Posted by: Nigel Praities Editor's Blog25 November 2015
Dead babies. An emotive subject at the best of times (as it should be).
That is why I was interested in reading a new study published in the BMJ that reports to show a ‘weekend effect’ in obstetric care.
The paper – published today – reports ‘approximately 770 perinatal deaths and 470 maternal infections per year above what might be expected if performance was consistent across women admitted, and babies born, on different days of the week’.
There could not be a more important subject for research and this particular paper was press-released by the BMJ to all the major media outlets with the headline: ‘Study finds higher risks for babies born at the weekend’.
This has resulted with two national newspapers splashing headlines such as ‘Betrayal of our Babies’ and ‘Risk of having a weekend baby’ across their front pages today.
These are front pages that will cause an enormous amount of upset and concern – I would not be surprised if GPs are faced with mothers terrified about being admitted on a weekend – and it will provide yet more grist to the health secretary’s seven-day working mill.
And I would not have such a problem with all of that if the result was as clear cut as it has been represented. Except, it isn’t.
The authors provide no real explanation at all for their figures, they suggest the weekend effect ‘might be amenable to the provision of healthcare’, but that is about it.
Even worse, the authors chose a midweek reference day of Tuesday for their comparisons, and this did show a significant increase in in-hospital neonatal mortality at weekends. Except if they had used Thursday as the comparator, then they would have shown a decrease.
Yes, you read that right. Neonatal mortality was actually higher on a Thursday compared with Saturday and Sunday.
If you look at puerperal infections then there is a peak on a Saturday, but again, they have chosen to compare it with Tuesday. The lowest point on the graph. And if you squint, then Sunday looks a lot like Friday.
So we are talking about a ‘Thursday-Friday-Saturday-Sunday’ effect? Well perhaps, but that does not make such a good headline. Why are they claiming there is a ‘weekend effect’ at all? I have no idea. In fact, I have no idea quite how this paper got through peer review at all given in the state it is in.
And it would be forgivable if it wasn’t the second paper the BMJ has published showing a ‘weekend effect’ that offered equivocal conclusions. After the previous paper, the editor of the BMJ had to write to the health secretary to tell him to stop using their data to justify his new junior doctor contract.
Hey, I have a radical idea. Perhaps we should ditch publishing this line of research altogether until it can come up with clear answers. It is just making things worse.
Nigel Praities is editor of Pulse