Wednesday November 25 2015
The weekend increase in mortality risk was small
"Babies delivered at the weekend are significantly more likely to die or suffer serious injury," the Daily Mail reports.
However, while the increase in risk is both significant and an obvious cause for concern, it should be noted that it is a very small increase.
Researchers looked at the outcomes of 1,349,599 births in the two years from April 1 2010, and found that 0.73% of all babies born at the weekend died, compared to 0.64% of babies born on weekdays. This is about an extra death per 1000 babies born. So the overall difference is small but this is still one death too many.
While it may be tempting to assume that the extra deaths are all down to staffing issues (e.g. consultants not working at weekends) other factors may be involved. For example, most women giving birth by planned caesarean section did so during the week, although the authors controlled for this in their analysis. Babies born this way may be lower risk, which could make the weekday births appear safer.
The study highlights that the overall risk of infant death is very low. However, the small difference in risk between those born on the weekend and on weekdays cannot be ignored.
The study raises important questions about the provision of maternity services at weekends, and whether changes to staff availability and resources might reduce the numbers of deaths among babies born at the weekend.
Where did the story come from?
The study was carried out by researchers from Imperial College London and was partially funded by Imperial College London’s research centre. The study was published in the peer-reviewed British Medical Journal (BMJ) on an open-access basis, which means it is free to read online.
The tone of the reporting varied sharply between different media outlets. The Daily Mirror went with the powerful headline: "Betrayal of our babies as weekend births puts hundreds of mums and newborns at risk". The emotive headline was followed by a story that misreported the study’s figures. The report said that 770 babies delivered at the weekend die each year, when that is the estimated increased number of deaths each year, compared to if all babies were born during the week
The Guardian took a more measured approach, with "Weekend-born babies slightly more likely to die in their first week", and like most other media sources, reported the study accurately and with context.
Unsurprisingly several sources, including the Daily Mail, The Daily Telegraph and BBC News, linked the study to the ongoing dispute between the government and junior doctors, over changes to doctors' contracts that would affect weekend working.
The dispute was further inflamed by a recent controversial study, published in the BMJ in September, which estimated that there were an extra 11,000 "weekend deaths" during 2013-14.
However, the researchers themselves warned: "It is not possible to ascertain the extent to which these excess deaths may be preventable; to assume that they are avoidable would be rash and misleading".
What kind of research was this?
This is an observational study which used a database of NHS statistics to look for differences in outcomes between babies born during the week and at the weekend.
Previous studies across various medical conditions have suggested that people admitted to hospital at the weekend have increased risk of death and other adverse outcomes, compared to if admitted on a weekday.
This study aimed to see whether the association may also be found in maternity care. However, a study of this nature cannot say what has caused these differences.
What did the research involve?
The researchers used a large database of NHS statistics to find information about outcomes for women and babies in English maternity units.
They looked at seven outcomes they said could be linked to quality of care, including overall infant mortality around the time of birth (including stillbirths and deaths within seven days), tears to women’s perineum (the area between the anus and the vulva), emergency re-admissions for mother or baby, and infections. They looked at rates of these outcomes on each day of the week, and compared weekend rates to overall weekday rates.
The authors also compared between the 7-days, choosing Tuesday as a "reference day", because women admitted in labour on a Tuesday are likely to give birth during the week, and babies born on Tuesdays are not likely to have been born after a labour starting at the weekend.
They compared weekend outcomes to outcomes on a Tuesday, after taking account of a number of factors (confounders) that might have affected the results. These included the mother’s age, and the baby’s gestational age, delivery method (including caesarean section) and birth weight. They then calculated how many extra deaths are likely to have occurred at the weekend, compared to if all births had the same risks as those happening on Tuesdays.
A number of checks and adjustments to the figures were carried out to try to account for missing information and for other things that could have affected the results. They also looked to see whether maternity units which complied with recommendations about how many hours consultants should be present had better outcomes than units which did not comply with these recommendations.
What were the basic results?
Overall, 0.73% of babies born at the weekend died around the time of birth, compared with 0.64% of babies born during the week. In other words, this meant that babies born at weekends had a 7.3 in 1,000 chance of dying, compared to babies born during the week, who had a 6.4 in 1,000 chance. After taking account of factors that could explain the difference, this means that babies born at the weekend had a 7% greater chance of death (odds ratio (OR) 1.07, 95% confidence interval (CI) 1.02 to 1.13).
Mothers had a 6% higher chance of getting an infection after giving birth if they were admitted at the weekend (95% CI 1.01 to 1.11), and babies had a 6% higher chance of being injured during birth if they were born at the weekend (95% CI 1.02 to 1.09).
There was a suggestion of a marginally increased chance of the baby being re-admitted as an emergency after a weekend birth, but this just fell short of statistical significance (OR 1.04, 95% CI 1.00 to 1.08). None of the other three outcomes measured showed a statistically significant difference between weekends and weekdays.
Women giving birth in hospital units which met the Royal College of Obstetricians and Gynaecologists’ guidelines on consultant staffing levels were slightly less likely to have a perineal tear, but consultant levels showed no other differences in outcomes.
How did the researchers interpret the results?
The researchers said their study had shown that "performance across four of the seven measures was significantly worse for women admitted, and babies born, at the weekend". They highlighted the increase in stillbirths or deaths within seven days of birth as being of particular concern.
They say that "further work is needed" to understand what lay behind their findings, and concluded: "Unless managers and practitioners work to better understand and tackle the problems raised in this paper, health outcomes for mothers and babies are likely to continue to be influenced by the day of delivery".
The media headlines resulting from this study sound alarming and could be worrying for pregnant women and their partners. However, there are some good reasons to be cautious.
Firstly, it's important to keep in mind that it is unusual for babies to be stillborn or die within a few days of birth. It is devastating when it does happen, but the risk is low. In this study, this happened to around seven in every 1,000 babies born at the weekend and six in every 1,000 born on a weekday. Therefore, the absolute risk is very low, but the small difference in rates between weekends and weekdays cannot be ignored.
The biggest difficulty is that we don't know what is behind the increased chances of certain problems at the weekend. We cannot say it is simply because care is less good in hospitals than during the week.
There are a number of important limitations to the study’s results. The database used, the Hospital Episode Statistics database, should include information about what happened to people from their admission to the maternity unit onwards.
However, the researchers found that much of the information they looked for was missing, including information about babies' birth weight (missing in almost 10% of cases) and whether they were born at full term (missing in 13% of cases). These are important factors that can affect whether a baby dies, and may have nothing to do with the care they receive during birth.
The timing of admission and birth may also have affected the results. Babies were counted as having been born at the weekend if they were born between midnight on Friday and midnight on Sunday, although their mother may have been admitted in labour before then. Women were counted as having been admitted at the weekend if they were admitted between midnight on Friday and midnight on Sunday, although they may have given birth after then.
This means that babies who died might have been classified as having been born at the weekend, even though the problems leading to their death might have happened during labour on the Friday.
Conversely, mothers who had problems after being admitted at the weekend might not have encountered those problems until giving birth on the Monday.
Although the researchers tried to make allowances for these issues, the amount of missing information from the database makes it harder to rely on the results.
In an editorial published with the study, two professors of obstetrics and gynaecology from Oregon, in the US, conclude that "the most likely mechanism underlying the weekend effect is systems factors (such as staffing, resource availability, hospital policies)". This may be the answer, at least in part, as has been suggested with other areas of medical or surgical care. However likely this may be, the study does not provide evidence to prove this is the case.
The availability of consultants did not seem to make a big difference to the outcomes, although we don’t know whether the numbers of nurses, junior doctors and midwives available might have made a difference.
Overall, this study raises a lot of questions about why certain outcomes, especially deaths of babies, were more common when babies were born at the weekend. We need more research to find out the answers.