November 24, 2015

expert reaction to study investigating the association between weekend and weekday births and obstetric outcomes

A paper published in The BMJ has examined differences in quality and safety of maternity services between different days, and reports that according to some of the measures which they assessed, weekends saw worse performances than weekdays.

Prof. Andrew Whitelaw, Emeritus Professor of Neonatal Medicine, University of Bristol, said:

“Several previous studies have found that babies born at weekends have slightly higher mortality than those born on weekdays and this has been a stimulus to increase the number of hours specialist obstetricians are physically present on delivery wards.

“The current study confirms this in nearly 1.4 million births in England over 24 months using data from the Hospital Episode Statistics Database. While the large number is an obvious strength, a weakness is the apparent sloppiness with which important information has been entered into the database (table 4). 13.5% of the babies were of unknown gestational age, 9.7% had unknown birth weight and in 7.4% of cases the authors did not know if the birth was of a single baby, twins or even triplets. Almost nothing is more important in deciding risk of death in a newborn baby than birth weight and gestational age.

“This missing information would have been recorded on the delivery ward in the original hospital records and would certainly have been on hospital databases but somehow the clerical staff of the Hospital Episode Statistics Database did not enter it. Why not? The authors of the current paper probably hope that the babies with missing data were no different from the rest of the population but they may well have been different. Excluding births with missing data does not get round this problem adequately when missing data are so numerous. The reason why information was “mislaid” may have been because an emergency or complication distracted staff so that they did not enter information at the usual time or in the usual place.

“Would a supermarket chain tolerate 9.7% of its checkout bills being incomplete or being unable to say whether 13.5% of the meat in its chilled section was beef, lamb or pork? Would a bank accept not knowing an accurate balance on 7.4% of accounts?

“96% of the elective Caesarean sections were carried out during weekdays. They carry a very low risk of the baby dying in the first week. The operation is planned and prepared for. All necessary trained staff know what is required and are present in good time. While elective Caesarean has its downsides, it virtually eliminates the risk of critical lack of oxygen to the fetus and of physical trauma, which are possible with spontaneous labour. I cannot see that the authors’ sophisticated statistical analysis has allowed for this big contribution of low risk babies during weekdays.

“While a comparison of weekends versus weekdays does show a statistically significant difference, the figure shows the highest perinatal mortality for an individual day was on Thursday not Saturday or Sunday.

“The main reason for the authors to use such an unreliable databank seems to have been low cost, one pound per case from the Hospital Episode Statistics Database against £10-£60 per case from clinical records. The NHS is a unique resource by which careful analysis of treatments and patient outcomes could benefit us all but it has to start with accurate collection of information. There is an overall increase in perinatal mortality at the weekend but we need much better clinical information to know how much is due to differences in the babies born at weekends and how much to differences in the quality of the clinical care.”

‘Association between day of delivery and obstetric outcomes: observational study’ by Palmer et al. published in The BMJ on Tuesday 24th November. 

Declared interests

Prof. Andrew Whitelaw: No conflict of interest