Intended for healthcare professionals

Feature Data Briefing

A 7/7 NHS: what price equity?

BMJ 2016; 352 doi: (Published 26 January 2016) Cite this as: BMJ 2016;352:i404
  1. John Appleby, chief economist, The King’s Fund, London, UK
  1. jappleby{at}

What are we willing to forgo to rectify unequal outcomes between weekend and weekday admissions? And, harder still, can the NHS justify spending the money to iron out these differences, asks John Appleby

A founding principle of the NHS is equal opportunity of treatment for those in equal need. It is this principle that underlies the way the NHS global budget is allocated across the country and the setting of nationwide standards and targets, for example. And it is this principle—expressed as “equality of treatment or clinical outcome regardless of the day of the week”—that underpins government policy to move towards a seven day NHS.1 But what does it mean in practical terms? What services need to be provided and at what cost?

There are two further fundamental questions. What is the evidence that there is inequality of treatment and outcomes? And, if inequality does exist, there is the standard economist’s question: given scarce resources that could be spent on something else to benefit patients, what are we prepared to forgo (for example, the reduction of other inequalities such as those arising from geographical variations in admission rates) to rectify the inequality?

All of these questions would be answered by a formal health impact assessment (usually required for major policy initiatives). Unfortunately there isn’t one. However, the NHS has produced a slew of policy documents and evidence reviews, and there is some (limited) research on mortality differences between patients admitted at weekends compared with those admitted during the week and currently just one published study of the cost effectiveness of seven day working in secondary care.2 3 4 5 6 7 8 9 10 11 12

Just as demand and need for hospital services vary by hour of the day (meaning that it would be inefficient to maintain a 24/7 service, with no variation in staffing or services between day and night time), so there is also variation between weekends and week days. There are fewer admissions at weekends than on week days (fig 1), and the nature of the patients and their illness varies too; non-emergency cases are generally admitted on a week day, for example. But is there evidence of avoidable differences in health outcomes across the week?


Fig 1 Emergency NHS admissions by day of the week, England, January-December 201413

Research suggests that even after standardising for various factors (age, sex, comorbidities, diagnosis, etc) there is a significant difference, both in England2 3 4 and in other countries,5 6 7 8 in death rates of patients admitted on weekends compared with those admitted on week days (figs 2 and 3). There is also evidence of differences between weekend and weekday admissions and the risk of an emergency readmission within seven days (fig 4). 13

The most recent study using data for NHS patients treated in England suggests a 15% higher risk of death for patients admitted on a Sunday compared with those admitted on a Wednesday (the least “risky” day).4 However, this finding relies on accurately ironing out factors unrelated to the provision of care. The statistical and interpretive difficulties of identifying a weekend admission effect is illustrated by Palmer et al’s research, which estimated a 7% higher relative risk of a still birth for mothers admitted at weekends compared with those admitted on a Tuesday.9 This finding was heavily criticised,10 and indeed, the paper itself noted extensive limitations in its analysis.


Fig 2 Crude (unadjusted) mortality risk within 30 days of discharge by day of the week, emergency admissions in England, April 2010 to March 201312


Fig 3 “Excess” deaths within 30 days of discharge derived from applying average week day crude mortality risk to total emergency admissions by day of the week, England April 2010 to March 201312


Fig 4 Risk of emergency readmissions within seven days by day of discharge, England, January to December 201413

It is also important when dealing with relative risks to bear in mind the absolute risk. In this case absolute risk is 1.8% for all deaths within 30 days of discharge4 so a 15% increase in the relative risk adds 0.27% to the absolute risk.

While many of the studies speculate on the causes of differences (consultant availability, access to diagnostic services, etc) none are able to categorically identify the reasons for the higher mortality among weekend admissions. Nevertheless, based on a small sample of eight hospitals in England, the costs of providing various forms of seven day services have been estimated at from 1.5% to 2% of a hospital’s income.11 Scaled up, this amounts to around £1.07bn (€1.4bn; $1.5bn) to £1.43bn across the NHS.

Taking these costs together with an estimate of 29 727 quality adjusted life years (QALYs) lost because of the weekend effect and a separate estimate of 36 539 QALYs based on Freemantle el al’s analysis,3 Meacock and colleagues calculate a cost per QALY saved ranging from £30 000 to £48 000.12 This assumes any new seven day service arrangement actually eradicates all excess deaths and that there are no knock-on costs to other services and patients. If seven day working was a new drug the National Institute for Health and Care Excellence (NICE) would have to think twice before recommending it to the NHS, given that it exceeds NICE’s value for money threshold (of £20 000-£30  000 per QALY).

This is a hard message, and maybe further analysis with more certain data and accounting for benefits other than averted deaths will show something different. But the difficult fact is that the existence of an inequality is not enough justification for spending money to reduce it.


Cite this as: BMJ 2016;352:i404


  • Competing interests: I have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Commissioned; externally peer reviewed.


View Abstract