Intended for healthcare professionals

Feature 24/7 NHS

The weekend effect—how strong is the evidence?

BMJ 2016; 353 doi: https://doi.org/10.1136/bmj.i2781 (Published 19 May 2016) Cite this as: BMJ 2016;353:i2781

Why are weekend emergency admissions lower than weekdays?

If we are to understand the ‘week-end effect’ this question could be the most important question to answer, yet it does not seem to have been adequately addressed by current research. Do patients really get to choose on which day they have an emergency? Why is there this difference in admission rates?

Consider the following two scenarios: a patient develops a dense hemiplegia affecting the right arm and face compared with a patient developing tingling in the face and right arm. Which of these two patients is most likely to present to hospital on the day the symptoms arise?

Any GP will tell you there are a proportion of patients who will develop symptoms at a weekend and not present to their GP until the Monday or Tuesday. This is far more likely if the symptoms are at the mild end of the spectrum when the patient interprets them as being less serious.

There are two points that arise from this. The first is that mild symptoms are associated with longer survival and these long survivors are in effect being shifted from the weekend statistics to the week days. This increases mortality rates for the weekends and decreases it for the weekdays. Is there any evidence for this sort of shift? The PLOS one paper [1], one of the few that gives a daily breakdown of admission, shows that admissions for strokes on a weekend are 16% lower than average, yet on a Monday and Tuesday they are 11% higher - although this does not characterise the types of patients being shifted.

The second point is that a few studies have claimed they have controlled for case-mix differences in their respective models [2,3]. This usually takes the form of including covariates that capture co-morbidities such as hypertension or diabetes, demographic factors such as age, numbers of previous admissions, deprivation scores, and so on. Such data are often well coded; however, the difference between the 2 scenarios described lies in the presenting symptomatology. It is very unlikely that presenting symptoms will be captured in the hospital coding. Thus, the models so far are unable to distinguish between cases such as the ones described. This is important, particularly if this could potentially affect the day the patient presents.

To their credit Meacock et al [4] have considered the question of why there may be difference between admission rates. They examined the differences between patients attending A&E and those being admitted for the weekend and weekdays and concluded that there was higher threshold for admission on a weekend. This would act as relative filter where patients presenting with mild symptoms on a weekend would be less likely to be admitted.

This would suggest that there are at least two mechanisms that would lead to a difference in admission rates. These need to be explored more extensively before the sort of strong assertions that have been made on the existence of a ‘weekend effect’ should have been made. It may end up being a statistical artefact.

The BMJ should bear some of the responsibility on this. The journal has published several articles in support of a ‘weekend effect’ one in which the statistical reviewer was asked to review on no fewer than three occasions despite expressing reservations on the analysis (the reviews are available on-line) [5]. The editor in an editorial said the evidence ‘sent a consistent message: an association between admission to hospital at weekends and a higher risk of death’ implying a ‘weekend effect’ [6]. It is incredible therefore that it did not publish the article by Meacock et al which offers a reasonable counter to the ‘weekend effect’ [4] and raises suspicions of an editorial policy that is in favour of the ‘effect’. This has been a politically charged issue which has received widespread coverage in the press and the journal editors need to re-examine whether they have prioritised sensationalism over objectivity.

Reference
1. : Roberts SE, Thorne K, Akbari A, Samuel DG, Williams JG (2015) Mortality following Stroke, the Weekend Effect and Related Factors: Record Linkage Study. PLoS ONE 10(6): e0131836. doi:10.1371/journal.pone.0131836

2. Freemantle N, Ray D, McNulty D, Rosser D, Bennett S, Keogh BE, Pagano D. Increased mortality associated with weekend hospital admission: a case for expanded seven day services? BMJ. 2015 Sep 5;351:h4596. doi: 10.1136/bmj.h4596

3. Aldridge C, Bion J, Boyal A, Chen YF, Clancy M, Evans T, et al. Weekend specialist intensity and admission mortality in acute hospital trusts in England: a cross-sectional study. Lancet. 2016 May 10. pii: S0140-6736(16)30442-1.

4. Meacock R, Anselmi L, Kristensen SR, Doran T, Sutton M. Higher mortality rates amongst emergency patients admitted to hospital at weekends reflect a lower probability of admission. J Health Serv Res Policy OnlineFirst, published on May 6, 2016 as doi:10.1177/1355819616649630

5. Palmer W, Bottle A, Aylin P. Association between day of delivery and obstetric outcomes: observational study. BMJ 2015;351:h5774

6. Godlee F. How Jeremy Hunt derailed clinician led progress towards a seven day NHS. BMJ. 2016 Jan 13;352:i187. doi: 10.1136/bmj.i187.

Competing interests: No competing interests

21 May 2016
Brian H Willis
NIHR Clinical Lecturer in Primary Care
University of Birmingham
Institute Of Applied Health Research, University of Birmingham, Edgbaston, Birmingham