The Daily Mail's front page warns that there's “82% more chance of dying in surgery at weekend”, after a major study examined whether death rates following planned surgery changed depending on what day of the week the patient had their operation.
The risk of dying after planned (elective) surgery is very small. The researchers looked at more than 4 million elective procedures conducted in NHS hospitals in England between 2008-2011 and there were 27,582 recorded deaths – an overall mortality risk of around 0.67%.
However, the researchers found a statistically significant increase in the death risk as the week progressed. Patients who had their operation on a Friday or Saturday were respectively 44% and 82% more likely to die within the next 30 days, than those having surgery on a Monday.
This is an important study which suggests evidence of a “weekday effect”, in which patients having operations closer to the end of the week, or at the weekend itself, have poorer outcomes. The causes of this weekend effect remain unclear, although the authors suggest it could be due to reduced staffing levels or less experienced staff working at the weekend.
It is possible that patients having elective procedures scheduled at the weekend have a different “risk profile” to others, but despite this potential limitation, the findings from this research raise important concerns for policymakers.
You can use NHS Choices to compare local hospitals’ surgical mortality rates.
Where did the story come from?
The study was carried out by researchers from the Dr Foster unit and St Mary’s Hospital, at Imperial College, London. The study was funded by Dr Foster Intelligence, an independent healthcare information company, and the National Institute of Health Research.
Not surprisingly, given the implications of the study, it was widely covered in the UK media. The reporting on the study was broadly accurate, though one fact not widely mentioned was that only about 4.5% of elective procedures are carried out at weekends. It is also worth noting that while the Mail’s headline may have suggested these were deaths on the operating table, the death rates were actually calculated to include the 30 days after the operation.
What kind of research was this?
This was a retrospective analysis of national hospital data, taken from all acute and specialist hospitals in England carrying out elective (planned) surgery from the years 2008-9 to 2010-11. Its aim was to examine the association between mortality and the day surgery was carried out.
The researchers point out that previous research has suggested a “weekend effect” – that is, worse outcomes for patients admitted at weekends rather than weekdays.
However, studies in other countries have found no such effect, suggesting that the “weekend effect” may only occur in certain national healthcare systems and is not universal.
The researchers also say that while some research has suggested a higher risk of death for patients admitted for emergencies at the weekends, this could be because those admitted at weekends are more severely ill. They therefore decided to concentrate on mortality rates for planned surgery only.
What did the research involve?
The researchers collected hospital administrative data on patient “episodes of care”, from acute admission to final discharge, for all English NHS hospitals for the three most recent financial years. The records included information on age, sex, source of admission, the patient’s diagnosis, length of stay, date of procedure and date of death. They also had information on any other illnesses (called comorbidity) and their social and economic deprivation scores.
The researchers extracted records of all the planned patient procedures over the three years. As few elective procedures are carried out at weekends (only 4.5% of the total in the UK), they analysed Saturday and Sunday together in one category. They excluded from their analysis any admissions with missing information on age, length of stay or surgery date.
Mortality was defined as any death occurring within 30 days of the procedure (whether in hospital or after discharge). Researchers also looked at deaths within two days of the procedure, to examine shorter-term outcomes.
As well as looking at all elective surgery, they focussed on patients undergoing five higher risk major surgical procedures:
- removal of oesophagus and/or stomach
- removal of colon and/or rectum,
- coronary artery bypass graft
- repair of abdominal aortic aneurysm
- removal of lung
The researchers analysed patient deaths within 30 days, by day of the week, both overall and for the selected procedures defined above. They adjusted the results for:
- ethnic group
- socioeconomic group
- number of emergency admissions in the past 12 months and year
They carried out further statistical tests to ensure the results were valid.
What were the basic results?
The researchers found that during this period:
- there were 4,133,346 inpatient admissions for elective surgery
- there were 27,582 deaths within 30 days of the date of the procedure (overall crude mortality rate 6.7 per 1,000)
- 4.5% of elective surgery was carried out at the weekend
- the risk of death within 30 days increased with each day of the week on which the procedure was performed (starting with Monday)
- the risk of death was 44% and 82% higher respectively, if the procedures were carried out on Friday (odds ratio (OR) 1.44, 95% confidence interval (CI) 1.39 to 1.50) or a weekend (OR 1.82, 95% CI 1.71 to 1.94) compared with Monday
- mortality within two days of the procedure was also 42% and 167% higher if it was carried out on Friday or the weekend respectively compared with Monday
- for four of the five high risk procedures, mortality was higher at the end of the working week and at weekends compared with Mondays
- the low risk procedures had higher mortality rates when carried out on Friday compared to Monday, although there was no difference in risk between the weekend and Mondays
How did the researchers interpret the results?
The researchers say their study suggests there is a higher risk of death for patients who have elective surgical procedures carried out later in the working week, as well as at the weekend. The reasons behind this remain unknown, they say, but point out that serious complications are likely to occur within the first 48 hours after an operation.
Failure to “rescue” a patient with complications after surgery could be due to reduced staffing over a weekend as well as a lack of more experienced staff working during that time.
This study suggests there is a higher death rate among patients undergoing planned surgery, both before and during weekends. As the authors say, the reason for this is unknown but it could be due to reduced staffing levels or other resources.
The study’s key strengths were its use of a large national database and its inclusion of all deaths within 30 days of an elective procedure, eliminating the potential bias of only counting deaths that occurred in hospital.
However, it is possible that other factors, called confounders might have affected the results of this study, although the researchers adjusted their results for several of these.
For example, the researchers found those who had planned surgery over the weekend had slightly longer waiting times than those having surgery in the week, which might indicate that their condition was more (or less) severe. However, as the researchers point out, this would not explain the increase in mortality rates from Monday to Friday.
Even though the overall risk of death from elective surgery was small, such a significant variation will be of concern to patients and policymakers alike.
The results of the study are likely to lead to further calls for changes to the patterns of working for healthcare teams in order to improve patient outcomes.
Analysis by Bazian
Edited by NHS Website
Links to the headlines
The Independent, 29 May 2013
BBC News, 29 May 2013
Daily Mail, 29 May 2013
The Guardian, 28 May 2013
ITV News, 29 May 2013
Daily Mirror, 29 May 2013
Channel 4 News, 29 May 2013
Links to the science
BMJ. Published online May 28 2013
December 3 2012
November 28 2011
November 28 2010