Friday, 8 July 2016

Read beyond the headlines!


Which of these statements about trisomy 18 are backed up by research?


Yes, I've given links to the research that apparently backs them up. However, when you read further into these studies, you realise that their conclusions are not in fact reliable.

  • Meyer et al is an invaluable study which certainly shows a high survival rate for children with trisomy 18. Higher than has ever been reported before. However, it is a dangerous leap, both for the authors and the lay person, to claim that the high reported survival rate reflects increased intensive care provided to newborns.


    This paper is a simple population study of survival. The authors do not have access to data showing whether the children in the study received intensive intervention or not. They don't even have a definition of ‘intensive intervention’.

    Not only that, they fail to differentiate between children with full trisomy 18 and partial or mosaic trisomy 18, even though the type of trisomy 18 is likely to have a significant impact on the chances of survival. If one was, for example, to assume the rates of partial and mosaic trisomy 18 were similar to that reported in Wu et al 2013 – a similar population study of England and Wales, where palliative care is often considered to be the norm – one might note that one-year survival is only slightly higher, and one-month survival markedly lower.

    If the newborns in the study are receiving intensive care, why are the one-month survival rates so low? Rather than focusing on the one-year survival rates, one could focus on the low one-month survival rates and – using the same theory that children in this study, being from the US, had relatively high rates of neonatal intervention – conclude that neonatal intervention has a detrimental effect on the survival of children with trisomy 18! (I'm not saying that it does, but it shows the danger of jumping to conclusions.)

  • Subramaniam is a study that concludes that lower levels of neonatal intervention might well increase survival, but ‘aggressive intervention does not’. I can't do better than to link to the Kosho/Carey comments on this. They do a better job at pointing out the flaws in this paper than I ever could.

  • Houlihan and O'Donaghue show very clearly that, in their Irish population, children with trisomy 13 had significantly longer survival than children with trisomy 18. As the opposite is true in almost every other study, I am not sure why the authors didn't question this, but reading the paper in full, it is possible to work out why in this particular group trisomy 18 survival is lower than trisomy 13 survival. Full marks if you spot it!

So what is my point? It has nothing to do with whether or not intensive interventions improve survival for infants with trisomy 18, or whether children with trisomy 18 survive longer or less long than children with trisomy 13. My point is simply that we need to be careful not to overstate what the research tells us.

When we overstate, we run the risk of being discredited very easily. Now, I'll concede, there is a good chance that if you quote research at medics, they're unlikely to read beyond the abstract either, so perhaps you'll get away with it. However, if you don't, it makes things much more difficult for you. They'll see you as either unable to understand, or attempting to deliberately mislead them. Neither helps your cause.

We need to be cautious about what we say, recognising that the research on many of the things we believe are beneficial is simply not there. For example, the evidence cannot be said to categorically support:
  • that active pregnancy management reduces the stillbirth rate in trisomy 18.
  • that elective c-section improves survival to hospital discharge in trisomy 18.
  • that prenatal diagnosis reduces survival rates.
  • that intensive neonatal intervention improves survival to one year and beyond in trisomy 18.

There is some  evidence to support some of these things, but it is not good quality, it is not definitive, and it could all be proven wrong tomorrow. Some of these things have little to no research to support them, but one could make a logical argument for them. We need to recognise the limits of our own research before someone else does, and we need to recognise our own biases. Sometimes, when you want a paper to say what you want it to say, you need to acknowledge that maybe it doesn't.

Remember always that the onus is not on you to prove beyond all doubt that the status quo is wrong; it is not necessary to prove that what you want is the best. Only that it is possible, may be beneficial and is a valid choice. The current research we have backs that up. Ask what opposing research supports the status quo.

A couple of things to finish with:
  • Joan Morris' excellent analysis on why we need to be cautious about saying that prenatal diagnosis harms survival chances for infants with trisomy 18.
  • And finally, what is wrong with using this study to show that aggressive intervention is futile?

No comments :

Post a Comment

Comments appear with some delay to allow moderation. Thanks for commenting!