It's not the co-sleeping

21:57, Jun 06 2013

The headline read "Co-sleeping baby deaths 'like an epidemic', says coroner". People who disapprove of co-sleeping glanced at the headline and felt vindicated; those of us who support safe co-sleeping practices gritted our teeth and read on.

And we found the factors we expected: a premature baby, formula-fed, covered by adult bedding. Oh, and his mother had drunk a bottle of wine and smoked cannabis before going to bed with him. 

It's too familiar a story.

There's "Infant died in bed with sibling", which apparently sparked "renewed warnings about the dangers of co-sleeping" (rather than the dangers of sleeping your infant on a pillow with adult bedding, in a household of smokers, next to a four-year-old). Or how about "Baby died needlessly, coroner finds"? "(The) baby boy who died sharing his parents' bed..." was a micro-preemie with parents who smoked. "Sleeping with baby earns coroner's ire", but not the drinking beforehand?

It's not co-sleeping that's the problem. (Particularly because co-sleeping refers to a child sleeping in the same room, as our Ministry of Health recommends. Bed-sharing is therefore a subset of co-sleeping.) 

The coroners and the headlines are quick to point the finger, but is it really bed-sharing that's the problem? 


There are well-documented risk factors for SUDI (Sudden Unexplained Death in Infancy) while bed-sharing: exposure to cigarette smoke, premature infants, alcohol or drug use, use of bedding or pillows, bed-sharing next to anyone but mum. 

Just saying "bed-sharing bad, mmkay?" is too simplistic. One Alaskan study found that 99 per cent of bed-sharing infant deaths involved one or more risk factors, and that bed-sharing without the risk factors did not increase the risk of infant death compared to sleeping alone. 

The NZ Herald article about the "co-sleeping epidemic" quotes another study which found otherwise, making much of the conclusion that bed-sharing with non-smoking parents increases the risk of SUDI five-fold in infants under three months. (It fails to mention that the risk increases 243 times when bed-sharing parents smoked and drank, a figure which provides valuable context for a fivefold increase.) The study was a meta-analysis of five large data sets, dating back to 1987, and at least one peer reviewer raised significant concerns:

"Maternal alcohol consumption prior to the last sleep was collected but only for 38.7 per cent of the mothers in the study. Imputing values for parental and alcohol drug consumption on a particular night from a single study when more than 50 per cent of the data is missing requires a fairly homogeneous population and good predictors of 'missingness'. Imputing values from a group of five studies, three of which did not even ask the question, is surely making unreasonable equivalence assumptions across studies conducted in different countries with different cultures in different time periods."

The Herald article concludes "previously there had been conflicting opinions about whether bed sharing represented a risk when these factors were not present." I'm pretty darn sure there are still conflicting opinions. 

» Follow PG on Facebook