – 1.31, p=0.13) and further analysis confirms a higher proportion of mild strokes admitted on weekdays.
Although the evidence for higher mortality at weekends is shaky, there is no doubt that 7-day consultant cover would have benefits, but these must be weighed against the disadvantages (particularly for continuity of care and supervision of trainees) of stretching the same workforce to provide full cover for 7 days instead of 5. In return for dropping resistance to the principle of 7-day contracts, consultants should demand that the proposed new way of working is evaluated in a large-scale, randomised cluster trial before being forced on the whole NHS. It is time to show that policy making can be based on reliable evidence rather than political expediency.
1. McCartney M. The zombie statistic behind the push for seven day working. BMJ 2015;351:h3575
2. Freemantle N, Richardson M, Wood J, et al. Weekend hospitalization and additional risk of death: an analysis of inpatient data. J R Soc Med 2012;105:74-84.
3. Aylin P, Yunus A, Bottle A, Majeed A, Bell D. Weekend mortality for emergency admissions. A large, multicentre study. Qual Saf Health Care 2010;19:213–7
4. Barer D, Cassidy T. Effects of diagnostic uncertainty and misclassification on hospital performance indicators for acute stroke care. Clin Med. 2014;14(6):597-600.
Competing interests: No competing interests