Two articles published in The BMJ in 2015 on the “weekend effect” have sparked considerable debate in The BMJ and on social media. [1,2] A number of previous studies have reported that hospital mortality is higher for patients admitted at weekends than for patients admitted on other days of the week. [3,4] This higher mortality for patients admitted at weekends has also been found in studies carried out in other countries as well as in England’s NHS.  However, the key question which this prior research has not answered is whether the higher mortality is “avoidable” and not simply due to the patients who are admitted on weekends having more complex health needs and a poorer health status than patients admitted on weekdays.
The research in England’s NHS on the weekend effect used Hospital Episode Statistics as the data source. The authors of these previous studies have tried to adjust for confounding factors that might affect mortality by using the data on secondary diagnoses in Hospital Episode Statistics. However, the extent to which we can adjust for comorbidity using these data is very limited. For example, let’s consider a patient admitted with an acute medical condition who also has type 2 diabetes. The patient’s type 2 diabetes could vary considerably in severity. The patient could have diet-controlled diabetes with good glucose control and with limited impact on their functional status; alternatively, the patient could have insulin-treated type 2 diabetes with poor glucose control and considerable impairment of their functional status. Using the secondary diagnosis code in Hospital Episode Statistics to adjust for comorbidity, these two patients would have the same risk of death even though the second patient is clearly at much higher risk of poor outcomes and death than the first.
Another key factor that has a major impact on health outcomes is a patient’s socio-economic status. In Hospital Episode Statistics, this is taken into account using postcode-based measures of area deprivation. All patients living in a postcode will be given the same level of deprivation even though these patients may differ markedly from each other in characteristics such as housing status, employment, and education level. This introduces a further limitation in research using Hospital Episode Statistics. Hence, because of the limitations of Hospital Episode Statistics data, we are not currently able to determine the extent to which higher mortality in hospital inpatients admitted at weekends is due to residual confounding rather than to any problems with the care that these patients receive.
Addressing the problem of residual confounding in research on the weekend effect requires collecting much more data on functional status, comorbidity, and socio-economic characteristics than is available in Hospital Episode Statistics. This kind of research won’t come cheap—the advantage of using Hospital Episode Statistics is that they contain data on millions of patients and it is possible to detect small absolute differences in death rates between different days of the week using them. Collecting much more detailed data on a large enough sample of patients to start to address the problem of residual confounding will be expensive. However, this research is essential if we are to try to understand the causes of higher mortality in patients admitted at weekends rather than just finding yet more associations that don’t bring us any closer to determining causes. For example, do delays in access to investigations at weekends play a part? Are there any delays in patients getting input into their care from consultants? What role do nurse staffing levels have in influencing patient outcomes at weekends? These questions can only be addressed in research that uses more detailed, higher quality, and more expensive studies than we have seen thus far.
The second area where we will need further research is in the evaluation of any initiatives that the Department of Health and NHS England introduce to improve NHS care at weekends. The history of the NHS is littered with examples of new models of care that have failed to improve health outcomes for patients.  At a time when the NHS is under considerable financial pressures, trying to implement seven day working (as the government describes it—the NHS of course already works seven days per week) may result in the weakening of clinical services during the week, with the effect that mortality may rise during Monday-Friday and cancel out any reductions in mortality through better staffing at weekends. Furthermore, even if additional resources were available to improve care at weekends, we may find that we could achieve more if these extra funds were used in different ways; for example, to improve discharge planning or give patients with long term conditions better access to rehabilitation services.
In conclusion, prior research on the weekend effect has shown only associations and not causation. It’s time for higher quality research to understand the causes of the weekend effect and for careful evaluation of any new initiatives that are introduced to improve NHS care at weekends.
1. Freemantle N, Ray D, McNulty D, et al. Increased mortality associated with weekend hospital admission: a case for expanded seven day services? BMJ 2015;351:h4596.
2. Palmer WL, Bottle A, Aylin P. Association between day of delivery and obstetric outcomes: observational study. BMJ 2015;351:h5774.
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. 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.
5. Bell CM, Redelmeier DA. Mortality among patients admitted to hospitals on weekends as compared with weekdays. N Engl J Med 2001;345:663-8.
6. Hawkes N. NHS truths that dare not speak their name. BMJ 2015;351:h4983.
Conflict of Interest: I have been a co-author of one paper on the “weekend effect” that was published in 2010. Staff in my department at Imperial College London have also published on this topic, including a paper published in The BMJ in November 2015.
Azeem Majeed is professor of primary care and head of the department of primary care and public health at Imperial College London, and a GP Principal at the practice of Dr Curran & Partners in Clapham, London. He can be followed on Twitter @Azeem_Majeed.