This supposed ongoing upward trend in birthweight has been used as an excuse for absurd levels of inductions of labour and skyrocketing C sections, and a springboard for blaming lazy, old, fat, neglectful mothers for the “obesity epidemic”, childhood cancer, asthma, diabetes, cats and dogs living together – you name it, someone screams “Inflating babies!”, and points the finger.
We all “know” that the population is getting bigger fast, and that overweight babies are something we should be chewing our fingernails about. We are supposed to be freaking out over where the numbers are going to be in five years, ten, twenty, if we continue on what we’re told is the current upwards course.
The Hadfield Paper
This paper hit this week’s MJA:
Are babies getting bigger? An analysis of birthweight trends in New South Wales, 1990–2005“
Ruth M Hadfield et al, Medical Journal of Australia (MJA), 2009; 190 (6): 312-315
The Conclusion is:
There is an increasing trend in the proportion of babies born LGA [large for gestational age], which is only partly attributable to decreasing maternal smoking, increasing maternal age and increasing gestational diabetes.
This paper’s argument is predicated on that so-called upward trend – which, when you dig deeper, is actually an increase in the percentage of LGA babies between two time points: 1990 and 2005.
Let’s have a look at Figure One, the one that purports to show this “increasing trend” in large birthweight babies. The one that the panic is about.
The table shows the mean birthweights for male and female live-born term singletons in NSW from 1990 to 2005. The mean birthweight rose from 1990 to a peak at 2000, from about 3510 to 3540 for boys, and from about 3375 to 3405 for girls (these numbers are eyeballed; the data isn’t included.) In the old language, that’s about an ounce. A good part of that increase, though not all, is attributable to decreasing levels of maternal smoking – as the authors note.
Another hole in the data is gestational age. Only those born 37-43 weeks were included, but gestational age within that range was not adjusted for. Was there an upward or downward trend in gestational age over this time period? If “routine” C sections were switched from 38 weeks to 39 weeks, something there has been a strong push for in the past five to ten years, how has this influenced the results? (I don’t know the answer to this – I’m raising the question.)
After the year 2000, the mean birthweight FELL. Significantly. Down to about 3525 for boys, and 3390 for girls. Then it plateaued. For the past four or five years, there has been no noticeable change, and definitely no trend, in mean birthweights for either boys or girls.
But if you just read the abstract, which is as much as the vast majority of people who will quote this paper will ever read, you’d think that was a progressive or accelerating increase.
I can’t see it.
Show Me The Data!
Here’s figure 2, which shows the percentage of LGA (large for gestational age) babies (the little squares), and the decrease in smoking and increase in maternal age and gestational diabetes over the study period:
Can you see a progressive trend in those LGA figures? Again, I see a very very slight rise to the year 2000, then a drop, then wobbling around a static-looking baseline. There are no confidence intervals marked. We don’t know whether any change in LGA numbers is statistically significant.
[ETA 19 Mar 09]: Just to make things perfectly clear, we’re looking at this line.
These are the numbers we’re supposed to be panicking about. The percentage of babies being born LGA (Large for Gestational Age). This is the figure we’re being told is going through the roof. This is the FAT BABY EPIDEMIC.
The authors present this in their Results (bold emphases are mine here, because they’re what appeared in the abstract):
The percentage of male infants with birthweight 4000g increased from 14.3% to 15.8% (10.5% increase), and the percentage with birthweight 4500g increased from 2.2% to 2.4% (9.0% increase); male infants born LGA increased from 9.2% to 10.8% (17.9% increase).
The percentage of female infants with birthweight 4000g increased from 8.3% to 9.5% (15.2% increase), and the percentage with birthweight 4500g increased from 1.0% to 1.2% (20.0% increase); female infants born LGA increased from 9.1% to 11.0% (21.0% increase).
Note the trick of presenting the change in relative percentage, which makes the numbers look much more dramatic than when presented in absolute form.
But there is no detailed statistical analysis! We are not even told whether these changes – from 14.3% to 15.8%, and from 8.3% to 9.5% – are statistically significant. What is the P value? But the newspapers have taken the relative “change” in LGA numbers – 17.9% more fat boy babies! TWENTY-ONE PERCENT more fat GIRL babies! – and run with them. Everyone, lose your head and run in circles: NOW!
There were some CI lines scratched on Figure One, but we don’t know who is comparing what to who, and what the numbers are. We can’t draw an informed conclusion from them, except that it looks like the numbers in 1990 and the numbers in 2005 are probably significantly different. And to my eyeballs, it looks like the mean birthweight in 2005 is significantly different from the number in 2000 – significantly SMALLER.
The only P values I can see in the paper are those applied with the adjustment protocol, not to the actual “increasing trend” that the authors are claiming as their primary result.
Show me the data. And give me an analysis of the past five years. I want to know what’s happening now, not what happened in the 1990s, and then stopped happening.
I may have been being a little optimistic when I suggest that people will read the abstract. Most will never get further than the mainstream media’s interpretation of the press release. Let’s have a look:
The Australian: “Australian babies larger than ever”
AUSTRALIAN babies are becoming increasingly big bundles of joy, and health experts warn it’s not always something to be happy about. […] Dr Ruth Hadfield, who conducted the study with her University of Sydney colleagues, said heavier babies had increased health risks, contrary to the popular notion that big babies were healthy babies.
“For example, there is evidence of a relationship between high birth weight and the increased future risk of asthma, type one diabetes and a number of cancers, including infant and childhood leukaemia, and breast, prostate and colon cancer,” said Dr Hadfield.
A new study has confirmed that babies are getting bigger in New South Wales.
Herald Sun: “Bulky bubs put at risk”
A generation of giant babies has doctors fearing the worst for the health of newborns and their unfortunate mothers.
Since the early 1990s there has been a huge increase in the size of Australian babies, with one in six boys and one in 10 girls weighing more than 4kg at birth.
Researchers warn a rise in obese mothers responsible for some of the bigger babies creates risks for children including asthma, diabetes and leukemia, and breast, prostate and colon cancer.
Do you know a big baby? Post your comment below.
(Those Herald Sun comments are worth reading, too. People are not buying into this panic quite as wholesale as the Obesity-Obstetric-Industrial Complex would like us to.)
Louise Hall and Kate Benson at the Sydney Morning Herald get it catastrophically wrong in the first sentence:
Babies in NSW are being born up to 21 per cent bigger than they were two decades ago […]
Um, no. Not even close. Not even in the same suburb. Are there no remedial journalism boot-camps we can send these folks to?
There is nothing to panic about. Babies did get bigger, on average. Then they got smaller, on average. Now, birthweights look about static.
But that wouldn’t give us anything much in the way of dramatic headlines now, would it?
And it wouldn’t attract much in the way of research funding.